Nanotechnology: Nanofabrication, Patterning, and Self Assembly 9781617617713

Citation preview

NANOTECHNOLOGY SCIENCE AND TECHNOLOGY SERIES

NANOTECHNOLOGY: NANOFABRICATION, PATTERNING AND SELF ASSEMBLY

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.

NANOTECHNOLOGY SCIENCE AND TECHNOLOGY SERIES Safe Nanotechnology Arthur J. Cornwelle 2009. ISBN: 978-1-60692-662-8

New Nanotechnology Developments Armando Barrañón (Editor) 2009. ISBN: 978-1-60741-028-7

National Nanotechnology Initiative: Assessment and Recommendations Jerrod W. Kleike (Editor) 2009. ISBN: 978-1-60692-727-4

Electrospun Nanofibers and Nanotubes Research Advances A. K. Haghi (Editor) 2009. ISBN: 978-1-60741-220-5

Nanotechnology Research Collection 2009/2010. DVD edition James N. Ling (Editor) 2009. ISBN: 978-1-60741-293-9

Electrospun Nanofibers and Nanotubes Research Advances A. K. Haghi (Editor) 2009. ISBN: 978-1-60876-762-5 (Online Book)

Nanotechnology Research Collection 2009/2010. PDF edition James N. Ling (Editor) 2009. ISBN: 978-1-60741-292-2 Strategic Plan for NIOSH Nanotechnology Research and Guidance Martin W. Lang 2009. ISBN: 978-1-60692-678-9 Safe Nanotechnology in the Workplace Nathan I. Bialor (Editor) 2009. ISBN: 978-1-60692-679-6 Nanotechnology in the USA: Developments, Policies and Issues Carl H. Jennings (Editor) 2009. ISBN: 978-1-60692-800-4 Nanotechnology: Environmental Health and Safety Aspects Phillip S. Terrazas (Editor) 2009. ISBN: 978-1-60692-808-0

Carbon Nanotubes: A New Alternative for Electrochemical Sensors Gustavo A. Rivas, María D. Rubianes, María L. Pedano, Nancy F. Ferreyra, Guillermina Luque and Silvia A. Miscoria 2009. ISBN: 978-1-60741-314-1 Nanostructured Materials for Electrochemical Biosensors Yogeswaran Umasankar, S. Ashok Kuma and Shen-Ming Chen (Editors) 2009. ISBN: 978-1-60741-706-4 Magnetic Properties and Applications of Ferromagnetic Microwires with Amorphous and Nanocrystalline Structure Arcady Zhukov and Valentina Zhukova 2009. ISBN: 978-1-60741-770-5

Electrospun Nanofibers Research: Recent Developments A.K. Haghi (Editor) 2009. ISBN: 978-1-60741-834-4

Bio-Inspired Nanomaterials and Nanotechnology Yong Zhou (Editor) 2009. ISBN: 978-1-60876-105-0

Nanofibers: Fabrication, Performance, and Applications W. N. Chang (Editor) 2009. ISBN: 978-1-60741-947-1

Nanopowders and Nanocoatings: Production, Properties and Applications V. F. Cotler (Editor) 2010. ISBN: 978-1-60741-940-2

Nanofibers: Fabrication, Performance, and Applications W. N. Chang (Editor) 2009. ISBN: 978-1-61668-288-0 (Online Book) Barrier Properties of Polymer Clay Nanocomposites Vikas Mittal (Editor) 2010. ISBN: 978-1-60876-021-3 Polymer Nanocomposites: Advances in Filler Surface Modification Techniques Vikas Mittal (Editor) 2009. ISBN: 978-1-60876-125-8

Nanomaterials: Properties, Preparation and Processes Vinicius Cabral and Renan Silva (Editors) 2010. ISBN: 978-1-60876-627-7 Nanomaterials Yearbook - 2009. From Nanostructures, Nanomaterials and Nanotechnologies to Nanoindustry Gennady E. Zaikov and Vladimir I. Kodolov (Editors) 2010. ISBN: 978-1-60876-451-8 Nanotechnology: Nanofabrication, Patterning and Self Assembly Charles J. Dixon and Ollin W. Curtines (Editors) 2010. ISBN: 978-1-60692-162-3

NANOTECHNOLOGY SCIENCE AND TECHNOLOGY SERIES

NANOTECHNOLOGY: NANOFABRICATION, PATTERNING AND SELF ASSEMBLY

CHARLES J. DIXON AND

OLLIN W. CURTINES EDITORS

Nova Science Publishers, Inc. New York

Copyright © 2010 by Nova Science Publishers, Inc.

All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works. Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA Nanotechnology : nanofabrication, patterning, and self assembly / editors, Charles J. Dixon and Ollin W. Curtines. p. cm. Includes index. ISBN 978-1-61761-771-3 (Ebook) 1. Nanostructured materials. 2. Nanotechnology. I. Dixon, Charles J. II. Curtines, Ollin W. TA418.9.N35N3574 2009 620'.5--dc22 2009004990

Published by Nova Science Publishers, Inc. Ô   New York

CONTENTS Preface

xi

Research and Review Studies

1

Chapter 1

Electrochemical Nanofabrication Di Wei

Chapter 2

Fabrication and Application of Novel Two-Dimensional Nanowebs via Electrospinning Bin Ding, Chunrong Li, Dong Wang and Seimei Shiratori

3

51

Chapter 3

Nano-scale Characterization and Spectroscopy of Strained Silicon Norihiko Hayazawa and Alvarado Tarun

Chapter 4

Nanotechnologies for Cancer Diagnostics and Treatment Phong Tran and Thomas J. Webster

Chapter 5

Mechanical Characterization at Nanometric Scale of Ceramic Superconductor Composites J.J. Roa, X.G. Capdevila and M. Segarra

151

ZnO Nanowire Arrays: Template-free Assembly Growth and Their Physical Properties Bingqiang Cao and Weiping Cai

237

Chapter 6

Chapter 7

Chapter 8

Spatially Resolved Control of Electrical Resistivity in Organic Materials —Development of a New Fabrication Method of Junction Structures Toshio Naito Fabrication of Electrical Contacts on Individual Metal Oxide Nanowires and Novel Device Architectures Francisco Hernandez-Ramirez, Juan Daniel Prades, Roman Jimenez-Diaz, Olga Casals, Albert Cirera, Albert Romano-Rodriguez, Joan Ramon Morante, Sven Barth and Sanjay Mathur

71 107

275

293

viii Chapter 9

Chapter 10

Chapter 11

Chapter 12

Chapter 13

Contents Functionalization of Nanoparticles, Nanotubes and Nanowires by Surface-Initiated Atom Transfer Radical Polymerization Jinying Yuan, Mi Zhou and Yingwu Yin

309

Synthesis and Applications of Nano-sized Ferroelectrics via Mechanochemical Activation L.B. Kong, Z. Xu and T.S. Zhang

331

Preparation and Characterization of Monoatomic Carbon Chains: Unraveling, Field Ion Microscopy, and Field Emission Igor M. Mikhailovskij

371

Sequential Nucleation and Growth of Complex Nanostructures by a Two-Step Strategy Li Yang, Paul W. May and Lei Yin

409

Progress of Self-standing Diamond Film Fabricated by DC Arc Jet Plasma CVD G.C. Chen, F.X. Lu, B. Li, C.M. Li, W.Z. Tang, J.H. Song, L.F. Hei and Y.M. Tong

Chapter 14

Nanoshell Arrays: Fabrication and Enhanced Photoluminescence Zhipeng Huang and Jing Zhu

Chapter 15

A Strategy for the Incorporation of Trivalent Lanthanide Ions into Anatase Tio2 Nanocrystals Wenqin Luo, Chengyu Fu, Renfu Li and Xueyuan Chen

Chapter 16

Chapter 17

Chapter 18

Nanocrystallite Superhard Titanium Nitride Film in Multi-arc Ion Plating Xiang Yu, Chengbiao Wang, Meng Hua, Yang Liu and Shengli Ma

435

459

479

509

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation X.T. Zu, X. Xiang, S. Zhu and L.M. Wang

525

Structural, Dynamical and Optical Properties of Self-assembled Porphyrins at the Mesoscopic Scale Valentina Villari, Norberto Micali and Luigi Monsú Scolaro

559

Short Communications Short Communication A The Influence of Thiophene Addition on Catalytic Pyrolysis of Poly (Dimethyl Siloxane) K.F. Cai, C.W. Zhou, A.X. Zhang and J.L. Yin

603

605

Contents

ix

Short Communication B Nanofinishing of Cotton Textiles N. Vigneshwaran and Virendra Prasad

615

Index

621

PREFACE This new book is dedicated to outstanding research in nanotechnology which is a “catchall” description of activities at the level of atoms and molecules that have applications in the real world. A nanometer is a billionth of a meter, about 1/80,000 of the diameter of a human hair, or 10 times the diameter of a hydrogen atom. Nanotechnology is now used in precision engineering, new materials development as well as in electronics; electromechanical systems as well as mainstream biomedical applications in areas such as gene therapy, drug delivery and novel drug discovery techniques. Nano- and micro-fabrications have been largely used in the applications such as integrated circuits, micro/nano electro-mechanical systems (M/NEMS), micro-optics and countless others. The methodology of nanofabrication can be divided into two types, topdown and bottom-up processes, which themselves can be further divided. Top-down process refers to approaching the nanoscale from the top (or larger dimensions), such as lithography, nanoimprinting, scanning probe and E-beam technique etc.. In bottom-up fabrication processes, the nanotechnology process builds nanoscale artifacts from the molecular level up, through single molecules or collections of molecules that agglomerate or self-assemble. Using a bottom-up approach, such as self-assembly enables scientists to create larger and more complex systems from elementary subcomponents (e.g. atoms and molecules). In general, top-down processes that transfer minute patterns onto material are more matured than bottomup processes. An exception is epitaxial processes that create layers through layer-by-layer growth with registry at the atomic level. Electrodeposition has actually been used for decades to form high quality, mostly metallic, thin films. It has recently been shown that high quality copper interconnects for ultra large scale integration chips can be formed electrochemically on Si wafer [1;2]. Electrodeposition has thus been shown compatible with state of the art semiconductor manufacturing technology. The largest semiconductor companies, for example, IBM, Intel, AMD, Motorola etc. are installing wafer-electroplating machines on their fabrication lines [1]. The electrodeposition of Cu with the line width 250 nm was used in the mass-production of micro-processor Pentium III in 1998. In 2003, the line width of the CPU was reduced to 130 nm in Pentium IV. Electrochemistry was largely used in chip fabrication [3] and the packaging of micro-electronics [4]. However, comparing with other nanofabrication techniques, electrochemical nanofabrication is still a maiden area which needs further development and fulfilment.

xii

Charles J. Dixon and Ollin W. Curtines

Chapter 1 summarized the most recent developments in electrochemical nanofabrications. It includes not only the conventional technique, under potential deposition (UPD), which deals with the deposition of a single metal-ion on a definite substrate but also some new developments using ultrashort voltage pulsing and template methods for 3D construction of nano-materials. Electrochemical nanofabrication is a versatile method, which includes both top-down nanofabrication (e.g. electrochemical lithography) and bottom-up process such as electrochemical atomic layer epitaxy (EC-ALE). Nano-templates including anodized aluminum oxide (AAO) membranes, colloidal polystyrene (PS) latex spheres, single/aligned carbon nanotubes, selfassembled monolayers (SAMs), blocked copolymers and cyclodextrin molecules can be used for the preparation of various types of nanowires, nanotubes, ordered arrays of nanoparticles and nanodots electrochemically. Combining electrochemistry with other nanofabrication techniques such as focused ion beam (FIB) and self-assembly provides many novel strategies in the fabrication of nanomaterials with specific design. Selective areas in the nanoscale can be modified by electrochemical nanostructuring with metals, metal oxides and conducting polymers using a bipolar electrochemical technique. The traditional lithography and pattern technique is costly. In the construction of soft matters such as conducting polymers, traditional spin casting cannot guarantee nanostructures due to the fast speed of solvent evaporation. Electrochemical technique provides an innovative, versatile and economic way of nanofabrication. It especially offers better alternative to construct the soft matter nano-structures in a controllable manner. In general, electrochemical nanofabrication offers simplicity, efficiency, low-temperature processing, cost-effectiveness, the possibility in preparing large area deposits and precise control of the deposit thickness, which are the essential advantages than other nanofabrication techniques till date. Additionally, it can be used to prepare a wide range of materials comprising the inorganic and the organic. The former includes quantum dots, metallic and semiconducting (e.g. ZnO, TiO2) nanotubes and nanorods. The latter includes conducting polymer nanotubes and nanowires. Chapter 2 reviews our recent progress on the novel two-dimensional nanowebs by the optimization of processing parameters during electrospinning. Using high applied voltage and low relative humidity in chamber, the by-product of micro-sized defect films can be splitted into nanowebs due to the fast phase separation of the charged droplets which flight with high moving speed in electric field from capillary tip to collector. The electrospun fibers act as a support for the “fishnet-like” nanowebs comprising interlinked one-dimensional nanowires. The average diameter of the nanowires contained in typical nanowebs is about one order of magnitude smaller than that of conventional electrospun fibers. Nanowebs together with common electrospun nanofibers can be assembled into a three-dimensional fibrous mat. So far, nylon-6, polyacrylic acid (PAA), poly(vinyl alcohol) (PVA)/SiO2 nanoparticles, and PVA/zinc acetate have been found to have the possibility forming nanowebs. The formation, morphology, and area density of the nanowebs in electrospun fibrous mats are strongly affected by the applied voltage, ambient relative humidity, kinds of solvents, solution concentration and conductivity, and distance between capillary tip to collector. The expanded applications of electrospun fibers are expected due to the formation of nanowebs, such as the nano-sized controllable filters, high efficient catalysts, catalyst supporter, and sensors. The preliminary data showing that the sensitivity of PAA nanowebs to ammonia is 2.5 times higher than that of electrospun PAA nanofibers. Additionally, PAA nanowebs show much

Preface

xiii

quicker absorption speed and larger capacities than that of PAA nanofibers during the ammonia absorption test. Strained silicon (ε-Si), the fundamental material of integrated circuit, is finding tremendous attention because it boosts the speed and reduces the power consumptions of electronic devices. However, poor homogeneity distribution of strain in ε-Si layers can degrade performance of electronic devices. Raman spectroscopy is used to study strain fluctuations in silicon because the optical phonons in Raman spectra are strongly influenced by strain. Though silicon are Raman active devices, the Raman efficiency of a nanometer layer of ε-Si is extremely weak and is often eclipsed under the Raman scattering of underlying buffer substrates. Micro Raman measurements show only uniform features in the nano-scale because of averaging effect from diffraction-limited spatial resolution. In Chapter 3, the authors utilized surface enhancement in Raman scattering to overcome weak emission problems and to suppress averaging effect. Thin ε-Si layers were covered with thin Ag layer to invoke surface enhanced Raman spectroscopy (SERS). Results show that SERS effectively enhanced the Raman signal from ε-Si layer and it stands distinctly apart from the Raman signal originating from the buffer layer. This technique is promising but it lacks the spatial resolution in the nano-scale due to diffraction limit from the probing light. In order to achieve nano-scale spectroscopy, point-surface-enhancement was used, rather than a large surface enhancement. The authors used a silver-coated sharp tip, just like SERS, but only the sample region very close to the tip apex is characterized. This technique, known as the tipenhanced Raman spectroscopy (TERS), provides nanometric resolution in our measurement. The authors observed localized strains by employing TERS. The TERS spectra revealed clear nano-scale variation in Raman frequency. Now that the authors can distinctively separate ε-Si from underlying buffer layer, signal-to-noise ration (SNR) needs further improvement. They improve TERS SNR in two ways: optical field enhancement using different metallic tip and background signals reduction arising from bulk materials. The tip-enhancement is more important for homogenous nano-materials or for samples with very weak signals whereas the background signal reduction is indispensable for nano-materials that consist of different thin layers with strong signals such as ε-Si or samples with strong signal level. Accordingly, the authors introduce several approaches mainly for the suppression of background signals arising from other bulk materials. The authors will discuss the utilization of UV light source, specialized tip, sample orientation relative to probing polarization, and depolarization configuration to obtain high contrast Raman signal. The characterization techniques describe above is applicable to other nano-materials. Cancer treatment usually uses drugs (chemotherapy) to reduce tumor size, followed by surgery to remove the tumor (if possible). Then, more chemotherapy and radiation therapy is used to kill as many tumor cells as possible. The goal of this collective treatment is to target and kill cancerous tissue while minimizing side effects on healthy cells. Due to their non specificity, current cancer therapies have poor therapeutic efficacy and can also have severe side effects on normal tissues and cells. In addition, cancer is often diagnosed and treated too late, i.e., when the cancer cells have already invaded and metastasized (i.e. spread) to other parts of the body. At this stage, treatment methods are highly limited in their effectiveness. Thus, scientists have been focusing efforts into finding alternative methods to detect cancer at earlier stages and kill such cancerous tissues more effectively. Nanoparticles (that is, particles with at least one dimension less than 100 nm) have become very attractive for improving

xiv

Charles J. Dixon and Ollin W. Curtines

cancer diagnosis and treatment due to their novel optical, magnetic and structural properties not available in conventional (or micron) particles or bulk solids. Nanoparticles have been extensively studied for various applications including delivering anti-cancer drugs to tumorous tissues and/or enhancing imaging capabilities to better diagnose and treat cancer. In Chapter 4, recent work related to the improved targeted therapy for specific cancers (whether by developing more specific anti-cancer agents or by altering delivery methods) are summarized. Discussions on the advantages and disadvantages of the most widely studied nanoparticles (i.e., liposome nanoparticles, polymer-based nanoparticles, quantum dots, nanoshells, and superparamagnetic particles) in cancer imaging followed by anti-cancer drug delivery are highlighted. Lastly, bone cancer and current research in using nanoparticles for treating bone cancer, with an emphasis on the novel use of selenium (a natural anti-cancer element found in our bodies), are addressed. The nanoindentation or indenter testing technique (ITT) is a functional and fast technique that can give us a lot of information about the mechanical properties of different materials at nanometric scale, from soft materials, such as copper, to brittle materials, such as ceramics. The principle of the technique is the evaluation of the response of a material to an applied load. In a composite material, if the size of the residual imprint resulting from a certain load is lower than the size of the studied phase, then is possible to determine its mechanical properties, and therefore its contribution to the global mechanical properties of the composite. Depending on the tipped indenter used, different equations should be applied to study the response of the material and calculate stress-strain curves and parameters such as hardness, Young’s modulus, toughness, yield strength and shear stress. These equations are related to the different deformation mechanisms (elastic, plastic or elastoplastic) that the material undergoes. In the case of most of the ceramic composites, when a spherical tipped nanoindenter is used, elastic deformation takes place, and Hertz equations can be used to calculate the yield stress, shear stress and the strain-stress curves. On the other hand, when a Berckovich indenter is used, plastic deformation takes place, then Oliver and Pahrr equations must be applied to evaluate the hardness, Young’s modulus and toughness. Nevertheless, in the hardness study, Indentation Size Effect (ISE) must be considered. In Chapter 5, the mechanical properties of a ceramic superconductor material have been studied. YBa2Cu3O7-δ (YBCO or Y-123) textured by Bridgman and Top Seeding Melt Growth (TSMG) techniques have been obtained and their mechanical properties studied by ITT. This material presents a phase transition from tetragonal to orthorhombic that promotes a change in its electrical properties, from insulating to superconductor, and that can be achieved by partially oxygenating the material. On the other hand, the structure of the textured material is heterogeneous, and two different phases are present: a Y-123 as a matrix and Y2BaCuO5 (Y-211) spherical inclusions. Moreover, the texture process induces an anisotropic structure, thus being the ab planes the ones that transport the superconductor properties while the c axis remains insulating. The purpose of this study is the characterization of the mechanical properties, in elastic and plastic range, of orthorhombic phases of YBCO samples textured by Bridgman and TSMG technique. With the ITT technique, the oxygenation process can be followed and its kinetics established. In Chapter 6, growth and physical properties of ZnO nanowire arrays were reviewed. It begins with some general remarks on semiconductor nanowires and basic properties of ZnO.

Preface

xv

In the second part, different kinds of growth methods that have been applied to grow ZnO nanowires are summarized. Vapor phase methods usually based on VSL or VS mechanism, depending on the presence or absence of a metal catalyst, were discussed in general. Typical solution methods for growth of ZnO nanowires were discussed separately as there is no common growth mechanism that can be applied to describe them. A new template-free strategy based on self-assembly process to grow ZnO nanowire into arrays were emphasized and discussed in detail. The obtained sample qualities were characterized with scanning electron microscopy, transmission electron microscopy, X-ray diffraction and energydispersive X-ray spectrum. The third part deals with the physical properties of ZnO nanowire arrays. Raman spectrum, including resonant Raman spectrum, was applied to test the crystal quality and phonon interaction of ZnO nanowires. Temperature-dependent photoluminescence spectra were measured to probe the intrinsic exciton and defect-related emission process of ZnO nanowires. Field emission properties of such ZnO nanowire arrays were also studied in view of the possible application for flat-panel displays. Some brief conclusions were summarized at the end. Organic materials (OMs) are diverse and are interesting in terms of application in electronic devices. In particular, organic charge transfer salts (OCTSs), which are typically composed of positive and negative ionic (and often radical) organic molecules, attract continued attention. Without carrier doping, they generally have high conductivity, magnetism, and well-defined unique nanostructures in their crystalline form. In order to apply the OCTSs to electronic devices, they should be made junction structures. Although there are established ways and advanced methods for doping and fabrication of junction structures in the current industrial techniques for the silicon devices, few of them are applicable to the OMs due to totally different chemical and physical properties between inorganic and organic materials. In Chapter 7, the authors would like to discuss a new method for simultaneous realization of doping and junction structures beginning with the single crystals of the OCTSs. The method utilizes a photo-induced chemical reaction, and produces a stable solid state composed of well-defined different parts of different conducting/magnetic properties. With reference to our recent and previous work as well as related studies of other groups, discussion will briefly cover experimental methods, preparation of materials, examination of irradiation conditions and resultant solids’ characterization, outline of mechanism of this photochemical modification, and remaining problems to be explained or overcome. Metal oxide nanowires exhibit novel properties due to their high surface-to-volume ratio and high surface stability. For this reason, they are considered excellent candidates to be incorporated into a new generation of devices with improved performance. Nevertheless, reaching complete control of their physical, chemical and electrical properties is needed before they can be widely used in our everyday life. This objective can be only fulfilled if reproducible electrical measurements on individual nanomaterials are performed. However, the fabrication of electrical nanocontacts in a fast and well-controlled process is still an unsolved issue. In Chapter 8, the main nanofabrication techniques that are commonly used to electrically access individual metal oxide nanowires, and to study their intrinsic properties are presented. Advantages and limitations of these methodologies are discussed in detail. By integrating bottom-up and top-down techniques, the first functional prototypes based on individual nanowires have already been implemented, paving the way to the future developments of nanoscale electronics, optoelectronics and chemical sensing devices.

xvi

Charles J. Dixon and Ollin W. Curtines

The inorganic-polymer hybrid nanomaterials have many excellent properties. So they are becoming increasingly important for various applications ranging from biomaterials to semiconductors in many fields and arouse much interest of scientists all over the world. Chapter 9 highlights the development of surface-initiated living radical polymerizations from the inorganic materials, including nanoparticles and one-dimensional (1D) nanostructures, by surface-initiated atom transfer radical polymerization (SI-ATRP). The emphasis is put upon the new developments of SI-ATRP taken to prepare hybrid nanomaterials in the recent years. Ferroelectric materials have been found to be promising candidates in applications of a wide range of electronic devices, such as high-dielectric constant capacitors, piezoelectric sonar or ultrasonic transducers, pyroelectric security sensors, medical diagnostic transducers, electro-optical light valves, and ultrasonic motors, and so on. Ferroelectric materials were conventionally fabricated via solid-state reactions at relatively and sometimes extremely high temperatures for calcining and sintering. Due to the presence of volatile components, such as lead (Pb), bismuth (Bi) or lithium (Li), in most ferroelectric compounds, high temperatures processing would brought out the problems of losing of the elements, which often resulted in the deteriorations in microstructures and thus electrical performances of the ferroelectric materials. To reduce the fabrication temperatures of ferroelectric ceramics, it is necessary to use ultrafine powders. High-energy mechanochemical technique, as an alternative method, has been used to synthesize nanosized ferroelectric powders directly from their oxide and other precursors. Chapter 10 serves as an overview of progress in the synthesis of various ferroelectric materials by using various mechanochemical milling facilities. In addition, applications of nanosized ferroelectric powders in materials preparation and device fabrication will be also be included. Linear forms of carbon are important in a wide variety of application, ranging from highly conducting interconnects to field emission materials. By methods of field ion microscopy (FIM) and mass-spectrometry, it was revealed the presence of linear carbon chains at the surface of carbon fibers after high-voltage treatment. The authors present in Chapter 11 a brief review of these research emphasizing recent developments. The carbon chains attached to the specimen tips can be produced in situ in a field ion microscope using low-temperature pulsed evaporation by electric fields of the order of 1011 V/m. Atomic Cchains are produced during the high-field unraveling of nanofibers. The experimental procedures used in FIM carbon chains studies are reviewed and the results in relation to the atomistics of unraveling processes are discussed. Molecular dynamics simulations and high resolution FIM experiments are performed to assess the evaporation of atomic chains under high-field conditions. Carbon exhibits a very rich dynamics of bond-breaking that allows transformation from graphenes to atomic chains. High-field experiments, theories leading to carbon chain formation, and methods to extract quantitative information on a variety of chainsurface interactions are described in detail. Isolated atomic carbon chains can be obtained at different temperatures, pulling speeds and forces. Current versus voltage field electron characteristics of monoatomic carbon wires were investigated. These results lend strong support to the conjecture of Smalley that linear carbon chains may provide the ultimate atomic-scale field emitters. Self-assembled nanostructures are new forms of materials which are interesting from a fundamental scientific perspective, as well as having many potential technological applications. As explained in Chapter 12, it is believed that the ability of nanostructures to self-assemble with controlled crystalline orientation, size, complexity and crystal

Preface

xvii

morphology, provide potential applications in data storage, functional devices, communications and technology. Recently, a two-step strategy was successfully developed in our lab to produce two-dimensional or three-dimensional carbon nitride well-defined hierarchical complex structures. This strategy is a combination of a novel laser-induced deposition technique followed by self-assembly. In the first step, a suspension of carbon nitride nanoparticles was prepared by liquid-phase pulsed laser ablation (LP-PLA). In the second stage, this suspension was deposited onto a silicon substrate to act as a ‘seed’ layer. Via controlling the rate of evaporation of the liquid phase part of the seed suspension, and the size and the quantity of nanocrystals within the droplet, it was possible to create a range of nanoscale structures, including dense nanospheres, highly-symmetric flowers, hollow coreshell and uniform grass-like structures. The growth of such complex structures is governed by an evaporation-driven self-assembly process. As the droplet dries, small building blocks, such as nanoparticles (NPs) or nanorods (NRs) nucleate upon the existing crystals and template, sharing the same edges, to form a close-packed arrangement. By varying the design of the building blocks, materials combination, interfacial chemistry, and confining dimensions, it is expected to extend this synthetic approach to a range of new structured materials with useful functional properties. Self-standing diamond films were fabricated by a 30 kW DC Arcjet CVD system. The novel progresses, including layer-structured film (nano-/micro-crystalline layer) fabrication, high orientated film deposition with high growth rate at very high ratio of CH4/H2, crack-free thick and large area films growth, and single crystal fabrication, were reported. Layer-structured self-standing films, 2- and 4-layered ones, were fabricated by fluctuating the ratio of methane to hydrogen with deposition time. Results of scan electronic microscopy (SEM) and Raman spectra showed that the layered films were constructed by the micro-crystalline grains layer / nano-crystalline grains layer. The residual stress within the films were balanced, and even diminished in the certain layer. The layer containing nanocrystalline grains due to a plenty of secondary nucleation could weakly inherit the columnar growth feature of the overlaid layer containing micro-crystalline grains. The grain size and growth orientation of the layer containing micro-crystalline grains could be adjusted by introduction a mid-layer containing nano-crystalline grains. Growth rate was over 10μm/hr in layered film fabrication. The effect of very high concentration of CH4 in H2, 10%≤CH4/H2≤ 25%, was studied on the film morphology and orientation. Diamond films with morphology containing nice faceted micro-sized grains were obtained with CH4/H2 up to 17%. The film composition change was found by Raman spectra. High (111)-oriented films were deposited under the condition of CH4/H2=15% at the maximum growth rate about 50μm/h. Deposition temperature could influence both the morphology and orientation of the diamond films. The higher deposition temperature, the higher CH4/H2 could be allowed to deposit micro-sized grain-containing films. However, high deposition temperature would spoil (111)-orientation. As a consequence, (220) and (311) would be enhanced. Crack patterns occurring in self-standing films were classified as network shape, river shape and circle shape. The distribution and style of dominating crystalline surface was found to influence the strength of self-standing film. The films with 60-120mm of diameter and 2mm of thickness were successfully deposited by controlling of dominating crystalline surface in the films.

xviii

Charles J. Dixon and Ollin W. Curtines

A new approach to single diamond crystal fabrication by arc jet was proposed and discussed in Chapter 13. This method was named as “stable-tip” method which was applied to overcome the morphology instability. Single crystal, 1×1×0.6mm3 in size, was successfully fabricated by this method. The synchrotron radiation topography was adopted to characterize this single crystal diamond. Low emission efficiency of silicon (Si) based light emission devices (LED) still blocks application of Si-LED. Therefore, studies focusing on improving light emission of Si-LED still attract researches’ passion. Recently, the authors have developed a new method combining nanosphere lithography and pulsed laser deposition to fabricate Si-based arrays nanostructures, and have obtained remarkably enhanced photoluminescence (PL) from these structures. The Si based nanostructures are hemisphere shell arrays (HSSAs) or nanoflower arrays assembled by silicon-germanium (SiGe) alloy. These structures include non-closepacked and close-packed ones, single layer and multilayer ones, as well as arrays on different substrates. In Chapter 14, the authors investigated the photoluminescence of these arrays structures, and found that all these structures could enhance the photoluminescence intensities. Among them, the enhancement of light emission from SiGe double layer HSSAs (DL-HSSAs), which is as high as 700 folds, is the highest among those of all structures. Employing transmission electron microscopy (TEM), scanning electron microscopy (SEM), time-resolved PL, and electromagnetic simulation etc, the authors found the enhancement of light emission in Si based nanostructures originated mainly from the increase of extraction efficiency of photons from the nanostructures. The electromagnetic simulation of enhancement matched well the experiment data. The authors also found that these enhancements are related to degree of order of arrays. In highly order arrays, the enhancement is higher than that in other arrays. Trivalent lanthanide (Ln3+) ion-doped semiconductor nanocrystals have attracted extensive attention due to the ability to tailor their optical properties via size control and to achieve highly efficient luminescence through sensitization by the host. To date, finding a way to dope the “undopable” Ln3+ ions into semiconductor nanomaterials via chemical methods remains a challenge. In Chapter 15, recent progress in the doping of Ln3+ ions in TiO2 nanomaterials has been reviewed. A novel sol-gel-solvothermal method has been developed to effectively incorporate Ln3+ ions (Eu3+, Er3+, Nd3+ and Sm3+) into anatase TiO2 nanoparticles via the self-assembly and crystallization process of previous amorphous nanoparticles, in spite of a large mismatch in ionic radius and charge imbalance between Ln3+ and Ti4+. The crystallization process of Ln3+ doped TiO2 nanoparticles were systematically studied by means of thermogravimetric-differential thermal analyses (TG-DTA), powder Xray diffraction (XRD), and transmission electron microscope (TEM). Photoluminescence (PL) spectra of Ln3+:TiO2 samples exhibit resolved and sharp emission and excitation lines from the intra f-f transitions of Ln3+ ions (Ln=Nd, Sm, Eu, Er), indicating regular crystalline surroundings of Ln3+ ions. Multiple sites of Eu3+, Sm3+ and Nd3+ ions in anatase TiO2 were detected by means of high- resolution site-selective spectroscopy at 10K, whereas only single site emission of Er3+ in TiO2 were observed. Very intense near-infrared luminescence around 1.53 µm was also observed, which originated from the single lattice site of Er3+ ions incorporated in TiO2 nanocrystals. The luminescence dynamics and CF levels of Ln3+ at different sites have been analyzed. Highly efficient emissions of Nd3+ and Sm3+ sensitized by the TiO2 host were observed upon the excitation above the TiO2 band gap energy at room

Preface

xix

temperature (RT), which is of particular interest for material applications. A growth mechanism for the incorporation of Ln3+ in the anatase lattice is also suggested. Titanium nitride (TiN) films synthesized by multi-arc ion plating (AIP) normally have a columnar microstructure, and are likely to induce surface defects due to the formation of macroparticles and neutral particles in the vicinity of cathode arc sources. Hence, the achievable microhardness of the normal AIP TiN films only ranges between 20~30 GPa. A systematic study for fabricating an adherent nano-superhard titanium nitride (TiN) film on M2 high speed steel substrate by a vacuum cathode multi-arc ion-plating (AIP) system was initiated. To understand the relationship of the film processing-structure-property, their microhardness, film-to-substrate adhesion, frictional property, and microstructure of the film were investigated using Vickers hardometer, scratch tester, ball-on-disc tester, X-ray diffractometer, and transmission electron microscope. Results in Chapter 16 show that: (i) the achievable film microhardness ranges between 35 GPa and 45 GPa; (ii) the critical load (Lc) of the superhard TiN film is at 64 N approximately; (iii) the friction coefficient, under a highload and a high rotating-speed, of the film is ranging from 0.5 to 0.8; and (iv) the nm scale mean main grain-sizes of the film are approximately 12.7 nm for TiN111, 19.7 nm for TiN200 and 9.6 nm for TiN220. The maximum achievable microhardness 45 GPa is more than twice of the 22 GPa for standard TiN film. Such hardness enhancement is anticipated as mainly due to: (a) the formation of nanoscaled crystalline grains; (b) the preferential orientation and growth of grains in the close-packed plane (111); and (c) the induced residual stress within the film by ion bombardment. In Chapter 17, nanoparticles embedded in insulators, e.g., Al2O3, MgO, YSZ and TiO2 single crystals, were fabricated by ion implantation and subsequent thermal annealing, including metallic Ni, Zn and their oxides, and intermetallic nanoparticles. Optical, magnetic and mircostructural properties of nanoparticles have been studied. The metallic nanoparticles have surface plasmon resonance absorption, and oxide nanoparticles show good photoluminenscence. The magnetic nanoparticles, e.g., metallic Ni and intermetallic CoxNi1-x nanoparticles, show strong ferromagnetism behaviors. The ion fluence can affect the concentrations and the intensities of the surface plasmon absorption of metallic nanoparticles. Ion flux is another important parameter to fabricate nanoparticles. An example of effects of ion flux on the nanoparticles has been presented in this data review. The relationship between annealing temperature and optical, magnetic and microstructural properties of nanoparticles has also been systematically studied. Ion implantation provides a versatile and powerful technique for synthesizing nanometerscale clusters embedded in the near-surface region of a variety of host materials. The embedded nanoparticles have attracted considerable attention because of their unique opticalelectrical properties that are different from those of the bulk matrix. Metallic nanoparticles embedded in insulators have pronounced optical effects, including surface plasma resonance (SPR) absorption, and strong third-order nonlinear optical (NLO) susceptibility. The former suggests applications as optical filters, including eye-glass coatings. The latter has potential application in all-optical-memory or switching devices. Oxide nanoparticles have good photoluminescence. They have promising application in light-emitting devices. Magnetic metallic nanoparticles often show a ferromagnetic behavior with a larger coercivity than that of the corresponding bulk materials, which may provide potential application of the nanocomposite as magneto-optical materials for a high density magnetic data storage device.

xx

Charles J. Dixon and Ollin W. Curtines

Organized self-assembly of molecules, driven by noncovalent intermolecular interactions, is the most versatile tool for accessing new materials with desired optical and electronic properties. Porphyrins are particularly attractive species to incorporate into supramolecular assemblies because their rich photochemistry may impart functionality, provide insight into the mechanisms of biological processes such as photosynthesis, serve as probes into the features of self-assembled structures and as models for molecular organization and energy/electron transfer processes. The close molecular packing in a self-assembled porphyrin aggregate leads to different electronic coupling and delocalization of the excitation energy, which can be exploited for applications in non-linear optical devices, photoelectric cells, recording devices. The possibility to control and tune either shape and size of the porphyrin clusters opens the way for their use as potential nanodevices. Chapter 18 aims to collect some recent developments in the field of porphyrin self-assembly and to frame all the reported topics into the current theories. The influence of thiophene addition on the pyrolysis of poly(dimethyl siloxane) catalyzed by ferrocene at ~1050 oC in Ar was studied in the first Short Communication. The assynthesized product was characterized by X-ray diffraction, scanning electron microscopy, transmission electron microscopy and high-resolution transmission electron microscopy. The thiophene addition caused several changes. Firstly, the yield of the product was increased by several times and the diameters of the product were somewhat increased. Secondly, the product was changed from only SiC/SiO2 nanocables to a mixture of SiC/SiO2 nanocables and SiC-SiO2 side-by-side nanowires. Thirdly, more “Y” type nanostructures were found. Finally, the growth process of the product was altered as the nanostructures each had a polyhedral FeS nanoparticle rather than spherical Fe nanoparticle. However, lengths of the product were still on the millimeter scale. The promotion mechanism of thiophene addition was also analyzed. As discussed the second short communication, nanotechnology revolutionized every field in science and technology. Recently, its usefulness in nanofinishing of cotton fabrics by imparting functional properties like antimicrobial, UV-resistance, self-cleaning and drugdelivery is well documented. In addition, enhancement in comfort properties of cotton textiles is also being evaluated with the help of nanofinishing. With judicial use of nanomaterials, keeping in view their bio-safety and environmental impact issues, nanofinishing will be a great boon to the users of cotton textiles.

RESEARCH AND REVIEW STUDIES

In: Nanotechnology… Editors: C.J. Dixon and O.W. Curtines, pp. 3-50

ISBN: 978-1-60692-162-3 © 2010 Nova Science Publishers, Inc.

Chapter 1

ELECTROCHEMICAL NANOFABRICATION Di Wei Nokia Research Centre, c/o Nanoscience Centre at University of Cambridge, 11 JJ Thomson Avenue, CB3 0FF, Cambridge, UK

Abstract Nano- and micro-fabrications have been largely used in the applications such as integrated circuits, micro/nano electro-mechanical systems (M/NEMS), micro-optics and countless others. The methodology of nanofabrication can be divided into two types, top-down and bottom-up processes, which themselves can be further divided. Top-down process refers to approaching the nanoscale from the top (or larger dimensions), such as lithography, nanoimprinting, scanning probe and E-beam technique etc.. In bottom-up fabrication processes, the nanotechnology process builds nanoscale artifacts from the molecular level up, through single molecules or collections of molecules that agglomerate or self-assemble. Using a bottom-up approach, such as self-assembly enables scientists to create larger and more complex systems from elementary subcomponents (e.g. atoms and molecules). In general, topdown processes that transfer minute patterns onto material are more matured than bottom-up processes. An exception is epitaxial processes that create layers through layer-by-layer growth with registry at the atomic level. Electrodeposition has actually been used for decades to form high quality, mostly metallic, thin films. It has recently been shown that high quality copper interconnects for ultra large scale integration chips can be formed electrochemically on Si wafer [1;2]. Electrodeposition has thus been shown compatible with state of the art semiconductor manufacturing technology. The largest semiconductor companies, for example, IBM, Intel, AMD, Motorola etc. are installing wafer-electroplating machines on their fabrication lines [1]. The electrodeposition of Cu with the line width 250 nm was used in the mass-production of micro-processor Pentium III in 1998. In 2003, the line width of the CPU was reduced to 130 nm in Pentium IV. Electrochemistry was largely used in chip fabrication [3] and the packaging of micro-electronics [4]. However, comparing with other nanofabrication techniques, electrochemical nanofabrication is still a maiden area which needs further development and fulfilment.

4

Di Wei This chapter summarized the most recent developments in electrochemical nanofabrications. It includes not only the conventional technique, under potential deposition (UPD), which deals with the deposition of a single metal-ion on a definite substrate but also some new developments using ultrashort voltage pulsing and template methods for 3D construction of nano-materials. Electrochemical nanofabrication is a versatile method, which includes both top-down nanofabrication (e.g. electrochemical lithography) and bottom-up process such as electrochemical atomic layer epitaxy (EC-ALE). Nano-templates including anodized aluminum oxide (AAO) membranes, colloidal polystyrene (PS) latex spheres, single/aligned carbon nanotubes, selfassembled monolayers (SAMs), blocked copolymers and cyclodextrin molecules can be used for the preparation of various types of nanowires, nanotubes, ordered arrays of nanoparticles and nanodots electrochemically. Combining electrochemistry with other nanofabrication techniques such as focused ion beam (FIB) and self-assembly provides many novel strategies in the fabrication of nanomaterials with specific design. Selective areas in the nanoscale can be modified by electrochemical nanostructuring with metals, metal oxides and conducting polymers using a bipolar electrochemical technique. The traditional lithography and pattern technique is costly. In the construction of soft matters such as conducting polymers, traditional spin casting cannot guarantee nanostructures due to the fast speed of solvent evaporation. Electrochemical technique provides an innovative, versatile and economic way of nanofabrication. It especially offers better alternative to construct the soft matter nano-structures in a controllable manner. In general, electrochemical nanofabrication offers simplicity, efficiency, low-temperature processing, cost-effectiveness, the possibility in preparing large area deposits and precise control of the deposit thickness, which are the essential advantages than other nanofabrication techniques till date. Additionally, it can be used to prepare a wide range of materials comprising the inorganic and the organic. The former includes quantum dots, metallic and semiconducting (e.g. ZnO, TiO2) nanotubes and nanorods. The latter includes conducting polymer nanotubes and nanowires.

Electrochemical Deposition Under Potential Deposition (UPD) Surface limited reactions are well known in electrochemistry and are generally referred to as under potential deposits [5;6]. Underpotential deposition (UPD) is the formation of an atomic layer of one element on a second element at a potential under, or prior to, that needed to deposit the element on itself. The shift in potential results from the free energy of the surface compound formation. ;Early UPD studies were carried out mostly on polycrystalline electrode surfaces [7]. This was due, at least in part, to the difficulty of preparing and maintaining single-crystal electrodes under well-defined (and controlled) conditions of surface structure and cleanliness [8]. For example, cadmium (Cd) can be underpotentially deposited on Cu(111) and Cu(100) [9]. There are a number of excellent reviews on this topic [5;6].

Electrochemical Atomic Layer Epitaxy (EC-ALE) Electrochemical atomic layer epitaxy (EC-ALE) is the combination of UPD and ALE, which uses UPD for the surface limited reactions in an ALE cycle [10]. Fundamental to forming high quality structures and devices with thin films of compound semiconductors is the concept of epitaxy. The definition of epitaxy is variable, but focuses on the formation of

Electrochemical Nanofabrication

5

single crystal films on single crystal substrates. This is different from other thin film deposition methods where polycrystalline or amorphous film deposits are formed even on single crystal substrates. Homoepitaxy is the formation of a compound on itself. Heteroepitaxy is the formation of a compound on a different compound or element, and is much more prevalent. The principle of ALE is to grow the deposit one atomic layer at a time [11]. Surface limited reactions are developed for the deposition of each component element in a cycle to directly form a compound via layer-by-layer growth, avoiding 3D nucleations. With cycle, a compound monolayer is formed, and the deposit thickness is controlled by the number of cycles. In an EC-ALE process, each reactant has its own solution and deposition potential, and there are generally rinse solutions as well. Control of growth at the nanoscale is a major frontier of materials science. The manipulation of a compound’s dimensions, or unit cell, at the nanoscale, can result in materials with unique properties. By constructing superlattices, nanowires and nanoclusters or forming nanocrystalline materials, the electronic structure (bandgap) of a semiconductor can be engineered. EC-ALE has been developed as an electrochemical methodology to grow compound semiconductors with nanoscale or atomic layer control. In an EC-ALE synthesis of CdTe, for example, atomic layers of tellurium and cadmium are alternatively electrodeposited to build up a thin layer of CdTe [12;13]. The necessary atomic level control over the electrodeposition of these two elements is obtained by depositing both elements using UPD. The thickness of the CdTe layer prepared by EC-ALE can be specified by controlling the number of Cd and Te layers that are deposited.

Figure 1. Transmission electron micrograph (TEM) of a CdTe deposit formed using 200 cycle of CdTe via EC-ALE [14]. Reproduced by the kind permission from the publisher.

6

Di Wei

Fig. 1 shows the TEM figure of the CdTe usng EC-ALE with 200 cycles [14]. The regular layered structure, parallel with the substrate Au lattice planes, suggests the epitaxial nature of the deposit. There are a number of ways to introduce dopants into an EC-ALE deposit and they can be introduced homogeneously throughout the deposit. Initial doping studies of ZnS were run with the idea of forming phosphor screens for flat panel display applications [15].

Electrochemical Deposition Methods for Semiconducting Nanocompounds Electrodeposition normally leads to small particle size, largely because it is a low temperature technique, thereby minimizing grain growth. It, however, possesses the additional feature of a very high degree of control over the amount of deposited material through Faraday’s law, which relates the amount of material deposited to the deposition charge. This feature is particularly desirable when isolated nanocrystals are to be deposited on a substrate. Several methods and variations have been developed to electrodeposit compounds. Oxides are probably the largest group of electrodeposited compounds (for example, aluminium anodization). The electrodeposition of II-VI compounds has been extensively studied and is well reviewed in a number of articles [16-18]. A number of reviews of semiconductor electrodeposition also exist which describe the various methods used [19;20]. The most prominent electrodeposition methods for semiconducting compounds include: codeposition, precipitation and various two-stage techniques. Semiconductor film can be electrodeposited either by EC-ALE or by co-deposition [21]. The most successful methodology to form II-VI compounds has been codeposition [22-26], where both elements are deposited at the same time from the same solution. Stoichiometry is maintained by having the more inactive element as the limiting reagent, and poising the potential where the less noble element will underpotentially deposit only on the more noble element. The classic example is CdTe formation [22], where the solution contains Cd2+ and HTeO2+, usually at pH 2. The potential is set to reduce HTeO2+ to Te on the surface at a limiting rate, while Cd2+ is reduced on the Te at an under potential, a potential where no bulk Cd is formed. Cd2+ ions are present in a large excess, to deposit quantitatively on Te as it is formed, resulting in stoichiometric CdTe. Although the structure and morphology of codeposited compounds are variable, some having been described as ‘cauliflower’ like, high quality deposits have been formed [27]. There are a number of papers in the literature concerning the formation of compound semiconductor diodes by electrodeposition, the most popular structure being a CdS-CdTe based photovoltaic. CdS was generally deposited first on an ITO/glass substrate, followed by a layer of CdTe, usually by codeposition [28-34]. Sailor and Martin et al. grew an array of CdSe-CdTe nano-diodes in 200 nm pore alumite [35-37], using a compound electrodeposition methodology called sequential monolayer electrodeposition [38]. A commercial process is being developed by BP Solar to form CdTe based photovoltaics using codeposition. Relatively rapid deposition rate has been achieved by codeposition and it is presently the most practical compound semiconductor electrodepostion methodology. Codeposition holds great promise if greater control can be achieved. At present the main parameters of control are solution composition and the deposition potential. There have

Electrochemical Nanofabrication

7

been a number of attempts to improve the process by using variations in reactant concentration, pH [39;40], and the potential program [38;41-45]. In most cases, the deposits are improved by annealing. In the application of photovoltaic applications, annealing is used to convert CdTe from the as-deposited n-type material to the desired p-type [34;46]. Semiconductors such as polycrystalline ITO on glass have been used to form deposits of ZnS with no obvious problems [15]. Ideally lattice matched semiconductor substrates could be used to form deposits. For instance, InSb is lattice matched with CdTe and could be used as a substrate. Good quality deposits of CdSe have been formed on InP and GaAs substrates using codeposition by Maurin and Froment et al. [47;48]. Their work clearly show the applicability of high quality commercial compound semiconductors wafers as substrates for compounds electrodeposition. The precipitation method involves electrochemical generation of a precursor to one of the constituent elements, in a solution containing precursors to the other elements [49-52]. The reaction is essentially homogeneous, but as one reactant is formed at the electrode surface, most of the product precipitates on the surface. This method resembles passive oxide film formation on reactive metals, where metal ions react with the solvent, oxygen or hydroxide. The film thickness is controlled by the amount of electrogenerated precursor. However, as the method resembles precipitation, the quality of the resulting deposit is questionable, and the process is difficult to control. Film thickness is necessarily limited by the need for precursor transport through the deposit. A classic example is the formation of CdS by oxidizing a Cd electrode in a sulphide solution [49;53-57]. Two stage methods are where thin films of the compound element, or an alloy, are first deposited, at least one by electrodeposition [58]. A second stage, annealing, then results in inter-diffusion and reaction of the elements to form the compound. The deposits are annealed in air, inert gas, or a gaseous precursor to one of the compound’s component elements. For instance, electrodeposited CuIn alloys have been annealed in H2S to form CuInS2 [59]. Given the need for annealing, this methodology has limitations for the formation of more involved device structures. In general, annealing has been used to either form or improve the structures of compound films formed by the electrodeposition methods described above. The primary tool for understanding compound electrodeposition and for improving control over the process has been the methodology of EC-ALE [60;61].

Electrochemical Synthesis of Quantum Dots Quantum dots are semiconductor particles having diameters that are smaller than about 10 nm. Such semiconductor nanoparticles exhibit a bandgap that depends on the particle diameter: the smaller nanoparticle, the larger the bandgap. Because quantum dots possess a ‘size-tunable’ bandgap, these diminutive particles have potential applications in detectors, light emitting diodes, electroluminescent devices, and lasers. Electrochemical methods can synthesize size-monodisperse quantum dots on graphite surfaces, which provide an electrical connection to the graphite in situ. The essential features of these methods can be depicted as in Fig. 2.

8

Di Wei

Figure 2. The electrochemical and chemical method to synthesize the quantum dots and other semiconductor nanocrystals on graphite.

The first step involves the electrodeposition of metal nanoparticles onto a graphite surface from a solution containing the corresponding metal ions. The metal nanoparticles are electrochemically oxidized to yield a metal oxide (MO), in which the oxidation state of the metal matches the oxidation state in the final product. Finally metal oxide nanoparticles are converted into nanoparticles of a semiconducting salt (MX) via a displacement reaction in which oxide or hydroxide is replaced by the desired anions (X). Examples of using these methods can be shown in the synthesis of CuI [62], CdS [63] and ZnO [64] quantum dots. Ultrathin films of quantum dots with deposits of non-connected nanocrystals and thick films of more than 10 nm in average thickness can be made by electrochemical methods [65-67].

Electrochemical Deposition Methods for Metallic Nanostructures Inorganic nanoparticles can be fabricated by many different techniques. Electrochemical and wet-chemical methods are demonstrated to be effective approaches to make metal nanostructures under control without addition of a reducing agent or protecting agent. An in situ electrochemical reduction method for fabricating metal nanoparticles on carbon substrates simultaneously assembling into ordered functional nanostructures was developed [68]. Ag+ was adsorbed on a highly oriented pyrolytic graphite surface modified by 4aminophenyl monolayer with coordination interaction, and then homogeneously dispersed Ag nanoparticles could be obtained through pulsed potentiostatic reduction. Multilayered metal nanostructures on glassy carbon electrodes have been obtained by extending this method [69;70]. The larger electrochemical window of ionic liquids in comparison of aqueous electrolytes enables the investigation of electrodeposition of metal and semi-conductor elements and compounds in nanoscale. Nanoscale electrocrystallization of metals such as Ni, Co and the electrodeposition of semiconductors (Ge) on Au (111) and Si (111):H have been studied in the underpotential and overpotential range from ionic liquids [71]. For example, 3D growth in Co electrodeposition on Au (111) from ionic liquids based on imidazolium cations starts at potentials below -0.17 V vs. Co/Co(II). 3D and 2D structures of Co and Ni deposition in the nanoscale were illustrated. In addition, nanocrystalline aluminium can be obtained by electrodeposition from ionic liquids containing imidazolium cations without additives[72].

Electrochemical Nanofabrication

9

The crystal refinement is due to a cathodic decomposition of the imidazolium ions to a certain extent giving rise to nanocrystalline aluminium. Metal nanowires can be obtained using solution phase reduction [73], template synthesis [74-77], and physical vapour deposition (PVD) [78] onto carbon nanotubes. Metal wires with widths down to 20 nm and lengths of millimetres can be prepared on silicon surface using electron beam lithography [79] or by PVD [80]. However, none of these methods are useful to prepare free-standing metal nanowires that are longer than 20 μm. Penner et al. [81] have used the step edge defects on single crystal surfaces as templates to form metal nanowires by electrochemical step edge decoration. Metallic molybdenum (Mo) wires with diameters ranging from 15 nm to 1 μm and lengths of up to 500 μm were prepared in a two-step procedure on freshly cleaved graphite surfaces [82]. Molybdenum oxide wires were electrodeposited selectively at step edges at -0.75 VSCE and then reduced in hydrogen gas at 500 °C to yield Mo metal. Such nanowires can be obtained size selectively because the mean wire diameter was directly proportional to the square root of the electrolysis time. Parrallel arrays of long (> 500 μm), dimensionally uniform nanowires composed of molybdenum, copper, nickel, gold, and palladium can also be electrodeposited by the same strategy [81]. They were firstly prepared by electrodepositing nanowires of a conductive metal oxide such as NiO, Cu2O or MoO2. Nanowires of the parent metal were then obtained by reducing the metal oxide nanowires in hydrogen at elevated temperature. Nanowires with diameters in the range from 15 nm to 750 nm were obtained by electrodeposition onto the step edges present on the surface of highly oriented pyrolytic graphite electrode. After embedding the nanowires in a polymer film, arrays of nanowires could be lifted off the graphite surface thereby facilitating the incorporation of these arrays in devices such as sensors. Vertical arrays of metal nanowire hold promise for making chemical and biological sensors in addition to electron emitters in field-emission displays. But the difficulty of growing well-defined arrays has kept these technologies at bay. Electrochemical nanofabrication using crystalline protein masks solved this problem [83]. A simple and robust method was developed to fabricate nanoarrays of metals and metal oxides over macroscopic substrates using the crystalline surface layer (S-layer) protein of deinococcus radiodurans as an electrodeposition mask. Substrates are coated by adsorption of the S-layer from a detergent-stabilized aqueous protein extract, producing insulating masks with 2-3 nm diameter solvent-accessible openings to the deposition substrate. The coating process can be controlled to achieve complete or fractional surface coverage. The general applicability of the technique was demonstrated by forming arrays of Cu2O, Ni, Pt, Pd, and Co exhibiting longrange order with the 18 nm hexagonal periodicity of the protein openings. This protein-based approach to electrochemical nanofabrication should permit the creation of a wide variety of two-dimensional inorganic structures.

Electrochemical Nanolithography In addition to its well-known capabilities in imaging and spectroscopy, scanning probe microscopy (SPM) has shown great potentials for patterning of material structures in nanoscales with precise control of the structure and location. Electrochemical nanolithography using SPM, which includes scanning tunnelling microscopy (STM) and atomic force microscope (AFM), has been used to fabricate of patterned metal structures

10

Di Wei

[84-87], semiconductors [87;88] and soft matters such as conducting polymers [89;90]. The electrochemical processing of material surfaces at nanoscale both laterally and vertically can be conducted by scanning probe anodization/cathodization, which used the tip-sample junction of a scanning probe microscope connected with an adsorbed water column as a minute electrochemical cell. A review on nanofabrication by scanning probe microscope lithography examines various applications of SPM in modification, deposition, removal, and manipulation of materials for nanoscale fabrication [91]. Comprehensive reviews of SPMrelated lithography can be found in the literature [92]. STM has a tremendous potential in metal deposition studies. The initial stages of metal deposition and the Ag adlayer on Au (111) have been studied by Kolb et al. [93].The inherent nature of the deposition process which is strongly influenced by the defect structure of the substrate, providing nucleation centres, requires imaging in real space for a detailed picture of the initial stage. This is possible with an STM, the atomic resolution helps to understand these processes on a truly atomistic level. The following figure demonstrates wealth of structural detail supplied by STM. In situ STM investigations for Ag UPD on Au (111) in the potential range from 600 to 200 mV vs. Ag/Ag+ were carried out [93]. Underpotentially deposited Ag revealed a series of ordered adlayer structures with an increasing density of adatoms when decreasing applied potentials. Nanostructuring can be also achieved by involving electrochemical processes into the overall procedures. For example, Li et al. [94;95] deposited Ag and Pt clusters on a graphite surface by applying positive voltage pulses to the STM tip in a solution containing the respective metal cations. This effect was attributed to nucleation within holes which were created on the graphite substrate surface by the voltage pulses. Kolb et al. [96], on the other hand, were able to detach Cu clusters from a tip, where they had been previously deposited electrochemically, onto a Au substrate by mechanical contact. However, all these techniques suffer from restrictions, which could be largely avoided if controlled nanostructuring could be achieved by a direct local electrochemical reaction on the substrate, with the geometry being determined by the location of the tip which acts as a local counter electrode. By applying ultrashort voltage pulses (≤ 100ns), holes of about 5 nm in diameter and 0.3 to 1 nm depth on Au substrate can be created by local anodic dissolution, while cathodic polarization led to the deposition of small Cu clusters [97]. The development of allowing the generation of small metal clusters, with the help of an STM tip, and placing them at will onto single crystal electrode surfaces was reported [96;98]. This so-called tip-induced metal deposition involves conventional electrolytic deposition of a metal onto the tip of an STM, followed by a controlled tip approach during which metal is transferred from the tip to the surface (so-called jump-to-contact [99]). Small copper clusters, typically two to four atomic layers in height, were precisely positioned on a Au(111) electrode by a process in which copper was first deposited onto the tip of the STM, which then acted as a reservoir from which copper could be transferred to the surface during an appropriate approach of the tip to the surface [100]. Tip approach and position were controlled externally by a microprocessor unit, allowing the fabrication of various patterns, cluster arrays, and ''conducting wires'' in a very flexible and convenient manner [96]. The formation of such clusters with the tip of a STM is simulated by atom dynamics and subsequently the stability of these clusters is investigated by Monte-Carlo simulations in a grand-canonical ensemble. It leads to the conclusion that optimal systems for nanostructuring are those where the metals participating have similar cohesive energies and

Electrochemical Nanofabrication

11

negative heats of alloy formation. In this respect, the system Cu-Pd(111) is predicted as a good candidate for the formation of stable clusters [101]. In addition to producing the metal nano-clusters, STMs can also be applied to form nanoscale pits in thin conducting films of thallium (III) oxide [102] as well as to write stable features on an atomically flat Au (111) surface [103]. Pit formation was only observed when the process was performed in humid ambient conditions. The mechanism involved in pit formation was attributed to localized electrochemical etching reactions beneath the STM tip. By applying voltage pulses (close to 3 V) across the tunnelling junction in controlled atmosphere with the presence of water or ethanol vapour, nano-hole can be produced. The smallest hole formed is 3 nm in diameter and 0.24 nm in depth. This nano-hole represents the loss of about 100 Au atoms in the top atomic layer of gold surface, there is no atomic perturbation seen inside and outside the nano-hole. Different nanostructures (lattice of dots, legends, map, etc.) can also be fabricated. The threshold voltage for the formation of a nanohole depends on the relative humidity, however, the relationship between the threshold voltage and the relative humidity is basically independent of the tip material. The application of conducting AFM probe anodization to nanolithography was also used in the fabrication and patterning of materials in a similar manner as STM probes. AFM-tipinduced and current-induced local oxidation of silicon and metals were reported and this novel local oxidation process can be used to generate thin oxide tunneling barriers of 10-50 nm.[88]. The direct modification of silicon and other semiconductor and metal surfaces by the process of anodization using the electric field from a SPM is one promising method of accomplishing direct-writing lithography for the electron device fabrication. This technique involves the application of an electrical bias to both the conducting probe and the sample substrate to locally oxidize selected regions of a sample surface. Since most of lithographic works with organic resists have been exclusively carried out on silicon surface for practical application, so several organic resists with different function groups have been studied in order to investigate the surface group effect on the anodic anodization using AFM. Silicon (Si) samples whose surfaces were terminated with hydrogen (Si-H samples) or covered with an organosilane monolayer were generally used to deposit metallic and/or semiconductor nanomaterials [86]. Silicon oxide (SiO2) patterns can be prepared on the Si-H sample surfaces by the use of anodization of Si, while in the second case the organosilane layer was selectively degraded by anodic corrosion. Furthermore, pattern transfer processes that fabricated metal nanostructures using these patterned SiO2 or organosilane layers as templates were developed. These processes are based on area-selective electroless plating where selectivity depends on the difference in the chemical reactivity between the surface modified by scanning probe anodization and the unmodified surface. Nanostructures down to a few ten nanometers in size have been fabricated with Langmuir–Blodgett (LB) films and self-assembled monolayers (SAMs) using SPM lithography [104]. The SAMs can be prepared with organosilane etc. as ultrathin resists on Si substrate. The effect of such functional groups of molecules on the AFM anodization, which was performed under contact mode has been studied in the optimized process conditions. Applied voltage between the AFM tip and sample, the scanning speed and the relative humidity in air are also important factors for nanometer-scale lithography of the ultrathin films. The high structural orderness and perfect thickness of ultrathin organic molecular assemblies are the major advantages as required for nanoscale lithography.

12

Di Wei

Cleaned p-type Si (100) wafer was firstly oxidized by piranha solution (3:1 mixture of sulfuric acid and 30% hydrogen peroxide) to make the surface hydrophilic. The octadecyldimethylmethoxysilane (ODMS) molecules react with OH groups on a silicon oxide surface resulting in the formation of a SAM in a thickness of 1.7 nm. The local degradation of the monolayer on selected area where the probe tip of the AFM was scanned with a bias voltage occurred due to the anodic reaction. The degraded regions became hydrophilic indicating that the ODMS molecules were decomposed and replaced with OH groups as a result of the probe scan. AFM images showed that the tip scanned regions were protruded compared to the surrounding regions as Fig.3 shows. This is due to volume expansion resulting from anodization of the Si substrate, which immediately followed the tip-induced degradation of the SAM.

Figure 3. Anodized pattern formation on Si-wafer. Si wafer was firstly oxidized to form hydrophilic oxide layer for the selfassembly of ODMS. ODMS SAMs act as the resist. Reproduced by the kind permission from the publisher.

Figure 4. The AFM image of ODMS monolayer after AFM anodization at the different applied currents [104]. Reproduced by the kind permission from the publisher.

Fig. 4 shows the AFM image of ODMS monolayer after AFM anodization at the different current conditions. Typically with ODMS monolayers, protruded patterns are fabricated with line-width of 70 nm at a scan rate of 120–200 μm/s and an applied voltage of 20–25 V. The

Electrochemical Nanofabrication

13

height of protruded lines can be controlled by changing the current between the tip and substrate. When the applied current was 5–7 nA, the height of protruded line was 1.0 nm. With increasing the current up to 20–23 nA, the anodized height increased to 3.0 nm. Thus, the height of the protruded lines increases due to the effective electric field strength for a given applied voltage and a scan rate. The electric field plays an important role in the formation of protrusion. AFM anodization was successfully carried out with SAMs on the silicon substrate. Applied voltage between the AFM tip and sample, the scanning speed, surface group, and the relative humidity in the laboratory are very important factors for nanometer-scale lithography of the ultrathin films. The high structural orderness and perfect thickness of ultrathin organic molecular assemblies are the major advantages as required for nanoscale SPM lithography. Soft matters such as conducting polymers can also be patterned in the nanoscale by such lithographic method. Cai et al. described the first observation of localized electropolymerization of pyrrole and aniline on highly oriented pyrolytic graphite (HOPG) substrates under AFM tip-sample interactions [89]. A scanning or oscillating AFM tip, providing the horizontal scratching force and the vertical tapping force, is essential as the driving force for the surface modification with the conducting polymer. It was shown that under the AFM tip interaction, the electropolymerization can be blocked on the bare HOPG substrate or enhanced on the as-polymerized film. The localized electropolymerization in selected surface areas enables the nanomodification of lines, square platforms, or hollows of polypyrrole and polyaniline on the substrates. The result indicates that AFM can be used as a unique tool for nanofabrication of conducting polymers. Nano-writing of intrinsically conducting polymer was also achieved via a novel electrochemical nanolithographic technique using tapping mode electrochemical AFM. Conducting polymer (polythiophene derivatives) nanolines as small as 58 nm in width were obtained and the line width is controlled as a function of the writing speed and writing potential [90].

Figure 5. Electrochemical nanolithography [90]. Reproduced by the kind permission from the publisher.

14

Di Wei

The electrochemical nanolithography process is shown in Fig. 5. Conductive AFM probes, gold coated silicon nitride (SiN4) are used as working electrode. Silver wire and platinum wire were used as reference electrode and counter electrode, respectively. Higher writing potential and slower writing speed produce wider conducting polymer nanolines due to enhanced propagation. The great benefit of this method lies in no specific restriction in the choice of substrates and the ease of controlling feature size, which is expected to facilitate to fabrication of all plastic nanoelectronic devices. As stated previously, many SPM lithography techniques based on anodization of Si surfaces [88], electrochemical reactions in solution using electrochemical STM tips [96;98] and electrochemical decomposition of self-assembled monolayers [104] have been developed in the past decade. More recently, a “dippen” nanolithography (DPN) method was invented that uses an AFM tip as a “nib” to directly deliver organic molecules onto suitable substrate surfaces, such as Au [105-107]. When AFM is used in air to image a surface, the narrow gap between the tip and surface behaves as a tiny capillary that condenses water from the air. This tiny water meniscus is actually an important factor that has limited the resolution of AFM in air. “Dip-pen” AFM lithography uses the water meniscus to transport organic molecules from tip to surface. By using this technique, organic monolayers can be directly written on the surface with no additional steps, and multiple inks can be used to write different molecules on the same surface. By coupling electrochemical techniques, the DPN are not limited to deliver organic molecules to the surface. Electrochemical “dip-pen” nanolithography (EC-DPN) technique can be used to directly fabricate metal and semiconductor nanostructures on surfaces.

Figure 6. Schematic sketch of the EC-DPN experimental setup [108]. Reproduced by the kind permission from the publisher.

The tiny water meniscus on the AFM tip was used as a nano-sized electrochemical cell, in which metal salts can be dissolved, reduced into metals electrochemically and deposit on the surface as shown in Fig. 6. In a typical experiment, an ultrasharp silicon cantilever coated with H2PtCl6 is scanned on a cleaned p-type Si (100) surface with a positive DC bias applied on the tip. During this lithographic process, H2PtCl6 dissolved in the water meniscus is electrochemically reduced from Pt(IV) to Pt(0) metal at the cathodic silicon surface and deposits as Pt nano-features as shown in the result below [108].

Electrochemical Nanofabrication

15

Figure 7. AFM image and height profile of two Pt lines drawn at different scan speed. a) line at 10 nm/s and b) line at 20 nm/s. The voltage applied at the tip is 3V for both lines and the relative humidity is 43% [108]. Reproduced by the kind permission from the publisher.

Electrochemical AFM "dip-pen'' nanolithography has significantly expanded the scope where DPN nanofabrication can be applied. It combines the versatility of electrochemistry with the simplicity and power of the DPN method to produce nanostructures with high resolution. Electrochemical STM-based methods require that the substrates be metallic, but substrates used in EC-DPN do not have to be metallic since the control feedback of the AFM does not rely on the current between the tip and surface. Si wafers coated with native oxide provides enough conductivity for the reduction of the precursor ions. This development significantly expands the scope of DPN lithography, making it a more general nanofabrication technique that not only can be used to deliver organic molecules to surfaces but is also capable of fabricating metallic and semiconducting structures with precise control over location and geometry. Local electrochemical deposition of freestanding vertically grown platinum nanowires was demonstrated with a similar approach, electrochemical fountain pen nanofabrication (ECFPN) [109]. The EC-FPN exploits the meniscus formed between an electrolyte-filled nanopipette ('the fountain pen') and a conductive substrate to serve as a confined electrochemical cell for reducing and depositing metal ions. Freestanding Pt nanowires were continuously grown off the substrate by moving the nanopipette away from the substrate while maintaining a stable meniscus between the nanopipette and the nanowire growth front. High quality and high aspect-ratio polycrystalline Pt nanowires with diameter of similar to 150 nm and length over 30 μm were locally grown with EC-FPN. The EC-FPN technique is shown to be an efficient and clean technique for localized fabrication of a variety of vertically grown metal nanowires and can potentially be used for fabricating freeform 3D nanostructures.

16

Di Wei

3D Electrochemical Nanoconstruction In combination with lithographic patterns, electrochemistry has taken a key position in products and manufacturing processes of microtechnology, which has established a multibillion dollar market with applications in information, entertainment, medical, automotive, telecom and many other technologies such as lab on a chip etc. [110]. Different strategies and techniques such as electrochemical etching, LIGA and ultrashort voltage pulsing, which have been used in microtechnology, were also applied to construct 3D structures in the nanoscale.

Electrochemical Etching and LIGA Technique Electrochemical etching with ultra-short voltage pulses allows to dissolve electrochemically active materials within an extremely narrow volume and to manufacture three dimentional (3D) microstructures. Micro- and nanoporous silicon can be generated by anodization of silicon wafers in hydrofluoric acid. Since etching proceeds preferably in the (100) direction of the single-crystalline silicon wafer, the pore shape is nearly straight and the depth is equal for all pores. The pores start to grow on a polished wafer in a random pattern, and their arrangement is usually defined by transferring a suitable lithographic pattern and generating a corresponding pattern of pits by alkaline etching as shows in Fig. 8. The pattern of the deep pores generated subsequently by the electrochemical etching process corresponds to the pattern of the shallow starter pits. The cross section of the pores is usually square with rounded corners and their size can be varied by changing the etching current. More complex cross sections can be generated by overlapping of pores using additional etching steps and corresponding pattern definition (e.g. by means of lithography).

Figure 8. Macroporous silicon structure generated by means of electrochemical etching of singlecrystalline silicon (Source: V. Lehmann, Siemens AG) [110]. Reproduced by the kind permission from the publisher.

Electrochemical Nanofabrication

17

The macroporous silicon structure generated by means of electrochemical etching of single-crystalline silicon was demonstrated above. In the electrochemical set-up, hydrofluoric acid (HF), with or without ethanol and/or water is used as electrolyte and platinum is the standard cathode. The etch rate of ca. 1 μm/min is observed. Both electro-polishing and pore formation take place in the anodic regime. Depending on current density, silicon can be etched in such two different modes: pore formation and electro-polishing. In pore formation, etching proceeds vertically downwards, leaving a silicon ‘skeleton’ with up to 80% empty space, whereas in electropolishing, the whole surface is being etched. Pore formation starts at the wafer surface from a defect or an intentional initial pit. Electronic holes from the bulk silicon are transported to the surface, and they react at the defect or pit. Further etching occurs at the newly formed pore tips, because they attract more holes due to higher electric field strength, and the process leads to a uniform porous layer depth as the holes are consumed by the growing tips and other surfaces are depleted of holes. This etching mode takes place under low hole concentration and it is limited by hole diffusion, and not by mass transfer in the electrolyte cell. If hole density increases, some holes reach the surface and react there, leading to surface smoothing. This is the electro-polishing regime, in which ionic transfer from the electrolyte plays a role. Illumination contributes to hole concentration in n-type silicon (but not in p-type silicon) and the anodic etching of n-type silicon happens under illumination whilst p-type silicon etches in dark. A very wide range of pore size from 0 to 20 μm can be etched by varying electrolyte concentration, current density and illumination [111]. As a rule of thumb, pore diameter in micrometre is half the resistivity in ohm/cm: for 1 um pore, 2 ohm/cm n-type silicon is suitable. For small pores, low resistivity is needed; for large pores, high resistivity material has to be used. If pore formation starts from an unobstructed surface, a random pore array results. If initial pits are prepared by lithography and etching, pores can be arranged at will [112]. When an n-type silicon wafer is an anode in an alkaline etching solution (e.g. KOH) biased positively above passivation potential, the surface will be oxidized, which stops silicon dissolution whilst p-type silicon was etched. The n-type layer of a p/n-structure can similarly be protected. Etching of p-type silicon continues until the diode is destroyed, and n-type silicon is then passivated. The confined etchant layer technique has been applied to achieve effective three-dimensional (3D) micromachining on n-GaAs and pSi. This technique operates via an indirect electrochemical process and is a maskless, lowcost technique for microfabrication of arbitrary 3D structures in a single step [113]. It has also been presented that free-standing Si quantum wire arrays can be fabricated without the use of epitaxial deposition or lithography by electrochemical and chemical dissolution of wafers [114]. This novel approach uses electrochemical and chemical dissolution steps to define networks of isolated wires out of bulk wafers. Electrochemical methods, either alone or in combination with other techniques, have been developed for shaping materials. 3D microstructures with extremely high precision and aspects ratio can be manufactured by means of LIGA technology, which combines deep lithography, electrodeposition and moulding process steps. The acronym LIGA is derived from the German expressions for these manufacturing steps and offer high potential regarding miniaturization, freedom of design and mass production. Micro-gear system produced from a nickel iron alloy by means of LIGA was shown in Fig. 9.

18

Di Wei

Figure 9. Micro-gear system produced from a nickel iron alloy by means of LIGA technology (Source: Micromotion, IMM) [110]. Reproduced by the kind permission from the publisher.

However, the ultra precise microstructures with extreme aspect ratio could only be generated by deep X-ray lithography. Difficulties, for example, the access to synchrotron radiation facilities have limited the commercialization of LIGA technique in mass fabrication.

Micro- and Nano-machining by Ultrashort Voltage Pulsing Technique The application of electrochemistry in micro-machining can be found in book [115]. In contrast to the conventional processes of electrochemical micromachining, where the gap between the electrodes is usually 0.1 mm and direct current is applied, a novel electrochemical micofabrication method using a gap in ‘μm’ range and very short voltage pulses of some tens of nanoseconds was developed. The short pulse confines electrochemical processes correspondingly, removal of material to a very narrow volume to enable a precise nanomachining. Ultrashort pulses can be employed to machine conducting materials with lithographic precision [116]. Resolution can be improved significantly through the use of ultrashort voltage pulses comparing to the use of conventional direct current anodization. Three dimensional complex nanostructures, lines, curved features, and arrays can be machined in substrates in single-step processing. The method is based on the application of ultrashort voltage pulses of nanosecond duration, which leads to the spatial confinement of electrochemical reactions, e.g. dissolution of material. The electrochemical dissolution rate of the material has to be intentionally varied over the workpiece surface by applying inhomogeneous current density distribution in the electrolyte and at the workpiece surface. This can be achieved by the geometric shape of the tool, locally very small tool-workpiece distances, partial insulation of the tool or workpiece, and high overall current densities etc. This situation is illustrated in Fig 10. The workpiece is preferentially etched within the gap region between the front face of the tool and the workpiece surface. This approach for local confinement of electrochemical reactions is based on the local charging of the electrochemical double layer (DL) and the resulting direct influence on the electrochemical reaction rates.

Electrochemical Nanofabrication

19

Figure 10. Sketch of the experimental setup and principle of electrochemical micromachining with ultrashort pulses. RE and CE are abbreviates of reference and counter electrode.

The potentials of the workpiece and tool are controlled by the low-frequency bipotentiostat. The voltage pulses are supplied by the high frequency pulse generator. An ultrashort voltage pulse limits the charging of double layer capacitance to the vicinity of the tool. The current distribution between the DL is also illustrated in Fig. 10. The pulsing time constant is given by the DL capacitance multiplied by the resistance of the electrolyte along the current path. The latter factor is locally varying, depending on the local separation of the electrode surfaces [116]. Therefore, upon proper choice of the pulse duration, DL areas where the tool and workpiece electrodes are in close proximity are strongly charged by the voltage pulses, whereas at further distances the charging becomes progressively weaker. The pulse duration provides a direct control for the setting the machining accuracy. Machining precisions below 100 nm were achieved by the application of 500 ps voltage pulses [117].

Figure 11. Spiral trough with a depth of 5 mm, machined into a Ni sheet with a W tool in 0.2 M HCl (3 ns, 2 V pulses, 33 MHz repetition rate, Φworkpiece≈ -0.1 VAg/AgCl, Φtool≈ -0.3 VAg/AgCl [117]. Reproduced by the kind permission from the publisher.

20

Di Wei

The application of ultrashort voltage pulses between a tool electrode and a workpiece in an electrochemical environment allows the three-dimensional machining of conducting materials with nanoscale precision. The principle is based on the finite time constant for DL charging, which varies linearly with the local separation between the electrodes. During nanosecond pulses, the electrochemical reactions are confined to electrode regions in close proximity. The performance of electrochemical micromachining with ultra-short voltage pulses was demonstrated in a number of experiments where microstructures were manufactured from various materials like copper, silicon and stainless steels [116]. Three dimentional structures with high aspect ratios can be achieved by using suitable microelectrodes and piezo-driven micropositioning stages. The spiral shown in Fig. 11 was manufactured by machining the Ni sheet with 3 ns pulses. Walls of similar thickness with surface roughness and radii of curvature less than 100 nm were readily machined [117].

Figure 12. Scanning electron micrographs. (a) the tool; (b) structure in Ni substrate. Experimental conditions: Usub=-0.35 V, Utool=-0.3 V, 2 ns pulse duration, 2.2 V amplitude, 1:10 pulse to pause ratio, and 0.05 M HCl electrolyte. The structure was machined 400 nm into the surface in less than 2 min [118]. Reproduced by the kind permission from the publisher.

Small tools can be used to make very small features. High aspect ratio nanometre accurate features were machined in nickel using ultrashort voltage pulse electrochemical machining [118]. Two tools (one is the rotunda tool, presented in Fig 12a and other is 2×2 array of cubes, presented in Fig. 13a) were firstly fabricated by focused ion beam (FIB) milling and then used in the machining. The potentials of the shaped tungsten tool and nickel substrate electrodes were controlled with a bipotentiosts which kept the potentials of the tool and substrate constant versus an Ag/AgCl reference electrode by applying a potential to Pt counter electrode. All experiments were conducted in an aqueous HCl solution with various concentrations. Supplementary circuitry was present to allow the additional of ultrashort (order of 1 ns) pulses to the potential of the tool electrode. The separation of the tool and the substrate workpiece electrodes was controlled with piezoactuators and the tool was fed into the workpiece with a constant feeding speed, avoiding mechanical contact between the electrodes by monitoring the dc current between tool and workpiece. Structures with 90 nm widths were made by applying 2ns voltage pulses for the parallel lines in the centre of the structure in Fig 12b. To reach a depth of 400 nm, total electrochemical machining time of 1

Electrochemical Nanofabrication

21

min 45s was applied. Examples of patterns made with the 2×2 array of cubes tool are shown in Fig. 13b. It indicates that the feature resolution improves with decrease in pulse duration.

Figure 13. Scanning electron micrographs. (a) Tungsten tool; (b) machined Ni substrate. Experimental conditions: Usub=-1.7 V, Utool=-1.0 V, pulse duration indicated, 2 V amplitude, 1:10 pulse to pause ratio, and 0.2 M HCl electrolyte. Feature resolution and edge sharpness increased as pulse duration decreased [118]. Reproduced by the kind permission from the publisher.

Three-dimensional machining of electrochemically active materials including construction of unconventional island patterns on a surface with nanoscale resolution was realized by this method [97;119-121]. Thus, electrochemical machining can be applied to microelectro-mechanical systems (MEMS) [122], and even in the nanoelectro-mechanical systems (NEMS). Electrochemical methods can realize the nanofabrication in a selective place and make the complicated 3D nanostructures. Conducting polymers can also be made in this way. Similar to the electrochemical machining, by application of short voltage pulses to the tool electrode in the vicinity of the workpiece electrode, the electropolymerization of pyrrole can be locally confined with micrometre precision [123]. As the produced nuclei of conducting polymers will grow preferentially vertically to the surface, fibre-like morphologies were found in the local polypyrrole electrosynthesis with short voltage pulses. A polypyrrole ring with needle-like feature can be selectively nanofabricated, which was grown with 1 μs pulses in 0.2 M H2SO4/0.2M Pyrrole. To obtain such structure, a 50 μm diameter flattened cylindrical Pt wire tool was brought to a distance of 1 μm from the workpiece substrate before applying the pulses. The small distance between tool and the substrate strongly hinders the supply of monomer at the reaction site. The consequent relatively low supersaturation leads to a low density of nuclei, which grow and form isolated polypyrrole fibres. With the increase of the distance, a more compact ring-like structure can be obtained.

22

Di Wei

Template Free Methods for Conducting Polymer Nano-architecture Although oriented carbon nanotubes, oriented metal and semiconductor nanowires have attracted wide attention, there have been few reports on oriented polymer nanostructures such as nanowires. The assembly of large arrays of oriented nanowires containing molecularly aligned conducting polymers without using a porous membrane template to support the polymer was reported recently [124]. The uniform oriented nanowires were prepared through controlled nucleation and growth during a stepwise electrochemical galvanostastic deposition process, in which a large number of nuclei were first deposited on the substrate using a large current density. After the initial nucleation, the current density was reduced stepwise in order to grow the oriented nanowires from the nucleation sites created in the first step. The usefulness of these new polymer structures is demonstrated with a chemical sensor device for H2O2, the detection of which is widely investigated for biosensors. It offers a general approach to control nucleation and and has potential for growing oriented nanostructures of other materials. Fig. 14 demonstrates the steps to electrochemically fabricate arrays of oriented conducting polyaniline (PANI) nanowires by well controlled nucleation and growth without templates. After initiating the nucleation of the conducting polymer at high current densities the current density is reduced to avoid formation of further nuclei. The existing nuclei grew preferentially vertically to the surface. A typical procedure involves electrochemical deposition in an aniline-containing electrolyte solution, by using the substrate as the working electrode. This process involves: 0.08 mAcm-2 for 0.5 h, followed by 0.04 mAcm-2 for 3 h, which was then followed by another 3 h at 0.02 mAcm-2. The stepwise growth produced uniform, oriented nanowires on a variety of flat and rough surfaces as Fig. 15 shows.

Figure 14. Schematic drawing of the steps for growing oriented polymer nanowires. a) Schematics of the reactions in the electrochemical cell. b) Schematics of the nucleation and growth.[124]. Reproduced by the kind permission from the publisher.

Electrochemical Nanofabrication

23

Figure 15. SEM micrographs of oriented polyaniline on Pt. a) Low magnification face-on. b) High magnification face-on. c) Tilted view, low magnification. d) Tilted view, high magnification. The insert in Figure 3 a is the image of the oriented nanowires on Si substrate[124]. Reproduced by the kind permission from the publisher.

The PANI films prepared by the above method appear to be fairly uniform with the diameters of the tips ranging from 50 nm to 70 nm. The direct electrochemical synthesis of large arrays of uniform and oriented nanowires of conducting polymers with a diameter much smaller than 100 nm, on a variety of substrates (Pt, Si, Au, carbon, silica), without using a supporting template was thoroughly studied [125]. Ordered PANI nanowires tailored by such stepwise electrochemical depositions showed remarkably enhanced capacitance [126]. The superior capacitive behaviors of PANI nanowires show great potential in the application of supercapacitors and rechargeable batteries. Conducting polymer nanowires are also promising one-dimensional nanostructured materials for application in nanoelectronic devices and sensors [127;128] due to their light weights, large surface areas, chemical specificies, easy processing with low costs and adjustable transport properties. The electrochemical growth of nanowire devices using e-beam-patterned electrolyte channels potentially enables the controlled fabrication of individually addressable sensor arrays [128]. This approach should be highly efficient and scalable, while meeting the current requirements for nanoelectronics technologies, i.e., an integration of bottom up production methods (electropolymerization of the nanoframeworks) and top-down fabrication (lithographic fabrication of Pt electrodes in array). The same galvanostastic template free, site specific electrochemical method was developed to precisely fabricate individual and addressable conducting polymer nanowires on electrode junctions in a parallel-oriented array [129]. Electrochemical polymerization at low and constant current levels was used to fabricate 10 nano-framework-electrode junctions simultaneously with uniform diameter (ca. 40 to 80 nm) PANI nanowires interwined to nanoframeworks. Electropolymerization was carried in an aqueous solution containing 0.5 M

24

Di Wei

aniline and 1.0 M HCl. Firstly, a constant current (50 nA) was applied for ca. 30 min to introduce the PANI nuclei onto the Pt working junction electrodes. The current was then reduced to 25 nA for 180 min. The PANI nanoframeworks begin to propagate from the working junction electrodes to the other set of the junction electrodes. In the last step, the current was decreased to 12 nA for 180 min, and the 10 polyaniline nano-frameworkelectrode junctions were obtained simultaneously with each PANI nanoframe work positioned precisely within the 2 μm gap between its electrodes. All these nanoframe work electrode junctions can be covered by such conducting polymer wires simultaneously and sitespecifically in a parallel fashion. This device can be used as miniaturated resistive sensors for real-time detection of NH3 and HCl gases. Such two-terminal devices can be easily converted to three-terminal transistors by simply immersing it into an electrolyte solution along with a gate electrode. Electrolyte-gated transistors based on conducting polymer nanowires junction arrays was developed and the field induced modulation can be applied for signal amplification to enhance the device performance [130]. Conducting polymer nanowires including PANI (ca. 50-80 nm), PPy (ca. 60-120 nm) and PEDOT(ca. 80-150 nm) were introduced to the 10 paralleled 2μm-wide gaps by the template free electropolymerization method. The preparation of electrolyte gated transistors can be completed simply immersing the conducting polymer nanowire-based two-terminal resistive device along with a gate electrode (a Pt wire or Ag/AgCl electrode) into a buffered electrolyte solutions containing NaCl. P-channel transistor characteristics at pH7 and n-type behaviour in basic media were observed for both PANI and PPy nanowires. Whereas PEDOT nanowire based device only exhibit depletion mode behaviors in neutral solutions. These open new opportunities to fabricate sensor arrays with conducting polymer nanowires to realize the ultrasensitive, realtime and parallel detection of analytes in solutions.

Template Methods The growth of thin films displaying special features like aligned pores perpendicularly to the substrate surface and nano-porous structures have attracted the attention of many research groups in the last decade and, with that aim several techniques such as ion beam lithography have been used [131]. On the other hand, a lot of bottom up techniques, particularly those in which, self-assembly processes play a relevant role in the growth mechanisms of that nanostructures have been reported. Among them, electrochemical techniques constitute one of the most used to fabricate highly ordered nanostructures to be used as templates for replicating other nanostructured materials and for growing functionalized material arrays. An overview on the nanofabrication techniques is done mainly of those related with the nanostructures fabrication based on ordered and nanoporous anodic aluminium oxide membranes (AAO), anodic titania membranes, colloidal polystyrene (PS) latex spheres. Templates from carbon nanotubes, self-assembled monolayers (SAMs) and selfassembled block copolymers etc. are also summarized. Template methods offer very important strategies for complicated nano-structures.

Electrochemical Nanofabrication

25

Anodized Aluminum Oxide (AAO) Membranes A simple and completely nonlithographic preparation technique for free-standing nanostructured films with a close-packed hexagonal array of nanoembossments has been developed by porous anodic aluminium oxide (AAO) membrane templates with different pore diameters. They have been largely used to construct different free-standing inorganic and organic nanowires [132-134], nanotubes [135] and ordered arrays of nanoparticles [136] since invented. Aluminium anodization provides a simple and inexpensive way to obtain nanoporous templates with uniform and controllable pore diameters and periods over a wide range. The usual electrochemical method for producing the AAO film is the anodization of high purity Al plates at constant voltage (e.g. anodized at 22 V from an Al foil and detached by the reverse-bias method) [137]. Membranes with several different pore sizes can be made, for example, in the following electrolytes: Aqueous solutions of H2SO4 at 10-20 V for pores ~1025 nm, H2C2O4 at 40-80 V for pores ~40-100 nm, and H3PO4 at 100-140 V for pores ~100170 nm. The pore diameter is linearly related to the anodizing voltage (1.2 nm/V). A voltage reduction was done to thin the barrier layer that inhibits anodic current during electrodeposition [138]. Other attempts have been made to create nano-porous symmetries other than hexagonal packing [139]. Recently, a novel AAO membrane with a six-membered ring symmetry co-existing with the usual hexagonal structure has been fabricated by constant current anodization [140].The pore sizes of this structure can be tailored by changing the processing conditions. Ordered arrays of nanodots with novel structure have been fabricated by this AAO template. In the final stage, the porous alumina substrate can be removed by etching in KOH. Moreover, one of the interesting possibilities afforded by the anodization process is that the anodization can take place on arbitrary surfaces, such as curved surfaces. Unique features including cessation, bending, and branching of pore channels are observed when fabricating AAO templates on curved surfaces [141]. The new structures may open new opportunities in optical, electronic and electrochemical applications. Many strategies have been ingeniously implemented to fabricate complicated nanostructures based on the AAO templates. For example, hexagonally ordered Ni nanocones have been fabricated using an a porous AAO template where the pores are of a cone shape [142].The conical AAO film was found to exhibit hexagonal order with a period of 100 nm. The Ni nanocones and the surface morphology of the nano-conical film exhibit the same periodic structure of the template as shown in Fig. 16. The hexagonally ordered Ni nanocones and nano-conical film were produced using anodization and metal plating techniques. The conical AAO template was produced using a process of repeated applications of the anodization and pore-widening steps, applying the two steps alternately. The Ni nanocones were produced by electroless Ni deposition onto the conical AAO template, with the pores filled with Ni particles. The resulting Ni nanocones exhibit the same ordered structure. The Ni nano-conical film was produced by detaching the deposited Ni layer, with the surface morphology also a hexagonally ordered array with a period of 100 nm. These nanostructures were produced using the wet-processes of anodization, a pore widening treatment, the pulsed deposition of Pd particles and electrochemical Ni deposition. Clearly, this complete process can produce good quality results. A double-templating approach using simple electrochemical methods to create aligned arrays of nanotubes on substrates was also introduced [143].

26

Di Wei

Figure 16. (A) Surface normal, (B) surface angled and (C) cross-section view of the hexagonally ordered conical Ni film [142]. Reproduced by the kind permission from the publisher.

Figure 17. Schematic of method to fabricate nanotube arrays on substrates. Nanorods are electrodeposited into nanoporous anodic alumina template films, the alumina is removed, and then the nanorods are used as secondary templates for nanotube electrodeposition [143]. Reproduced by the kind permission from the publisher.

Electrochemical Nanofabrication

27

The method used to fabricate nanotube arrays is shown schematically in Fig. 17. Initially, nanorod arrays are fabricated by electrodeposition into AAO templates. By varying the anodization conditions, the pore diameter, spacing, and height can be tuned, and the pore ordering can also be controlled. Nickel nanorods are then electrodeposited into the pores of the alumina. After deposition, the exposed ends of the nanorods are modified by anodization in a dilute KOH solution. Because the anodization is performed when the alumina is still in place, only the top ends of the nanorods are anodized. Then the alumina template is removed by a selective chemical etch, leaving an array of nickel nanorods with anodized tips. This nanorod array is then used for electrodeposition of gold nanotube arrays. The nanotube material deposits uniformly across the entire surface of the nanorod arrays, except at the anodized tips of the nanorods. Finally, the nickel nanorod array template is selectively removed, resulting in an array of open-ended nanotubes on the substrate. This approach allows both fabricate and organize nanostructures over large areas on substrates in the same process. This method is demonstrated to prepare arrays of electrodeposited, open-ended nanotubes aligned vertically on substrates. The nanotube inner diameter, spacing, and height are determined by the nanorod dimensions, which in turn correspond to the alumina film characteristics. The nanotube morphology and wall thickness can be controlled by the electrodeposition parameters. Utilizing electrochemically prepared textured aluminum sheets as a replication master in conjunction with electrochemical deposition of metals revealed a highly facile and economical way for the production of periodic metallic nanostructures in a large area with high fidelity in pattern transfer as well as with a good degree of flexibility in materials [144]. The growth of metal nanowires using AAO membranes as hard templates has been reviewed [145]. These kinds of metal nanowires can be applied to superconductivity, optical spectroscopy, sensing, and catalytic conversion, and energy harvesting. The AAO template method provides access to arrays of single-crystal metal nanowires and to quasi-onedimensional metal nanostructures with controlled compositional variation along their length. Important semiconductors such as ZnO and TiO2 with nanostructures can also be manufactured in such template or related anodization techniques.

Zinc Oxide (ZnO) ZnO exhibits many unusual properties including uniaxial piezoelectric response and ntype semiconductor characteristics. Such properties can be used in applications such as fieldemission materials [146], light emitting diodes (LEDs) [147], solar cells [148] and gas sensors [149]. Electrodeposition of ZnO films has been reported by several groups and used in fabrication of oriented nanowire and nanorod arrays [150-155]. ZnO films can be electrodeposited cathodically in aqueous chloride solutions using dissolved oxygen as a precursor. The deposition reaction in the electrolyte is: Zn 2+ + 1 O 2 + 2 e − → ZnO . The 2

deposition mechanism was analyzed in terms of electrochemical induced surface precipitation due to an increase of local pH resulting from the oxygen reduction reaction. A dramatic effect of temperature on the formation of ZnO was observed. When the temperature increases (Ttransition≈ 50˚C), a transition between amorphous insulating zinc oxyhydroxide (ZnOx(OH)y) to well-crystallized and conducting ZnO happens [156]. The dimensions and the deposition

28

Di Wei

rate of electrodeposited ZnO nanowire arrays can be controlled by changing the chemical nature of the anions in solution. A significant variation of the diameter (65 to 110 nm) and length (1.0 to 3.4 μm) of the nanowires can be obtained by changing only the nature of the anions in solution [157]. Recently, large-scale, single-crystalline ZnO nanotube arrays were directly fabricated onto F-doped SnO2 (TCO) glass substrate via an electrochemical deposition method from an aqueous solution [158]. The tubes had a preferential orientation along the [0001] direction and hexagon-shaped cross sections. The novel nanostructure could be easily fabricated without a prepared layer of seeds on the substrate. The surface condition of substrate material and the experimental conditions played a key role in the nanotube formation. The growth of ZnO arrays and the deposition reaction take place by applying a cathodic potential (-0.7 V vs. SCE) to the TCO substrate at 80˚C. The dissolved oxygen serves as the oxygen source for the growth of ZnO arrays in the electrolyte. SEM image of the result ZnO nanotubes are shown in Fig. 18. The electrochemically deposited single crystalline ZnO nanowires can be applied in LEDs [154;155].The ZnO nanowire films can be embedded in an insulating spin-coated polystyrene layer. The spin-coating parameters are carefully finetuned to completely fill out the space between the ZnO nanowires and produce only a very thin coverage of the nanowire tips. The polystyrene layer thickness at the tips can be further reduced by the plasma etching treatment to make the n-type ZnO tip junctions outside. A top contact consisting of a thin ptype poly(3,4-ethylene dioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) layer and an evaporated Au film are provided to serve as the hole injection anode in the LED. A flexible LED can be realized when electrochemical depositing the ZnO nanowire on flexible transparent polymer substrates (e.g. polyethylene terephtalate, PET) coated with indium-tinoxide (ITO) [147]. The infrastructure of such flexible LED is illustrated in Fig. 19.

Figure 18. SEM image of the ZnO nanotube film [158]. Reproduced by the kind permission from the publisher.

Electrochemical Nanofabrication

29

Figure 19. Design scheme for a flexible LED structure consisting of vertically oriented single crystalline nanowires grown electrochemically on a polymeric ITO-coated substrate. The top contact consists of p-type polymer (PEDOT:PSS) and an evaporated Au layer. Light is emitted through the transparent polymer [147]. Reproduced by the kind permission from the publisher.

This device exhibits electroluminescence over most of the visible spectrum at moderate forward bias.

Titanium Dioxide (TiO2) Another semiconductor which has similar bandgap with ZnO, titanium dioxide (TiO2) thin films have been widely exploited in many applications such as microelectronics [159], highly efficient catalysts [160], microorganism photolysis [161], antifogging and selfcleaning coatings [162], biosensors [163], gate oxides in metal-oxide-semiconductor field effect transistors (MOSFET) [164] and more recently in dye-sensitized solar cells (DSSCs) [165]. The geometry of anodic TiO2 nanotubes can be controlled over a wide range by the applied potential in H3PO4/HF aqueous electrolytes in contrast to other electrolytes. It was found that for potentials between 1 and 25 V, tubes could be grown with any desired diameter ranging from 15 to 120 nm combined with tube length from 20 nm to 1 μm. The diameter and the length depend linearly on the voltage [166]. Aqueous HCl electrolyte can be used as an alternative to fluoride containing electrolytes to obtain the vertically oriented TiO2 nanotube arrays by anodization of titanium thin films [167]. Nanotube arrays upto 300 nm in length, 15 nm inner pore diameter and 10 nm wall thickness can be made using 3 M HCl aqueous electrolyte for anodization potentials between 10 and 13 V. A highly ordered TiO2 nanotube array with a unique surface morphology can be fabricated by electrochemical anodization of titanium in an organic electrolyte (e.g. 1:1 mixture of DMSO and ethanol) containing 4% HF. The TiO2 nanotube arrays with improved

30

Di Wei

photochemical response can be obtained using electrochemical anodization of titanium in fluorinated organic electrolytes by optimizing etching time, applied potential, solvents and the HF concentration [168]. Using an electrochemical approach in organic electrolytes, the growth of more than 250 μm thick self-organized TiO2 nanotube layers is possible [169]. Combining the electrochemical parameters in an optimal way, ordered TiO2 nanotube layers with a length of over 250 μm have been obtained at 120 V with 0.2 M HF. The tubes can grow as a hexagonal close packed pore array. Although the TiO2 nanotubes fabricated in organic solution have the longest length and the largest surface area, its conductivity may be lower than the one synthesized in aqueous solutions [163]. Crucial parameters that decide on the dimensions are the fluoride ion concentration, the voltage and the anodization time. The different length of TiO2 nanotubes synthesized by electrochemical anodization can be controlled, and is shown in Fig. 20.

Figure 20. SEM images of TiO2 nanotubes prepared in (A) 0.1M HF acid solution at 20 V (B) 1.0M NaHSO4 containing 0.1M KF at 20 V and (C) ethylene glycol containing 0.25% NH4F at 60 V for 1 h [163]. Reproduced by the kind permission from the publisher.

Highly ordered transparent TiO2 nanotube arrays produced by electrochemical anodization have been used in DSSCs. It suggests superior electron transport in nanotubular TiO2 based DSSCs [170]. Remarkable photoconversion efficiencies were expected to be obtained with increase in the length of the nanotube arrays. Carboxylated polythiophene derivatives can be selfassembled onto the TiO2 arrays produced by anodizing titanium foils in

Electrochemical Nanofabrication

31

ethylene glycol based electrolytes [171]. Such self-assembled hybrid polymer-TiO2 nanotube array heterojunction solar cells can yield power efficiency of 2.1% under AM 1.5 without dyes. It was found that the formation mechanism of TiO2 nanotubes is similar to the porous alumina case under high electrical field. TiO2 nanotube arrays can be fabricated by anodic oxidation of titanium foil in different electrolytes. The produced TiO2 nanotube arrays possess large surface area and good uniformity and are ready for enzyme immobilization [172-174], which can be used as biosensors. Furthermore, different length of TiO2 nanotube arrays fabricated by anodic oxidation in different electrolytes were studied for their sensitivities to hydrogen peroxide after co-immobilized horseradish peroxidase (HRP) and thionine chloride. The nanotube arrays fabricated in potassium fluoride solution has the best sensitivity to H2O2 with a detection range from 10-5 to 3×10-3 M [163]. With the use of the template of an AAO, TiO2 nanowires can be obtained by cathodic electrodeposition [175;176] where the metallic ions are attracted to the AAO cathode electrode and reduced to metallic form. For example, in a typical process, the electrodeposition is carried out in 0.2 M TiCl3 solution with pH=2 with a pulsed electrodeposition approach, and titanium and/or its compound are deposited into the pores of the AAO. By heating the above deposited template at 500 °C for 4 hours and removing the template, pure anatase TiO2 nanowires can be obtained. Highly ordered TiO2 single crystalline (pure anatase) nanowire arrays can also be fabricated within the pores of anodic aluminum oxide (AAO) template by a cathodically induced sol-gel method [177]. During this electrochemically induced sol-gel process, both the formation of sol particles and the gelation process take place in the AAO pores. Therefore, TiO2 nanowires with very small diameters (less than 20 nm or even smaller) can be obtained by this technique. In addition, the length of the nanowires can be well controlled by varying the deposition time and potential of the working electrode.

Electrochemical Fabrication of Soft Matters in Nanoscale Nanofiber, nanospheres and other nanoscale of soft matters such as conducting polymers have been fabricated in traditional chemistry way and/or via selfassembly [178-187]. Basically, 1D conducting polymer nanostructures can also be synthesized chemically or electrochemically by using “hard” and “soft” template methods. Obviously, the hard-template method (e.g. AAO) is an effective and straightforward route for fabricating conducting polymer nanostructures with diameters determined by the diameter of the pores in the template. Controlled conducting poly(aniline) nanotubes and nanofibers have been fabricated in the AAO templates and find promising applications in lithium/poly(aniline) rechargeable batteries [188]. Nanotubes and nanowires of conducting polymers including polypyrrole (PPy), polyaniline (PANI) and poly(3,4-ethylenedioxythiophene) (PEDOT), can be synthesized by electrochemical methods using the AAO templates [189]. By changing the doping level, dopant and template-dissolving solvents, the electrical and optical properties of the nanotubes and nanowires can be controlled. The diameter of the conducting polymer nanotubes and nanowires are in the range from 100 to 200 nm, depending on the diameter if the nanoporous template used. It was found that the polymerization was initiated from the wall-side of the AAO template. The synthesized nanotubes have an open end at the top with the filled end at the bottom. As polymerization time increased, the nanotubes will be filled and nanowires will be formed with the length increased. For example, PPy nanotube can be

32

Di Wei

synthesized by applying current of 2-3 mA for 1 min. When the time is increased to 15-40 mins, PPy nanowires will be produced. Conducting polymer nanotube and nanowires prepared by this electrochemical method using AAO templates can be applied in field emitting applications [190;191]. Fig. 21 below shows an uniform of polyaniline nanowires produced at constant potential at 1.0 V for 10 min through the AAO template [192]. PPy nanotubes and nanowires can be also electrochemically synthesized through nanoporous AAO template in ionic liquids[193]. The electrolyte consisted of pyrrole monomer, solvent, and ionic liquid dopant such as 1-butyl-3-methyl imidazolium tetrafluoroborate (BMIMBF4) or 1-butyl-3-methyl imidazolium hexafluorophosphate (BMIMPF6), which is stable in air and moisture. The length and diameter of PPy nanotubes and nanowires were determined by the synthetic conditions such as polymerization time, current, and dopant. The formation of nanotube and nanowire of PPy sample was confirmed by using field emission scanning electron microscope and transmission electron microscope. Formation of PANI nanotubules in room temperature ionic liquids by means of electrochemical polymerization without any template has also been reported [194]. PANI nanotubules were synthesized electrochemically on a modified ITO glass in BMIMPF6 containing 1M trifluoroacetic acid. Tubular structures of PANI with the diameter of ca.120 nm were shown by scanning electron microscopy.

Figure 21. SEM of electrochemically synthesized (a) PANI nanowires and (b) composite nanowires in AAO membranes (etched with NaOH) [192]. Reproduced by the kind permission from the publisher.

Electrochemical Nanofabrication

33

Variable inorganic nanoparticles of different sizes can be combined with conducting polymers, giving rise to a novel composite material with interesting physicochemical properties and possibilities for important applications. Electrochemical methods have proved to be effective in incorporating metal nanoparticles in either pre-deposited polymers or in growing polymer films. Metals can be electrodeposited at conducting polymer electrodes [195]. Nanocables consisting of Ag nanowires sheathed by polyaniline were fabricated in porous AAO template [192]. Silver/polyaniline (Ag/PANI) nanowires were prepared by simultaneous oxidative electropolymerization of aniline and reduction of Ag+ in porous AAO from an acidic electrolyte containing Ag+ and aniline. One-step electrochemical fabrication of devices based on such inorganic-organic hybrid materials offers a new strategy to make electronic devices. Electrochemical fabrication of non-volatile memory device based on polyaniline and gold particles was reported [196]. Au nanoparticles are synthesized and embedded into the PANI polymer matrix simultaneously during the electropolymerization. The impedance states of the PANI:Au composite films are nonvolatile in nature and can be read and switched several times with minimal degradation in air. Such electrochemical fabrication method can produce multi-stable nonvolatile memory device in one step, which simplifies fabrication of the memory device significantly.

Carbon Nanotube Templates Carbon nanotubes (CNTs) can be used as templates for soft matters such as conducting polymers. Composite films of CNTs with PANI, PPy or PEDOT were prepared via electrochemical co-deposition from solutions containing acid treated CNTs and the corresponding monomer. The capacitance of such composite was studied [197].

Figure 22. Continued on next page.

34

Di Wei

Figure 22. SEM image of (a) modified ITO glass surface (MITO); (b) purified SWNTs; (c) SWNT modified with PANI (electropolymerization for 50 cycles); (d) SWNT modified with PANI (electropolymerization for 300 cycles); (e) bare ITO glass surface [198]. Reproduced by the kind permission from the publisher.

Electrochemical functionalization of single walled carbon nanotubes (SWNTs) with PANI has also been done in ionic liquids [198]. SWNTs are covalently functionalized during the electropolymerization of aniline in ionic liquids. This methodology provides a novel way by which large amount of SWNTs (15 mg/ml) can be modified by aniline electrochemically. Fig. 22 shows the processes of coating CNTs with PANI by increasing the scan cycles during cyclic voltammetry from 50 cycles (Fig. 22c) to 300 cycles (Fig. 22d). Electrochemical synthesis of PPy/CNT nanoscale composites using well aligned carbon nanotube arrays was reported [199]. The thickness of the PPy film can be easily controlled by the value of the film-formation charge. Aligned coaxial nanowires of carbon nanotubes can be sheathed with conducting polymers shown in the following figures (Fig. 23 and Fig. 24) [200]. In addition, the aligned MWNT electrode arrays can be given additional robustness by electrodepositing conducting polymer around the tubes and then employing them as enzyme sensors. Nanostructuring electrodes with carbon nanotubes for its sensing applications have been reviewed [201]. Glucose oxidase, the most popular oxidase enzyme can be immobilized onto CNTs by electrodeposition within conducting polymers [202].

Figue 23. Continued on next page.

Electrochemical Nanofabrication

35

Figure 23. SEM images of a) aligned nanotunes after transfer onto a gold foil (a small piece of the assynthesized aligned nanotue film is inlded at the bottom-left corner to show the amorphous carbon layeras well) and b) the CP-NT coaxial nanowires produced by cyclic voltammetry (25 mV/s) on the aligned carbon nanotube electrode in an aqueous solution of 0.1M NaClO4 containing 0.1 M pyrrole [200]. Reproduced by the kind permission from the publisher.

Figure 24. TEM images of the CP-CNT coaxal nanowire formed from the cyclic voltammetry method. The images are in the tip region (left) and on the wall (right) [200] Reproduced by the kind permission from the publisher.

Colloidal Polystyrene (PS) Latex Templates In nanoelectrodeposition, the aim is to place only a single layer or more of coverage on a surface in a very controlled way. Colloidal crystal such as polystyrene (PS) latex templates have been widely used to synthesise highly ordered macroporous ceramics [203], metals [204] and polymers [205;206] where a particular interest has been in the optical, magnetic and photonic band gap properties of the resultant structures. Studies of the magnetic properties of the macroporous films show a large coercivity enhancement in comparison to

36

Di Wei

the corresponding plain films and it was found that the coercive field gradually increases as the diameter of the spherical voids decreases for films of a constant thickness [207]. PS latex nanospheres can be self-assembled on hydrophobic surfaces such as unoxidised silicon or gold and used as templates [207;208]. Two or three dimensional highly ordered macroporous cobalt, iron, nickel, gold, platinum and palladium films containing close packed arrays of spherical holes of uniform size (an inverse opal structure) can be prepared by such simple and versatile template technique [207-209]. The films were prepared by electrochemical reduction of metal cations (e.g. gold ions) dissolved in aqueous solution within the interstitial spaces of pre-assembled colloidal templates assembled on gold surfaces. The templates were assembled from colloidal polystyrene latex spheres assembled onto gold electrode surfaces from aqueous solution by slow evaporation. The aqueous layer was allowed to evaporate naturally so that the PS nanospheres assemble by capillary force on the substrate surface. Following the electrochemical deposition of the metal films, the polystyrene templates were removed by dissolution in toluene. The resulting films will show well-formed two or three dimensional porous structures consisting of interconnected hexagonal close packed arrays of spherical voids. The diameter of the spherical voids within the structures can be varied by changing the diameter of the PS latex spheres used to form the template. The thickness of the films is controlled by varying the charge passed in their deposition.

Figure 25. SEM images of regions of macroporous gold films grown with a thickness gradient by electrochemical deposition through templates assembled from either 750 nm-diameter polystyrene spheres [209]. Φ is the diameter of the sphere on top layer. The three dotted circle lines represent the spheres beneath. Reproduced by the kind permission from the publisher.

Electrochemical deposition is ideal for the production of thin supported layers for applications such as photonic mirrors, because the surface of the electrochemically deposited

Electrochemical Nanofabrication

37

film can be very uniform. Electrochemical deposition occurs from the electrode surface out through the overlying template, the first layer of templated material, deposited out to a thickness comparable with the diameter of the template spheres used, has a different structure from subsequent layers. The subsequent growth of the film by electrodeposition out through the template leads to a modulation of the surface topography of the film in a regular manner that will depend on the precise choice of deposition bath and deposition conditions. This is clearly shown in the following figure if more than one layer of PS is used in templates. The electrodeposition was performed at a potential of -0.90 VSCE. The SEM image shows that the spherical voids left in the gold films after the removal of the PS spheres are arranged in well-ordered, single domain, close-packed structures. Fig. 25 gives an image for a macroporous gold film prepared through a template of 750 nm diameter spheres in a region where more than one layer of PS is selfassembled. Within each hemispherical void in the top layer there are again three smaller dark circles (diameter ca.100 nm). These correspond to the interconnections to the three spherical voids in the layer below (marked as dotted circles in Fig. 25) that are left around the regions where the original polystyrene spheres in the two layers were in contact. Nanostructured macroporous semiconductors such as PbO2 can also be electropolymerized by such template methods [211]. Other nanostructured patterns such as nanodots can also be synthesized by this template method. Highly ordered magnetic nanoscale dot arrays of Ni can be fabricated from double-templated electrodeposition [212;213]. Patterns of ordered arrays of spheres with controlled spacing can be electrochemical deposited by two steps. The double templates were firstly prepared by selfassembly of PS latex spheres on a gold coated glass substrate. This primary template was used for the electrodeposition of the conducting polymer resulting in a macroporous polymer template. After the deposition of PPy, the PS spheres were dissolved in toluene leaving a secondary polymer template. The PPy was then converted into an insulator either by over-oxidation or by undoping at a sufficiently negative potential. This insulating structure was used as the template for electrochemical deposition of magnetic material such as Ni. Electrodeposition of magnetic materials gradually fills the spherical cavities of the polypyrrole template. Ordered arrays of Ni dots with quasi-spherical geometry can be fabricated by this way and the diameter of the dots can be set from 20 nm [213]. Ordered 3D arrays of polyaniline (PANI) inverse opals can also fabricated via electrochemical methods by using colloidal crystals of polystyrene beads as sacrificial templates as shown in Fig. 26 [214].

Figure 26. Schematic illustration of the procedure used for fabricating PANI inverse opals. Reproduced by the kind permission from the publisher.

Fig. 27 shows the PANI inverse opals prepared by cyclic voltammetry in 3D PS colloidal crystal template. Compared with films obtained by chemical synthesis, the inverse opaline samples obtained by electrochemistry had a much higher structural quality.

38

Di Wei

Figure 27. SEM images of PANI inverse opals prepared via cyclic voltammetry, at low (A) and higher (B) magnification. scan rate 20 mV/s, 10 scan cycles [214]. Reproduced by the kind permission from the publisher.

PANI inverse opals were also prepared by a galvanostatic method at a current density 0.05 mA/cm2. By adjusting the polymerization time and applied current, this method allows for the exact control over the structure formation and film thickness of the obtained PANI inverse opline films. To explore potential biosensing applications, PANI composite inverse opals were fabricated by modifying the structure with different dopants, such as poly(acrylic acid) (PAA) and poly(styrene sulfonate) (PSS). It was found that these dopants had a significant effect on the structure and the mechanical stabilities of the obtained opaline films. With selection of suitable dopants, PANI composite inverse opals could be fabricated with very high quality. The obtained films remained electroactive in buffer solutions of neutral pH. Together with their huge surface area, they would be ideal candidates for biosensing applications. Such macroporous structures were used as electrocatalysts for the oxidation of reduced β-nicotinamide adenine dinucleotide (NADH). It showed that the electrocatalytic efficiency of the inverse opline film was much higher compared with that of an unpatterned film. Using such templates is a clear example to show the significant advantages of electrochemical deposition methods. It produces a high density deposited material and no shrinkage of the material takes place when the template is removed. Also it can be used to prepare a wide range of materials and allows fine control over the thickness of the resulting macroporous film through control of the total charge passed to deposit the film.

Electrochemistry and Self-assembled Monolayers (SAMs) Electron transfer cannot occur in blocking films. However, for very thin self assembled monolayers (SAMs) of alkane thiols or oxide films, electrons can tunnel through the film and cause Faradaic reactions. Monolayer of alkane thiols can form spontaneously ordered adlayers on substrate like Au(111) due to a strong interaction between the sulphur of the thiol and the gold substrate. These SAMs have received tremendous attention in recent years [215;216]. Electrochemical deposition onto self-assembled monolayers gives new insights into nanofabrication [217]. Pattern transfer with high resolution is a frontier topic in the emerging

Electrochemical Nanofabrication

39

field of nanotechnologies. Electrochemical molding is a possible route for nanopatterning metal, alloys and oxide surfaces with high resolution in a simple and inexpensive way. This method involves electrodeposition onto a conducting master covered by a alkanethiolate SAMs. This molecular film enables direct surface-relief pattern transfer from the conducting master to the inner face of the electrodeposit, and also allows an easy release of the electrodeposited film due their excellent anti-adherent properties. Replicas of the original conductive master can be also obtained by a simple two-step procedure. SAM quality and stability under electrodeposition conditions combined with the formation of smooth electrodeposits are crucial to obtain high-quality pattern transfer with sub-50 nm resolution. Fig. 28 demonstrated the steps involved in metal electrodeposition on SAMs covered substrates. Further in-depth investigations are required for improving SAM quality reducing the defect size and density, and accordingly increasing the lateral resolution of the method.

Figure 28. Scheme showing the steps involved in metal electrodeposition on SAMs covered substrates. a) Defective sites at SAMs (the molecules are indicated in black); b) nucleation and growth of the metal (grey) within the SAM; c) three dimensional growth outside the SAM; d) formation of a continuous metallic film on the SAM [217]. Reproduced by the kind permission from the publisher.

Electrochemical work comprised mainly copper deposition onto alkanethiol SAMs. Under UPD, Cu would go down on the bare substrate only [218]. The chain length of the alkane thiol on gold has an influence on the deposition over potential [219]. The long hain alkane thio (C18) on Au (111) is demonstrated to have a high blocking power [220]. However, it appears to remain one of the challenges for the near future to find the experimental conditions, under which metal can be deposited on top of a SAM, preferably as a 2D overlayer.

40

Di Wei

On the other hand, electrochemistry can induce self-assembly of surface-templated (organo)silica thin films on various conducting supports, with mesopore channels oriented perpendicular to the solid surface over wide areas [221]. This method is intrinsically simple, very fast and does not require any pre-treatment of the support. It consists of combining the electrochemically driven self-assembly of surfactants at solid-liquid interfaces and an electroassisted generation methods to produce sol-gel films. The method of electrochemically assisted self-assembly of mesoporous silica thin films involves the application of a suitable cathodic potential to an electrode immersed in a surfactant-containing hydrolysed sol solution to generate the hydroxyl ions that are necessary to catalyse polycondensation of the precursors and self-assembly of hexagonally packed one-dimentional channels that grow perpendicularly to the electrode surface. The method opens the way to electrochemically driven nanolithography for designing complex patterns of widely accessible meso-structured materials.

Other Template Methods Molecular templates such as modified cyclodextrin have been used for electrochemical nanofabrication. Conducting polymer nanowires and nanorings can be electrochemically synthesized using the molecular templates (thiolated cyclodextrins) on gold [222]. The strategy is to apply electrochemical growth on gold electrodes modified with SAMs of wellseparated thiolated cyclodextrins in an alkanethiol forest. Thiolated aniline monomer is anchored to the surface within the cyclodextrin cavity and forms an initiation point for polymer wire growth. Nanosized conducting polymer wires several tens of nanometer thick and a few micrometers long can be synthesized electrochemically by this template. Even though the polymer wires appear to be made of numerous single strands, it was the first time that a single nanowire thread of a conducting polymer has been isolated. Block copolymers are a class of materials that selfassemble on macromolecular dimensions and have enormous potential in nanoscale patterning and nanofabrication as templates [223;224]. Some complicated nanoporous structures can be replicated by electrochemical deposition with the help of selfassembly of block-copolymers, for example, poly(4-fluorostyrene)-b-poly(D,L-lactide) [PFSPLA] [225]. By this means, very large arrays of possible materials can be manufactured. Free standing nanowires were obtained after removal of the block copolymer template by either dissolution or by UV irridation. Such mild etch method is generally useful to the nanostructures that are sensitive to more aggressive template removal processes. Furthermore, the electrochemical deposition method guaranties a conformable and stable interface between the electrode and the electrodeposited materials. PPy nanotubules have been chemically and electrochemically synthesized inside the pores of nanoporous polycarbonate (PC) particle track-etched membranes (nano-PTM) [226]. Other templates to produce the nano-structured conducting polymers such as DNA [227] and living neural tissue [228] have been reported for its potential applications in biosensors. Nanofabrication technologies are useful for developing highly sensitive, reproducible nanobiosensors. A nanometric system that is composed of well-oriented nanowell arrays can be used for highly sensitive electrochemical DNA detection, it obtained a two-orders-ofmagnitude enhancement in sensitivity [229].

Electrochemical Nanofabrication

41

The step edge defects on single crystal surfaces can be also used as templates to form conducting polymer nanostructures. Polypyrrole nanostructures with diameters less than or equal to 10 nm have been electropolymerized using step and pit defects on highly ordered pyrolytic graphite (HOPG) as templates for electropolymerization [230]. Step defects were naturally occurring, and pits were formed via oxidation of freshly cleaved surfaces of an HOPG wafer by heating at similar to 640 °C. Underpotential deposition of approximately similar to 80 mV caused polypyrrole to form only on the step and pit edges of HOPG at and not on the basal plane. The size of these nanostructures could be controlled by limiting the pyrrole polymerization time at anodic potentials. Recent modeling results allow the morphology of the deposition to be inferred, and the wire-shaped growth is up to 30 s at constant potential, after which the growth changes morphology. Scanning tunneling microscopy data confirm this result. These polypyrrole nanostructures can be removed by sonication.

Others Nanoscale Electrochemistry Individual carbon nanotubes have been modified selectively on one end with metal using a bipolar electrochemical technique [231]. A stable suspension of nanotube was introduced in a capillary containing 10 mM HAuCl4 aqueous electrolyte, and a high electric field is applied to orientate and polarize the individual tubes. During their transport through the capillary under sufficient polarization (30 kV), each nanotube is the site of water oxidation on one end and the site of metal ion reduction on the other end with the size of the formed metal cluster being proportional to the potential drop along the nanotube. Bipolar electrochemistry occurs when an external electrical field polarizes an object that is not physically connected to the electrodes and thus generates an anodic and a cathodic area on the same object. The substrate can be any kind of material, but its conductivity must be higher than that of the surrounding medium. The induced potential difference between the two extremities of the object, and therefore the kinetics of the associated redox reactions, is directly proportional to the particle’s effective length [232]. When the capillary is exposed to a high electric field (10-30 kV), an electroosmotic flow is generated inside the capillary, transporting the CNT/AuCl4 suspension from the anodic capillary inlet towards the cathodic compartment. A 1 mm long carbon fiber was inserted into a glass capillary connected to two reservoirs filled with HAuCl4 electrolyte. The fiber was observed under the microscope during the application of different potential values. It was found that potentials higher than 40 V are needed to generate a visible metal deposit on the cathodic side of the fiber. After less than 5 min, a visible gold deposit was clearly formed on the negativel polarized end of the fiber. In the course of 1 h, the metal continued growing and its morphology, as revealed by SEM, was dominated by an agglomeration of small crystallites. The Au particle formed after 45min was illustrated in Fig 29.

42

Di Wei

Figure 29. Capillary filled with an aqueous CNT/AuCl4 suspension dipping in the two reservoirs of a capillary lectrophoresis setup. A high electric field is applied, leading to the polarization of the individual nanotubes, thus triggering different electrochemical reactions on either end. Optical micrographs of a carbon fiber inside a glass capillary during dissymmetric gold deposition by bipolar electrochemistry. The applied voltage was 70 V for a capillary with a length of 10 cm, filled with 10 mM HAuCl4 aqueous electrolyte [231]. Reproduced by the kind permission from the publisher.

Looking to the future, this capillary assisted bipolar electrodeposition can be generalized to other types of nano-objects and also deposits of a very different nature such as other metals, semiconductors, or polymers. The approach therefore opens up the way to a whole new family of experiments leading to complex nano-objects with an increasingly sophisticated design allowing original applications.

Sonoelectrochemistry Single crystalline CdSe nanotubes have been successfully synthesized by a sonoelectrochemical route in aqueous solution at room temperature [233]. The sonoelectrochemical method is accomplished by applying an electric current pulse to nucleate and perform the electrodeposit, followed by a burst of ultrasonic energy that removes the products from the sonic probe cathode [234;235]. The growth progress suggested that the CdSe nanotubes were fabricated by sonication-induced rolling-up of CdSe nanosheets and the resulting CdSe products are in tubular structure. This method is simple, convenient and environmentally benign. The sonoelectrochemical synthesis of other nanomaterials is being investigated currently. In addition, traditional top-down nanofabrication methods such as focused ion beam (FIB), can be used to fabricate nanopore array electrodes [236]. FIB milling thus represents a simple and convenient method for fabrication of prototype nanopore electrode arrays. These electrode nano-arrays can be used in electrochemical nanofabrication for applications in sensing and fundamental electrochemical studies. Various organic and inorganic materials with nanoscale architectures can be fabricated by electrochemical nanofabrication. It is a versatile method for fabricating nanostructures with its simplicity, low-temperature processing, low equipment cost and precise control of the deposit thickness through control of the total charge passed. It is an exciting era to witness the

Electrochemical Nanofabrication

43

emerging of nanotechnology. Electrochemistry will definitely contribute to its development independently and interdisciplinarily with other nanofabrication methods.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32]

P. C. Andricacos, C. Uzoh, J. O. Dukovic, J. Horkans, and H. Deligianni, IBM J. Rev. Development, 42 (1998) 567. P. C. Andricacos, Interface, 8 (1999) 32. M. Datta, Electrochim. Acta, 48 (2003) 2975. J. J. Sun, K. Kondo, T. Okamura, S. Oh, M. Tomisaka, H. Yonemura, M. Hoshino, and K. Takahashi, J. Electrochem. Soc., 150 (2003) G355. E. Herrero, L. J. Buller, and H. D. Abruna, Chem. Rev., 101 (2001) 1897. A. A. Gewirth and B. K. Niece, Chem. Rev., 97 (1997) 1129. D. M. Kolb, in H.Gerisher and C.W.Tobias (Eds.), Advances in Electrochemistry and Electrochemical Engineering, Vol. 11, Wiley Interscience, New York, 1978. E. Herrero, L. J. Buller, and H. D. Abruna, Chem. Rev., 101 (2001) 1897. S. Humann, J. Hommrich, and K. Wandelt, Thin Solid Films, 428 (2003) 76. B. M. Huang, T. E. Lister, and J. L. Stickney, Surf. Sci., 392 (1997) 27. S. Bedair, Atomic Layer Epitaxy, Elsevier, Amsterdam, 1993. D. W. Suggs and J. L. Stickney, Surf. Sci., 290 (1993) 362. D. W. Suggs, I. Villegas, B. W. Gregory, and J. L. Stickney, J. Vac. Sci. Technol. A, 10 (1992) 886. I. Villegas and P. Napolitano, J. Electrochem. Soc., 146 (1999) 117. L. P. Colletti, R. Slaughter, and J. L. Stickney, J. Soc. Info. Display, 5 (1997) 87. G. Hodes, Sol. Energy Mater., 32 (1994) 323. K. Rajeshwar, Advanced Materials, 4 (1992) 23. G. F. Fulop and R. M. Taylor, Ann. Rev. Mater. Sci., 15 (1985) 197. R. C. DeMattei and R. S. Feigelson, in J.McHardy and F.Ludwig (Eds.), Electrochemistry of Semiconductors and Electrodes, Noyes, Park Ridge, NJ, 1992. G. Hodes, in I.Rubinstein (Ed.), Electrochemistry: Principles, Methods and Applications, Marcel Dekker, 1995. J. L. Stickney, in R.C.Alkire and D.M.Kolb (Eds.), Advances in Electrochemical Science and Engineering, Vol. 7, Wiley-VCH, New York, 2001. M. P. R. Panicker, M. Knaster, and F. A. Kroger, J. Electrochem. Soc., 125 (1978) 566. W. J. Danaher and L. E. Lypns, Nature, 271 (1978) 139. G. Hodes, J. Manassen, and D. Cahen, Nature, 261 (1976) 403. J. J. Cuomo and R. J. Gambino, J. Electrochem. Soc., 115 (1968) 755. H. Gobrecht, H. D. Liess, and A. Tausend, Ber. Bunsenges. Phys. Chem., 67 (1963) 930. M. Tomkiewicz, I. Ling, and W. S. Parsons, J. Electrochem. Soc., 129 (1982) 2016. S. Dennison, J. Mater. Chem., 4 (1994) 41. S. K. Das and G. C. Morris, J. Appl. Phys., 73 (1993) 782. S. K. Das, Thin Solid Films, 226 (1993) 259. T. Yoshida, J. Electrochem. Soc., 139 (1992) 2353. G. C. Morris, S. K. Das, and P. G. Tanner, J. Cryst. Growth, 117 (1992) 929.

44

Di Wei

[33] S. K. Das and G. C. Morris, J. Appl. Phys., 72 (1992) 4940. [34] R. N. Bhattacharya and K. Rajeshwar, J. Appl. Phys., 58 (1985) 3590. [35] R. M. Metzger, V. V. Konovalov, M. Sun, T. Xu, G. Zangari, B. Xu, M. Benakli, and W. D. Doyle, IEEE Transactions on Magnetics, 36 (2000) 30. [36] G. Zangari and D. N. Lambeth, IEEE Transactions on Magnetics, 33 (1997) 3010. [37] J. W. Diggle, T. C. Downie, and C. W. Goulding, Chem. Rev., 69 (1969) 365. [38] J. D. Klein, R. D. Herrick, D. Palmer, M. J. Sailor, C. J. Brumlik, and C. R. Martin, Chem. Mater., 5 (1993) 902. [39] K. Murase, H. Watanabe, S. Mori, T. Hirato, and Y. Awakura, J. Electrochem. Soc., 146 (1999) 4477. [40] K. Murase, H. Uchida, T. Hirato, and Y. Awakura, J. Electrochem. Soc., 149 (1999) 531. [41] A. M. Kressin, V. V. Doan, J. D. Klein, and M. J. Sailor, Chem. Mater., 3 (1991) 1015. [42] S. K. Das and G. C. Morris, Sol. Energy Mater., 30 (1993) 107. [43] M. W. Verbrugge and C. W. Tobias, Am. Inst. Chem. Eng. J., 33 (1987) 628. [44] C. D. Lokhande, J. Electrochem. Soc., 134 (1987) 1728. [45] M. W. Verbrugge and C. W. Tobias, J. Electrochem. Soc., 132 (1985) 1298. [46] B. M. Basol, J. Appl. Phys., 58 (1985) 3809. [47] L. Beaunier, H. Cachet, M. Froment, and G. Maurin, J. Electrochem. Soc., 147 (2000) 1835. [48] H. Cachet, R. Cortes, M. Froment, and G. Maurin, Philos. Mag. Lett., 79 (1999) 837. [49] B. Miller, S. Menezes, and A. Heller, J. Electroanal. Chem., 94 (1978) 85. [50] J. Vedel, M. Soubeyrand, and E. Castel, J. Electrochem. Soc., 124 (1977) 177. [51] M. H. Miles and W. S. McEwan, J. Electrochem. Soc., 119 (1972) 1188. [52] A. J. Panson, Inorg. Chem., 3 (1964) 940. [53] M. Krebs and K. E. Heusler, Electrochim. Acta, 37 (1992) 1371. [54] V. I. Briss and L. E. Kee, J. Electrochem. Soc., 133 (1986) 2097. [55] L. M. Peter, Electrochim. Acta, 23 (1978) 165. [56] L. M. Peter, Electrochim. Acta, 23 (1978) 1073. [57] B. Miller and A. Heller, Nature, 262 (1976) 680. [58] V. K. Kapur, V. Choudary, and A. K. P. Chu, Patent, (1986) 4581108. [59] G. Hodes, T. Engelhard, C. R. Herrington, L. L. Kazmerski, and D. Cahen, Progr. Cryst. Growth Charact., 10 (1985) 345. [60] B. W. Gregory and J. L. Stickney, J. Electroanal. Chem., 300 (1991) 543. [61] B. W. Gregory, D. W. Suggs, and J. L. Stickney, J. Electrochem. Soc., 138 (1991) 1279. [62] G. S. Hsiao, M. G. Andersen, S. Gorer, D. Harris, and R. M. Penner, J. Am. Chem. Soc., 119 (1997) 1439. [63] M. G. Andersen, S. Gorer, and R. M. Penner, J. Phys. Chem., 101 (1997) 5859. [64] R. M. Nyffenegger, B. Craft, M. Shaaban, S. Gorer, and R. M. Penner, Chem. Mater., 10 (1998) 1120. [65] A. S. Baranski, W. R. Fawcett, and A. C. McDonald, J. Electroanal. Chem., 160 (1984) 271. [66] A. S. Baranski and W. R. Fawcett, J. Electrochem. Soc., 131 (1984) 2509. [67] D. Gal, G. Hodes, D. Hariskos, D. Braunger, and H. W. Schock, Appl. Phys. Lett., 73 (1998) 3135.

Electrochemical Nanofabrication

45

[68] Z. Tang, S. Liu, S. Dong, and E. Wang, J. Electroanal. Chem., 502 (2001) 146. [69] Y. Shen, J. Liu, A. Wu, J. Jiang, L. Bi, B. Liu, Z. Li, and S. Dong, Langmuir, 19 (2003) 5397. [70] J. Liu, L. Cheng, Y. Song, B. Liu, and S. Dong, Langmuir, 17 (2001) 6747. [71] W. Freyland, C. A. Zell, S. Zein El Abedin, and F. Endres, Electrochim. Acta, 48 (2003) 3053. [72] Q. X. Liu, S. Zein El Abedin, and F. Endres, J. Electrochem. Soc., 155 (2008) D357. [73] N. R. Jana, L. Gearheart, and C. J. Murphy, Adv. Mater., 13 (2001) 1389. [74] A. Loiseau and H. Pascard, Chemical Physics Letters, 256 (1996) 246. [75] C. R. Martin, Adv. Mater., 3 (1991) 457. [76] L. Sun, P. C. Searson, and C. L. Chien, Appl. Phys. Lett., 74 (1999) 2803. [77] C. J. Brumlik and C. R. Martin, J. Am. Chem. Soc., 113 (1991) 3174. [78] Y. Zhang, N. W. Franklin, R. J. Chen, and H. Dai, Chemical Physics Letters, 331 (2000) 35. [79] M. Peckerar, R. Bass, and K. W. Rhee, Journal of Vacuum Science & Technology B, 18 (2000) 3143. [80] F. J. Himpsel, T. Jung, and J. E. Ortega, Surface Review and Letters, 4 (1997) 371. [81] E. C. Walter, M. P. Zach, F. Favier, B. Murray, K. Inazu, J. C. Hemminger, and R. M. Penner, in J.Z.Zhang and Z.L.Wang (Eds.), Physical Chemistry of Interfaces and Nanomaterials, Proc. SPIE 2002, 2002. [82] M. P. Zach, K. H. Ng, and R. M. Penner, Science, 290 (2000) 2120. [83] D. B. Allred, M. Sarikaya, F. Baneyx, and D. T. Schwartz, Nano Letters, 5 (2005) 609613. [84] D. M. Kolb, Angew. Chem. Int. Ed., 40 (2001) 1162. [85] D. M. Kolb, R. Ullmann, and J. C. Ziegler, Electrochim. Acta, 43 (1998) 2751. [86] H. Sugimura and N. Nakagiri, Thin Solid Films, 282 (1996) 572-575. [87] A. I. Mardare, A. D. Wieck, and A. W. Hassel, Electrochim. Acta, 52 (2007) 7865. [88] P. Avouris, R. Martel, T. Hertel, and R. Sandstrom, Applied Physics A-Materials Science & Processing, 66 (1998) S659-S667. [89] X. W. Cai, J. S. Gao, Z. X. Xie, Y. Xie, Z. Q. Tian, and B. W. Mao, Langmuir, 14 (1998) 2508-2514. [90] S. Y. Jang, M. Marquez, and G. A. Sotzing, Synthetic Metals, 152 (2005) 345. [91] A. A. Tseng, A. Notargiacomo, and T. P. Chen, Journal of Vacuum Science & Technology B, 23 (2005) 877-894. [92] R. M. Nyffenegger and R. M. Penner, Chem. Rev., 97 (1997) 1195. [93] M. J. Esplandiu, M. A. Schneeweiss, and D. M. Kolb, Phys. Chem. Chem. Phys., 1 (1999) 4847. [94] W. Li, J. A. Virtanen, and R. M. Penner, Appl. Phys. Lett., 60 (1992) 1181. [95] W. Li, G. S. Hsiao, D. Harris, R. M. Nyffenegger, J. A. Virtanen, and R. M. Penner, J. Phys. Chem., 100 (1996) 20103. [96] D. M. Kolb, R. Ullmann, and T. Will, Science, 275 (1997) 1097. [97] R. Schuster, V. Kirchner, X. H. Xia, A. M. Bittner, and G. Ertl, Phys. Rev. Lett., 80 (2008) 5599. [98] R. Ullmann, T. Will, and D. M. Kolb, Chem. Phys. Lett., 209 (1993) 238. [99] U. Landman, W. D. Luedtke, N. A. Burnkam, and R. J. Colton, Science, 248 (1990) 454.

46

Di Wei

[100] D. M. Kolb, R. Ullmann, and T. Will, Science, 275 (1997) 1097-1099. [101] M. G. Del Popolo, E. P. M. Leiva, M. Mariscal, and W. Schmickler, Surface Science, 597 (2005) 133-155. [102] C. J. Hung, J. N. Gui, and J. A. Switzer, Applied Physics Letters, 71 (1997) 1637-1639. [103] C. Lebreton and Z. Z. Wang, Microelectronic Engineering, 30 (1996) 391-394. [104] H. Lee, Y. K. Jang, E. J. Bae, W. Lee, S. M. Kim, and S. H. Lee, Current Applied Physics, 2 (2002) 85. [105] S. H. Hong and C. A. Mirkin, Science, 288 (2000) 1808. [106] R. D. Piner, J. Zhu, F. Xu, S. H. Hong, and C. A. Mirkin, Science, 283 (1999) 661. [107] S. H. Hong, J. Zhu, and C. A. Mirkin, Science, 286 (1999) 523. [108] Y. Li, B. W. Maynor, and J. Liu, Journal of the American Chemical Society, 123 (2001) 2105-2106. [109] A. P. Suryavanshi and M. F. Yu, Nanotechnology, 18 (2007). [110] W. Ehrfeld, Electrochim. Acta, 48 (2003) 2857. [111] V. Lehmann, Proc. IEEE MEMS, (1995) 1. [112] S. D. Collins, J. Electrochem. Soc., 144 (1997) 2242. [113] L. Zhang, X. Z. Ma, M. X. Lin, Y. Lin, G. H. Cao, J. Tang, and Z. W. Tian, Journal of Physical Chemistry B, 110 (2006) 18432-18439. [114] L. T. Canham, Appl. Phys. Lett., 57 (1990) 1046. [115] J. W. Schulze and T. Osaka, Electrochemical Microsystem Technologies, Taylor & Francis, London, 2002. [116] R. Schuster, V. Kirchner, P. Allongue, and G. Ertl, Science, 289 (2000) 98-101. [117] M. Kock, V. Kirchner, and R. Schuster, Electrochim. Acta, 48 (2003) 3213. [118] A. L. Trimmer, J. L. Hudson, M. Kock, and R. Schuster, Appl. Phys. Lett., 82 (2003) 3327. [119] V. Kirchner, X. Xia, and R. Schuster, Acc. Chem. Res., 34 (2001) 371. [120] R. Schuster, Chem. Phys. Chem., 2 (2001) 411. [121] R. Schuster, Chem. Phys. Chem., 8 (2007) 34. [122] A. Cohen, in M.Gad el Hak (Ed.), The MEMS Handbook, CRC Press, Boca Raton, Florida, 2001. [123] M. T. Giacomini and R. Schuster, Phys. Chem. Chem. Phys., 7 (2005) 518. [124] J. Liu, Y. Lin, L. Liang, J. A. Voigt, D. A. Huber, Z. R. Tian, E. Coker, B. Mckenzie, and M. J. Mcdermott, Chemistry-A European Journal, 9 (2003) 605. [125] L. Liang, J. Liu, C. F. Windisch Jr, G. J. Exarhos, and Y. Lin, Angew. Chem. Int. Ed., 41 (2002) 3665. [126] X. Zou, S. Zhang, M. Shi, and J. Kong, J. Solid State Electrochem., 11 (2007) 317. [127] H. Liu, J. Kameoka, D. A. Czaplewski, and H. G. Craighead, Nano Letters, 4 (2004) 671. [128] M. Yun, N. V. Myung, R. P. Vasquez, C. Lee, E. Menke, and R. M. Penner, Nano Letters, 4 (2004) 419. [129] J. Wang, S. Chan, R. R. Carlson, Y. Luo, G. Ge, R. S. Ries, J. R. Heath, and H. R. Tseng, Nano Letters, 4 (2004) 1693. [130] M. M. Alam, J. Wang, Y. Guo, S. P. Lee, and H. R. Tseng, J. Phys. Chem. B, 109 (2005) 12777. [131] M. Hernandez-Velez, Thin Solid Films, 495 (2006) 51-63.

Electrochemical Nanofabrication

47

[132] J. Qin, J. Nogues, M. Mikhaylova, A. Roig, J. S. Munoz, and M. Muhammed, Chem. Mater., 17 (2005) 1829. [133] D. S. Xu, Y. J. Xu, D. P. Chen, G. L. Guo, L. L. Gui, and Y. Q. Tang, Chem. Phys. Lett., 325 (2000) 340. [134] Y. Huang, X. F. Duan, Y. Cui, L. J. Lauhon, K. H. Kim, and C. M. Lieber, Science, 294 (2001) 1313. [135] M. Lahav, E. A. Weiss, Q. Xu, and G. M. Whitesides, Nano Letters, 6 (2006) 2166. [136] H. Masuda, K. Yasui, and K. Nishio, Adv. Mater., 12 (2000) 1031. [137] K. Nielsch, J. Choi, K. Schwirn, R. B. Wehrspohn, and U. Gosele, Nano Letters, 2 (2002) 677. [138] R. C. Furneaux, W. R. Rigby, and A. P. Davidson, Nature, 337 (1989) 147. [139] Z. Sun and H. K. Kim, Appl. Phys. Lett., 81 (2002) 3458. [140] S. Zhao, K. Chan, A. Yelon, and T. Veres, Adv. Mater., 19 (2007) 3004. [141] A. J. Yin, R. S. Guico, and J. Xu, Nanotechnology, 18 (2007). [142] T. Nagaura, F. Takeuchi, Y. Yamauchi, K. Wada, and S. Inoue, Electrochem. Commun., 10 (2008) 681. [143] M. S. Sander and H. Gao, J. Am. Chem. Soc., 127 (2005) 12158. [144] W. Lee, M. K. Jin, W. C. Yoo, E. S. Jang, J. H. Choy, J. H. Kim, K. Char, and J. K. Lee, Langmuir, 20 (2004) 287-290. [145] T. R. Kline, M. Tian, J. Wang, A. Sen, M. W. H. Chan, and T. E. Mallouk, Inorg. Chem., 45 (2006) 7555. [146] C. J. Lee, T. J. Lee, S. C. Lyu, Y. Zhang, H. Ruh, and H. J. Lee, Appl. Phys. Lett., 81 (2002) 3648. [147] A. Nadarajah, R. C. Word, J. Meiss, and R. Koenenkamp, Nano Letters, 8 (2008) 534. [148] M. Law, L. E. Greene, J. C. Johnson, R. Saykally, and P. D. Yang, Nature Mater., 4 (2005) 455. [149] Y. Hu, X. Zhou, Q. Han, Q. Cao, and Y. Huang, Mater. Sci. Eng. B, 99 (2003) 41. [150] L. F. Xu, Y. Guo, Q. Liao, J. P. Zhang, and D. S. Xu, J. Phys. Chem. B, 109 (2005) 13519. [151] J. B. Cui and U. J. Gibson, Appl. Phys. Lett., 87 (2005) 133108. [152] R. Koenenkamp, K. Boedecher, M. C. Lux-Steiner, M. Poschenrieder, and S. Wagner, Appl. Phys. Lett., 77 (2000) 2575. [153] Z. G. Chen, Y. W. Tang, L. S. Zhang, and L. J. Luo, Electrochim. Acta, 51 (2006) 5870. [154] R. Koenenkamp, R. C. Word, and C. Schlegel, Appl. Phys. Lett., 85 (2004) 6004. [155] R. Koenenkamp, R. C. Word, and M. Godinez, Nano. Lett., 5 (2005) 2005. [156] S. Peulon and D. Lincol, J. Electrochem. Soc., 145 (1998) 864. [157] J. Elias, R. Tena-Zaera, and C. Levy-Clement, J. Phys. Chem. C, 112 (2008) 5736. [158] Y. Tang, L. Luo, Z. Chen, Y. Jiang, B. Li, Z. Jia, and L. Xu, Electrochem. Commun., 9 (2007) 289. [159] G. P. Burns, J. Appl. Phys., 65 (1989) 2095. [160] T. Carlson and G. L. Giffin, J. Phys. Chem., 90 (1986) 5896. [161] T. Matsunaga, R. Tomoda, T. Nakajima, N. Nakamura, and T. Komine, Appl. Environ. Microbiol., 54 (1988) 1330. [162] R. Wang, K. Hashimoto, and A. Fujishima, Nature, 388 (1997) 431.

48

Di Wei

[163] P. Xiao, B. B. Garcia, Q. Guo, D. Liu, and G. Cao, Electrochem. Commun., 9 (2007) 2441. [164] P. S. Peercy, Nature, 406 (2000) 1023. [165] M. Gratzel, Nature, 421 (2003) 586. [166] S. Bauer, S. Kleber, and P. Schmuki, Electrochem. Commun., 8 (2006) 1321. [167] N. K. Allam and C. A. Grimes, J. Phys. Chem. C, 111 (2007) 13028. [168] C. Ruan, M. Paulose, O. K. Varghese, G. K. Mor, and C. A. Grimes, J. Phys. Chem. B, 109 (2005) 15754. [169] S. P. Albu, A. Ghicov, J. M. Macak, and P. Schmuki, Phys. Stat. Sol. (RRL), 1 (2007) R65. [170] G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, and C. A. Grimes, Nano Letters, 6 (2006) 215. [171] K. Shankar, G. K. Mor, H. E. Prakasam, O. K. Varghese, and C. A. Grimes, Langmuir, 23 (2007) 12445. [172] H. Tsuchiya, J. M. Macak, L. Muller, J. Kunze, F. Muller, P. Greil, S. Virtanen, and P. Schmuki, J. Biomed. Mater. Res., 77A (2006) 534. [173] S. Liu and A. Chen, Langmuir, 21 (2005) 8409. [174] D. V. Bavykin, E. V. Milson, F. Marken, D. H. Kim, D. H. Marsh, D. J. Riley, F. C. Walsh, K. H. EI-Abiary, and A. A. Lapin, Electrochem. Commun., 7 (2005) 1050. [175] Y. Lei, L. D. Zhang, and J. C. Fan, Chem. Phys. Lett., 338 (2001) 231. [176] S. Liu and K. Huang, Sol. Energy Mater. Sol. Cells, 85 (2004) 125. [177] Z. Miao, D. Xu, J. Ouyang, G. Guo, X. Zhao, and Y. Tang, Nano. Lett., 2 (2002) 717. [178] Z. Zhang, J. Deng, J. Shen, M. Wan, and Z. Chen, Macromolecular Rapid Communications, 28 (2007) 585. [179] X. Li, J. Shen, M. Wan, Z. Chen, and Y. Wei, Synthetic Metals, 157 (2007) 575. [180] Y. Zhu, D. Hu, M. Wan, L. Jiang, and Y. Wei, Advanced Materials, 19 (2007) 2092. [181] Y. Long, Z. Chen, Y. Ma, Z. Zhang, A. Jin, C. Gu, L. Zhang, Z. Wei, and M. Wan, Appl. Phys. Lett., 84 (2004) 2205. [182] L. Zhang and M. Wan, Thin Solid Films, 477 (2005) 24. [183] H. Qiu, M. Wan, B. Matthews, and L. Dai, Macromolecules, 34 (2001) 675. [184] Z. Zhang, M. Wan, and Y. Wei, Advanced Functional Materials, 16 (2006) 1100. [185] L. Zhang and M. Wan, Nanotechnology, 13 (2002) 750. [186] L. Zhang, Y. Long, Z. Chen, and M. Wan, Advanced Functional Materials, 14 (2004) 693. [187] H. Ding, M. Wan, and Y. Wei, Advanced Materials, 19 (2007) 465. [188] F. Cheng, W. Tang, C. Li, J. Chen, H. Liu, P. Shen, and S. Dou, Chemistry-A European Journal, 12 (2006) 3082. [189] B. H. Kim, D. H. Park, J. Joo, S. G. Yu, and S. H. Lee, Synthetic Metals, 150 (2005) 279. [190] J. Joo, B. H. Kim, D. H. Park, H. S. Kim, D. S. Seo, J. H. Shim, S. J. Lee, K. S. Ryu, K. Kim, J. I. Jin, T. J. Lee, and C. J. Lee, Synthetic Metals, 153 (2005) 313. [191] K. Kim and J. I. Jin, Nano Letters, 11 (2001) 631. [192] A. Drury, S. Chaure, M. Kroell, V. Nicolosi, N. Chaure, and W. J. Blau, Chem. Mater., 19 (2007) 4252. [193] Y. K. Koo, B. H. Kim, D. H. Park, and J. Joo, Molecular Crystals and Liquid Crystals, 425 (2004) 55.

Electrochemical Nanofabrication

49

[194] D. Wei, C. Kvarnstroem, T. Lindfors, and A. Ivaska, Electrochem. Commun., 8 (2006) 1563. [195] N. F. Atta, A. Galal, and F. Khalifa, Applied Surface Science, 253 (2007) 4273. [196] D. Wei, J. K. Baral, R. Oesterbacka, and A. Ivaska, J. Mater. Chem., 18 (2008) 1853. [197] C. Peng, J. Jin, and G. Z. Chen, Electrochimica Acta, 53 (2007) 525. [198] D. Wei, C. Kvarnstroem, T. Lindfors, and A. Ivaska, Electrochem. Commun., 9 (2007) 206. [199] J. H. Chen, Z. P. Huang, D. Z. Wang, S. X. Yang, J. G. Wen, and Z. F. Ren, Appl. Phys. A, 73 (2001) 129. [200] M. Gao, S. Huang, L. Dai, G. Wallace, R. Gao, and Z. Wang, Angew. Chem. Int. Ed., 39 (2000) 3664. [201] J. J. Gooding, Electrochim. Acta, 50 (2005) 3049. [202] M. Gao, L. M. Dai, and G. G. Wallace, Electroanalysis, 15 (2003) 1089. [203] B. T. Holland, C. F. Blanford, and A. Stein, Science, 281 (1998) 538. [204] O. D. Velev, P. M. Tessier, A. M. Lenhoff, and E. W. Kaler, Nature, 401 (1999) 548. [205] S. H. Park and Y. Xia, Adv. Mater., 10 (1998) 1045. [206] S. A. Johnson, P. J. Ollivier, and T. E. Mallouk, Science, 283 (1999) 963. [207] P. N. Bartlett, M. A. Ghanem, I. S. EI Hallag, P. de Groot, and A. Zhukov, J. Mater. Chem., 13 (2003) 2596. [208] P. N. Bartlett, P. R. Birkin, and M. A. Ghanem, Chem. Commun., (2000) 1671. [209] P. N. Bartlett, J. J. Baumberg, P. R. Birkin, M. A. Ghanem, and M. C. Netti, Chem. Mater., 14 (2002) 2199. [210] M. E. Abdelsalam, P. N. Bartlett, J. J. Baunberg, and S. Coyle, Advanced Materials, 16 (2004) 90. [211] P. N. Bartlett, T. Dunford, and M. A. Ghanem, J. Mater. Chem., 12 (2002) 3130. [212] A. Zhukov, M. A. Ghanem, A. V. Goncharov, P. N. Bartlett, and P. de Groot, Journal of Magnetism and Magnetic Materials, 272-276 (2004) e1369. [213] A. V. Goncharov, A. Zhukov, P. N. Bartlett, M. A. Ghanem, R. Boardman, H. Fangohr, and P. de Groot, Journal of Magnetism and Magnetic Materials, 286 (2005) 1. [214] S. Tian, J. Wang, U. Jonas, and W. Knoll, Chem. Mater., 17 (2005) 5726. [215] G. E. Poirier, Chem. Rev., 97 (1997) 1117. [216] A. Ulman, Chem. Rev., 96 (1996) 1533. [217] P. L. Schilardi, P. Dip, P. C. D. Claro, G. A. Benitez, M. H. Fonticelli, O. Azzaroni, and R. C. Salvarezza, Chemistry-A European Journal, 12 (2005) 38-49. [218] L. Sun and R. M. Crooks, J. Electrochem. Soc., 138 (1991) L23. [219] J. A. M. Sondag-Huethorst and L. G. J. Fokkink, Langmuir, 11 (1995) 2237. [220] M. A. Schneeweiss and D. M. Kolb, Phys. Stat. Sol. (A), 173 (1999) 51. [221] A. Walcarius, E. Sibottier, M. Etienne, and J. Ghanbaja, Nature Mater., 6 (2007) 602. [222] S. J. Choi and S. M. Park, Advanced Materials, 12 (2000) 1547. [223] D. Grosso, C. Boissiere, B. Smarsly, T. Brezesinski, N. Pinna, P. A. Albouy, H. Amenitsch, M. Antonietti, and C. Sanchez, Nature Mater., 3 (2004) 787. [224] M. Park, C. Harrison, P. M. Chaikin, R. A. Register, and D. H. Adamson, Science, 276 (1997) 1401. [225] E. J. W. Crossland, S. Ludwigs, M. A. Hillmyer, and U. Steiner, Soft Matter, 3 (2007) 94. [226] S. Demoustier-champagne, J. Duchet, and R. Legras, Synthetic Metals, 101 (1999) 20.

50

Di Wei

[227] X. Lin, G. Kang, and L. Lu, Bioelectrochemistry, 70 (2007) 235. [228] S. M. Richardson-Burns, J. L. Hendricks, and D. C. Martin, Journal of Neural Engineering, 4 (2007) L6. [229] H. Y. Lee, J. W. Park, and J. M. Kim, Appl. Phys. Lett., 89 (2006) 113901. [230] J. D. Noll, M. A. Nicholson, P. G. Van Patten, C. W. Chung, and M. L. Myrick, Journal of the Electrochemical Society, 145 (1998) 3320-3328. [231] C. Warakulwit, T. Nguyen, J. Majimel, M. H. Delville, V. Lapeyre, P. Garrigue, V. Ravaine, J. Limtrakul, and A. Kuhn, Nano Letters, 8 (2008) 500. [232] J. C. Bradley, S. Babu, and P. Ndungu, Fullerenes, Nanotubes and Carbon Nanostructures, 13 (2005) 227. [233] Q. Shen, L. Jiang, J. Miao, W. Hou, and J. Zhu, Chem. Commun., (2008) 1683. [234] X. F. Qiu, Y. B. Lou, A. C. S. Samia, A. Devadoss, J. D. Burgess, S. Dayal, and C. Burda, Angew. Chem. Int. Ed., 44 (2005) 5855. [235] X. F. Qiu, C. Burda, R. L. Fu, L. Pu, H. Y. Chen, and J. J. Zhu, Journal of the American Chemical Society, 126 (2004) 16276. [236] Y. H. Lanyon, G. De Marzi, Y. E. Watson, A. J. Quinn, J. P. Gleeson, G. Redmond, and D. W. M. Arrigan, Analytical Chemistry, 79 (2007) 3048-3055.

In: Nanotechnology… Editors: C.J. Dixon and O.W. Curtines, pp. 51-69

ISBN: 978-1-60692-162-3 © 2010 Nova Science Publishers, Inc.

Chapter 2

FABRICATION AND APPLICATION OF NOVEL TWO-DIMENSIONAL NANOWEBS VIA ELECTROSPINNING Bin Ding*,1, Chunrong Li2, Dong Wang3 and Seimei Shiratori2 1

Modern Textile Institute, Donghua University, Shanghai 200051, China 2 SNT Co. Ltd., Kawasaki 212-0054, Japan 3 Fiber and Polymer Science, University of California, Davis, CA 95616, USA

Abstract This chapter reviews our recent progress on the novel two-dimensional nanowebs by the optimization of processing parameters during electrospinning. Using high applied voltage and low relative humidity in chamber, the by-product of micro-sized defect films can be splitted into nanowebs due to the fast phase separation of the charged droplets which flight with high moving speed in electric field from capillary tip to collector. The electrospun fibers act as a support for the “fishnet-like” nanowebs comprising interlinked one-dimensional nanowires. The average diameter of the nanowires contained in typical nanowebs is about one order of magnitude smaller than that of conventional electrospun fibers. Nanowebs together with common electrospun nanofibers can be assembled into a three-dimensional fibrous mat. So far, nylon-6, polyacrylic acid (PAA), poly(vinyl alcohol) (PVA)/SiO2 nanoparticles, and PVA/zinc acetate have been found to have the possibility forming nanowebs. The formation, morphology, and area density of the nanowebs in electrospun fibrous mats are strongly affected by the applied voltage, ambient relative humidity, kinds of solvents, solution concentration and conductivity, and distance between capillary tip to collector. The expanded applications of electrospun fibers are expected due to the formation of nanowebs, such as the nano-sized controllable filters, high efficient catalysts, catalyst supporter, and sensors. The preliminary data showing that the sensitivity of PAA nanowebs to ammonia is 2.5 times higher than that of electrospun PAA nanofibers. Additionally, PAA nanowebs show much quicker absorption speed and larger capacities than that of PAA nanofibers during the ammonia absorption test.

*

E-mail address: [email protected] (B. Ding) Tel & Fax: +86-21-62378392; Corresponding author.

52

Bin Ding, Chunrong Li, Dong Wang et al.

Introduction The process of electrospinning was first studied by Zeleny [1] in 1914 and patented by Formhals [2] in 1934. In the past decade, this technique regained a great deal of attention due to a surging interest in nanotechnology as continuous ultra-fine fibers or fibrous structures of various polymers with diameters in the range from several micrometers down to tens of nanometers [3-6]. The ultra-fine fibers can be easily fabricated under the driving force of an external electric field imposed on a polymer solution or melt. Electrospinning can be considered as a special case of electrospray process which uses electrostatic fields to form and accelerate liquid jet from the tip of a capillary [1,7]. The surface of a hemispherical liquid drop suspended in equilibrium at the end of a capillary will be distorted into a conical shape in the presence of an external electric field. This distortion is caused by a balancing of the repulsive force resulting from the induced charge distribution on surface of droplet with the surface tension of liquid [8]. Taylor showed that at a critical voltage, the equilibrium shape of the suspended meniscus was a cone with a semi-vertical angle of 49.3o [7]. A stable jet of liquid could be ejected and accelerated if the applied voltage exceeded this critical voltage. The jet breaks up into small droplets as a result of the longitudinal Rayleigh instability caused by surface tension in the case of low viscosity liquids. This process is known as electrospray for applications to obtain aerosols composed of sub-micron droplets with narrow distribution [9,10]. For high viscosity liquids, polymer solutions or melts, the jet does not break up, but travels as a jet to the grounded collector. The transverse instability or splaying of the jet into two or more smaller jets is observed due to the radial charge repulsion [11]. This process is termed as “electrospinning” and it produces the polymer fibers with diameter in sub-micron scale [8]. The experimental observation of micro-sized films among the electrospun fibers has been reported [8, 12, 13]. The sub-micron or micron sized droplets, which produced by the high electrically driven instability of the liquid droplet suspended at the tip of capillary, was considered to form micro-sized films among the electrospun fibers. These unexpected microsized films with uncompleted solvent evaporation led the melt of their covered dry fibers, and destroyed the uniformity of fibers. Therefore, they were called defect films. The variation in frequency of defect films formation, size and area density of defect films could be controlled by adjusting the processing parameters during electrospinning. The defect film density increases with increasing instability of the jet at the capillary tip. The one-dimensional (1D) electrospun fibers has a typical length > 100 μm and diameters in the range of 30-2000 nm [3]. The three-dimensional (3D) non-woven mats composed of electrospun fibers are considered to be have a larger surface-to-volume ratio and smaller pore size compared to commercial textiles, making them excellent candidates for applications in sensor system [14-16], filtration [17,18], tissue engineering [19-21], dye-sensitized solar cells [22, 23], super-hydrophobic surfaces [24-28], etc. Considerable recent progress in electrospinning were made that are expected to have lasting impact on the quality and scope of the applications, such as the alignment of electrospun fibers [29-31], fabrication of continuous carbon [32], polyoxometalate [33, 34], ceramic [35-37], coated [38-41], and hollow [34, 42, 43] fibers. Additionally, Yu et al. reported two major strategies to decrease the fiber diameter from the low concentration of polymer in solution and the high chargecarrying capacity of solution [44]. However, the reduction of electrospun fiber diameter is

Fabrication and Application of Novel Two-Dimensional Nanowebs…

53

still a serious challenge. Electrospun fibers with average diameters below 50 nm are still difficult to be produced repeatedly and uniformly for most materials with electrospinnability until now. This chapter reviews our recent new observation of nanowebs, which widely distributed among electrospun fibers during electrospinning with optimized processing parameters. The nanowires within a nanoweb, not electrospun fibers, can be easily fabricated with diameters below 50 nm. Poly(acrylic acid) (PAA) and nylon-6 are selected as a template material for demonstration of the nanoweb formation process and a typical nanoweb-rich example, respectively. Additionally, other solution systems such as poly(vinyl alcohol) (PVA)/ZnO and PVA/SiO2 also formed the nanowebs after electrospinning. Preliminary comparative study of PAA nanofibers and nanowebs was carried out by ammonia gas absorption and detection.

Experimental Preparation of Polymer Solutions for Electrospinning The starting materials included PAA (Mw 250 000, Wako), Nylon-6 (Aldrich), PVA (Mn 66 000, Wako), zinc acetate (Wako), SiO2 nanoparticles (12 nm), ethanol (Wako), distilled water, and formic acid (Wako). Four kinds of polymer solutions were prepared as the following procedures. PAA powder was dissolved in H2O, ethanol, and their blends, respectively, with concentrations of 6, 8, and 10 wt%. Nylon-6 was dissolved in formic acid with concentration of 10, 15, 20, and 25 wt%. A 10 wt% PVA solution was prepared from PVA powder and distilled water at 80 oC with vigorous stirring. The electrospinning PVA/ZnO solution can be obtained by blending 10 g PVA solution, 1 g zinc acetate, and 1 g distilled water at room temperature under stirring for 6 h. 0.2 g SiO2 nanopartilces was blended with 10 g PVA solution (10 wt%) under vigorous stirring and ultrasonic treatment. The viscosity and conductivity of blend solutions were measured by a viscotester (6L/R, Hakke, USA) and electric conductivity meter (CM-40G, TOA•DKK Co., Japan), respectively.

Fabrication of Nanofibrous Membranes via Electrospinning The polymer solutions were loaded into plastic capillaries which immersed with a copper wire. The copper wire was connected to a high voltage power supply (FC30P4, Glassman High Voltage Inc., USA) that was capable of generating DC voltage up to 30 kV. The ambient relative humidity was used as 20, 45, and 75 %, respectively. The spinning distance was regulated in the range of 5-25 cm. Various samples were obtained by adjusting the processing parameters during electrospinning. The fibrous samples can be deposited on Al foil and quartz crystal microbalance (QCM, 10 MHz, AT-cut quartz crystal with Ag electrodes) electrodes. The resultant samples were conducted by 5 s of Os coating under vacuum using an Os coater (HPC-1S, Vacuum Device Ltd., Japan). The thickness of Os coating layer on samples was less than 2 nm. Morphological observations of samples were made by scanning electron

54

Bin Ding, Chunrong Li, Dong Wang et al.

microscopy (SEM) (S-4700, Hitachi Ltd., Japan). The diameters of samples were measured using image analyzer (Adobe Photoshop 7.0).

Results and Discussion Finding of Nanowebs in Electrospun Fibrous Membranes SEM images of PAA samples produced with concentration of 6 wt%, voltage of 30 kV, distance of 15 cm, and relative humidity of 20 % as a function of kinds of solvents are shown in Figure 1. As shown in Figures 1A and B, several discontinuous irregular defect films with maximum length up to 20 μm were found among the electrospun PAA fibers, which were spun from the solvent of H2O and H2O/ethanol = 3/1 (w/w). The same phenomenon for observation of defect films was reported with electrospinning of other materials [8,12,13]. The defect films were formed from the charged droplets which originated from the jet at the capillary tip with increased instability. However, the defect films were partly split into webs when the PAA spun from volatile solvent of H2O/ethanol = 1/1 and ethanol (Figures 1C and D). The formation of webs from films was considered due to the fast phase separation of polymer and solvent in charged droplet caused by solvent evaporation during the flight in high electric field [45,46]. PAA has a relatively faster phase separation with ethanol than with H2O because of the low boiling point of ethanol [47].

(A)

(B)

(C)

(D)

Figure 1. Continued on next page.

Fabrication and Application of Novel Two-Dimensional Nanowebs…

55

(E)

Figure 1. SEM images of PAA fibers produced with PAA solution concentration of 6 wt%, voltage of 30 kV, distance of 15 cm, and relative humidity of 20 % as a function of kinds of solvents: (A) H2O; (B) H2O/ethanol = 3/1; (C) H2O/ethanol = 1/1; and (D) ethanol. (E) High magnification SEM image taken from the sample shown in D.

Figure 1E showed the high magnification SEM image of sample, which spun from ethanol. The strong bonding between nanowebs and electrospun fibers was found. The trace solvent left in the nanowebs caused by the uncompleted solvent evaporation led the bonding between nanowebs and electrospun fibers. Additionally, the nanowebs that covered on electrospun fibers still kept the morphology of nanoweb to form a porous surface on fibers indicating the formation of the nanowebs was performed during the flight before reaching collector.

Relative Humidity Effect on PAA Nanoweb Formation The ambient relative humidity in electrospinning chamber was important to affect the evaporation speed of solvent in charged droplets during the flight [48]. As the relative humidity increased to 45% and 75 %, the PAA fibers (Figures 2A and B) showed a relatively larger fiber diameter than that of fibers produced under a low relative humidity of 20 % (Figure 1D). And the adjacent fibers stuck together to form a porous film without the appearance of nanowebs.

(A)

(B)

Figure 2. SEM images of PAA fibers formed with PAA solution concentration of 6 wt%, voltage of 30 kV, distance of 15 cm, solvent of ethanol, and relative humidity of 45% (A) and 75 % (B).

56

Bin Ding, Chunrong Li, Dong Wang et al.

The high relative humidity retarded the evaporation of solvents from jets which led an increased fiber diameter and the wet fibers linked together on collector.

Effect of Applied Voltage on PAA Nanoweb Formation Another processing parameter which affected the PAA nanoweb formation was the applied voltage. As shown in Figure 3, there were no nanowebs for the sample produced with the applied voltage of 20 kV. As reported by Deitzel [8], the applied voltage was one of the key parameters to initial the bead defect formation. In this study, the applied voltage was further proved to be one of the key parameters to affect the formation of nanowebs.

Figure 3. SEM image of PAA fibers formed with PAA solution concentration of 6 wt%, voltage of 20 kV, distance of 15 cm, relative humidity of 20 %, and solvent of ethanol.

Proposed Mechanism The forces acting on the charged droplet which flight with a high moving speed in electric field are shown in Figure 4A. The forces included electrostatic force, drag force, gravity, coulombic repulsion force, surface tension and viscoelastic force. The electrostatic force carried the charged droplet from capillary tip to collector. The drag force happened between the surrounding air and charged droplet with high moving speed. And the drag force was attributed to deforming the droplets into films. The coulombic repulsion force tried to expand the droplet. The surface tension and viscoelastic forces led the contraction of charged droplet [49]. The electric field could be increased on increasing the applied voltage with a constant distance. Consequently, the electrostatic and coulombic repulsion forces of charged droplet were reinforced with increasing of electric field. The increased electrostatic force further accelerated the moving of charged droplet, which led an increased drag force. The distortion and expand of charged droplet from spherical-like to spindle-like in electric field during electrospraying was reported by Grimm and Beauchamp [50]. The further expand could happened when the electric field increased to form thin films from droplets with the effect of increased coulombic repulsion and drag forces. Moreover, the increased radial charge repulsion force also has trend to expand the charged films. As a result, the deformation of charged droplet was strongly affected by the electric field.

Fabrication and Application of Novel Two-Dimensional Nanowebs…

57

(A)

(B) Figure 4. (A) Forces acting on the charged droplet. (B) Schematic diagram illustrating the possible mechanism of nanoweb formation during electrospinning.

The schematic diagram illustrating the possible mechanism of nanoweb formation during electrospining was shown in Figure 4B. Due to the high viscosity of polymer solutions, the major jets ejected from the tip could be continuous (as in conventional electrospinning). We considered that the defect films or nanowebs were not formed from the charged droplets which were obtained by breaking jets (as in electrospraying). The defect films or nanowebs just could be regarded as a by-product caused by a high electric field induced instability of suspended charged droplets during electrospinning [8]. The microsized charged droplets [51] could be initialed together with the common electrospun fibers from the electrospining tip with a high instability. During the flight of charged droplet from capillary tip to collector, the charged droplet bore the comprehensive effects of the forces acting on it. On increasing the moving distance, the droplet was distorted and expanded into thin film by the coulombic repulsion and drag forces in the strong electric field. The splitting of thin film into nanoweb performed due to the fast phase separation between polymer and solvent which caused by the fast solvent evaporation under a low relative humidity. The fast phase separation led the

58

Bin Ding, Chunrong Li, Dong Wang et al.

spinodal or binodal types of phase morphologies within the fibers and the solvent rich regions were apparently transformed into pores [45]. As the electrospun fibers and nanowebs were formed at the same time, the 2D nanowebs stacked into 3D fibrous mats in a layer-by-layer structure as same as 1D electrospun fibers. The increasing of formation frequency and area density of nanowebs could be realized with increasing the instability of droplet at the electrospinning tip, such as increasing the applied voltage [8, 52].

PAA Solution Concentration Effect The morphologies of PAA nanowebs produced with various concentrations of PAA solution are shown in Figure 5. It could be observed that all the samples showed the uncompleted splitting of defect films. As shown in Figures 5A and B, the nanowebs produced with 6 wt% of PAA showed the largest average hole diameter and diameter deviation of nanowebs among three samples. As the PAA concentration increased to 8 wt% (Figures 5C and D), the hole diameter and shape of nanoweb became uniform and the average hole diameter reduced compared with the sample in Figures 5A and B. Moreover, as shown in Figures 5B and D, the shapes of holes in nanowebs included round, triangle, quadrangle, pentagon, and hexagon. Figures 5E and F showed the nanowebs produced with 10 wt% of PAA. The nanowebs showed almost only round shape of holes with the smallest average hole diameter among three samples. As a result, the shape, diameter, and uniformity of holes in PAA nanowebs could be adjusted by changing the solution concentrations.

(A)

(B)

(C)

(D)

Figure 5. Continued on next page.

Fabrication and Application of Novel Two-Dimensional Nanowebs…

(E)

59

(F)

Figure 5. SEM images of PAA fibers produced with applied voltage of 30 kV, distance of 15 cm, relative humidity of 20 %, and solvent of ethanol as a function of PAA concentrations: (A) 6 wt%; (C) 8 wt%; and (E) 10 wt%. (B, D, and F) High magnification SEM images taken from the sample shown in A, C, and E, respectively.

Effect of Voltage and Relative Humidity on Nylon-6 Nanoweb Formation The example for relatively completed splitting of films into nanowebs was demonstrated using nylon-6 dissolved in formic acid. As the key parameter for the formation of nanowebs, the influence of applied voltage was investigated during fabrication of nylon-6 nanowebs. When the applied voltage was 10 kV (Figure 6A), a web comprising several ultrathin nylon-6 nanowires appeared in a small region. As the voltage increased gradually from 15 to 25 kV (Figures 6B to D), the 2D web comprising many interlinked 1D nanowires appeared. The nanowebs almost covered all the regions in SEM image when the applied voltage was 30 kV (Figure 6E). It was found that the area density of nanowebs in fabric was sharply increased with increasing the applied voltage. This result was identical to the increase of bead defect density [8] with increasing the applied voltage. (A)

(B)

Figure 6. Continued on next page.

60

Bin Ding, Chunrong Li, Dong Wang et al.

(C)

(D)

(E)

Figure 6. SEM images of nylon-6 fibers formed with nylon-6 solution concentration of 15 wt%, distance of 15 cm, and relative humidity of 20 % as a function of applied voltage: (A) 10 kV; (B) 15 kV; (C) 20 kV; (D) 25 kV; and (E) 30 kV.

Unlike the PAA, the nylon-6 nanowebs could be fabricated under a high relative humidity of 75 % (Figure 7). This was attributed to the nylon-6 has faster solidification [53] speed than that of PAA [45]. Despite the observation of nylon-6 nanowebs under the high relative humidity of 75%, the area density of nylon-6 nanowebs was largely decreased compared with the sample fabricated under the low relative humidity of 20 % (Figure 6E).

Figure 7. SEM image of nylon-6 fibers formed with nylon-6 solution concentration of 15 wt%, voltage of 30 kV, distance of 15 cm, and relative humidity of 75 %.

Fabrication and Application of Novel Two-Dimensional Nanowebs…

61

Effect of Distance Figure 8 showed the influence of electrospinning distance to the morphology of nylon-6 nanowebs. As shown in Figure 8A (distance of 5 cm) and 8C (distance of 25 cm), the 2D nanowebs distributed randomly among electrospun fibers as the way that electrospun fibers stacked. The electrospun fibers acted as supporter to support the nanowebs. From the high magnification SEM image (Figure 8B), the short distance (5 cm) led a relatively larger wire diameter about 30-50 nm in nanowebs because of the uncompleted expand of the nanowebs. Meanwhile, the strong bonding between electrospun fibers and nanowebs was found due to the uncompleted solvent evaporation in this short spinning distance. As the distance increased to 25 cm (Figure 8D), the thinner wire diameter about 10-20 nm in nanowebs was found due to the relatively sufficient expand of nanowebs in this long spinning distance. As a result, the nanowire diameter in nanowebs and bonding between fibers and nanowebs were strongly affected by spinning distance.

(A)

(B)

(C)

(D)

Figure 8. SEM images of nylon-6 fibers formed with nylon-6 solution concentration of 15 wt%, voltage of 25 kV, relative humidity of 20 %, and distance of (A) 5 and (C) 25 cm. (B and D) High magnification SEM images taken from the samples shown in A and B, respectively.

Nylon-6 Solution Concentration Effect Figure 9 showed the SEM images of nylon-6 samples produced from various concentrations. At 10 wt% (Figure 9A), the electrospun fibers have the thinnest mean fiber

62

Bin Ding, Chunrong Li, Dong Wang et al.

diameter of 66 nm (Figure 10A) and a few nanowebs appeared. On increasing the concentration from 10 to 20 wt% (Figures 9A to C), the fiber diameter increased gradually up to 184 nm (Figures 10A to C) as well as increased the area density of nanowebs. However, at 25 wt%, the fiber diameter was sharply increased to 1 μm and the area density of nanowebs decreased (Figures 9D and 10D).

(A)

(B)

(C)

(D)

Figure 9. SEM images of nylon-6 fibers produced with applied voltage of 25 kV, distance of 15 cm, and relative humidity of 20 % as a function of nylon-6 concentrations: (A) 10 wt%; (B) 15wt%; (C) 20 wt%; and (D) 25 wt%. 100

100

(A)

Frequency distribution (%)

(B)

90

Average diameter: 66 nm Standard deviation: 14 nm

80

Frequency distribution (%)

90

70 60 50 40 30

70 60 50 40 30

20

20

10

10

0

Average diameter: 106 nm Standard deviation: 17 nm

80

0 0

25

50

75

100

125

150

175

200

225

250

0

25

50

Fiber diameter (nm)

75

100

125

150

Fiber diameter (nm)

Figure 10. Continued on next page.

175

200

225

250

Fabrication and Application of Novel Two-Dimensional Nanowebs…

100

100

(C)

90

(D)

90

Average diameter: 184 nm Standard deviation: 14 nm

Average diameter: 1.06 μm Standard deviation: 0.22 μm

80

Frequency distribution (%)

80

Frequency distribution (%)

63

70 60 50 40 30 20 10

70 60 50 40 30 20 10

0

0

0

25

50

75

100

125

150

Fiber diameter (nm)

175

200

225

250

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

Fiber diameter (μm)

1.5

1.6

1.7

1.8

Figure 10. Diameter distributions of nylon-6 fibers produced with applied voltage of 25 kV, distance of 15 cm, and relative humidity of 20 % as a function of nylon-6 concentrations: (A) 10 wt%; (B) 15wt%; (C) 20 wt%; and (D) 25 wt%.

The typical nanoweb-rich example of nylon-6 (shown in Figure 9C) formed from 20 wt% of nylon-6 solution were found like fishing net to cover the electrospun fibers completely and to form strong bonding with electrospun fibers. The high magnification of same sample was shown in Figure 11A. The dense nylon-6 nanowebs with uniform pore were formed. The pore diameter of nanowebs was in the range of 10-80 nm, which was much less than that of pores among electrospun fibers. The nanowire diameter distribution of nylon-6 nanowebs was shown in Figure 11B. The major distribution region (over 80 %) of nanowire diameters was 10-20 nm with average diameter of 17 nm. The standard deviation of the wire diameters in nanowebs was 5 nm. The average diameter of nanowires was one order of magnitude smaller than that of common electrospun fibers. Therefore, the surface-to-volume ratio of fibrous mats was expected to be increased with the appearance of nanowebs [36]. (A)

Figure 11. (A) High magnification SEM images of nylon-6 nanowebs formed with solution concentration of 20 wt%, voltage of 25 kV, distance of 15 cm, and relative humidity of 20 %. (B) Histogram showing the diameter distribution of nanowires shown in A.

64

Bin Ding, Chunrong Li, Dong Wang et al.

Gas Absorption and Sensing Properties of PAA Nanofibers and Nanowebs The gas absorption test was carried out in a closed box (9 L) which installed a fan (Figure 12A). Each 0.25 g fibrous PAA sample was examined with 10.5 ppm ammonia gas. The remaining concentration of ammonia was determined by ammonia gas testing tube. Figure 12B presents the ammonia gas absorption by PAA fibers and nanowebs. Due to the higher surface area of PAA nanowebs, the nanowebs showed the faster absorption speed and higher absorption ability than those of nanofibers. After 55 min absorption, the concentration of ammonia decreased from 10.5 ppm to 3.5 and 1.0 ppm for nanofibers and nanowebs, respectively.

(A)

(B) Figure 12. (A) Schematic of gas absorption system. (B) Gas absorption efficiency of PAA fibers and PAA webs.

Fabrication and Application of Novel Two-Dimensional Nanowebs…

65

(A)

(B) Figure 13. (A) Schematic of a gas testing system for NH3 detection. (B) Response of sensors containing PAA nanofibers and nanowebs exposed to 1 ppm NH3 at the relative humidity of 30 %.

Nanofibrous membranes were incontinuously collected on the surface of QCM until the required frequency shift (coating load) was got. The resonance frequencies were measured by a frequency counter (Hewlett Packard 53131 A). Then, the nanofibrous membranes coated QCMs were dried at 80 oC in vacuum for 2 h to remove the trace solvent prior to the subsequent characterizations. The QCM frequency shifts caused by the deposition of PAA fiber and nanowebs were 28 000 and 22 000 Hz, respectively. In the current work, the Sauerbrey equation for used QCM can be drawn as Δf = —Δm/0.96×10-9. It means that the frequency is decreased for 1 Hz in the case of 0.96 ng of gas molecules were absorbed. The flow-type experimental setup for measuring the sensing properties of sensors is shown in Figure 13A. The sensor was installed in the chamber which kept with constant temperature (25 oC) and relative humidity (30%). The N2 was used as carrier gas. The flow rates of dry N2, wet N2, and target gases were kept constant by mass flow controllers (MFCs, Estec, SEC-

66

Bin Ding, Chunrong Li, Dong Wang et al.

400 MK3). During the measurement, the concentration of NH3 was regulated at 1 ppm and the absorption time was 30 min. The sensor responses to target gases were examined by measuring the resonance frequency shifts of QCM which due to the additional mass loading. The resonance frequencies were measured by the frequency counter. The data from the sensors were recorded by a personal computer. After stabilization in N2 at relative humidity of 30%, both samples showed the response the 1 ppm ammonia gas (Figure 13B). The PAA nanofibers showed the maximum 6 Hz frequency shift for 30 min ammonia absorption and returned to the original frequency after the ammonia gas desorption. For nanoweb sample, it showed the faster response speed and higher maximum frequency shift of 15 Hz to 1 ppm ammonia gas, 2.5 times higher than that of nanofibers. However, it needs more time to get the maximum frequency shift and desorb the ammonia gas. As the nanofiber sensor already showed much higher sensitivity than flat sensing films [15,16], the nanoweb structure would be the better candidate for the future highly sensitive material mode.

Formation of Nanowebs in Other Solution Systems Figure 14 showed the morphology of the electrospun pure PVA fibrous films. As a typical electrospun fibrous film, the PVA fibers were randomly oriented as a porous film with a wide fiber diameter distribution. The PVA fiber was straight with an average fiber diameter of 239 nm.

(A)

(B)

Figure 14. SEM images of pure PVA electrospun nanofibers.

The SEM images of composite PVA/ZnO fibrous membranes are shown in Figure 15. It can be observed that the composite fibers have many junctions among the fibers, showing a poor fiber uniformity compared with the pure PVA fibers. The average diameter of PVA/ZnO fibers (258 nm) was larger than that (239 nm) of pure PVA fibers due to its increased viscosity (Table 1). It is well known that the morphology and properties of electrospun nanofibers are strongly influenced by the solution properties such as viscosity and conductivity [54]. Table 1 shows the viscosity and conductivity of the electrospinning solutions. Additionally, in Figure 15, the formation of nanowebs was observed among the fibers. The electrospun PVA/ZnO fibers acted as a support for the “fishnet-like” nanowebs comprising interlinked one-dimensional nanowires. The average diameter of the PVA/ZnO

Fabrication and Application of Novel Two-Dimensional Nanowebs…

67

nanowires (10 nm) contained in this nanoweb was about one order magnitude less than that of conventional electrospun fibers. The formation of the nanowebs was considered to be due to the electrically forced fast phase separation of the charged droplets which move at high speed between capillary tip and the collector. The phenomenon of nanoweb formation has been reported in our previous study [55]. Here, the observation of PVA/ZnO nanowebs was ascribed to the sharply increased conductivity from 0.019 to 0.712 S/m (Table 1).

(A)

(B)

Figure 15. SEM images of composite PVA/ZnO electrospun nanofibers.

Table 1. Properties of PVA and PVA/zinc acetate solutions Sample PVA PVA/zinc acetate

Viscosity (centipoises) 420 600

Conductivity (S/m) 0.019 0.712

Another example for nanoweb formation is the PVA/SiO2 nanowebs by electrospinning the suspension of PVA and SiO2 nanoparticles. SEM images of PVA/SiO2 nanowebs are shown in Figure 16. The formation of nanowebs probably was due to the increased instability by blending the high content nanoparticles in polymer solution.

(A)

(B)

Figure 16. SEM images of composite PVA/SiO2 electrospun nanofibers.

68

Bin Ding, Chunrong Li, Dong Wang et al.

Concusion In this chapter, we have reviewed the fabrication of PAA, nylon-6, PVA/ZnO, and PVA/SiO2 nanowebs that stacked as layer-by-layer and widely distributed in the 3D structure of fibrous mats. The development of PAA nanowebs indicated the formation process of nanowebs. The splitting of defect films into nanowebs occurred under the extreme processing conditions, such as the high applied voltage, low relative humidity, and fast phase separation between polymers and solvents. Meanwhile, the morphology and area density of nanowebs were found to be the comprehensive effect of various electrospinning processing parameters of relative humidity, applied voltage, kinds of solvents, distance, and solution concentration. The typical nanoweb-rich sample nylon-6 showed that the nanowire diameter in nanowebs was one tenth of common electrospun fibers. As the formation of defect films was usual phenomenon during electrospinning, other materials with electrospinability also have the possibilities to form nanowebs. Additionally, the PAA nanowebs showed the larger gas absorption capacity and higher sensitivity to ammonia compared with PAA nanofibers due to its relatively high specific surface. This review collected some data from our previous publication of reference 55.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]

J. Zeleny, Phys. Rev. 3 (1914) 69. A.Formhals, U.S. Patent Specification. 1934-1975504. D.H. Reneker and I. Chun, Nanotechnology, 7 (1996) 216. Y. Dzenis, Science, 304 (2004) 1917. Z. Huang, Y. Zhang, M. Kotaki and S. Ramakrishna, Compos. Sci. Technol. 63 (2003) 2223. B. Ding, H. Kim, S. Lee, C. Shao, D. Lee, S. Park, G. Kwag and K. Choi, J. Polym. Sci. B 40 (2002) 1261. G.I. Taylor, Proc. R Soc. Lond. A 280 (1964) 383. J.M. Deitzel, J. Kleinmeyer, D. Harris and N.C. Beck Tan, Polymer. 42 (2001) 261. L. Dejuan and J. Delamora, J. Colloid Interf. Sci. 186 (1997) 280. D. Chen and D. Pui, Aerosol Sci. Tech. 27 (1997) 367. J. Doshi and D. Reneker, J. Electrost. 35 (1995) 151. X. Zong, K. Kim, D. Fang, S. Ran, B.S. Hsiao and B. Chu, Polymer 43 (2002) 4403. C.J. Buchko, L.C. Chen, Y. Shen and D.C. Martin, Polymer. 40 (1999) 7397. X. Wang, C. Drew, S. Lee, K. Senecal, J. Kumar and L. Samuelson, J. Macromol. Sci. Pure, 39 (2002) 1251. B. Ding, J. Kim, Y. Miyazaki and S. Shiratori, Sensor Actuat. B 101 (2004) 373. B. Ding, M. Yamazaki and S. Shiratori, Sens. Actuator B 106 (2005) 477. M. Hajra, K. Mehta and G. Chase, Sep. Purif. Technol. 30 (2003) 79. S. Shiratori, B. Ding and M. Yamazaki, Japanese Patent Specification 2006-234386. W. Li, C. Laurencin, E. Caterson, R. Tuan and F. Ko, J. Biomed. Mater. Res. 60 (2002) 613. H. Yoshimoto, Y. Shin, H. Terai and J. Vacanti, Biomaterials 24 (2003) 2077. G. Wnek, M. Carr, D. Simpson and G. Bowlin, Nano Lett. 3 (2003) 213.

Fabrication and Application of Novel Two-Dimensional Nanowebs…

69

[22] K. Onozuka, B. Ding, Y. Tsuge, T. Naka, M. Yamazaki, S. Sugi, S. Ohno, M. Yoshikawa and S. Shiratori, Nanotechnology, 17 (2006) 1026. [23] H. Kokubo, B. Ding, T. Naka, H. Tsuchihira and S. Shiratori, Nanotechnology 18 (2007) 165604. [24] B. Ding, C.R. Li, Y. Hotta, O. Kuwaki and S. Shiratori, Nanotechnoligy 17 (2006) 4332. [25] Y. Miyauchi, B. Ding and S. Shiratori, Nanotechnology 17 (2006) 5151. [26] T. Ogawa, B. Ding, Y. Sone and S. Shiratori, Nanotechnology 18 (2007) 165607. [27] B. Ding, T. Ogawa, J. Kim, K. Fujimoto and S. Shiratori, Thin Solid Films 516 (2008) 2495. [28] M. Kanehata, B. Ding and S. Shiratori, Nanotechnology, 18 (2007) 315602. [29] D. Li, Y. Wang and Y. Xia, Adv. Mater. 16 (2004) 361. [30] P. Katta, M. Alessandro, R.D. Ramsier and G.G. Chase, Nano Lett. 4 (2004) 2215. [31] Theron, E. Zussman and A. Yarin, Nanotechnology 12 (2001) 384. [32] K. Yang, D. Edie and D. Lim, Carbon 41 (2003) 2039. [33] J. Gong, C. Shao, G. Yang, Y. Pan and L. Qu, Inorg. Chem. Commun. 6 (2003) 916. [34] B. Ding, J. Gong, J. Kim and S. Shiratori, Nanotechnology, 16 (2005) 785. [35] D. Li and Y. Xia, (2003) Nano Lett. 3 555. [36] B. Ding, H. Kim, C. Kim, M. Khil and S. Park, Nanotechnology 14 (2003) 532. [37] B. Ding, C. Kim, H. Kim, M. Seo and S. Park, Fiber. Polym. 5 (2004) 105. [38] R. Caruso, J. Schattka and A. Greiner, Adv. Mater. 13 (2001) 1577. [39] B. Ding, J. Kim, E. Kimura and S. Shiratori, Nanotechnology 15 (2004) 913. [40] B. Ding, K. Fujimoto and S. Shiratori, Thin Solid Films, 491 (2005) 23. [41] B. Ding, C.R. Li, S. Fujita and S. Shiratori, Colloids and Surfaces A: Physicochem. Eng. Aspects 284-285 (2006) 257. [42] D. Li, J. Mccann and Y. Xia, Small, 1 (2005) 83. [43] D. Li and Y. Xia, Nano Lett. 4 (2004) 933. [44] J.H. Yu, S.V. Fridrikh and G.C. Rutledge, Adv. Mater. 16 (2004) 1562. [45] M. Bognitzki, W. Czado, T. Frese, A. Schaper, M. Hellwig, M. Steinhart, A. Greiner and J. Wendorff, Adv. Mater. 13 (2001) 70. [46] S. Megelski, J. Stephens, D. Chase and J. Rabolt, Marcromolecules, 35 (2002) 8456. [47] B. Ding, M. Yamazaki and S. Shiraori, Sensor. Actuat. B 106 (2005) 477. [48] C. Casper, J. Stephens, N. Tassi and J. Rabolt, Marcromolecules, 37 (2004) 573. [49] C. Mit-uppatham, M. Nithitanakul and P. Supaphol, Macromol. Chem. Phys. 205 (2004) 2327. [50] R.L. Grimm and J.L. Beauchamp, J. Phys. Chem. B 109 (2005) 8244. [51] Z. Olumee, J. Callahan and A. Vertes, J. Phys. Chem. A 102 (1998) 9154. [52] S. Hong, J. Moon, J. Lim, S. Kim and S. Yang, Langmui. 21 (2005) 10416. [53] T. Young, D. Lin, J. Gau, W. Chuang and L. Cheng, Polymer. 40 (1999) 5011. [54] H. Fong, I. Chun and D. Reneker, Polymer. 40 (1999) 4582. [55] B. Ding, C.R. Li, Y. Miyauchi, O. Kuwaki and S. Shiratori, Nanotechnology, 17 (2006) 3685.

In: Nanotechnology… Editors: C.J. Dixon and O.W. Curtines, pp. 71-105

ISBN: 978-1-60692-162-3 © 2010 Nova Science Publishers, Inc.

Chapter 3

NANO-SCALE CHARACTERIZATION AND SPECTROSCOPY OF STRAINED SILICON Norihiko Hayazawa* and Alvarado Tarun Nanophotonics Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan

Abstract Strained silicon (ε-Si), the fundamental material of integrated circuit, is finding tremendous attention because it boosts the speed and reduces the power consumptions of electronic devices. However, poor homogeneity distribution of strain in ε-Si layers can degrade performance of electronic devices. Raman spectroscopy is used to study strain fluctuations in silicon because the optical phonons in Raman spectra are strongly influenced by strain. Though silicon are Raman active devices, the Raman efficiency of a nanometer layer of ε-Si is extremely weak and is often eclipsed under the Raman scattering of underlying buffer substrates. Micro Raman measurements show only uniform features in the nano-scale because of averaging effect from diffraction-limited spatial resolution. Here, we utilized surface enhancement in Raman scattering to overcome weak emission problems and to suppress averaging effect. Thin ε-Si layers were covered with thin Ag layer to invoke surface enhanced Raman spectroscopy (SERS). Results show that SERS effectively enhanced the Raman signal from ε-Si layer and it stands distinctly apart from the Raman signal originating from the buffer layer. This technique is promising but it lacks the spatial resolution in the nano-scale due to diffraction limit from the probing light. In order to achieve nano-scale spectroscopy, point-surface-enhancement was used, rather than a large surface enhancement. We used a silver-coated sharp tip, just like SERS, but only the sample region very close to the tip apex is characterized. This technique, known as the tip-enhanced Raman spectroscopy (TERS), provides nanometric resolution in our measurement. We observed localized strains by employing TERS. The TERS spectra revealed clear nano-scale variation in Raman frequency. Now that we can distinctively separate ε-Si from underlying buffer layer, signal-to-noise ration (SNR) needs further improvement. We improve TERS SNR in two ways: optical field enhancement using different metallic tip and background signals reduction arising from bulk materials. The tip-enhancement is more important for homogenous nano-materials or for samples with very weak signals whereas the background signal reduction is indispensable for nano-materials that consist of different thin layers with strong signals such as ε-Si or samples *

E-mail address: [email protected]. phone: +81-48-467-9339, fax: +81-48-467-9170. (Author for Correspondence)

72

Norihiko Hayazawa and Alvarado Tarun with strong signal level. Accordingly, we introduce several approaches mainly for the suppression of background signals arising from other bulk materials. We will discuss the utilization of UV light source, specialized tip, sample orientation relative to probing polarization, and depolarization configuration to obtain high contrast Raman signal. The characterization techniques describe above is applicable to other nano-materials.

1. Introduction Strained Si (ε-Si) on silicon germanium (Si1-xGex) substrate has now been widely used in state of the art electronic devices such as Intel dual core products primarily due to the mobility enhancement of electrons and holes in the ε-Si layer [1,2]. As ε-Si technologies become more complicated and transistors become smaller, in the order of nanometer (~45nm), issues involving strain defects and the characterization of key properties in silicon substrates are expected to increase. Thus, development of advanced technique for nanoscale characterization and investigation of localized strain in the ε-Si substrate are important in the current semiconductor technology. Raman spectroscopy has been widely used for the investigations of strain in semiconductors [3-5] because of its nondestructive capability to observe lattice vibrations sensitive to strains. However, due to the low Raman scattering efficiency, it becomes difficult to observe the localized strain in nanometer scale. Moreover, in case of thin ε-Si layer fabricated on Si1-xGex substrate, the Raman scattering of Si-Si vibrational mode from Si1-xGex buffer layer can easily eclipse the vibrational signals originating from the ε-Si layer. Because of these two difficulties, conventional Raman scattering is not suitable for strain measurements in these thin samples. Nanometer scale and surface selective sensitivities are required, and hence surface enhanced Raman scattering (SERS) [6] and tip-enhanced Raman spectroscopy (TERS) [7-13] turn out to be exceptionally promising spectroscopic techniques. In this chapter, we demonstrate how effectively and selectively enhance the Raman signal from the thin ε-Si layer on a thick underlying Si1-xGex layer using SERS. We show that the proposed technique can be straightforwardly used for TERS microscopy to investigate localized nano-scaled strain in ε-Si. In section 2, we discuss SERS from the thin ε-Si covered with thin Ag layer. Results show that SERS effectively enhanced the Raman signal originating from ε-Si layer and it stands distinctly apart from the Raman signal originating from the buffer layer. This technique is promising but it lacks the spatial resolution in the nano-scale due to diffraction limit from the probing light. In section 3, in order to achieve nano-scale spectroscopy, point-surface-enhancement was used, rather than a large surface enhancement. We used a sharp Ag-coated tip placed over the ε-Si and only the sample region very close to the tip apex is characterized. This technique, known as the tip-enhanced Raman spectroscopy (TERS), provides nanometric resolution in our measurement. We observed localized strains by employing TERS. The TERS spectra revealed clear nano-scale variation in the Raman frequency. Now that we can distinctively separate ε-Si from underlying buffer layer, signal-to-noise ration (SNR) needs further improvement. We improve TERS SNR in two ways: optical field enhancement using different metallic tip and background signals reduction arising from bulk materials. The tipenhancement is more important for homogenous nano-materials or for samples with very weak signals whereas the background signal reduction is indispensable for nano-materials that consist of different thin layers with strong signals such as ε-Si or samples with strong

Nano-scale Characterization and Spectroscopy of Strained Silicon

73

Raman signal level. Accordingly, we introduce several approaches mainly for the suppression of background signals arising from other bulk materials. We will discuss the utilization of UV light source and specialized tip in section 4 and then introduce depolarization configuration in TERS for the efficient suppression of background signals in section 5.

2. Surface Sensitive Detection by Surface Enhanced Raman Scattering (SERS) Recent developments in surface enhanced Raman scattering (SERS) has gained attention not only due to enormous (1014) level of Raman enhancement capable of single-molecule detection but also for surface sensitive characterization of nanometer structured surfaces. It was observed that huge increase in the Raman cross section was result of excitation of surface plasmon in metals. The most popular metal used in SERS studies is Ag due to intense plasmon resonance excitation produced by nanometer sized Ag in the visible wavelength. In this section, we utilize SERS to obtain strain information in a thin layer, ~30nm, of ε-Si over silicon germanium substrate.

2.1. Experimental Configuration for SERS Figure 1 illustrates the concept of the SERS experimental system for highly sensitive detection of strain in the ε-Si surface. The substrate used for growing ε-Si layer was prepared by doping Ge in a pure 〈100〉 oriented silicon substrate. The concentration of Ge in the host Si lattice was gradually increased up to 25 % to expand the lattice constant of the host Si lattice. In order to minimize inhomogeneous strain, a layer with 1.0 μm thick Si1-xGex buffer layer was grown on the prepared substrate at constant 25% Ge concentration. Finally, a 30-nmthick Si layer was epitaxially grown on the buffer layer to fabricate ε-Si. The crystal face for both the buffer and silicon layer was maintained at 〈100〉. Due to lattice parameter mismatch at the interface, the thin Si layer was under a constant strain. The thickness of this ε-Si layer has to be low otherwise the induced strain could be gradually relaxed with increasing the thickness. To detect Raman from a very thin layer, we invoked SERS effects by evaporating Ag with 8-10 nm thickness on the ε-Si layer under vacuum (~10-6 Torr). The topographic image is shown in the inset of Fig. 1. This image was obtained by contact-mode atomic force microscope (AFM) using a silicon cantilever (Nanosensor ATEC-CONT). The Ag film consists of plenty of Ag grains, with typical grain size of 20~40 nm in diameter and 10 nm in height. These fractal-like "pancake” structures of Ag grains provide almost uniform surface enhancement factor averaged within the diffraction limited focused spot of micro-Raman experiments. A CW blue laser (wavelength: 488 nm, power: 9 mW) was focused by an oilimmersion high numerical aperture (NA) objective lens (NA=1.4; 100x). The gap of Ag film layer and the sample was filled with oil that has a matching refractive index (1.515) as the objective lens. The Raman and SERS signals from the sample were collected by the same objective lens and detected by confocal Raman microscope (pinhole: 30 μm), in which the spectrometer (focal length: 560 mm, 1800 grooves/mm, blaze wavelength: 500 nm) was equipped with both, a cooled charge coupled device (CCD) camera for Raman spectra

74

Norihiko Hayazawa and Alvarado Tarun

measurements (such as the results in Fig. 2), and an avalanche photodiode for Raman image acquisitions (such as in Fig. 3).

Figure 1. Schematic of SERS microscopy for ε-Si and AFM topographic image of the Ag island film coated on the ε-Si.

2.2. Comparison of Raman and SERS Spectra Vibrational modes located at the center of the Brillouin zone can only be observed in the first-order Raman scattering due to the momentum conservation. Further, the selection rules for the 〈100〉 oriented Si crystal in backscattering geometry allow only LO phonon to be Raman active. Therefore, we expect to observe only LO phonon in first-order Raman scattering experiments, which should appear at 520 cm-1 for an unstrained or pure Si crystal. Since the sample consists of a ε-Si layer above Ge-doped Si buffer layer, we expect shift from the original frequency of LO phonon in both layers. The LO phonons in doped crystals can be easily modified depending on the atomic mass as well as the amount of the dopant. In Gedoped Si crystal, the LO phonon can have three different vibrations, namely the Si-Si mode arising from the vibrations of neighboring Si atoms, the Ge-Ge mode arising from the vibrations of neighboring Ge atoms, and the Si-Ge mode arising from the vibrations involving the bonds between a Si and a Ge atom. Since the amount of Ge is much smaller than the amount of Si in our samples, the LO mode corresponding to the Si-Si vibration is dominant. We need to concentrate only on Si-Si mode because it interferes with the Si-Si vibration coming from the thin ε-Si layer in the experiments and we will be discussed throughout the chapter. The vibrational frequency of this mode decreases almost linearly with increasing Ge concentration [14]. The bonds between two neighboring Si atoms are stretched when Ge is doped introducing strain in the lattice. This strain is responsible for the shift of LO mode frequency originating from the Si-Si vibration in the buffer layer. Similarly, when a thin Si layer is grown on Si1-xGex buffer layer, the Si-Si bonds in Si layer are stretched due to the lattice mismatch at the interface. If the thickness of the layer is not too large, the strain produced in the Si layer is transferred vertically throughout the surface. The amount of

Nano-scale Characterization and Spectroscopy of Strained Silicon

75

developed strain at surface can be different from that of the strain in the buffer layer even if the strain in both layers is caused by the presence of Ge atoms in the buffer layer. The penetration depth of blue-green probing light in Si is much larger than the thickness of the ε-Si layer in our sample [15]. Therefore, in Raman scattering experiments, the probing laser penetrates deep through the buffer layer. As a result, Raman scattering in the buffer layer is strongly excited. The intensity of LO phonon mode coming from the buffer layer is much more strong than the one coming from the ε-Si layer. This is a problem if the vibrational frequencies of the two phonons are close to each other. It becomes difficult to clearly observe the LO phonon corresponding to the ε-Si layer. To overcome this problem, we used the SERS technique to selectively enhance Raman at the surface and compared the results from those obtained by normal Raman spectroscopy. Figure 2 shows SERS (with a Ag film deposited on ε-Si layer) and Raman spectra (without a Ag film) of the sample. Both spectra were fitted by double Lorentzian functions as shown by dashed lines. The lower and higher Raman modes in both spectra are the LO phonon modes arising from the Si-Si vibration in Si1-xGex (x=0.25) and in ε-Si, respectively. The Raman spectrum is dominated by a strong peak at 504.9 cm-1, corresponding to the background LO phonon signal from Si1-xGex substrate, together with a very weak shoulder at 513.8 cm-1 that arise from the LO phonon in ε-Si [16-20]. Since the Raman signal from the ε-Si is very close to the Si1-xGex buffer layer and considering 30 nm thickness of the ε-Si layer, the Raman signal from ε-Si is almost overwhelmed by the signal from the Si1-xGex layer. On the other hand, in SERS spectrum, the Raman signal from ε-Si is remarkably enhanced. The signal becomes even stronger than the Si1-xGex buffer layer due to the SERS effect. Even though the ε-Si Raman signal in SERS is enhanced relative to the signal from the buffer layer, we noticed reduction in the scattering efficiency in SERS with a signal bias of ~20000 counts/60s compared to the Raman spectrum. The reduction in overall scattering efficiency is due to the lower transmission of the probing light caused by Ag island film whereas the bias signal is due to the white continuum emission from the Ag island film, which is often accompanied in SERS measurements [21].

Figure 2. SERS (with Ag film) and Raman (without Ag film) spectra of ε-Si. Both spectra are fitted by double Lorentzian functions representing Raman modes of Si1-xGex (lower frequency) and ε-Si (higher frequency). The meshed areas are used for Raman imaging in Fig. 3.

As seen from Fig. 2, we resolved Raman signal from the ε-Si in SERS. This suggests that SERS effect is very powerful to observe the surface condition of the thin ε-Si layer. Both SERS and Raman measurements were done in the same condition except the existence of Ag

76

Norihiko Hayazawa and Alvarado Tarun

island film to invoke SERS. It should be noted here that we did not use any analyzer for polarization dependent detection in either SERS or Raman detection. The ratio of the intensities between Si1-xGex and ε-Si would change when measured through an analyzer, which will be discussed later in Section 5.

2.3. Comparison of Raman and SERS Images Next, we carried out confocal micro-Raman imaging of the same sample at specific vibrational modes from Si1-xGex and ε-Si. Figure 3(a) shows the reflection confocal image of the sample from a region close to the edge of the evaporated Ag film. As expected, the Ag covered layer appears to be brighter in the reflection image. Figure 3(b) was obtained by setting the spectrometer at the lower frequency in Fig. 2, so that the avalanche diode detects only the Raman signal from Si1-xGex buffer layer. Since the intensities level of SERS and Raman from Si1-xGex buffer layer is almost similar from Fig. 2, Raman image in Fig. 3(b) also reflects low contrast between Ag coated and uncoated area. However, when the spectrometer is set to detect only the higher frequency shift arising from ε-Si (see Fig. 2), the enhanced Raman signal from Ag coated ε-Si has high contrast as depicted in Fig. 3(c). Fig. 3(c) does not reflect any particular information about the possible nanometric strain distribution in ε-Si, which could be varying within areas smaller than the focal spot. This is because the enhanced scattering signal in this experiment is averaged within the diffraction

Figure 3. (a) A reflection confocal and Raman images obtained at Raman shift of (b) Si1-xGex and (c) εSi as indicated by meshed area in Fig. 2. (d) is obtained at Raman shift of ε-Si, the same as (c) but with different polarization of excitation light. An arrow in the figure indicates polarization of an excitation light in each image. The dimension of each image is 45 μm by 45 μm consisting of 100 pixels by 100 pixels.

Nano-scale Characterization and Spectroscopy of Strained Silicon

77

limited focused spot. Accordingly, when polarization of the excitation light is rotated 90 from the previous polarization as indicated in the Fig 3(d), no significant change in the contrast could be observed. The entire signal level in Fig. 3 (d) is decreased compared to that in Fig 3(c) due to the polarization dependence of crystalline silicon 〈100〉 facet [22,23] and also due to the different detection efficiency of the experimental system. Here, we note that the temperature dependence of Raman scattering in silicon [24] is not observed in our experiments under room temperature conditions. However, the temperature dependence is also an important factor that can influences the properties of ε-Si substrates and could be observed under the experimental conditions of SERS.

3. Surface Sensitive and High Spatially Resolved Detection by Tip-Enhanced Raman Scattering (TERS) In SERS measurement, high surface sensitivity and signal enhancement was successfully achieved, however, the spatial resolution is still diffraction limited. Moreover, SERS is an irreversible or destructive detection process because it requires Ag coating on the sample. For non-destructive and higher spatial resolution measurement capable of detecting localized nanometric strain, a sharpened metallic tip [25-27] can be utilized instead of Ag island film. In this section we introduce the use of “tip-enhancement” instead of “surface enhancement” to go beyond the diffraction limit of probing light. The concept of TERS is rather straightforward that can be understood by considering the sharp apex of a metallic tip as a point source surface enhancer [7-10].

3.1. Experimental Configuration of Reflection-Mode TERS Figure 4 shows the concept of reflection mode TERS microscopy [28-31] for opaque sample such as ε-Si layer assembled on thick substrate. There are generally two configurations for TERS, transmission mode and reflection mode. In transmission mode [713,32], the metallic probe tip is illuminated through the sample using objective lens. This configuration has been widely used in TERS because of its high efficiency in both illumination and collection, especially when using a high numerical aperture (NA) objective lens. However, transmission mode TERS cannot be used for opaque or thick samples. In reflection mode TERS [28-31], the tip is illuminated from the same side in which it is approached towards the sample (Fig. 4). This is advantageous and promising for observing opaque samples such as silicon substrates. A schematic of the experimental configuration is shown in Fig. 5 [33,34]. An incident CW YAG laser (532 nm, JDS Uniphase) is directed to the optical set-up through a single mode optical fiber. The light coming out of fiber is set to p-polarization, parallel to the tip axis, for optimum tip-enhancement [28] from the combination of a half-waveplate and a polarizer. The beam diameter of the laser is then expanded to ~10 mm. This expanded and polarized light is focused on the tip and sample using a long-working-distance (LWD) objective lens (Mitsutoyo, NA=0.28, 20x, WD=30.5 mm). The illumination optics mentioned above, including the LWD objective lens, is precisely positioned by three-axis actuators (New Focus Inc. Picomotor, accuracy 45o – implying the instrumental broadening is more significant for the (200) plane spectrum rather than for the (111) plane spectrum (Figure 7); (c) there is not any obvious plane distortion and bending respectively for the (111) and (220) planes in the HR-TEM micrograph (Figure 8(a)) – implying the negligible attribution from microstrain broadening. The above analyses therefore suggest that the peak broadening is mostly attributed to the smaller-grain size. They also affirm that the calculation by Eq.(2) really gives a reliable guide to the trend of the grain-sizes. The layer number N lapped normal to a crystalline plane can also be calculated theoretically. The calculated N lapped normal to the (111) plane of the TiN grain by N111=D111/d111=12.7/0.244 =52 (layer) implies the structure of a TiN111 grain in a crystallographic plane normal to the (111) plane to be constituted by 52 layers. Such structure gives the highest density for the atomic arrangement and induces hindering to the dislocation within the TiN nano-crystallines so as to enhance the film macroscopic mechanical behaviors.

Figure 7. XRD spectra of nanocrystalline superhard TiN film.

520

Xiang Yu, Chengbiao Wang, Meng Hua et al.

The XRD spectra (Figure 7), HR-TEM micrograph (Figure 8(a)) and electron diffraction pattern (Figure 8(b)) confirmed the three main crystallographic planes (111), (200) and (220) of the TiN films. Our study (Table 2) showed that: the obtained value of the relative intensity I/Io of the maximum characteristic peak for a standard TiN film was 100 for the (200) plane at 2θ ≈ 42.4o, 75 for (111) plane, and 55 for (220) plane; whereas the obtained intensity value Ik for a superhard TiN film was 100 for the (111) plane at 2θ ≈ 36.5o, and the minimum characteristic peak of 2.5 for the (200) plane. Hence, it is derived that the superhard TiN films deposited by the ion-plating system have a preferred orientation in (111) plane. Moreover, the maximum characteristic peak is shifted to the plane (111) at 2θ ≈ 36.5o for the superhard TiN film from the (200) plane at 2θ ≈ 42.4o for the standard TiN film in PDF card 6-642. Comparing the relative intensity I/Io values in the column 6-642 for the standard TiN film and in the column Ik for the superhard TiN film (Table 2), it shows a striking difference of the intensity between their maximum and minimum characteristic peaks. Typically, their respective planes have mutually swapped over. The preferential crystalline orientation of the nano-crystalline TiN film is strong on the plane (111) and weak on the planes (200) and (220). With the aim of alleviating the effect of steel substrate on the diffraction peaks, STD diffraction at αo＝0.6º, 3º and 5º were respectively conducted and their XRD spectra were correspondingly labeled as (2), (3) and (4) in Figure 7. The location of lower azimuth angle and the shallower penetration depth in Figure 7 suggests that the elimination of the substrate effect was advantageous for the analyses. It can be observed from Figure 7 that: (i) there is an obvious peak for the substrate in the case (1); (ii) a reduction in substrate peak intensity, in cases (2) to (4), can be clearly seen; (iii) the lowest X-ray depth observed in the case (4), which suggests its substrate effect is minimum. The difference in intensities for crystallographic planes (200), (220) and (311) of the superhard TiN film as seen from the STD diffractions confirms that preferred growth has taken place at the close-packed plane (111) for the superhard TiN film. The TiN material belongs to NaCl type face-centred cubic structure that has a skeleton of Ti atoms as the crystal lattice and N atoms in its interstices. The densest film structure is in the (111) plane rather in the planes (200) and (220). The preferred orientation and preferred growth being in the close-packed plane (111) of the superhard TiN film may mainly be due to the concentration of crystalline grains at the (111) plane and its associated uniform arrangement. Such a high degree of order plane subsequently provides the densest and hardest structure when compared with the planes (200) and (220). Figure 8 shows the HRTEM micrograph and the electron diffraction pattern of a selected area of the multi-grains on a deposited TiN film. It illustrates the existence of preferred crystalline orientation on plane (111) and the occurrence of local orientation at the grain interface on (220) (Figure 8a). The electron diffraction pattern of the correspondingly selected area (Figure 8b) confirms the existence of an obvious nano-polycrystalline character for the deposited TiN film. As distinguished from the ring-shaped diffraction rings obtained from amorphous grain planes, Figure 8b shows a distinct arc-shape cluster of diffraction rings for the crystalline planes (111), (200) and (220) with the brightest ring being for the (111) plane.

Nanocrystallite Superhard Titanium Nitride Film in Multi-arc Ion Plating

521

( c)

1μm

Figure 8. (a) HR-TEM micrograph, (b) selected area electron diffraction pattern and (c) cross-sectional micrograph of the TiN film.

It is thus reasonable to derive that these planes, particularly the close-packed (111) plane, are those for the TiN nanocrystalline grains to grow. Figure.8c shows a cross-sectional micrograph of a TiN film. It can be seen that the film has a dense and fine microstructure, and there is no sight of a columnar character. The HRTEM micrograph shows a fine, uniform and nanometer-scale crystalline size for the superhard TiN film, which is consistently in agreement with the traces of XRD analysis in Figure 7. During the formation of the film, the high flux and energy of the energetic ion bombardment enhances the mobility and dispersive capacity of the adsorptive particles and consequently facilitates the arrangement of the orderly atoms. Calculation by Eq.(1) with the relevant measurements gave the degree of preferred orientation (POD) on the plane (111) of the deposited superhard TiN film as 3.59. This calculated POD value is much higher than that for the reported sputtered TiN film [17] and serves to verify the methods used for the accomplishment of a preferred orientation of the TiN film.

522

Xiang Yu, Chengbiao Wang, Meng Hua et al.

Figure 9. Correlation among film hardness, PODs and gas pressures & bias voltages.

To reveal the natural characteristics of the nanocrystalline TiN film, a correlation of the experimental hardness and structure to their process parameters was performed. Experimentally, it was found that by varying the process parameters of substrate bias voltage and gas pressure a superhard TiN film was produced which exhibited a microstructure with (i) a preferred orientation and preferred growth in the close-packed plane (111) (that can be expressed in terms of POD) and (ii) gave mean grain-sizes on a nm scale. A typical correlation among the TiN film microhardness, the PODs of TiN111, the gas pressure as well as the bias voltages is shown in Figure 9, on which an indent image illustrates the geometry and dimensions of an indented superhard TiN film having hardness of 45 GPa. The depth of indentation was made specifically in the range of 0.140 ∼ 0.34 μm so as to minimize the influence of substrate on the measured film hardness. The deviation of the microhardness measurements so obtained was estimated and found in an order of 10 %. The result in Figure 9 shows that: (i) the superhard TiN films have the structure character of POD beyond 3.1; and (ii) suitable selection and simultaneous adjustment of gas pressure (in the range of 0.10 Pa ~ 0.30 Pa) as well as pulsed bias voltages (in the range of -100 ~ -250 V) allowed the formation of superhard TiN films with the required structure. The AIP deposition method is likely to activate high ionization in the vacuum chamber. Hence, biasing the substrate leads to the generation of a glow discharge surrounding the ion sources and the substrate holder. The effect is to enhance further the ionization and reactivity of nitrogen gas. The majority of the nitrogen under such circumstance reacts with titanium atoms in substrate surface, and with

Nanocrystallite Superhard Titanium Nitride Film in Multi-arc Ion Plating

523

discharging plasma and target surface. Although the application of biasing increases the density of excited radicals on the substrate and suitably activates high-energy ion bombardment to the substrate, it may however increase the residual stress in the TiN films and result in the loss of the required special structural properties when the bias voltage is above -350 V. Consequently, it degrades the mechanical properties of the film. In our experiment, the AIP deposition was set to a steadily low gas pressure of 0.1 Pa. It thus gave a relatively pure deposition environment and it also increased the mean free paths of the favorably excited particles. Optimization of the process parameters like the substrate bias voltage and the gas pressure thus facilitates the formation of superhard TiN films. The results of this study therefore indicate that the PODs of TiN111 are a good indication for judging whether the microhardness of the TiN films will increase or decrease. The phenomenon of hardness enhancement of the nanocrystalline TiN film is closely related with the ion flux and energy associated with the bombardment by the energetic particles in the film growing process. Estimation by Veprek et al [18] indicated that such a type of induced film superhardness would degrade the TiN film intrinsic hardness to, or below 22 GPa when its residual stress was reduced after the deposition period by an annealing treatment. Data in our experiments suggested that there was no visible decrease in the hardness values of the TiN films after one-month deposit under room temperature. Our experiments showed that a 45 GPa TiN film reduced its hardness to 36 GPa (i.e. 21 percent decrease) when it was annealed in a furnace at 600 oC for 0.5 h. It was only 21 percent decrease and the hardness value after annealing was still much higher than the intrinsic hardness of a standard TiN counterpart. The measured value of the residual stress of the film before and after annealing was 2.43 GPa and 1.54 GPa, respectively. Such a reduction suggests that annealing may result in grain growth and an associated decrease in hardness of the film. Consequently, it leads us to conclude that the residual stress induced in the film by the energetic ion bombardment can also make a partial contribution to the film hardness enhancement.

4. Conclusion An adherent nanocrystalline superhard TiN film was deposited on the substrate of M2 high-speed steel using a multi-arc ion plating system that was operated under (i) a low depositing pressure of 0.2 Pa and (ii) the energetic bombardment of the high ion flux and ion energy. The nanocrystalline superhard TiN film so deposited possesses a microstructure characteristic of nanometer scale preferred crystalline orientation on the plane (111), which is subsequently inducing superhardness of the film, as confirmed by the analyses of both XRD and TEM. Our studies have illustrated that it is possible to produce superhard naonosrystallite TiN films with microhardness of 45 GPa. The studies have also indicated that it is possible to correlate the preferred orientation on the close-packed plane (111) and the growth of TiN layers with the film hardness. Furthermore, the results of TEM studies on superhard TiN film are consistent with that of XRD analysis –This supports that our approach in using AIP system for producing superhard TiN film is feasible and that the relevant mechanisms found are valid. We believe that such findings contribute significantly to the scientific understanding in the field of thin film hardness enhancement.

524

Xiang Yu, Chengbiao Wang, Meng Hua et al.

Acknowledgments The support from (i) Supported by Program for New Century Excellent Talents in University (NCET), (ii) A Foundation for the Author of National Excellent Doctoral Dissertation of PR China, (iii) Strategic Research Grant of City University of Hong Kong (CityU: 7002235), (iv) Major Project of International Scientific Cooperation Plan of China (2006DFB51260) is greatly acknowledged and (v) Open Research Foundation of National Scientific Drilling Laboratory of China University of Geosciences (NLSD200705).

References [1] G.S. Kim, S.Y. Lee, J.H. Hahn, B.Y. Lee, J.G. Han, J.H. Lee and S.Y. Lee, Surf. Coat. Technol. 2003, 171, 83-90. [2] W.M. Posadowski, Thin Solid Films 2001, 392 (2), 201-207. [3] U. Krause, M. List and H. Fuchs, Thin Solid Films 2001, 392, 196-200. [4] X.Yu, C. B.Wang, Y. Liu and D. Y. Yu, Acta. Metall. Sin. 2006, 42 (6), 662-666. [5] T. Hanabusa, K. Kusaka, T. Matsue, M. Nishida, O. Sakata and T. Sato, Vacuum 2004, 74, 571-575. [6] T. Nishikiori, T. Nohira and Yasuhiko Ito, Thin Solid Films 2002, 408, 148-154. [7] P. F. Mcmillan, Nature 2004, 430, 738. [8] S. Q. Hao, B. Delley, S. Veprek and C.Stampfl, Phys. Rev. Lett. 2006, 97, 086102. [9] K. Reuter and M. Scheffler, Phys. Rev. B 2002, 65, 035406. [10] S. Ma, J. Prochazkab, P. Karvankova, Q. Ma, X. Niu, X. Wang, D. Ma and K. Xu, Surf. Coat. Technol. 2005, 194, 143-148. [11] M. D. Huang, G. Q. Lin, Y. H. Zhao, C. Sun, L. S. Wen and C. Dong, Surf. Coat. Technol. 2003, 176, 109-114. [12] Novikov, R. Riedel, R. Solozhenko and Y. Zhao, Nat. Mater. 2004, 3, 576. [13] X.Yu, C. B.Wang, M. Hua, P. Tam, Y. Liu and D. Y. Yu, Surf. Rev. Lett. 2007, 14(4), 789-793. . [14] Ming-Hua Shiao, Sui-An Kao and Fuh-Sheng Shieu, Thin Solid Films 2000, 375, 163169. [15] T. Matsue, T. Hanabusa and Y. Ikeuchi, Vacuum 2004, 74, 647-653. [16] X.Yu, M.Hua, C. B. Wang, Z. Q. Fu and Y. Liu, Appl. Surf. Sci. 2007, 253 (7), 37053711. [17] X.Yu, C. B.Wang, M. Hua, Y. Liu and D. Y. Yu, Nanotechnology 2007, 18: 355710. [18] S. Veprek, G. J. Maritza, V. Heijman, P. Karvankova and J. Prochazka, Thin Solid Films 2005, 476, 1-29

In: Nanotechnology… Editors: C.J. Dixon and O.W. Curtines, pp. 525-557

ISBN: 978-1-60692-162-3 © 2010 Nova Science Publishers, Inc.

Chapter 17

EMBEDDED OPTICAL-ELECTRICAL NANOMATERIALES FABRICATED BY ION IMPLANTATION X.T. Zu and X. Xiang Department of Applied Physics, University of Electronic Science and Technology of China, Chengdu, 610054, People’s Republic of China

S. Zhu and L.M. Wang Department of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor, MI 48109-2104, USA

Abstract Ion implantation provides a versatile and powerful technique for synthesizing nanometer-scale clusters embedded in the near-surface region of a variety of host materials. The embedded nanoparticles have attracted considerable attention because of their unique optical-electrical properties that are different from those of the bulk matrix. Metallic nanoparticles embedded in insulators have pronounced optical effects, including surface plasma resonance (SPR) absorption, and strong third-order nonlinear optical (NLO) susceptibility. The former suggests applications as optical filters, including eye-glass coatings. The latter has potential application in all-optical-memory or switching devices. Oxide nanoparticles have good photoluminescence. They have promising application in light-emitting devices. Magnetic metallic nanoparticles often show a ferromagnetic behavior with a larger coercivity than that of the corresponding bulk materials, which may provide potential application of the nanocomposite as magneto-optical materials for a high density magnetic data storage device. In this data review, nanoparticles embedded in insulators, e.g., Al2O3, MgO, YSZ and TiO2 single crystals, were fabricated by ion implantation and subsequent thermal annealing, including metallic Ni, Zn and their oxides, and intermetallic nanoparticles. Optical, magnetic and mircostructural properties of nanoparticles have been studied. The metallic nanoparticles have surface plasmon resonance absorption, and oxide nanoparticles show good photoluminenscence. The magnetic nanoparticles, e.g., metallic Ni and intermetallic CoxNi1-x nanoparticles, show strong ferromagnetism behaviors. The ion fluence can affect the concentrations and the intensities of the surface plasmon absorption of metallic nanoparticles. Ion flux is another important parameter to fabricate nanoparticles. An example of effects of ion flux on the nanoparticles has been presented in this data review. The relationship between

526

X.T. Zu, X. Xiang, S. Zhu et al. annealing temperature and optical, magnetic and microstructural properties of nanoparticles has also been systematically studied.

I．Introduction Nanostructured materials are promising to play a dominant role in future technology as they possess different, and often unique, properties relative to their macroscopic counterparts. There has been a great deal of recent interest in the incorporation of nanoparticles into dielectric and semiconductor materials to form nanocomposites. Metallic nanoparticles embedded in insulators have been extensively studied because of pronounced optical effects, including surface plasmon resonance (SPR) absorption and strong third-order nonlinear optical (NLO) susceptibility [1]. These composites have drawn much attention due to applicability for all-optical-memory or switching devices and single electron transistors [2], etc. Magnetic metallic nanoparticles often show a ferromagnetic behavior with a larger coercivity than that of the corresponding bulk materials. Ferromagnetic nanoparticles have potential application in magnetic storage devices [3, 4]. Oxide nanoparticles have good photoluminescence. They have promising application in the light-emitting devices. Various synthesis methods have been attempted to synthesize these nano-phases, such as, surface sputtering [5], pulsed laser deposition [6, 7], sol–gel [8] and ion implantation [9–12]. Among these techniques, ion implantation is one of the most reliable and effective methods for the synthesis of nano-scaled particles. The implanted ions frequently precipitate as nanoparticles in controlled concentrations in the near surface regions of the host materials. Almost any element in the periodic table can be implanted into virtually any selected host material. The average precipitate size can be controlled by varying the implantation and annealing conditions (such as dose, dose rate, energy, temperature and annealing time) at pre-calculated depths of the host matrices [1]. Oxide crystals are often stable substrates used in a large range of technological applications. These high stabilities make them suitable candidates to allow the controlled formation of colloidal dispersions of metallic precipitates using ion implantation. In general, ion implantation techniques used to form nanoclusters may be categorized as follows [13]: (1) room temperature implantation, followed by high temperature annealing; (2) room temperature implantation at dosage above the threshold dose for spontaneous nanocrystals formation; (3) ion implantation at elevated temperatures. In this data review, metallic nanoparticles were prepared by means of the second; oxide nanoparticles were prepared by the second method of ion implantation and thermal annealing; the intermetallic nanoparticles participated spontaneously by sequential implantation of two ions at a high dose. This data review will be concerned primarily with the optical, magnetic and microstructural properties of metallic, oxide and intermetallic nanoparticles prepared in single crystals by room temperature ion implantation combined with subsequent thermal annealing. The substrates, metallic ion sources and implantation parameters used in this data review are listed in Table 1. X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM) analyses are used to characterize microstructural properties of nanoparticles. Optical absorption and room temperature photoluminescence (PL) measurements are used to obtain the optical properties of nanoparticles. The magnetic

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

527

properties of the samples were characterized by a MPMS superconducting quantum interference device (SQUID) magnetometer. Table 1. Substrate and ion implantation parameters Material α-Al2O3 (0001) TiO2 (001) MgO (100) YSZ (001)

Ion Ion energy Ion fluence (cm-2) species (keV) Ni 64 1×1017 Zn 48 0.1, 0.5, 1, 5×1017 Ni 64 1×1017 Ni 64 1×1017 Ni 64 1×1017 Co+Ni

90

1×1017

Flux (μA/cm2) or dose rate Reference (ions/cm2 s) 2 5, 10 μA/cm [14, 15] 5 μA/cm2 [16, 17] 5 μA/cm2 [18] 5 μA/cm2 [19,20] 5 μA/cm2 [21] 2.93×1013 ions/cm2s for Co, [22] 2.63×1013 ions/cm2s for Ni

Abbreviations in This Data Review BF EDS EELS FC HAADF HREM MEVVA MPMS NLO PL SAED SPR SQUID STEM TEM UV-VIS XPS XRD YSZ ZFC

bright-field energy dispersive spectroscopy energy electron-loss spectroscopy field-cooling high-angle annular dark-field high-resolution electron microscopy metal vapor vacuum arc magnetic property measurement system nonlinear optical photoluminescence selected area electron diffraction surface plasma resonance superconducting quantum interference device scanning transmission electron microscopy transmission electron microscopy ultraviolet-visable X-ray photoelectron spectroscopy X-ray diffraction yttria-stabilized zirconia (with 9.5 mol. % Y2O3 in this datareview) Zero-field-cooling

II．Conclusion 2.1. Ni and NiO Nanoparticles Embedded in Single Crystals Ion implantations were conducted at room temperature using a metal vapor vacuum arc (MEVVA) implanter. The samples were tilted off-axis by around 7 degrees to avoid channeling implantation. After Ni ion implantation, the Al2O2 and YSZ samples were

528

X.T. Zu, X. Xiang, S. Zhu et al.

annealed in oxidization in order to study the formation of oxide nanoparticles, and the MgO sample was annealed in reducing atmosphere to study growing up of metallic nanoparticles.

2.1.1. Chemical Charge States of As-Implanted and Annealed Samples XPS measurements were performed to characterize chemical charge states of Ni element in as-implanted and annealed crystals. The C1S peak at 285.0 eV is used to calibrate the spectra. The samples had been etched 2 nm with Ar+ ion before measurements in order to remove surface contamination. 2.1.1.1. XPS Results of Al2O3 Figure 1 shows the XPS spectra of Ni2p3/2 energy level of the as-implanted and annealed Al2O3 crystals at etching depth of 2 nm, respectively [14]. The as-implanted spectra can be resolved into two Gaussian components. The peak at 852.3 eV (Figure 1a) is attributed to metallic Ni0, and the peak at 854.4 eV (Figure 1b) is due to Ni2+ (NiO). It is clear that the Ni element is prominent in charge state of metallic Ni0 in as-implanted crystals and Ni2+ in annealed crystals at 900 oC in ambient atmosphere. After etching 12 nm, there is no obvious change for the as-implanted and annealed samples.

a

864

b

856

848

864

856

848

Binding Energy/eV Figure 1. XPS spectra of Ni2p3/2 core level of Ni+-implanted Al2O3 at an ion flux of 5 μA/cm2 before (a) and after annealing at 900o for 1h in ambient atmosphere.

The XPS results above are obtained from the Ni-ion-implanted samples at ion flux of 5μA/cm2. The XPS result from Ni-ion-implanted Al2O3 at ion flux of 10μA/cm2 is shown in Figure 2. The peak at a binding energy of 856.6 eV can be attributed to Ni2+ in NiAl2O4 and the one at 862.8 eV is assigned to the well known shake-up satellite peak of Ni2+. The two peak positions and their energy difference 6.2 eV are just consistent with the previous study [23]. Thus, the XPS result indicates the formation of NiAl2O4 with the spinel structure when ion implantation conducted at a flux of 10μA/cm2.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

873

864

855

529

846

Binding Energy/eV Figure 2. XPS spectra of Ni2p3/2 energy level of Ni+-implanted Al2O3 crystals at a flux of 10μA/cm2.

2 nm

a

12 nm

864

2 nm

b

12 nm

855

846 864 Binding Energy/eV

855

846

Figure 3. XPS spectra of Ni2p3/2 energy level of Ni ion-implanted (a) and annealed (b) YSZ crystals at 900oC in ambient atmosphere.

530

X.T. Zu, X. Xiang, S. Zhu et al.

2.1.1.2 XPS Results of YSZ Figure 3 shows the XPS spectra of Ni2p3/2 energy level of the as-implanted and annealed YSZ crystals at different etching depths, respectively [21]. The spectra can be resolved into Gaussian components. In the as-implanted spectrum at etching depth of 2 nm, the peak at 852.5 eV is attributed to metallic Ni0, the peak at 855.6 eV may be due to Ni2+ (NiO) or Ni3+ (Ni2O3), and the peak at 860.6 eV could be assigned to the shake-up satellite peak of Ni0, Ni2+ or Ni3+. But after etching 12 nm, the charge state of Ni is only Ni0. These results show that the implanted Ni in the surface is easy to be oxidized. This is not same as the Ni ion implanted αAl2O3 single crystals, whose Ni on the surface are mainly in charge state of metallic Ni0 [14]. In the annealed spectrum at etching depth of 2 nm, the peak at 855.5 eV shows Ni is only in charge state of Ni2+ (NiO) or Ni3+ (Ni2O3). The implanted Ni ion had entirely been oxidized in the surface of YSZ matrix. However, the metallic Ni0 (the Gaussian peak at 853.0 eV) appeared after etching 12 nm. This result shows that there is still some metallic Ni0 retained in the YSZ matrix even after annealing at 900 oC in ambient atmosphere. For the as-implanted and annealed TiO2 crystals, the XPS spectra of Ni are similar to those of YSZ. 2.1.1.3. XPS Results of MgO Figure 4 shows the XPS spectra of Ni2p3/2 energy level of the Ni ion implanted MgO crystals after annealing at 700 and 900 ºC in Ar+4% H2 atmosphere, respectively [20]. There is no obvious difference between the Ni2p3/2 energy level spectra at 700 and 900 ºC, i.e., the charge state of Ni did not change with the increasing annealing temperature. The peak at 852.9 eV is attributed to metallic Ni0, and the peak at 859.1 eV could be assigned to the shake-up satellite peak of Ni0. These results show that the implanted Ni is only in the charge state of Ni0 and the charge state of Ni was still metallic Ni0 after annealing at 900 ºC in Ar+4% H2 atmosphere.

a

864

b

855

846 864

855

846

Binding energy/eV Figure 4. XPS spectra of 1×1017 cm-2 Ni ion implanted MgO samples after annealing at 700 oC (a) and 900 oC (b) in Ar+4% H2 atmosphere, respectively.

531

5µA/cm2 Ni

Al 2O3

unidentified

900oC

NiO

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

Intensity/ a.u.

800 oC

700 oC

600 oC

as-implanted

as-received

30

40

50

2θ/deg Figure 5. XRD traces (θ-2θ) of Al2O3 single crystals implanted at 5μA/cm2 and after annealing at different temperatures.

2.1.2. XRD Spectra of As-Implanted and Annealed Al2O3 Samples X-ray diffraction measurement was used to clarify the formation of metallic Ni, NiO and NiAl2O4 at ion flux of 5 and 10μA/cm2, respectively [15]. XRD traces (θ-2θ) of some samples were collected with a Cu Kα line of 1.54056 Å. Figure 5 shows XRD traces (θ-2θ) of asreceived and as-implanted crystals at a flux of 5μA/cm2 and annealed crystals at temperatures from 600 to 900 oC. In all the spectra, two diffraction peaks can be observed at ~40.6˚ and ~41.7˚, which were assigned to unidentified and (006) planes of the as-received Al2O3 crystals, respectively. For the as-implanted sample, a broad diffraction peak of metallic Ni appeared at ~44.3˚ indicating the formation of metallic Ni nanoparticles during ion implantation. After annealing at 600 oC in air, the XRD spectra show coexisted diffraction peaks of both Ni and NiO nanoparticles. As the annealing temperature increased, NiO nanoparticles grew up at the expense of Ni nanoparticles. When the annealing temperature reached 900 oC, most of the Ni nanoparticles were oxidized into NiO nanoparticles. The trace of the Ni nanoparticles was hardly detectable based on the XRD spectrum. This is consistent with the XPS result above.

X.T. Zu, X. Xiang, S. Zhu et al.

20

10µA/cm2

30

40

NiAl2 O3

Al 2O3

NiAl 2O3

Intensity/a.u.

532

50

60

70

2θ/deg Figure 6. XRD traces (θ-2θ) of Al2O3 single crystals implanted at 10μA/cm2 and after annealing at different temperatures.

In order to evaluated the mean grain size of Ni and NiO nanoparticles, the Scherrer formula [24] was used

D=

0.9λ B cos(θ B ) ,

where λ, θB, and B are the X-ray diffraction wavelength (1.54056 Å), Bragg diffraction angle and the full width at half maximum (FWHM) of diffraction peaks, respectively. The calculated grain sizes of Ni and NiO nanoparticles are listed in Table 2 for as-implanted and annealed samples. For the as-implanted sample, Ni nanoparticles have mean dimension of 4.9 nm, which will be proved by the following TEM measurement. The annealing effect on the growth of Ni nanoparticles is not obvious before the annealing temperature up to 800 oC, which is similar to Ni nanoparticles in the silica glass [9, 25]. The growth of NiO nanoparticles mainly occurred at annealing temperature above 800 oC. These results just explained the shift towards longer wavelength of UV absorption band, i.e., the optical band gap of NiO nanoparticles shifted towards lower energy due to quantum confinement effect as the grain size increased. Table 2. Average dimensions of Ni and NiO nanoparticles calculated from the XRD spectra after Ni ion implantation with an ion flux of 5μA/cm2. samples as-implanted annealed at 600°C annealed at 700°C annealed at 800°C annealed at 900°C

Ni particle size (nm) 4.9 5.1 5.5 9.2 —

NiO particle size (nm) — 10.8 11.6 19.9 21.5

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

533

Figure 6 shows XRD traces (θ-2θ) of as-implanted crystal at an ion flux of 10μA/cm2. After ion implantation, two new diffraction peaks appeared at ~20.58˚ and ~64.56˚, respectively. According to the JCPDS card (No. 78-1601), these two diffraction peaks may be corresponding to (111) and (440) planes (2θ = 19.08˚ and 65.545˚) of NiAl2O4. The peak shifts with respect to the powder diffraction data may suggest an existing stress due to the lattice distortion after the Ni ion implantation. This XRD result proved the formation of NiAl2O4 spinel structure.

2.1.3. TEM Results of As-Implanted and Annealed Samples Nanoparticles of Ni precipitated spontaneously during ion implantation. In TEM measurements, a HAADF STEM (high-angle annular dark-field scanning transmission electron microscopy) technique was used besides a conventional bright-field imaging technique. For HAADF STEM imaging, intensity in the image is approaching a Z2 dependence on atomic number Z [26]. As an example, for the Ni ion implanted Al2O3, the local area where Ni element distributed will show brighter contrast because Ni has a larger Z than both Al and O. This suggests a HAADF STEM image provides chemical information on element distribution by its contrast (so called Z- contrast image). 2.1.3.1. Ni and NiO Nanoparticles in Al2O3 Figure 7 shows a bright-field and a HAADF STEM cross-sectional image indicating size and distribution of Ni nanoparticles embedded in Al2O3 formed by ion implantation at an ion flux of 5μA/cm2. In this study, Ni has a much higher Z than both Al and O; therefore the nanoparticles of Ni show bright contrast. As is shown in Figure 2, nearly spherical embedded nanoparticles are distributed from the surface to 30 nm below the surface, consistent with calculation results by SRIM 2000 code [27]. The size of nanoparticles ranges from 1 to 5 nm in diameter, which is consistent with the XRD result. Figure 8 is a HREM image showing the crystalline structure of nanoparticles of Ni in the surface of Al2O3 matrix. The image shows clearly the Ni-ion implanted area is amorphized entirely [14]. surface a

surface

b

Ni Ni

Al 2O3

15 nm

Al2 O3

15 nm

Figure 7. A bright-field (a) and a HAADF STEM (b) cross-sectional image in the near surface of asimplanted Al2O3 matrix.

534

X.T. Zu, X. Xiang, S. Zhu et al.

Ni

Ni

Al2O3

22nm nm

Al2O3

Figure 8. A HREM image showing structure of Ni nanoparticles in the near surface of Al2O3 matrix. NiO

a

b

nanovoids

Al2O3 matrix

NiO nanoparticle

25 nm

5 nm Al2O3 matrix

Figure 9. A cross-sectional bright-field and a high-resolution TEM image from a sample annealed at 900 oC after Ni ion implantation with an ion flux of 5μA/cm2.

Figure 9 shows a cross-sectional bright-field (BF) and a high-resolution electron microscopy (HREM) image from a sample annealed at 900 oC after Ni ion implantation with an ion flux of 5μA/cm2. In the BF imaging (Figure 9a), the NiO nanoparticles grew to 6-20 nm in diameter with irregular shape, which is consistent with the XRD result. In addition, the nanoparticles migrated towards the surface of the crystal after annealing. A high density of voids formed below the nanoparticles during the thermal annealing process. The amorphous area of Al2O3 matrix was partially recrystallized. The HREM image (Figure 9b) demonstrates the single crystalline nature of the NiO nanoparticle [15]. 2.1.3.2. Ni and NiO Nanoparticles in YSZ TEM measurement showed that no obvious nanoparticles precipitated after Ni ion implantation. And YSZ matrix did not amorphize after ion implantation. Although XPS measurements have detected the metallic Ni, TEM measurement did not observe nanoparticles precipitated in as-implanted YSZ matrix. This may be due to the very small crystals or none, and even show amorphous conditions. Figure 10 shows a cross-sectional

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

535

bright-field transmission electron microscopy image of the annealed sample at 900 oC. After thermal annealing at 900 oC for 0.5 h, nanoparticles with size ranging from 4 to 12 nm can be observed. The large particles distributed in the surface of the YSZ crystal. Figure 11 is a highresolution electron microscopy (HREM) image clearly showing the crystalline structure of nanoparticles in the surface of YSZ matrix. The nanoparticles are nearly in shape of sphere. In order to clarify which oxide the nanoparticles are, the selected area electron diffraction (SAD) pattern had been obtained in Ni-implanted region in YSZ sample after annealing at 900 oC (shown in Figure 12). After indexing the SAD pattern (the circled spots in Figure 12) the nanoparticles were identified to be NiO [21].

surface

NiO

25 nm Figure 10. A cross-sectional bright-field transmission electron microscopy imaging of the annealed sample at 900 oC.

NiO

YSZ

3 nm Figure 11. A HREM image showing the crystalline structure of nanoparticles in the surface of annealed YSZ matrix.

536

X.T. Zu, X. Xiang, S. Zhu et al.

200YSZ 022YSZ 111NiO

202NiO

Figure 12. A SAD pattern of the annealed YSZ matrix showing the formation of NiO nanoparticles.

2.1.3.3. Ni Nanoparticles in MgO

Figure 13. A cross-sectional HAADF STEM images indicating the nanoparticles of Ni in the surface region of the MgO single crystal after Ni ion implantation (a) and followed by annealing at 900oC for 0.5 h under Ar+4%H2 atmosphere.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

537

Figure 14. Composite selected area electron diffraction (SAED) patterns indicating orientation relationship between Ni nano-particles and the MgO matrix (a); A high resolution TEM (HREM) micrograph showing the crystalline characteristics of Ni nano-particles in the annealed MgO single crystal (b).

Figure 13 shows HAADF images in MgO matrix after Ni ion implantation (a) and subsequent annealing at 900 oC under Ar + 4% H2 atmosphere (b), respectively. In asimplanted MgO, the formation of Ni nano-particles is evident and the particle ranges 3–5 nm in size. The grown and coalescence of these nanoparticles occurred, after thermal annealing and the particles size increased to 8–10 nm. Moreover, larger rods of Ni with 20 nm in length precipitated in the surface of MgO matrix. Besides, the nanoparticles spread into the deeper regions of MgO matrix after thermal treatment. The Ni nanoparticles have specific orientation relationship with the MgO matrix in both of as-implanted sample and thermal annealed sample after ion implantation. As indicated by the composite selected area diffraction (SAD) pattern in Figure 14 (a), the orientation relationship between the Ni particles and MgO matrix was determined as: Ni║MgO and {010}Ni║{010}MgO. Extra satellite spots also appear around the strong spots from the original MgO single crystal in the composite SAD patterns. These satellite spots result from reflections attributable to double diffraction since Ni particles were embedded in the MgO single crystal with the same orientation relationship but different lattice spacing. The double-diffraction reflections are directly responsible for the Moiré fringes in the high resolution TEM images (Figure 14 (b)). As shown in Figure 14 (b), Morié fringes appeared in the place where embedded Ni particles were overlapped with the MgO matrix, due to lattice spacing mismatch between the two phases. The spacing of Morié fringes, dm, can be calculated by the following equation: dm =

d2 d 1− 2 d1

,

538

X.T. Zu, X. Xiang, S. Zhu et al.

where d1 and d2 are lattice spacings of two overlapping crystals. The lattice spacing for Ni(002) is dNi,002 = 0.176 nm (lattice parameter aNi = 0.352 nm), and for MgO(002) dMgO,002 = 0.211 nm (lattice parameter aMgO = 0.422 nm). Thus according to the equation above, dm is 1.06 nm, which is consistent with the measured value in the image [19]. 2.1.3.4. Ni Nanoparticles in TiO2 Ni Nanoparticles precipitated spontaneously during ion implantation from the surface of TiO2 matrix to a depth of 50 nm, as revealed by the cross-sectional HAADF Z-contrast image shown in Figure 15(a). In this HAADF STEM image, Ni has a higher Z than both Ti and O, so the nanoparticles of Ni show a brighter contrast. The dimensions of nanoparticles ranged from 3 nm to 10 nm. Some elongated precipitates up to 20 nm in length were observed in the near surface of TiO2. These elongated Ni precipitates are observed with the round shape in a plan-view TEM image. The implantation region of TiO2 matrix has been damaged and amorphized to 60 nm in the depth below the surface. Figure 15(b) is a HREM image of Ni nanoparticles embedded in the TiO2 matrix, revealing a well-developed crystalline structure of Ni nanoparticles after ion implantation without a thermal treatment [18].

Figure 15. A cross-sectional HAADF STEM image of Ni nanoparticles in the near surface of a TiO2 single crystal (a); and a HRTEM micrograph showing a crystalline Ni nanoparticle in the amorphous TiO2 matrix (b).

2.1.4. Optical Absorption of As-Implanted and Annealed Samples The optical absorption spectra were measured by a SHIMADZU UV-2550 spectrophotometer at room temperature, with a deuterium lamp for UV and a tungsten halogen lamp for the visible region. The wavelength used in the experiment ranged from 200 to 1000 nm. 2.1.4.1. Optical Absorption of Ni-Implanted and Annealed Al2O3 The optical absorption spectra of as-received, as-implanted, and annealed Al2O3 single crystals with an ion flux of 5μA/cm2 are shown in Figure 16. The spectra have vertically been offset to avoid overlapping except that of the as-received crystal. The absorption spectrum of

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

539

the pure crystal was a smooth line in the visible waveband due to a wide band gap ~9 eV. A broad absorption band peaked at 400 nm in the as-implanted crystal [14]. According to the XPS, XRD and TEM results above, the absorption band can be ascribed to the surface plasma resonance absorption of metallic Ni nanoparticles. As the annealing temperature increased, the absorption band shifted towards the longer wavelength. After annealing at 800 oC, this absorption band was absent. At the same time, the crystals turned colorless. However, a new absorption shoulder in the UV region began to appear after annealing at 600 oC. As the annealing temperature reached 800 oC, the UV absorption shoulder peak evolved into an absorption band and its peak position shifted to the longer wavelength of 306 nm (4.05 eV). There is no detectable change for the peak position of the absorption band after annealing at higher temperatures. This UV absorption band was related to the formation of NiO, since NiO is an insulator with a band gap of ~4eV (310 nm) [9]. These results have been proved by the XPS and XRD results. And TEM showed the microstructure of NiO and the recrystalization of Al2O3 matrix [15]. 0.6 5μA/cm2

as-implanted 0.4 400 oC

Absorbance/a.u.

500 oC 600 oC 700 oC 0.2

800 oC 900 oC 1000 oC as-received

0 250

400

600

800

Wavelength/nm Figure 16. Optical absorption spectra of Al2O3 single crystals after Ni ion implantation at 5μA/cm2 and annealing at different temperatures.

2.1.4.2 Optical Absorption of Ni-Implanted and Annealed YSZ The optical absorption spectra of Ni-implanted and annealed YSZ single crystals are partly shown in Figure 17. The spectra have vertically been offset to avoid overlapping except

540

X.T. Zu, X. Xiang, S. Zhu et al.

that of the annealed crystal at 900 ºC. In the spectrum of as-implanted crystal, there is a very broad and weak absorption band ranging from 400 to 700 nm. The absorption curves of annealed crystals are similar to that of as-implanted crystal up to 200 oC. The absorption intensity begins to decrease after annealing at 200 oC, and the absorption band disappears after 250 oC. There is no change for the absorption spectra at annealing temperature of 300~900 oC, similar to that of as-received YSZ [21]. 1.2 as-implanted 200 oC 250 oC 300 oC 900 oC

absorbance/a.u.

1.0 0.8 0.6 0.4 0.2

0 200

400

600

800

wavelength/nm Figure 17. Optical absorption spectra of Ni ion-implanted and annealed YSZ crystals at different annealing temperatures.

YSZ is known to have a band gap of 5.6 eV (222 nm). So its absorption spectrum is nearly a line in the visible region. According to the TEM results above, the broad and weak absorption band ranging from 400 to 700 nm is not due to the SPR absorption of metallic Ni nanoparticles. Although XPS measurements have detected the metallic Ni, TEM measurement did not observe nanoparticles precipitated in as-implanted YSZ matrix. This may be due to the very small crystals or none, and even show amorphous conditions. So the metallic Ni in the as-implanted crystals does not show metallic behavior. In addition, the broad absorption band disappeared just after annealing at 250 oC. At this temperature the metallic Ni clusters did not begin to grow. So the broad absorption band can be ascribed to the point defects and their clusters induced by ion implantation. Just as the Xe ion-implanted YSZ crystals [28], this absorption band may be associated with the combination of electrons trapped at oxygen vacancies and oxygen ions with trapped holes. Metallic Ni clusters grow with the increasing annealing temperature. At the same time, metallic Ni clusters are partly oxidized into NiO nanoparticles. However, the optical absorption spectra did not detect the absorption of Ni or NiO nanoparticles. A possible reason is that the Ni and NiO coexist in the YSZ single crystals and each concentration is low.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

541

2.1.4.3. Optical Absorption of Ni-Implanted and Annealed MgO The optical absorption spectra of as-implanted and annealed MgO single crystals are shown in Figure 18. The spectra have vertically been offset to avoid overlapping except that of the annealed crystal at 900 ºC. In the spectrum of as-implanted crystal, there are two weak absorption bands centered at ~360 nm and ~575 nm, respectively. They have been identified with the F2 centers and the V-type centers in magnesium sublattice (magnesium vacancies), respectively, which is consistent with the Ag+ and Ni+ ion implanted MgO single crystals [29,30]. Upon heat treatment above 700 °C these two absorption bands are completely annihilated by the recombination of point defects. During the annihilation of centers, a new absorption band at ~430 nm formed gradually. Form the XPS and TEM results, it can be concluded that the broad absorption band is related to the metallic Ni nanoparticles, i.e., the surface plasmon resonance (SPR) absorption. After annealing above 700 °C, the absorption maximum shifts to a longer wavelength with the increasing annealing temperature, which is related to the growth of Ni nanoparitcles [20]. 1.5

Absorbance/a.u.

1.0

as-implanted 200 °C 0.5

400 °C 600 °C 700 °C 800 °C 900 °C

0

200

400

600

800

1000

Wavelength/nm Figure 18. Optical absorption spectra of MgO single crystals after Ni ion implantation and annealing at different temperatures.

542

X.T. Zu, X. Xiang, S. Zhu et al.

2.1.5. Magnetic Properties of Ni Nanoparticles The magnetic properties of the samples were characterized by a magnetic property measurement system (MPMS) superconducting quantum interference device (SQUID) magnetometer at 10 and 300 K. The zero-field-cooling (ZFC) and field-cooling (FC) curves for detecting superparamagnetism of the nanoparticles were measured in an applied magnetic field of H = 100 Oe. The ZFC curve was achieved by cooling the sample initially in a zero field to 10 K, and magnetization was recorded in an applied magnetic field where H = 100 Oe as the temperature increased. The FC magnetization was measured by gradually cooling the sample from 300 K to 20 K, and the magnetization was recorded in the presence of a 100 Oe field. 2.1.5.1. Magnetic Ni Nanoparticles in MgO Figure 19 shows a magnetization plot as a function of magnetic field at 10 K in the Niimplanted MgO sample. The applied magnetization field H is parallel to (100) plane of the MgO single crystal, i.e. (100) of Ni nano-particles (Figure 14), owing to their orientation alignment. As shown in the hysteresis loop measured at 10 K, the coercivity, Hc, was about 195 Oe, which is larger than ~150 Oe of randomly oriented Ni particles in Al2O3 [31], because crystallographically oriented particles have stronger tendency to retain its magnetic moments than that of randomly oriented particles under a reversing magnetic field [31]. The magnetization curves at 10 K were completely saturated as the applied field increased to H = 5000 Oe (not shown in the figure). No coercive force Hc was observed in the sample at 300 K [19].

Figure 19. A magnetic hystersis loop of Ni nanoparticles in the MgO single crystal at 10 K after ion implantation. The coercivity Hc was about 195 Oe at this temperature.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

543

Figure 20. ZFC and FC magnetizations as a function of temperature for Ni nanoparticles in asimplanted and after subsequent annealed MgO samples. Curves were taken in the ZFC and FC processes at H=100 Oe.

The zero-field-cooled (ZFC) and field-cooled (FC) magnetization as a function of temperature are shown in Figure 20 for both as-implanted and the annealed samples. As shown in Figure 20, the FC magnetization increases monotonically with the decrease of temperature. However, the magnetization increases at first, then decreases with an increasing of temperature in ZFC curve. The temperature, at which the maximum in ZFC magnetization occurs, is characterized as the blocking temperature (TB). The blocking temperature TB of Ni nanoparticles was determined to be ~35 K in the as-implanted sample and above 300 K in the annealed MgO sample, respectively. These Ni nanoparticles exhibit superparamagnetic properties above the blocking temperature. This behavior is consistent with the result from the as-implanted sample that shows almost immeasurable coercivity and remanence at 300 K. However, superparamagnetism of the Ni nanoparticles in the annealed MgO sample persists to above 300 K, and these Ni nanoparticles remain ferromagnetic at room temperature [19]. TB has a relation with particle size, based on the equation: T B = Keff V/25kB, where Keff the effective anisotropy related to the magnetocrystalline anisotropy and to the shape anisotropy, V is the volume of a particle, kB is Boltzmann constant [32]. If we use the magnetocrystalline anisotropy K1 value (-8×105 erg/cm3) in [32] to replace Keff, the Ni particles size that contributed to the measured magnetic property can be roughly calculated to be ~3.3 nm in as-implanted sample and 6.8 nm in thermal annealed sample. This is close to

544

X.T. Zu, X. Xiang, S. Zhu et al.

the TEM observations above. It should be noted that the calculation neglects the shape anisotropy caused by large rod-shape Ni precipitates in the annealed sample, which may cause magnetic anisotropy and increase average particles size contributing to the magnetic property. 2.1.5.2. Magnetic Ni Nanoparticles in TiO2 Figure 21 shows the ZFC and FC magnetization as a function of temperature of Ni nanoparticles in TiO2 matrix. Figure 21 clearly shows the non-zero difference between the FC and ZFC data, indicating the hysteresis while eliminating any para- and diamagnetic contributions. In these ZFC/FC curves, the blocking temperature, TB, is ~85 K. Above the blocking temperature, the magnetization is unstable and the sample loses all its hysteric responses. The two insets in Figure 21 showing magnetization plots as a function of magnetic field (M-H) at 10 K and 300 K, respectively. The applied magnetization field H is parallel to (100) of the TiO2 single crystal. The magnetic hysteresis loop at 10 K after correcting paramagnetic contribution from TiO2 matrix shows a ferromagnetic behavior with coercivity, Hc, equal to ~ 210 Oe [18]. This value is larger than that of Ni nanoparticles in MgO (195 Oe) above. In addition, a coercivity, Hc, equal to ~ 270 Oe was obtained at 10 K in Ni-implanted YSZ crystal when the applied magnetization field H is parallel to (001) of the YSZ single crystal [21]. The variation of coercivity strongly depends on the grain size. Since nanoparticles less than a critical size are in single domain states, their coercivities are higher than that of a common bulk sample [33]. This large coercivity can be explained by the nanosize effect and the single domain structure (the critical size of the single domain for spherical Ni particles was ~ 42 nm) [34, 35]. The magnetization curves at 10 K was completely saturated at applied fields H = 4000 Oe. The remnant magnetization of the saturation magnetization (Mr/Ms) is about 40% at this temperature. The coercive force and remnant magnetization of Ni nanoparticles were not observed as the temperature increased to 300 K, confirming the superparamagnetic behavior of the nanoparticles above the blocking temperature. Since some nano-particles are not truly spherical, the shape anisotropy has contribution to the superparamagnetism. Langevin function can be used to calculate the true magnetic moment of each particle for superparamagnetic particles as [36]:

M ( H / T ) = M 0 [coth( M 0 mH / k BT ) − (k B T / M 0 mH )] where, M0 is the saturation magnetization (emu/g), m is the mass of individual particle (gram), and kB is Boltzmann constant. The saturation magnetization in our sample was about 25 emu/g from the hysteresis loop at 300 K. Figure 22 shows magnetization vs. applied magnetic field of Ni nanoparticles at 300 K (solid circles) and the best fit for the Langevin function (solid line). From the data fitting, we calculate the average grain size of 9.4 nm for the Ni nanoparticles contributing to magnetic moment, which is in the range of particle size measured by TEM observation. The mean-magnetic moment per particle of the sample was calculated to be 11064μB [18].

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

M (×10-4 emu)

1

FC 4.0

0.5 0 T=300K

-0.5 -1 -3 -2 -1

ZFC

2.0

0

1

2

3

H (kOe)

M (×10 -4 emu)

Magnetization (×10-5, emu)

6.0

545

1 0 T= 10 K

1

-3 -2 -1 0 1 2 3

0

H (kOe)

0

50

100

150

200

250

300

Temperature (K)

Magnetization (emu/g)

Figure 21. Temperature dependence of the ZFC and FC magnetization curves for the Ni implanted TiO2 sample with magnetic hysteresis loops at 10 K (bottom inset) and 300 K (top inset).

20 10 0 -10 -20 -30 -20

T=300K

-10 0 10 Applied Magnetic field (kOe)

20

Figure 22. Measured (solid circle) and the Langevin function fitted (solid line) magnetization (M) vs. magnetic filed (H) at the room temperature.

546

X.T. Zu, X. Xiang, S. Zhu et al.

2.2. Zn and ZnO Nanoparticles Embedded in Al2O3 Ion implantations were conducted at room temperature using a metal vapor vacuum arc (MEVVA) implanter. The samples were tilted off-axis by around 7 degrees to avoid channeling implantation. After Zn ion implantation, the Al2O2 samples were annealed in oxygen atmosphere in order to fabricate zinc oxide nanoparticles. ZnO is well-known as a versatile wide band gap (~3.3 eV) semiconducting material with a large exciton binding energy of 60 meV, which allows excitonic recombination and optically pumped laser oscillations even at the room temperature. Following the demonstration of blue-green light emitting diodes (LEDs) and lasers using II–VI compounds, ZnO has been intensively studied for optoelectric applications in both the visible and ultraviolet (UV) regions. Optical absorption and TEM measurements are same as those of the Ni-ion-implanted samples above.

2.2.1. Optical Absorption of Zn Nanoparticles Fabricated with Different Fluences The optical absorption spectra of as-implanted α-Al2O3 single crystals at fluences of 1×1016, 5×1016, 1×1017, and 5×1017 cm-2 are shown in Figure 23. The spectra have been offset to avoid overlapping except that of the as-implanted spectrum at fluence of 1×1016 cm-2. After Zn+ ion implantation at a fluence of 1×1016 cm-2, a weak absorption peak appeared at ~260 nm. This absorption peak becomes clear gradually with the increasing fluences. At the same time, the peak wavelength shift to longer wavelength linearly (dash line in Figure 23). The absorption peak shifts to ~285 nm when the ion fluence up to 5×1017 cm-2. The strong and broad peak at 260-285 nm is due to surface plasma resonance absorption of metallic Zn nanoparticles [16]. Actually, the similar absorption peaks have been observed in SiO2 glass and MgO single crystal matrices (listed in Table 3). 0.8 5×1017 1×1017 5×1016 1×1016

Absorbance/a.u.

0.6

0.4

0.2

0 200

400

600

800

Wavelength/nm Figure 23. Optical absorption spectra of Zn+-implanted Al2O3 single crystals at different fluences.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

547

As a whole, the SPR peaks in α-Al2O3 and MgO lie at longer wavelength than that in SiO2 matrix. This is because that α-Al2O3 and MgO have larger refractive indexes (1.76 for αAl2O3 and 1.74 for MgO) than SiO2 (1.52). According to the Maxwell-Garnett (MG) [41] and Mie [42] theory, the SPR peak would shift towards low energy with increasing refractive index of surrounding medium. The size of nanoparticles is another factor to influence the SPR peak wavelength, which will shift to a longer wavelength with the increasing crystalline size. The SPR absorption peaks in Table 3 are observed in the as-implanted matrices except that in MgO which was observed after annealing at 1150 K, which shift to a longer wavelength due to the growth of nanoparticles. Table 3. SPR peaks of metallic Zn nanoparticles in several insulator matrices matrix

SiO2 ∗

MgO

α-Al2O3

Zn+ ion/cm-2

SPR peak

reference

1×1017

4.8 eV/259 nm

[10]

1×1017

4.86 eV/255 nm

[37]

1×10

17

5.3 eV/234 nm

[38]

3×10

17

4.86 eV/255 nm

[39]

1×1017

4.2 eV/295 nm

[40]

1×1017

4.56 eV/272 nm

17

4.35 eV/285 nm

5×10

[17]

2.2.2. Optical Absorption of Zn-Ion-Implanted and Annealed Al2O3 The optical absorption spectra of the as-received, as-implanted with a fluence of 1×1017 cm-2 and annealed α-Al2O3 single crystals are shown in Figure 24. The spectra have been shifted vertically to avoid overlapping except that of the as-received sample. The SPR absorption peak slightly shifted towards the longer wavelength due to the growth of nanoparticles with the increasing annealing temperature. The spectrum of annealed sample at 500 °C is a transition one because it includes both the absorption peak of Zn nanoparticles and the absorption edge of ZnO nanoparticles. After annealing at 600 °C for 1 h, the SPR absorption peak disappears and a clear absorption peak appears at ~360 nm, which is consistent with the exciton absorption of ZnO [10, 17, 40]. The intensity of this exciton absorption peak decreases with the further increasing annealing temperature. Apparently, this intensity decrease does not indicate a decrease of Zn content in the sample. It may be due to the decreased concentration of ZnO nanoparticles because of the formation of ZnAl2O4 spinel during thermal annealing in O2 atmosphere [43]. The microscopic morphology of Zn and ZnO nanoparticles has been characterized by TEM imaging with the main results shown below.

∗

Observed after annealing at 1150 K.

548

X.T. Zu, X. Xiang, S. Zhu et al. 1.0 as-implanted 300 oC 400 oC 500 oC

Absorbance/a.u.

0.8

0.6

0.4 600 oC 700 oC 0.2

800 oC 900 oC

0

400 600 Wavelength/nm

200

800

Figure 24. Optical absorption spectra of Zn+-implanted and annealed Al2O3 single crystals at different temperature. surface

amorphous

Crystalline matrix

5 nm

20 nm

Figure 25. A cross-sectional bright-field TEM image (a) and a high resolution TEM (HREM) image (b) of Zn nanoparticles embedded in α-Al2O3 formed by Zn+ ion implantation at a dose of 1×1017 cm-2. The polycrystalline ring in the bottom inset in (a) showing the random orientation of the Zn nanopartilces precipitated after the ion implantation.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

549

2.2.3. TEM Results of Zn and ZnO Nanoparticles Figure 25 shows a cross-sectional bright-field TEM image (a) and a high resolution TEM (HREM) image (b) of Zn nanoparticles embedded in α-Al2O3 formed by Zn+ ion implantation at a dose of 1×1017 cm-2. Nearly spherical embedded Zn nanoparticles of 3-10 nm in diameter are observed and the ion-implanted area is amorphized. The insert one in Figure 25a is a selected area electron diffraction (SAD) pattern. The polycrystalline ring can be observed in the SAD pattern showing the random orientation of the Zn nanopartilces precipitated after the ion implantation [16]. surface

ZnO

voids recrystallized Al2O3 100 101

20 nm

crystalline matrix

b a

Figure 26. A cross-sectional bright-field TEM image (a) and selected area electron diffraction (SAED) pattern (b) of ZnO nanoparticles embedded in α-Al2O3 formed after annealing at 600 °C.

Figure 26 contains a cross-sectional bright-field TEM image (a) and selected area electron diffraction (SAED) pattern (b) of ZnO nanoparticles embedded in α-Al2O3 formed after annealing at 600 °C. The ZnO can be confirmed by the SAED pattern. Figure 27 is a high resolution TEM image of ZnO nanoparticles, showing the nanoparticles of 10-12 nm in dimensions. The morié fringes indicate the precipitation of ZnO nanoparticles after annealing. It is clear that the ZnO nanoparticles formed close to the surface of the α-Al2O3 single crystal, with a depth shallower than the projectile range of implanted Zn atoms. It indicates that the Zn atoms migrated towards the surface of the crystal during the annealing in oxygen atmosphere. At the same time, some high density of large voids (labeled in Figure 26a) is observed in the near-surface region due to the migration and precipitation of irradiation induced vacancies. The recrystallized Al2O3 grains have different orientation relationship with the original matrix, which can be observed in Figure 28. However, the recrystallization was not complete after annealing at 600 °C for 1 h. The front surface area (labeled in Figure 28) remained amorpous. Apparently, longer annealing time or higher annealing temperatures are needed to complete the recrystallization [17].

550

X.T. Zu, X. Xiang, S. Zhu et al.

5 nm

Figure 27. A high resolution TEM image of ZnO nanoparticles, showing the nanoparticles of 10-12 nm in dimensions.

amorphous Al2O3

recrystallized Al2O3

2 nm

Al2O3 matrix

Figure 28. A high resolution TEM image showing the recrystallized Al2O3 with different orientation relation with the matrix.

2.2.4. PL of ZnO Nanoparticles Figure 29 shows the photoluminescence (PL) spectra of the as-implanted crystal and the annealed crystal at 600 °C using a He–Cd laser excitation at 325 nm line at room temperature. There is a very weak PL band peaked at ~470 nm in the as-implanted crystal. It may be due to the photoluminescence combination of both F (PL at 3.0 eV) and F2 (PL at 2.4 eV) centers coexisted in Al2O3 induced by ion implantation [44]. PL spectrum of the annealed crystal shows two PL peaks, one at 370 nm, and the other at 500 nm, which have also been observed

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

551

in Zn+ ion implanted SiO2 and CaF2 [37, 39]. The UV emission is the characteristic PL peak ascribed to ZnO free-exciton recombination at room temperature, which confirms the formation of ZnO nanoparticles after thermal annealing. The green emission at ~500 nm may originate from the deep levels [45, 46]. The deep-level emission at around 2.5 eV is associated with either surface state emission or excess Zn interstitials (or oxygen vacancies). Up to now, both UV and green PL peaks of embedded ZnO nanoparticles have been observed in SiO2, CaF2 and Al2O3 matrices by ion implantation and thermal annealing (listed in Table 4). 400

PL intensity/a.u.

300

b

200

100

a 0 350 400 450 500 550 600 650 700 750 800

Wavelength/nm Figure 29. Photoluminescence spectra of the as-implanted (a) and the annealed (b) crystal at 600 °C at room temperature.

Table 4. PL of ZnO nanoparticles embedded in several insulator matrices

SiO2 CaF2 Al2O3

700 oC 1h 700 oC 2h 400 oC 45min 500 oC 45min 700 oC 45min 600 oC 1h

375 377 384 372 379 370

500 500 500 500

1:2 1:7 2:1 3:1

[37] [39] [47] [17]

2.3. Intermetallic CoxNi1-x Nanoparticles Embedded in YSZ Sequential ion implantation in dielectric matrix determines three different cluster morphologies: separated systems, alloy clusters and core–shell clusters. Sequential ion implantations with different elements were performed mostly on silica substrates, but also on quartz and sapphire [48]. In this data review, the intermetallic CoxNi1-x nanoparticles embedded in YSZ were synthesized by sequential implantation of 90 keV Co and Ni ions at room temperature. TEM and magnetic measurements were utilized to analyses the magnetic CoxNi1-x nanoparticles [22].

552

X.T. Zu, X. Xiang, S. Zhu et al.

2.3.1. TEM Images of CoxNi1-x Nanoparticles A bright field cross-sectional TEM image (Figure 30a) shows that the nanoparticles precipitated spontaneously during ion implantation in the near surface of the YSZ matrix. The typical nanoparticle size ranged from 3 to 10 nm. In the high concentration region of implanted Co and Ni ions (~25–40 nm below the surface, consistent with the results from SRIM 2000 code [27]), some large precipitates up to 25 nm in length formed parallel to the surface of YSZ. High-resolution TEM images clearly revealed that the elongated particles consisted of individual crystals joined across twin boundaries. The (111) twin planes are shown in Figure 30b. The selected area diffraction pattern inserted in Figure 30a exhibits obvious extra diffraction spots from the precipitates, which are circled in the SAD pattern. The precipitates are cubic with a>0.35 nm. The phases consistent with this lattice parameter include cubic Co, Ni, solid solution CoxNi1-x, as well as intermetallic phases, CoNi and Co3Ni7 [49].

Figure 30. Bright-field cross-sectional TEM (a) and high resolution TEM (b) micrographs showing nanoparticles of CoxNi1-x in YSZ single crystal; SAD pattern (inset in (a)) indicating the orientation of the nanoparticles to the matrix.

Energy filtered elemental mapping images of Co, Ni, and O using the L2,3 or K edges in a corresponding energy electron-loss spectroscopy (EELS) spectrum, are shown in Figure 31. The bright contrast of the nanoparticles in the Co and Ni maps indicates that the nanoparticles contained both Co and Ni. These nanoparticles do not contain oxygen as indicated by the dark contrast in nanoparticles in the O map image. Energy dispersive spectroscopy (EDS) analysis showed the composition ratio of Co over Ni ranges from 0.8 to 1. Therefore, the nanoparticles are CoxNi1-x solid solution. The orientation relationship between aligned nanoparticles and the YSZ matrix are (200)YSZ║(200) CoxNi1-x, and [011]YSZ║[011] CoxNi1-x. Based on calculations using the SRIM code, the maximum level of damage reached 300 dpa in the YSZ matrix; however, the YSZ matrix still retained its crystallinity.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

553

Figure 31. (a) Bright-field image and elemental mapping images of (b) Co, (c) Ni, and (d) O, respectively, indicating the compositional distribution of the nanoparticles.

2.3.2. Magnetic Properties of CoxNi1-x Nanoparticles Magnetization plots as a function of magnetic field at both 10 and 300 K are shown in Figure 32. The magnetization field H applied is parallel to (001) of the YSZ single crystal, i.e. (001) of CoxNi1-x, due to its orientation alignment. The coercivity, Hc, at 300 K was measured to be ~100 Oe. The coercivity increased to 260 Oe as the temperature decreased to 10 K.

Figure 32. Magnetization vs field applied at temperatures of 10 and 300 K. The field applied is parallel to the (001) plane of the YSZ matrix.

554

X.T. Zu, X. Xiang, S. Zhu et al.

These values are consistent with previously reported particle-size dependent coercive forces in CoNi alloys [50, 51]. The magnetization curves at 10 and at 300 K were completely saturated at applied fields of H = 800 and 5000 Oe, respectively. The saturation magnetization, Ms, increased 15% as the temperature decreased from 300 to 10 K. The remanent magnetization Mr /Ms was 0.41 at 300 K and its value increased to 0.77 at 10 K. The variation of coercivity strongly depends on the grain size. Since nanoparticles less than a critical size are in single domain states, their coercivities are higher than that of a common bulk sample [33]. Previous investigations reported that the critical size with the largest coercivity lies in the 30–40 nm range for Co50Ni50 [50]. The coercivity of our samples can be attributed to this nanostructural effect. The mismatch between the matrix and the implanted CoxNi1-x particles probably induces stress, which also affects the coercivity [52]. Nanoparticles are expected to illustrate superparamagnetic properties.

Figure 33. Temperature dependence of the magnetization. The curves show that in the ZFC and FC processes at H = 100 Oe.

The blocking temperature can be readily characterized by ZFC and FC magnetization. Figure 33 shows the temperature dependent magnetization curves under ZFC and FC processes at H=100 Oe. No blocking temperature within 300 K was observed. Since TEM analysis has confirmed that the nanoparticles are not oxidized, the results from the ZFC and FC curves suggest that the sample might not be blocked at low temperature. This is reasonable if particle-size and shape effects are considered. As observed in the TEM micrograph (Figure 30), the particle size is widely distributed from 3 to 10 nm, and some longer particles up to 25 nm have formed from a sequence of twin planes roughly perpendicular to the length of the particle. They might not be in single domain states. The blocking temperatures therefore are different and may overlap. The critical temperature, i.e., the blocking temperature, TB, is also roughly given as the equation: TB = Keff V/25kB [32]. Based on this equation, TB = ~227 K, if we use the magnetocrystalline constant K1 value (1.5×105 J/m3) from Ref. [53] for Keff and a maximum diameter of 10 nm for the nanoparticles. The result of this calculation differs from the experimental observation because K1 only represents the magnetocrystalline anisotropy contribution. Since some nanoparticles are not truly spherical, the contribution by shape

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

555

anisotropy to Keff should also be considered. On the other hand, the elongated particles formed by repeating twins may result in interparticle dipolar interactions. Thus, supermagnetism of the CoxNi1-x nanoparticles may persist to above 300 K, and the nanoparticles will remain ferromagnetic at room temperature.

References [1] Meldrum, A.; Haglund Jr., R. F.; Boatner, L. A.; White C. W. Adv. Mater. 2001, vol 13, 1431-1444. [2] Chakraborty, P. J. Mater. Sci. 1998, vol 33, 2235-2249. [3] Beecroft, L. L.; Ober,C. K. Chem. Mater. 1997, vol 9, 1302-1317. [4] Meldrum, A.; Boatner, L. A.; White, C. W. Nucl. Instrum. Meth. B, 2001, vol 178, 7-16. [5] Yang, C. J.; Kim, K. S.; Wu, J. M. J. Appl. Phys. 2001, vol 90, 5741-5746. [6] Gonzalo, J.; Perea, A.; Babonneau, D.; Afonso, C. N.; Beer, N.; Barnes, J. P.; PetfordLong A. K.; Hole, D. E.; Townsend, P. D. Phy. Rev. B 2005, vol 71, 125420. [7] Dempsey, N. M.; Ranno, L.; Givord, D.; Gonzalo, J.; Serna, R.; Fei, G. T.; PetfordLong, A. K.; Doole, R. C.; Hole, D. E. J. Appl. Phys. 2001, vol 90, 6268-6274. [8] Fonseca, F. C.; Goya, G. F.; Jardim, R. F.; Carreno, N. L. V.; Longo, E.; Leite, E. R.; Muccillo R. Appl. Phys. A: Mater. 2003, vol 76, 621-623. [9] Amekura, H.; Umeda, N.; Takeda, Y.; Lu J.; Kishimoto, N. Appl. Phys. Lett. 2004, vol 85, 1015-1017. [10] Amekura, H.; Umeda, N.; Sakuma, Y.; Kishimoto, N.; Buchal, Ch. Appl. Phys. Lett. 2005, vol 87, 013109. [11] Vallet, C. E.; White, C. W.; Withrow, S.P.; Budai, J. D.; Boatner, L. A.; Sorge, K. D.; Thompson, J. R.K.; Beaty, S.; Meldrum, A. J. Appl. Phys. 2002, vol 92, 6200-6204. [12] Sorge, K. D.; Thompson, J. R.; Schulthess, T. C.; Modine, F. A.; Haynes, T. E.; Honda, S.; Meldrum, A.; Budai, J. D.; White, C. W.; Boatner, L. A. IEEE Trans. Mag. 2001, vol 37, 2197-2199. [13] Ila, D.; Williams, E. K.; Zimm erman, R. L.; Poker D. B.; Hensley, D. K. Nucl. Instrum. Meth. B, 2000, vol 166-167, 845-850. [14] Xiang, X.; Zu, X.T.; Zhu, S.; Wang, L. M. Appl. Phys. Lett. 2004, vol 84, 52-54. [15] Xiang, X.; Zu, X. T.; Bao, J. W.; Zhu S.; Wang, L. M. J. Appl. Phys. 2005, vol 98, 073524. [16] Xiang, X.; Zu, X. T.; Zhu S.; Zhang, C. F.; Wang, L. M. Nucl. Instrum. Meth. B, 2006, vol 250, 192–195. [17] Xiang, X.; Zu, X. T.; Zhu S.; Wei, Q. M.; Zhang, C. F.; Sun K.; Wang, L. M. Nanotech, 2006, vol 17, 2636-2640. [18] Zhu, S.; Wang, L. M.; Zu X. T.; Xiang, X. Appl. Phys. Lett. 2006, vol 88, 043107. [19] Zhu, S.; Xiang, X.; Z u, X.T.; Wang, L.M.; Nucl. Instrum. Meth. B, 2006, vol 242, 114– 117. [20] Xiang, X.; Zu, X. T.; Zhu S.; Zhang, C. F.; Wang, L. M. Nucl. Instrum. Meth. B, 2006, vol 250, 229–232. [21] Xiang, X.; Zu, X. T.; Zhu S.; Wang, L. M. Physica B, 2005, vol 368, 88-93. [22] Zhu, S.; Sun, K.; Zhang, Q. Y.; Zu, X. T.; Wang, L. M.; Ewing, R. C. J. Appl. Phys. 2003, vol 94, 5648-5651.

556

X.T. Zu, X. Xiang, S. Zhu et al.

[23] Słoczyński, J.; Ziόłkowski, J.; Grzybowska, B.; Grabowski, R.; Jachewicz, D.; Wcisło K.; Gengembre, L. J Catal. 1999, vol 187, 410-418. [24] Cullity, B. D. Elements of X-Ray of Diffractions; Addition-Wesley: Reading, MA, 1978, pp 102. [25] Amekura, H.; Takeda, Y.; Kishimoto, N. Nucl. Instrum. Meth. B, 2004, vol 222, 96-104. [26] Treacy, M. M. J.; Howie, A.; Wilson, C. J. Phil. Mag. A, 1978, vol 38, 569-585. [27] Ziegler, J. E.; Biersack J. P.; Littmark, U. The Stopping and Range of Ions in solids Pergamon Press: New York, NY, 1985. [28] Zhu, S.; Zu, X. T.; Xiang, X.; Wang, Z. G.; Wang, L. M.; Ewing, R. C. Nucl. Instrum. Meth. B, 2003, vol 206, 1092-1096. [29] Mitamura, T.; Kawatsur, K.; Takahashi, R.; Adachi, T.; Igarashi, T.; Arai, S.; Masuda, N.; Aoki, Y.; Yamamoto S.; Narumi, K.; Naramoto, H.; Horino, Y.; Mokuno, Y.; Fujii, K. J. Nucl. Mater. 1999, vol 271-272, 15-20. [30] Qian, Y.; Ila, D.; Zimmerman, R. L.; Poker, D. B.; Boatner, L. A.; Hensley, D. K. Nucl. Instrum. Meth. B, 1997, vol 127-128, 524-527. [31] Kumar, D.; Zhou, H.; Nath, T. K.; Kvit, A. V.; Narayan, J. Appl. Phys. Lett. 2001, vol 79, 2817-2819. [32] Fonseca, F. C.; Goya, G. F.; Jardim, R. F.; Muccillo, R.; Carreno, N. L. V.; Longo, E.; Leite, E. R.; Phys. Rev. B, 2002, vol 66, 104406. [33] Leslie-Pelecky, D. L.; Rieke, R. D. Chem. Mater. 1996, vol 8, 1770-1783. [34] Du, Y. W.; Xu, M. X.; Jian W.; Shi Y. B.; Lu, H. X.; Xue R. H. J. Appl. Phys. 1991, vol 70, 5903-5905. [35] Zhao, Z. J.; Wang, H. Q.; Wang, B.; Hou, J. G.; Liu G. L.; Jin, X. F. Phys. Rev. B 2002, vol 65, 235413. [36] Bucher, J. P.; Douglass D. C.; Bloomfield, L. A. Phys. Rev. Lett. 1991, vol 66, 30523055. [37] Lee, J. K.; Tewell, C. R.; Schulze, R. K.; Nastasi, M.; Hamby, D. W.; Lucca, D. A.; Jung H. S.; Hong, K. S. Appl. Phys. Lett. 2005, vol 86, 183111. [38] Chen J. L.; Mu, R.; Ueda, A.; Wu, M. H.; Tung, Y.-S.; Gu, Z.; Henderson, D. O.; White, C. W.; Budai, J. D.; Zuhr, R. A. J. Vac. Sci. Technol. A 1998, vol 16, 1409-1413. [39] Liu, Y. X.; Liu, Y. C.; Shao C. L.; Mu, R. J. Phys. D: Appl. Phys. 2004, vol 37, 30253029. [40] van Huis M. A.; van Veen, A.; Schut, H.; Kooi, B. J.; De Hosson, J. Th. M.; Du, X. S.; Hibma, T.; Fromknecht, R. Nucl. Instrum. Meth. B, 2004, vol 216, 390-395. [41] Maxwell-Garnett J. C. Philos. R. Soc. London 1906, vol 205, 237. [42] Mie G. Ann. Phys. 1908, vol 25, 377 (in German). [43] White, C.W.; Meldrum, A.; Sonder, E.; Budai, J. D.; Zuhr, R. A.; Withrow, S. P.; Henderson, D. O. Materials Research Society Symposium-Proceedings, 1999, vol 540, 219-224. [44] Kotomin, E. A.; Popov, A. I. Nucl. Instrum. Meth. B, 1998, vol 141, 1-15. [45] Zhao, Q.; Xu, X. Y.; Song, X. F.; Zhang, X. Z.; Yu, D. P.; Li C. P.; Guo, L. Appl. Phys. Lett. 2006, vol 88, 033102. [46] Zhang, D. H.; Wang Q. P.; Xue, Z. Y. Appl. Surf. Sci. 2003, vol 207, 20-25. [47] Liu, Y. C.; Xu, H. Y.; Mu, R.; Henderson, D. O.; Lu, Y. M.; Zhang, J. Y.; Shen, D. Z.; Fan, X. W.; White, C. W. Appl. Phys. Lett. 2003, vol 83, 1210-1212. [48] Mattei G. Nucl. Instrum. Meth. B, 2002, vol 191, 323–332.

Embedded Optical-electrical Nanomateriales Fabricated by Ion Implantation

557

[49] Villars P.; Calvert, L. D. Peason’s Handbook of Crystallographic Data for Intermetallic Phases, ASM, Materials Park, OH 1991, Vol. 2. [50] Toneguzzo, P.; Acher, O.; Viau, G.; Pierrard, A.; Fievet, F.; Rosenman, I. IEEE Trans. Magn. 1999, vol 35, 3469-3471. [51] de Julián C.; Sangregorio, C.; Mattei, G.; Battaglin, G.; Cattaruzza, E.; Gonella, F.; Lo Russo S.; D'Orazio, F.; Lucari, F.; De, G.; Gatteschi, D.; Mazzoldi, P. J. Magn. Magn. Mater. 2001, vol 226–230, 1912-1914. [52] Morales, M. P.; Munoz-Aguado, M. J.; Garcia-Palacios, J. L.; Lazaro, F. J.; Serna, C. J. J. Magn. Magn. Mater. 1998, vol 183, 232-240. [53] Wijn H. P. J. Magnetic Properties of Metals, Springer: Berlin, 1991, pp 22.

In: Nanotechnology... ISBN 978-1-60692-162-3 c 2010 Nova Science Publishers, Inc. Editors: C.J. Dixon and O.W. Curtines, pp. 559-602

Chapter 18

S TRUCTURAL, DYNAMICAL AND O PTICAL P ROPERTIES OF S ELF - ASSEMBLED P ORPHYRINS AT THE M ESOSCOPIC S CALE Valentina Villari 1,∗, Norberto Micali1 and Luigi Monsu´ Scolaro2 1 CNR-Istituto per i Processi Chimico-Fisici, S.ta Sperone C.da Papardo, I-98158, Messina, Italy 2 Dipartimento di Chimica Inorganica, Chimica Analitica e Chimica Fisica, Universit´a di Messina Salita Sperone 31, I-98166 Vill. S. Agata, Messina, Italy

Abstract Organized self-assembly of molecules, driven by noncovalent intermolecular interactions, is the most versatile tool for accessing new materials with desired optical and electronic properties. Porphyrins are particularly attractive species to incorporate into supramolecular assemblies because their rich photochemistry may impart functionality, provide insight into the mechanisms of biological processes such as photosynthesis, serve as probes into the features of self-assembled structures and as models for molecular organization and energy/electron transfer processes. The close molecular packing in a self-assembled porphyrin aggregate leads to different electronic coupling and delocalization of the excitation energy, which can be exploited for applications in non-linear optical devices, photoelectric cells, recording devices. The possibility to control and tune either shape and size of the porphyrin clusters opens the way for their use as potential nanodevices. This review aims to collect some recent developments in the field of porphyrin self-assembly and to frame all the reported topics into the current theories.

1.

Introduction

Porphyrins constitute a wide class of natural and synthetic molecules whose photophysical properties can be modulated by changing the peripheral substituent groups and/or by ∗

E-mail address: [email protected]; [email protected]. (Corresponding author)

560

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

inserting metal ions in the central core of the macrocycle.[1, 2] In nature, for instance, porphyrins are present under different forms and are responsible of many biochemical processes in animal and vegetable kingdom: specific examples are furnished by hemoglobin or by chlorophyll, as well as by many other biological molecules whose building blocks are porphyrin residuals, like cytochromes and hemocyanines. Along with natural porphyrins many synthetic porphyrins are exploited in biological and medical research for the study of nucleic acid conformation, of intercalation phenomena[3, 4, 5, 1, 6] and as drugs in anticancer photodynamic therapy.[7, 8] Moreover, the stability of these systems makes them especially interesting for photoionization processes, energy/electron transfer, photocatalysis and nonlinear optical properties.[9, 10, 11, 12] Numerous studies, for instance, have been carried out to investigate the DNA structure, by exploiting the ability of some porphyrin derivatives of intercalating into DNA of appropriate composition, while other porphyrins, depending on the nature of peripheral substituents or inserted metals, are limited to external, groove binding.[3, 4, 5, 1] The interaction of porphyrin-based moieties with different chemical species makes them useful for molecular recognition processes in the sensor field.[13, 14, 15] One important, promising, and newly developed practical application is the determination of the absolute configuration of various chiral compounds, even of biological importance like amino acids, and of the enantiomeric excess.[16, 17] In addition, the chemo-responsive behaviour of metalloporphyrins provides a way of reporting the presence of odors by changes in color; two-dimensional display of metalloporphyrins, as an example, was employed as sensor for the visual identification of a wide range of olfactants and solvent vapors.[18, 19] One of the most challenging aspect for specific application in materials science, condensed matter science, engineering, farmaceutics (drug delivery) as well as in nano-science and nanotechnology, consists in using porphyrins and their derivatives as building blocks for designing and accessing to supramolecular systems, by means of controlled self-assembly. From a general point of view, self-assembly is the autonomous organization of components into patterns or structures; it involves components from the microscopic to the macroscopic scale and many different kinds of non-covalent interactions, like van der Waals, electrostatic, hydrophobic interactions, hydrogen and coordination bonds. Thanks to the wide range of interactions when using components larger than molecules, often it is possible to adjust interactions themselves over wide ranges of strength and selectivity. Nonmolecular systems are, thus, in many aspects more versatile in their design than molecular systems. It is often easier to build-up nonmolecular components than it is to synthesize molecules, and easier to observe the process and products of self-assembly using conventional experimental techniques (like, for instance, scattering). Depending on their geometrical disposition in the assembly, porphyrins can form differerent kinds of aggregates called H- (face-to-face interactions) and J-type (side-by-side interactions). J aggregates have attracted a great deal of interest for their nonlinear optical properties originating from the close molecular packing: the stacking interactions together with electrostatic and hydrogen bonding interactions, in fact, leads to electronic coupling and delocalization of the excitation energy. Among different kinds of porphyrins the water soluble moieties are very interesting because their self-aggregation can be conveniently controlled by screening the charge repulsion through the ionic strength and pH and by varying concentration.[20, 21] More precisely,

Self-assembled Porphyrins at the Mesoscopic Scale

561

acting on the intermolecular interaction potential leads to structures which are different not only at a mesoscopic scale but also locally. Furthermore, porphyrin self-assembly on the DNA surface[22, 23, 24, 25, 26] or engineered viral particles[9] has been reported as a convenient method to control the size and extent of supramolecular porphyrin assemblies. Resonant energy transfer to the interacting porphyrins in these self-assembled structures has been also observed, occurrence which is of crucial importance in the design of new efficient artificial antenna systems. The local structural organization of self-assembled porphyrin aggregates can be conveniently studied by exploiting the features of the supramolecular chirality, which can arise from intrinsically chiral assemblies[27, 28, 29, 30, 31, 32] or from aggregation onto chiral templates.[34, 35, 36, 37] Selection of the chirality of a supramolecular structure, in the absence of any templating agents, was carried out by means of macroscopic chiral fields (i.e. vortex motion) during the aggregation process.[32, 33] Such an occurrence suggested to speculate about the role played by the sign of vorticity in relation to the origin of biological chirality. Also intriguing is the possibility to induce chirality on porphyrin aggregates which self-replicate in solution, that is retain the memory of their imprinted chirality.[37] These systems are useful in understanding the transfer of information in biologically relevant aggregation processes. The huge literature reporting the mesoscopic self-assembly of porphrins suggests that the shape and size of the mesoscopic (and also local) structure, as well as the aggregation kinetics and physico-chemical properties, of porphyrin aggregates in solution are not an intrinsic property of porphyrin itself. Rather they depend on the proper choice of the thermodynamic parameters of the solution and templating agent, which can be easily tuned and controlled. For these reasons porphyrin assemblies are extremely good candidates for applications in nanodevices, photoelectric cells, recording devices and as model for light harvesting in antenna systems. In this review article some recent developments in the field of porphyrin self-assembly is presented within the following topics: • theoretical and experimental aspects related to the characterization of structural and dynamical properties of the aggregating species; • kinetic mechanisms of the aggregation process and the consequent structure of the final aggregate; • dependence of the mesoscopic structure and geometry of the aggregates on the thermodynamic parameters of the solution (like porphyrin concentration, ionic strength and pH); • analogy between nonlinear optical properties of porphyrin aggregates and those of metal composites: delocalization of the excitons in submicrometric zones (”hot zones”) generating Raman and Rayleigh scattering enhancement; • supramolecular chirality induction by a templating agent or by an external field and dependence of the symmetry factor on aggregate’s size. All these subjects will be discussed and supported by numerous examples and experimental techniques (Static and Dynamic Light Scattering, Raman Scattering, Uv-Vis, Circular

562

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Differential Extinction and Circular Intensity Differential Scattering). Attention will be devoted to the experimental aspects concerning the mesoscopic systems like for instance the influence of the scattering on the absorption and circular dichroism measurements.

2.

Exciton Delocalization

In a porphyrin assembly the strong interactions give rise to the coupling between excited states (excitons) of the contituent molecules. The optical properties of these structures are described by the Frenkel exciton model, according to which the extent of the exciton wave functions is determined by the competition between intermolecular transfer interactions and (static) disorder. According to a purely exciton model, in fact, the energy and the broadening of the band for an aggregate of N two-level molecules can be obtained by the excitation Hamiltonian: H=

N X

(Emon + Em)b+ m bm +

m=1

N −1 X

+ J(b+ m bm+1 + bm+1 bm )

(1)

m=1

Here, b+ m and bm are the Pauli creation and annihilation operators for de-excitation and excitation on the site m, J is the nearest-neighbour excitation transfer interaction, Emon = hcνmon is the molecular (monomer) two level excitation energy and Em represents the energy offsets introduced by the static disorder. For perfectly ordered aggregates (distribution width of Em being zero), the exciton wave functions are delocalized along all the length of the aggregate. The one-exciton states, |ki (the ground state being |0i), which determine the absorption spectrum, and their k-th exciton energy are given by: |ki =

N X

sin

m=1



πkm b+ |0i N +1 m

νk = νmon + 2Jcos





πk N+1



(2) (3)

If molecules are arranged in a J-structure the exciton coupling constant J is negative, which leads to red-shifts in the spectrum relative to the monomer. By contrast, J is positive for H-type aggregates, with resultant blue-shifts in the spectrum. In disordered aggregates the exciton states are mixed due to the broken translational simmetry, and localize on a part of the aggregate; these states, which can be visible as separate lines when the disorder is small, merge into one broader line when the disorder increases. The width (half-width-at-half-maximum) of this absorption line, Γ, is directly related to the finite number of molecules over which delocalization occurs:[39, 38] Ndel =

2.1.

q

3π 2 |J| /Γ − 1

(4)

Resonance Light Scattering Effects

Exciton coupling and electronic communication among the molecules in an aggregate (especially in a J-type aggregate) can cause extraordinary optical effects: one example is resonance light scattering (RLS), i.e., an increase in the intensity of scattered light at the

Self-assembled Porphyrins at the Mesoscopic Scale

563

Figure 1. Experimental extinction (plot a) and scattering (plot b) spectra of an aggregated H2 T P P S4 solution ([H2 T P P S4 ]=3 µM , pH=1).

wavelength where the aggregate has an electronic absorption transition. For aggregates of sufficient size, the enhanced scattering overwhelms the absorption and a peak in the scattering spectrum appears. Besides in solutions of porphyrin aggregates[40], the phenomenon is observed also in a variety of small chromophores in solution[41, 42], chromophore-protein complexes[43, 44], chromophore-nucleic acid complexes[45, 46] and chlorophyll a aggregates[47]. Figure 1 displays, as examples, the absorption and scattering spectra of an aggregated porphyrin solution (H2 T P P S4 of figure 2). The isolated porphyrin has a Soret band at the frequency νmon , corresponding to 1/νmon = λmon = 434nm, whose width is Γmon = 875cm−1, and the J-aggregate a peak at λJ = 490nm (J = −1275cm−1 ), with Γ ≈ 100cm−1. According to the exciton theory the resonant band width leads to a delocalization over few tens of porphyrins. Even though the exciton theory allows for a calculation of the delocalization length from the spectral width of the absorption band of the aggregate, for large aggregates the scattering component affects the absorption spectra significantly. As a consequence the evaluation of the band width can be misleading. The distinction between scattering and absorption can be conveniently done by measuring independently the extinction and RLS spectra on the same sample and by performing a linear regression analysis of the data in the region close to the resonance.[48, 49, 50] By assuming that only a single excited state of the monomer contributes to the polarizability around the lower energy absorption frequency, α(ν), the quantum mechanical expressions

564

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 2. Scheme of 5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrin ( H2 T P P S4 ) structure. for the absorption and scattering cross-section are: Cabs = 2πνIm(α) h i Csca = 8/3π 3ν 4 (Re(α))2 + (Im(α))2

(5)

Re(α) and Im(α) being the real and imaginary parts of the polarizability for a system where the dipole moment vector has one component[48, 51, 52]. According to this theory, the scattering cross-section Csca depends on the number N of interacting chromophores and the calculation for a J-aggregate can be performed according to the point dipole approximation as described in the literature.[51] Figures 3a) and b) describe the calculated dependence of the absorption and scattering cross-section, respectively, on the number of interacting chromophores. This approach was particularly useful for obtaining the size and the structure of a J-aggregate inside the water pool of a microemulsion on increasing the size of the inner core[48], as shown in the next subsection.

2.2.

Geometrical Arrangement of Porphyrins in a J-aggregate

Porphyrin molecules are optically isotropic, but when they aggregate in a Jarrangement, the large interaction between porphyrins (coherence length of the exciton coupling) makes the free displacements of electrons under incident radiation asymmetric, giving rise to optical anisotropy[21, 48, 51] and depolarization ratio ρdep = IV H /IV V 6= 0 (IV V and IV H being the polarized and depolarized scattered intensity, respectively). Unlike the scattering cross-section, the depolarization ratio is not dependent on the size of the aggregate, but it is related only to the principal values of the polarizability tensor at the resonance wavelength[51, 52]: ρdep =



3 α// − α⊥ h



45 1/3 α// + 2α⊥

i2

2



+ 4 α// − α⊥

2

(6)

being α// and α⊥ the parallel and ortogonal component of the polarizability, respectively, for an axially-symmetric molecule.

Self-assembled Porphyrins at the Mesoscopic Scale

565

Figure 3. Dependence of the absorption (upper plot) and scattering (lower plot) crosssection on the number of constituting monomers (curve a: N=1, curve b: N=2, curve c: N=5, curve d: N=10, curve e: N=20), calculated by using the RLS theory[51] for a J-type aggregate of H2 T P P S4 .

In practice, the depolarization ratio depends on the slip angle φ between adjacent porphyrin planes (see inset of figure 4) and can give information on the geometry of the excited state of an aggregated species. By assuming a parallel arrangement of the transition moments of the exciton-coupled chromophores, in the case of J-aggregates confined in a water pool of a microemulsion the slip angle decreases on increasing the droplet size, as reported in Figure 4. On considering that aggregates grow in the water pool and the dimensions are limited by the confined environment, it is possible to combine the calculated angles with simple geometrical analysis and estimate the aggregation number, Naggr , and the length of the porphyrin clusters in the inner water compartments. Assuming the model recently proposed in the literature[53], in which the porphyrins form monodimensional arrays, we can calculate the coherence length as L = (Ndel + 1)R (where R is the radius of a single molecule). The ratio r = (Ndel + 1)/(Naggr + 1) indicates the quality of the exciton delocalization, being close to unity for an ideal J-aggregate. The maximum value for this parameter reported so far has been 0.25, for J-aggregates obtained in acidic aqueous solutions, using ammonium chloride as the nucleating agent.[53] For J-aggregates in a microemulsion the r values obtained range between about 1 and 0.35 for the smallest core and the largest one, respectively, suggesting a high level of coherence for these aggregates.[48]

566

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 4. Slip angle between porphyrins as a function of the droplet radius of the water/AOT/decane microemulsion (experimental conditions: volume fraction φ = 0.05, [H2 O/AOT ]=5 ÷ 65), [acetate buffer]=25mM, pH=2.7, TPPS stock solution at 80µM ). The inset reports the depolarization ratio dependence on the slip angle, calculated from the RLS theory for a TPPS4 solution with N=50.

3.

Kinetics of Self-assembly

Aggregation of porphyrins (particularly in water solutions) can be induced acting on many parameters, so that the mesoscopic and local arrangement of the final structures vary significantly depending on the induced kinetic mechanism and rate. The rate of aggregation of particles depends on frequency of collision and probability of sticking after collision. These factors are related to the solution parameters such as temperature, pH or other chemical properties that affect particle correlations, and on the presence of an external field. In 1917 Marian Smoluchowski introduced his kinetic expression for the time evolution of particle aggregation; it relates the concentration changes of the species i and j to form species i+ j due to binary collisions.[54] This equation is based on a mean field theory of identical particles (namely it assumes that the probability of two particles meeting is simply proportional to the product of their concentration) and includes a kernel, which depends on the physical properties of the system like shape and size of particles, composition, and interaction potential. The kernel can take many forms, depending on collision interactions among the particles, and in recent years many different forms of the kinetic profile have been obtained analytically and numerically for different systems[55, 56]. For a finite system the Smoluchowski equation takes the form of a system of coupled ordinary differential equations: K(i,j)

Pi + Pj → Pi+j P∞ dck 1P j=1 K(k, j)cj i+j=k ci K(i, j)cj − ck dt = 2

(7)

where Pj denotes an aggregate containing j monomers, K(i, j) is the reaction kernel of creation and destruction of the aggregate and ck is the concentration of the species k. By

Self-assembled Porphyrins at the Mesoscopic Scale

567

considering dilute solutions, in which only binary collisions are relevant, Kernels are configurational and orientational averages of the reaction rates between two colliding species. Smolukowski equation is analytically solvable only for few kernels like the constant kernel K(i, j) = K(1, 1) (i.e. same probability of collision for all clusters and monomers), the sum kernel K(i, j) ≈ i + j, the product kernel K(i, j) ≈ ij and their linear combinations. Another particular case was suggested by Smoluchowski, who obtained an explicit form for the Kernel K(i, j) ∝ 4π(Ri + Rj )(Di + Dj ), by solving the diffusion equation, Ri and Di being the radius and the diffusion coefficient of the aggregate constituted by i monomers, respectively. No general solution exists and for real systems the reaction kernel depends on many parameters, making it difficult to know its form. However, in the case of kinetic mechanism of aggregation driving to clusters with specific properties, like scaling, self-similarity and universality, some considerations and simplifications can be done. For self-similar structures, for example, namely those structures which are space scaleinvariant (like fractals for which Ri = Rmon i1/Df , with Df defined as fractal dimension), the homogeneity relation for the kernel holds:[57, 58] K(ai, aj) = aλ K(i, j)

(8)

with the condition that the kernel for the collision between the species i and j is K(i, j) = iµ j ν , with j >> i and ν ≤ 1. The quantity λ = µ + ν ≤ 2 describes the tendency of a large cluster to join with another large cluster, so determining the overall rate of aggregation; µ, on the other hand, governs the aggregation rate between big clusters and small clusters. The occurrence λ > 1 describes a gelation process which gives rise to an infinite-size cluster in finite time. In this frame, for different values of the quantities λ and ν, different fractal structure of the aggregate is obtained. Among the numerous collision mechanisms which can be considered, particular attention has to be devoted to the diffusion driven collisions. If the sticking probability between smaller and bigger clusters is predominant ( λ = 0 or, more generally λ < ν) the aggregation process gives rise to monodisperse Diffusion Limited Aggregates (DLA) with Df = 2.5, whereas higher sticking probability between bigger clusters ( µ > 0) generates polydisperse Diffusion Limited Cluster-Cluster Aggregates (DLCCA) characterized by Df = 1.75. In the case of the Smoluchowski kernel, for example, the self-preserving shape implies: K(i, j) ≈ (i1/Df + j 1/Df )(i−1/Df + j −1/Df ).

(9)

where the Einstein-Stokes relation D = kB T /(6πηR) was used (however, a more general dependence Di ∝ i−φ can be also considered[59]). The kernel used by Smoluchowski represents an example of DLA mechanism in which λ = 0. From the mean field and Smoluchowski treatment it is clear that in colloidal systems the kinetics of growth and the relative mechanism drive the morphology of the resulting aggregate. Examples of fractal structures built through different aggregation mechanisms in a lattice can be obtained by numerical calculation and molecular dynamics, giving to each lattice site a probability p to be occupied. In a real system p is represented by the probability that,

568

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 5. Numerical calculation for fractal structure originating from DLA (at left) and percolation (at right) mechanism. under collision, two species stick together. In figure 5a) the fractal structures originating by the diffusion-limited aggregation mechanisms is sketched, as an example. A quantitative description of the aggregation process can be done by considering that the distribution of fractal clusters with different mass (or size) is an homogeneous function evolving in a self-preserving form[60], independently of the initial distribution:[61] ¯ ci (t) = i−2 f (i/S(t))

(10)

¯ ∝ tz is the where ci is the concentration of clusters constituted by i monomers and S(t) ¯ average aggregation number of clusters (or in terms of mean cluster radius R(t) ∝ tz/Df ), with z = (1 − λ)−1 . For the Smolukowski kernel given in eq. 9 it is z = 1 which corresponds to the result otained for a DLA growth mechanism. Because it is easy from an experimental point of view to monitor the time dependence of the monomer concentration (an example is reported in the following subsection), let us consider the kernel for the reaction involving only collisions between clusters and monomers: K(1, j) = j ν . The Smoluchowski equation 7 becomes: ∞ X dc1 ¯ ν−1 j ν cj ∝ S(t) = −c1 dt j=1

(11)

the last proportionality being obtained by using eq. 10 for cj (t). By integrating previous equation, the time dependence of the monomer concentration results in a stretched exponential: h i (12) c1 (t) = Aexp −t−(λ−ν)/(1−λ)

According to the mean-field theory[54] and to Molecular Dynamics simulations[57, 62], a reaction-limited (RLA) early stage of the fractal aggregation (with Df = 2.1) is expected to precede the diffusion-limited regime described above. In this early stage the clustering process is hindered by short-range high repulsive barrier, so that many collisions are required before aggregation occurs (low sticking probability). For reaction limited aggregation the

Self-assembled Porphyrins at the Mesoscopic Scale

569

dynamic scaling law of the mean cluster radius is described better by an exponential in¯ ∝ exp(bt), where b depends on the experimental conditions. crease S(t) Due to its transient regime, it is difficult to observe the RLA early stage experimentally, but in the subsection 4.2 and 6.2 it will be shown how to obtain the required information in the case of porphyrin solutions. Another mechanism originating fractal structures is percolation, a model which represents a powerful tool for the study of many physical processes like second order phase transitions or gelation in complex fluids, as well as of porous materials, biological structures, polymers, just to cite some examples. This model assigns a critical threshold concentration (called site percolation probability) at which the largest cluster includes almost all the molecules and spatially extends to all the space[63], as sketched in Figure 5b).

3.1.

Experimental Observation of Porphyrin Fractal Structures

The fractal structure of colloidal aggregates can be studied experimentally by means of Elastic Light Scattering experiment, which measures the zero-time autocorrelation function of the scattered field, ES :[65] G1(Q, 0) = hES∗ (Q, 0)ES (Q, 0)i = I(Q) = KMw cP (Q)S(Q)

(13)

in which P(Q) and S(Q) are the normalized form factor and the structure factor, respectively, Mw the weight average molecular weight of the particle, c the mass concentration, Q = (4πn/λ0)sin(θ/2) and K = [4π 2n2 /(NA λ40)](dn/dc)2 the optical constant (λ0 being the incident wavelength, θ the scattering angle, NA the Avogadro number, n the refractive index and dn/dc the refractive index increment)[64]. The factorization in the product between P(Q) and S(Q) is performed under the hypothesis of independence of intermolecular and intramolecular averages. The form factor, P(Q), represents the intra-particle interference and is related to the spatial Fourier transform of the particle mass distribution, w(r), all over the volume of the particle: R 2 (iQ · r) dr V w(r)exp R P (Q) = V w(r)dr

(14)

For particles smaller than the wavelength (tipically smaller than λ0/20) P (Q) ≈ 1. The structure factor, S(Q), is defined as the Fourier transform of g(r1, r2) − 1, where g(r1, r2) = hρ(r1)ρ(r2)i is the number density correlation function. In the case of selfsimilar systems this function is homogeneous, that is: ′

hρ(σr1)ρ(σr2)i = σ −A hρ(r1)ρ(r2)i

(15)

where ρ(r1) is the number density of monomers at the position r1. For an homogeneous and isotropic system the number density correlation function depends only on the distance between two particles g(r1, r2) = g(|r2 − r1| , 0), simply indicated as g(r). Therefore, the structure factor can be written as: S(Q) = 1 + ρ

Z

[g(r) − 1]exp (iQ · r) dr

(16)

570

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro ′

For self-similar structures eq.15 implies that hρ(r)ρ(0)i ≈ r−A , A′ being related to the space dimension d and to the fractal dimension Df as A′ = d − Df . As already seen in the previous section, the fractal dimension is defined through a scaling law between the number of monomers constituting the cluster with radius R and the radius itself: N (R) = RDf . By using the power-dependence of the number density correlation function on r, the structure factor is easily calculated:[60] S(Q) = Q−Df (17) Previous equation is valid for describing ideal fractals, i.e. clusters extending all over the space scale; real fractal systems, however, have a finite size and at Q values low enough they must obey a Gaussian law (Guinier limit)[66]. Chen and Teixeira[67] proposed a structure factor suitable for describing a finite fractal aggregate constituted by m′ monomers, encompassing Gaussian and fractal behavior. The finite extension of the aggregate is taken into account by an exponential cut-off exp(−r/ξ), with ξ a cutoff correlation length: S(Q) = m′

sin[(Df − 1)arctan(Qξ)] (Df − 1)Qξ(1 + Q2 ξ 2)(Df −1)/2

(18)

Previous equation reduces to eq. 17 if Qξ >> 1 and ξ/Rmon > 50. The theoretical approach described above, along with the experimental results obtained by light scattering, turns out to be extremely powerful in showing that the kernel of the kinetic equation can be easily modified by means of pH, ionic strength and concentration to give rise to aggregates with different fractal morphology. In figure 6, as examples, the intensity profiles of porphyrin solutions at different pH values are reported. By considering that the smallness of porphyrin monomers makes the form factor be approximately unity, the measured intensity profile is directly related to the structure factor. The porphyrin aggregation process is driven by the interparticle potential which can be described through the well-known Derjaguin-Landau-Verwey-Overbeek potential (DLVO),[68] related to the presence of a diffuse double layer surrounding colloidal particles. For H2 T P P S44− solutions containing [HCl] < 0.01 M the protonation of the core (giving rise to the diacid form H4T P P S42− ) originates the zwitterionc form. Under these conditions the electrostatic repulsion between negative charges of the SO3− perypheral groups is still high and is responsible for the high kinetic barrier to the approach of monomers and the low probability of aggregation, so that light scattering experiments do not reveal any detectable amount of aggregates. On increasing the acid concentration (0.1M < [HCl] < 4M ) negative charges are progressively screened by H+ ions and the sticking probability increases accordingly. At [HCl] ≈ 0.1M the interaction between diacid porphyrins is still repulsive (charge 2-) with a long screening length and the weak attractive van der Waals forces ( ∝ r−6 ) are scarcely effective, leading to a small probability of interaction. Therefore, after an induction period due to the small sticking probability, the DLVO potential, whose attractive component depends on the third inverse power of the interparticle distance, favors the adhesion between clusters (µ > 0). As shown in Figure 6, the scattered intensity profile follows a power law with Df = 1.75, indicating that small clusters are able to self-interact

Self-assembled Porphyrins at the Mesoscopic Scale

571

Figure 6. Absolute scattered intensity profile of aggregated H2T P P S4 in water solution under strongly acidic conditions (at the end of kinetics). Experimental conditions: [H2 T P P S4 ]=3µM and [HCl]=0.1 M (squares), [HCl]=1 M (circles), [HCl]=2 M (triangles), λ0 = 532nm. The continuous line has slope equal to -1.75 and the dashed one slope equal to -2.5. leading to larger aggregates with a DLCCA mechanism. For [HCl] ≥ 1M the net charge of the molecules is small enough to lower the screening length and to increase the sticking probability between clusters and monomers ( µ < 0); aggregation takes place according to a DLA mechanism (Df = 2.5). At [HCl] ≈ 4M , the net charge on the monomer becomes very small and the ionic strength is large enough to almost reduce to zero the screening length. The aggregation process is driven almost entirely by attractive forces and it takes place by nucleation, without any fractal arrangement. Any further increase of the acid concentration leads to the full protonation of the sulfonate end groups and to the disruption of the aggregates. Because of the presence of the two positive charges, the H8 T P P S42+ porphyrin does not self-interact even under these extreme ionic strength conditions.

3.2.

Monitoring of the Fractal Growth

As the structural information on the formed fractal cluster can be obtained by Static Light Scattering, the size distribution and dynamic scaling properties can be studied by Quasi-elastic Light Scattering. In fact, by considering that the form factor of monomers is unity, the autocorrelation function of the scattered field contains information on the dynamics of the growing aggregates: G1 (Q, τ ) = hES∗ (Q, 0)ES (Q, τ )i = a

X

Mi2Ni Si (Q)exp(−Γi τ )

(19)

i

where a is a factor depending on the experiment geometry, Mi , Ni and Si represent the mass, number and structure factor of the i-th cluster with radius Ri and the decay rate Γi =

572

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 7. Diffusion coefficient of a fractal aggregate of H2T P P S4 at 3µM with [HCl] = 1M as a function of the exchanged wave vector (at the end of kinetics). The dashed line indicates the Q2 dependence. The incident wavelength is λ0 = 532nm. Di Q2 + A(QR)Θi (Di and Θi being its translational and rotational diffusion, respectively, and A(QR) being related to the anisotropy of the cluster). The initial decay rate of the correlation function is given by: Γ=

1 G1 (Q, 0)

Z

Mi2Ni (t)Si (Q)(DiQ2 + A(QR)Θi)dMi

(20)

In the QR 1, F (QR) ∝ Q3 and no diffusion coefficient can be extracted. For fractal systems, for which the condition QR >> 1 is fullfilled, the diffusion coefficient (and hence the size) are obtained only for rigid structures (absence of internal motions). Thanks to the strong inter-porphyrin interactions inside the aggregate, fractals constituted of porphyrins are rigid, as put in evidence by the dynamic light scattering measurements (see Figure 7). From a practical point of view, although dynamic light scattering is able to measure the mean radius of the growing clusters for the cheking of the dynamic scaling, in the case of porphyrin solutions the rate of the kinetic process is faster than the time required for data aquisition. Alternatively, UV-vis measurements guarantee a higher dynamic range and allow for the determination of the time-dependence of the monomer concentration and an estimation of that of the mean cluster mass. Figure 8a) reports the time evolution of the spectral contribution of monomeric t − H2 Pagg

Self-assembled Porphyrins at the Mesoscopic Scale

573

Figure 8. Integrated area of the absorption band of the monomeric t − H2Pagg , centered at 420 nm (plot a), and of the extinction band of the aggregated form, centered at 452 nm (plot b). Experimental conditions: [t − H2Pagg ]=5µM , [N aCl] = 0.1M . The continuous line in plot a) represents the fit according to eq. 21 and that in plot b) a power law with exponent z ≈ 1. porphyrins (see figure 9 for the structure) which are depleted from solution to originate clusters. The theoretical prediction of a stretched exponential (eq. 12) is valid in the late aggregation stage and it cannot describe the whole kinetic. As already pointed out a complete description of the measured monomer concentration must take into account the RLA early stage, during which the depletion of monomers follows an exponential form[70]. So that the time dependence of the monomer concentration can be represented by:[71, 72] c1 (t) = A0 + A1exp −(k1 t)−γ + A2 exp(−k2t) 



(21)

where k1 and k2 are two observed time constants, γ = (λ − ν)/(1 − λ) and each process is weighted by a proper amplitude factor. The continuous line in Figure 8a) represents the fit

Figure 9. Scheme of the trans-bis(N-methylpyridinium-4-yl)diphenylporphyine ( t − H2 Pagg ) structure.

574

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

with eq. 21 and shows a very good agreement with the experimental behaviour. On the other hand, by following the time evolution of the increasing spectral contribution ascribed to porphyrins in the aggregated form, it appears that it is obeying (at least before the levelling off due to finite concentration constraints) the power law of the mean cluster mass, with z ∼ = 1 (see Figure 8b). This implies a value λ ≈ 0 and suggests that, after the very early stage driven by RLA, a DLA growth mechanism occurs. The dynamic scaling of the cluster mass is fullfilled likely due to the predominance of resonant scattering, which depends on the weight average molecular mass of the aggregate, on the absorption at the characteristic wavelength of the J aggregate.

4.

Tuning and Control of the Aggregate Mesoscopic Structure

A careful survey of the literature about aggregation of the anionic TPPS4 porphyrin shows that different authors found different structural arrangements for the final aggregate[73, 74, 75, 76], especially regarding structure and size. In this section it will be shown that structural changes of porphyrin aggregates can occur in aqueous solutions by simply changing porphyrin concentration and ionic strength. These parameters, acting on the interaction potential and on the aggregation kinetic rate, lead to final aggregates with different structure and size. The resonant scattering allows for collecting experimental data with a good signal to noise ratio even in the depolarized configuration and for determining the reorientational motions of the porphyrin aggregates. To this aim all the components of the polarizability tensor, in principle, must be measured, but, for sake of simplicity, let us assume a cylindrical symmetry of the polarizability tensor. For a dilute solution of monodisperse noninteracting N particles, resulting from the aggregation of small molecules (small with respect to the wavelength of radiation), the Rayleigh-Debye-Gans approximation can be used. Thus, the normalized polarized and depolarized field autocorrelation functions in the absence of correlation between translational and rotational motions are:[64, 66, 77, 78] 

GV V (Q, t) = N S(Q) α2iso +

  4 2 β exp (−6Θt) exp −DQ2 t 45 

(22)

  β2 exp (−6Θt) exp −DQ2 t (23) 15 where αiso is the isotropic excess polarizability of the particle ( 1/3T r(α)) with respect to the solvent, β the anisotropy of the particle polarizability and D and Θ the translational and rotational diffusion coefficient, respectively. The latter two coefficients can be obtained experimentally from the initial decay rate of the field correlation functions, as follows:

GV H (Q, t) = N S(Q)

ΓV V (Q) = DQ2 +

6Θ 45α2iso 4β 2

(24)

+1

ΓV H (Q) = DQ2 + 6Θ

(25)

From equations 22 and 23 taken at t=0, the depolarization ratio is written as ρdep = β 2 /(15α2iso + 4/3β 2), indicating that the intensity of the depolarized scattering becomes

Self-assembled Porphyrins at the Mesoscopic Scale

575

relatively more important the larger the optical anisotropy of the scattering object. The hypothesis of independence of rotational and translational motions leading to eqs. from 22 to 25 remains valid for dilute solutions until the coupling parameter Q2 ∆D/Θ < 5 (where ∆D = D// − D⊥ is the anisotropy in the translational diffusion motion).

4.1.

From Fractal to Rod-Like Structures

As it was already shown, in dilute solution of water soluble porphyrin aggregation is triggered by lowering pH enough to shield the charged sulfonate groups and allow a closer approach of molecules to occur; the same effect is obtained by increasing ionic strength. In both cases extended fractal structures are formed.[20, 21] In the presence of salt the autocorrelation function relaxation rate in polarized and depolarized configuration does not indicate any difference, as shown in Figure 10a) and b). This finding indicates that rotational motions do not contribute to the width of the quasielastic optical spectrum and eqs from 22 to 25 are not distinguishable one from another. Indeed, although aggregates are optically anisotropic, the rotational diffusion of the whole aggregate is too slow to be detected. Moreover, the Q2-dependence of the relaxation rate proves that, despite their large size (RH ≈ 0.85µm), the strong intermolecular interaction gives rise to internally rigid structures. The intensity profile shown in Figure 10c) clearly indicates a fractal arrangement of porphyrins. The Q-range investigated by a combination of data at wide and small angles is wide enough to display the bending at small Q values, representing the effect of the cutoff of the density correlation function described in eq. 18. From the fit it results that this cutoff is ξ ≈ 0.7µm and the fractal dimension is Df ≈ 2.2 (consistent with that found for an analogous system[79]). It has been shown recently[80] that the fractal dimension of a DLCCA and an RLCA aggregate is an increasing function of the monomer aspect ratio, so that aggregation of rod-shaped building blocks leads to a loss of distinction between the two kinetic mechanisms of growth as the axial asymmetry increases. By considering that the building blocks of the fractal are the strongly exciton coupled porphyrins, their nonspherical shape can be responsible of the underestimated value of the fractal dimension deriving from a DLA growth mechanism. √ ∼ The gyration radius of the aggregates, Rg = 3ξ = 1.2µm, along with the value of the hydrodynamic radius, allows for an estimation of the fractal porosity through their ratio RH /Rg ∼ = 0.7; this value is in good agreement with that expected for fractals with an exponential cutoff in the density correlation function.[81, 82] When salt is not added to the solution, aggregation of T P P S4 is fostered at mild acidic conditions by increasing porphyrin concentration[21, 75]; under this condition porphyrins aggregate in rod-like mesoscopic structures. In the concentration range from 40 up to 500 µM , polarized and depolarized autocorrelation functions have significatively different relaxation rates one from another, as shown with some examples in Figure 11. From eqs.24 and 25 it is clear that the zero-Q value of the depolarized relaxation rate gives the rotational diffusion coefficient, whereas the slope of the Q2-dependence of both polarized and depolarized relaxation rate furnishes the translational diffusion coefficient. In order to adopt a reasonably corrected model for calculating the aggregate size, let us consider that the structure factor of molecules in the aggregate can be regarded as the form

576

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 10. Relaxation rate of the polarized (plot a) and depolarized (plot b) correlation function; the continuous lines represent the Q2 -dependence with the same slope. Plot c) reports the whole intensity profile along with the fit with eq. 18, typical of finite fractals. Experimental conditions: [H2 T P P S4 ]=3µM , pH=2.7, [NaCl]=1.5M, λ0 = 532nm factor P(Q) of the aggregate itself; moreover, if solution is diluted enough to neglect interactions between aggregates the structure factor can be approximated to unity. In this frame the form factor is simply proportional to the measured intensity profile through eq.13. Figure 12 displays that the scattered intensity of the concentrated porhyrin solutions (in the absence of added salt) can be well fitted by a rod form factor with length L, described by: P (Q) =

2 QL

Z

0

QL

2 QL sin(x) dx − sin x QL 2 



2

(26)

Then, for a diffusion coefficient of a rod the following Broesma’s equations hold:[83, 84] Θ= D=


3kB T ′ δ −ζ 3 πηL

  kB T  ′ δ − 1/2 γ// + γ⊥ 3πηL

(27) (28)

with D=

D// + 2D⊥ 3

(29)

Self-assembled Porphyrins at the Mesoscopic Scale

577

Figure 11. Relaxation rate of the polarized (plot a) and depolarized (plot b) correlation functions at pH=2.7 for two concentration values: [ H2 T P P S4 ]=40µM (squares), λ0 = 532nm and [H2 T P P S4 ]=500µM (circles), λ0 = 780nm. The continuous lines represent the Q2 -dependence with the same slope for the same symbols and the dashed lines indicate the zero-Q value of the depolarized relaxation rate.

Figure 12. Absolute scattered intensity profile for the H2 T P P S4 solutions for different concentration values: 40µM (squares) at λ0 = 532nm, 250µM (diamonds) at λ0 = 780nm, 500µM (circles) at λ0 = 780nm, all at pH=2.7. The continuous lines are the fit according to the rod form factor (eq. 26). kB T ′ (δ − γ//) 2πηL kB T ′ (δ − γ⊥ ) D⊥ = 4πηL   2 ′ δ = ln ψ

D// =

(30) (31) (32)

578

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

ψ being the ratio between diameter and length of the rod and ζ, γ//, γ⊥ functions of the parameter δ ′ [83, 84]. By inserting the measured values of the translational and rotational diffusion coefficients in the Broesma’s equations, the solution of the non-linear systems of equations gives the geometrical parameter of the rod (L and ψ). Both static and dynamic measurements agree in indicating that the length of the rod depends on porphyrin concentration and in particular from 40µM to 250µM it is about 0.5 µm and then decreases to about 0.2 µm at 500µM ; moreover, the dynamic measurements add information on the rod diameter, ψL ∼ = 10nm. The decrease of the length with increasing concentation can be attributed to the lower steric hindrance of a large number of shorter rods with respect to a small number of longer rods. The increase of absorbance, and hence of the local heating of the sample upon irradiation, makes it more difficult to investigate higher concentration values by means of light scattering; however, X-ray measurements on the same porphyrin by Gandini et al.[75] showed that, at concentration values ten times higher and above, porphyrins are arranged in even shorter rods (about 35 nm). These authors proposed two plausible structures for the rods: one consisting in planar association (sheets) into a ringlike configuration (hollow cylinders) and the other deriving by the formation of a continuous shallow helix without disruption into separate layers. This tubular structure has been found also for other self-assembled bio-systems like bacterial antenna chlorophyll[85], guanosine nucleoside[86] and melanin particles[87]. Also porphyrin derivatives under proper conditions were proved to form fiber-like aggregates[88, 89] with only one molecule per cross section. In section 7.1 it will be shown that, in the presence of a chiral templating agent, rod-like H2 T P P S4 J-aggregates are more consistent with an helical arrangement for which the projection of the transition dipole moment is mostly perpendicular to the long axis of the rod,[90] as also observed for other J-aggregate wires.

4.2.

Effects of Initial Conditions in the Diffusion-Limited Aggregation Kinetics

The kinetic profile of a real diffusion controlled aggregation process is strongly dependent not only on the initial concentration of the reactant but also on their order of mixing (i.e. initial spatial distribution of reactants).[91] Another important factor influencing the aggregation kinetic profile is the presence of nucleation centres, for example small aggregates (seeds) already formed in an aged stock reactant, which represent a divergence in the spatial distribution of reactant itself. As far as self-assembly of porphyrins in solution are concerned, dramatic effects on kinetics (and also on the structure) are observed varying the initial spatial distribution of reactants (or using aged stock concentrated porphyrin solution). It was suggested that a protocol to mix the reagents must to be set up in order to get reproducible results, besides a careful control of the thermodynamic conditions.[92, 93] Pasternack et al., for instance, put in evidence the importance of the mixing order in studying the kinetic of supramolecular assembling of the t − H2Pagg porphyrin onto the surface of DNA[94] and the formation of J-aggregates from H2 T P P S4 [93]. Also the relative concentrations of the stock solution of reagents play an important role, as well as the lag time between injecting a reagent and mixing the solu-

Self-assembled Porphyrins at the Mesoscopic Scale

579

tion. This peculiar behaviour is also characteristic of most simple reactions as bimolecular diffusion limited reaction (A + B → C) were an ”anomalous” rate law is observed[95, 96, 97]. The anomalous effects result from the preservation of a ”memory” of the initial spatial reactant distribution, which can be minimized by thorough continuous randomization (stirring). In many diffusion limited reactions, however, the reaction progresses very quickly (at finite reactant concentrations), so that it is difficult to randomize the reactant distribution. In such a case it is difficult to fix a well defined initial time for the reaction since the mixing itself takes time. Also for this reason in realistic case of diffusion limited reaction, the inadequate mixing and the initial reactant distribution effects are dramatic. The effects of the initial distribution of reactants on the kinetics is explicity taken into account in the differential equation describing the time evolution of the density of a pair of species in the simpler case of the bimolecular diffusion limited reaction (e.g. recombination):[98] ∂ρ (33) = −Lρ = − (L0 + kr S) ρ ∂t where kr is the rate of reaction of pairs, S is a term containing information on the efficiency of reaction at different distances of the two species and their mutual orientation, and, for diffusive processes it can be written L0 = −D∇2 ρ. In the absence of reaction the density of pairs in the volume V is at the equilibrium ρ = ρeq ≈ g(r)/V and L0 ρeq = 0. By transferring the space-time dependence of the density of pairs to the probability that the R pair survive at time t[98], P (t) = drρ (r, t), equation 33 becomes: dP =− dt

Z

t

0

V −1 kr (τ )P (t − τ )dτ − keq I(t)

(34)

with I(t) a term depending on the rate constant at equilibrium, on the efficiency of reaction and on the ratio ρ(0)/ρeq (ρ(0) being the initial distribution). Equation 34, therefore, takes into account the initial distribution of reactants through the term I(t),which is equal to zero only when ρ(0) = ρeq . In the next subsections some examples of the effects of zero-time conditions on the aggregation of porphyrins are reported. 4.2.1. Effects of Mixing Order and Aging in Porphyrin Aggregation Two different mixing methods can be adopted in preparing H2 T P P S4 aqueous solutions in the presence of other reagents required for triggering aggregation: adding a stock solution of concentrated porphyrin as last reagent (porphyrin-last mixing, PL) or as first reagent at the desired concentration in water (porphyrin-first mixing, PF). Although UV-vis measurements do not distinguish between the two final solutions, being sensitive only to the local molecular arrangement, light scattering puts in evidence two significatively different mesoscopic structures. In fact, as previously seen, the kinetics of growth and the relative mechanism drive the morphology of the resulting aggregate. The intensity profiles reported in Figure 13 put in evidence that at higher pH aggregates exhibit a fractal structure when the porphyrin is added in an acidic environment (PL mixing), while they form statistical isotropic objects using the PF mixing. The structure of such

580

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 13. Absolute scattered intensity profile for the H2 T P P S4 solutions ([H2 T P P S4 ]=3µM ) at different concentration of hydrochloric acid ([HCl]=0.1 M (plot a) and [HCl]=2 M (plot b)) for PL (squares) and PF (circles) mixing order (λ0 = 532nm). isotropic objects can be described by the Ornestein-Zernike behavior: 

I(Q) ∝ 1 + ξc2Q2

−1

(35)

ξc being the correlation length. In the PF mixing case there is no fractal arrangement in the whole investigated Q range. At lower pH, but still high enough not to break the aggregates (see section 4.1), the distinction between the structures formed with the two mixing methods is progressively lost. All these occurrences can be ascribed to the inevitable concentration gradient of the added last reagent, which causes different nucleation probability in some parts of the system. In the case of the acidic porphyrin solutions this effect becomes more important when the final concentration of acid is not high and is mainly due to the large volumetric ratio between the reagent solutions to be mixed. Another example is given by the porphyrin aggregation in the presence of a polyamine, spermine; also in this case the mixing order is a key parameter. In figure 14 the intensity profiles shows that, in the PF mixing, aggregates take a fractal structure with a fractal dimension of Df ≈ 2.5 and radius of 3µm (as obtained by the fitting with eq. 18), whereas in the PL mixing no fractal arrangement is observed (aggregate radius ≈ 0.6µm). The presence of small aggregates pre-formed in the porphyrin solution before adding the reactant and inducing aggregation can act as nucleation centers. The number and the size of these pre-existent aggregates affect, in a uncontrolled way, the features of the aggregation

Self-assembled Porphyrins at the Mesoscopic Scale

581

Figure 14. Absolute scattered intensity profile for the H2 T P P S4 solutions ([H2 T P P S4 ]=3µM , λ0 = 532nm) in the presence of 30-fold higher spermine concentration for PF (plot a) and PL (plot b) mixing order. kinetics, e.g. the induction time and rate. The induction time is the apparent initial reaction inactivity, often observed experimentally in porphyrin aggregation[94], which makes sigmoidal-shaped the kinetic profile. An example of the aging effect on induction time and rate of aggregation is displayed in figure 15.

5.

Formation of Organic Fractal Composites

The resonance properties of porphyrin J-aggregates have proved to be extremely useful for the study of their structure and dynamics, as well as for insights in the aggregation kinetic mechanisms of fractals. But porphyrin fractals reserve other interesting electromagnetic phenomena, like the scattering enhancement with scaling properties analogous to fractal metal composites. In composites (thin films, colloidal aggregates, etc...), which are formed by non-linear material embedded in a host medium (either linear or non-linear), the enhancement of the linear and non-linear optical response originates from the strongly fluctuating local fields, Ei , which exceed the applied one, E0.[99, 100, 101, 102] The local field enhancement is described by the factor G=

N X 1 |Ei|2 . N |E0 |2 i=1

(36)

N being the number of monomers constituting the cluster. By considering the composite constituted by N polarizable small (with respect to incident wavelength) monomers with short range dipole-dipole coupling, the local field, acting on the i-th monomer is a superposition of the incident and all the scattered waves:[103, 104] Eis = α−1 mon dis = E0s exp(ik0 · ri ) −

N X

j6=i=1

Wsl (ri − rj )djl

(37)

582

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

Figure 15. Kinetic profiles of t − H2 − Pagg aggregation ([t − H2 − Pagg ]=5µM with [NaCl]=0.1M) obtained by starting from a freshly prepared (within a week) stock solution (continuous curve) and from the same stock aged more than one month (dashed curve). with αmon the isolated monomer polarizability, and s and l indicate the different orientation in the space. Previous coupled-dipole equation means that the amplitude of the transition dipole moment di at the i-th monomer within the composite is related to the external field E0 through a light-induced dipolar interaction W . The condition of short range dipole-dipole coupling is not so restrictive for fractal composites because of the strong localization which makes the interaction of monomers at distances greater than the excitation wavelength negligible. The dipole-dipole interaction, expressed in eq. 37, can be synthetically written as: (Z + W )|d >= |E0 >

(38)

where Z = α−1 mon and |d > is the state of the light-induced dipolar moments. Because polarizability is, in general, complex, let us define X = −Re (1/αmon ) and δ = −Im (1/αmon ) its real and imaginary part, respectively. The parameter X has the meaning of a relative frequency detuning and δ determines the resonance width (related to the dissipation within a cluster). The optical properties are easily related to the polarizability α of the system, α = P s and l for sake of brevity) being obtained through the 1/N N i=1 αi , αi (omitting indeces P solution of equation 38 as αi = nj [(< is|n >< n|jl)/(Z + wn )]. The eigenvalues wn of W (with eigenvectors |n > traditionally called surface plasmons) contribute to the width of the resonant band. With these definitions the enhancement described by equation 36 becomes:[104] h

i

G = δ 1 + (X/δ)2 Im(α).

(39)

The ratio |X| /δ is called quality factor, because, when |X| >> δ, the enhancement factor can be very large. Dynamic simulations showed that enhancements and fluctuation of local fields in non-fractal composites are significantly lower than for fractals.[105]

Self-assembled Porphyrins at the Mesoscopic Scale

583

The enhancement in a composite concerns both non-linear (e.g. harmonic generation phenomenon and Kerr effect) and linear (e.g. scattering, absorption and fluorescence) optical response and is defined as the ratio between the optical response from monomers in a cluster and that of the same number of isolated monomers. As far as linear properties are con  (0) cerned, Rayleigh scattering enhancement factor, for instance, is GR = σs / N σs , where (0)

σs is the scattering cross-section of the composite and σs that of the single monomer. For Raman scattering, on the other hand, the enhancement factor depends also on the Raman (S) shift. By considering that the dipole moment di induced at the Stokes-shifted frequency (S) νS on an isolated monomer (with Raman polarizability χi ) interacts, besides with the local field, with other dipoles, the enhancement of the Raman scattering is:

GRS =

  P (S) 2 i di

(40)

N |χ|2 |E0|2

where |χ| is the Raman polarizability of the monomer (due to the incoherent nature of the Raman scattering hχ∗i χj i = χ2 δij ) and averages are over orientations. For very large Stokes shift, outside the aborption band of the cluster the enhancement factor is due only to the local field and GRS reduces to eq. 39. But when the Stokes shift is small also the Raman amplitudes are enhanced according to: GRS ≈

*

|Ei |4

|E0 |4

+

E

D

4

≈ |αi |4 α−1 0 .

(41)

More in general for a non-linear optical process ∝ E n the enhancement can be written as: G=

N X 1 |Ei |n ≈ |X|n δ 1−n Im(α) n N |E0| i=1

(42)

where the right-hand of the previous equation follows from eq.39 under the condition |X| >> δ. The enhancement of the optical response is expecially high in composites with fractal morphology for which the breaking of the translational invariance causes the localization of the excitations (eigenmodes) in subwavelength regions; in these ’hot zones’ absorption by monomers is much higher than by other monomers in a fractal composite. Moreover, the scale invariance gives to the optical response of these composites interesting scaling properties provided that the eigenmodes delocalization (i.e. coherence length) occurs over a distance in between the characteristic spacing between the nearest monomers and the radius of the cluster (R0 δ, the following scaling law was derived:[100, 104] 

d0 −1

Im(α) ≈ R30 R30 |X|

(43)

with 0 ≤ d0 ≤ 1 the optical spectral dimension, which is analogous to the spectral dimension appearing in the scaling law for the vibrational density of states. Also the coherence

584

Valentina Villari, Norberto Micali and Luigi Mons´u Scolaro

length of dipolar excitations obeys a scaling law involving the same optical spectral dimen
(d −1)/(3−Df ) . sion L ≈ R0 R30 |X| 0 From eqs. 42 for Rayleigh scattering (n=2) and 43 it follows[103] that the Rayleigh scattering enhancement of a fractal composite obeys the scaling law GR ≈

d0 +1 N  3 R |X| 0 R30 δ

(44)

for clusters smaller than the excitation wavelength ( kR