Functionalized Nanoscale Materials, Devices and Systems
NATO Science for Peace and Security Series This Series presents the results of scientific meetings supported under the NATO Programme: Science for Peace and Security (SPS). The NATO SPS Programme supports meetings in the following Key Priority areas: (1) Defence Against Terrorism; (2) Countering other Threats to Security and (3) NATO, Partner and Mediterranean Dialogue Country Priorities. The types of meeting supported are generally "Advanced Study Institutes" and "Advanced Research Workshops". The NATO SPS Series collects together the results of these meetings. The meetings are coorganized by scientists from NATO countries and scientists from NATO's "Partner" or "Mediterranean Dialogue" countries. The observations and recommendations made at the meetings, as well as the contents of the volumes in the Series, reflect those of participants and contributors only; they should not necessarily be regarded as reflecting NATO views or policy. Advanced Study Institutes (ASI) are high-level tutorial courses intended to convey the latest developments in a subject to an advanced-level audience Advanced Research Workshops (ARW) are expert meetings where an intense but informal exchange of views at the frontiers of a subject aims at identifying directions for future action Following a transformation of the programme in 2006 the Series has been re-named and re-organised. Recent volumes on topics not related to security, which result from meetings supported under the programme earlier, may be found in the NATO Science Series. The Series is published by IOS Press, Amsterdam, and Springer, Dordrecht, in conjunction with the NATO Public Diplomacy Division. Sub-Series A. B. C. D. E.
Chemistry and Biology Physics and Biophysics Environmental Security Information and Communication Security Human and Societal Dynamics
http://www.nato.int/science http://www.springer.com http://www.iospress.nl
Series B: Physics and Biophysics
Springer Springer Springer IOS Press IOS Press
Functionalized Nanoscale Materials, Devices and Systems
edited by
A. Vaseashta Nanomaterials Processing and Characterization Laboratories, Marshall University, Huntington, WV, U.S.A.
I.N. Mihailescu National Institute for Lasers, Plasma and Radiation Physics, “Laser-Surface-Plasma Interactions” Laboratory, Bucharest-Magurele, Romania
Published in cooperation with NATO Public Diplomacy Division
Proceedings of the NATO Advanced Study Institute on Functionalized Nanoscale Materials, Devices and Systems for Chem.-bio Sensors, Photonics, and Energy Generation and Storage Sinaia, Romania 4–15 June 2007
Library of Congress Control Number: 2008931994
ISBN 978-1-4020-8902-2 (PB) ISBN 978-1-4020-8901-5 (HB) ISBN 978-1-4020-8903-9 (e-book)
Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. www.springer.com
Printed on acid-free paper
All Rights Reserved © 2008 Springer Science + Business Media B.V. No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work.
TABLE OF CONTENTS Preface................................................................................................................xi Part I. Invited Contributions Nanoscale Materials, Devices, and Systems for Chem.-Bio Sensors, Photonics, and Energy Generation and Storage A. Vaseashta...................................................................................................3 Nanostructured Thin Optical Sensors for Trace Gas Detection C. Ristoscu, I.N. Mihailescu, D. Caiteanu, C.N. Mihailescu, Th. Mazingue, L. Escoubas, A. Perrone, and H. Du....................................29 X-Ray Photoelectron Spectroscopy and Tribology Studies of Annealed Fullerene-like WS2 Nanoparticles F. Kopnov, R. Tenne, B. Späth, W. Jägermann, H. Cohen, Y. Feldman, A. Zak, A. Moshkovich, and L. Rapoport .................................51 The Development and Application of UV Excimer Lamps in Nanofabrication I.I. Liaw and I.W. Boyd ................................................................................61 Functionalization of Semiconductor Nanoparticles M.-I. Baraton ...............................................................................................77 Flexoelectricity: A Universal Sensoric Mechanism in Biomembranes and in Chem.-Biosensors A.G. Petrov...................................................................................................87 Carbon Nanotubes: From Fundamental Nanoscale Objects Towards Functional Nanocomposites and Applications W. Maser, A.M. Benito, E. Muñoz, and M. Teresa Martínez .....................101 Ultrashort Pulse PLD: A Technique for Nanofilm Fabrication T. Szörényi and Zs. Geretovszky ................................................................121 Laser Ablation and Laser Induced Plasmas for Nanomachining and Material Analysis D. Batani ....................................................................................................145
v
vi
TABLE OF CONTENTS
Photo-, Dual- and Exoelectron Spectroscopy to Characterize Nanostructures Y. Dekhtyar ................................................................................................169 Laser Interaction with Nano-Spheres: Applications in Sub-Micron Particles Removal and Nanodot Array Fabrication M. Sentis, D. Grojo, Ph. Delaporte, and A. Pereira ..................................185 Clean Fossil Fuels: Advanced Membrane Reactors T. Tran, K. Stoitsas, and J. Schoonman .....................................................199 Nanocrystalline Diamond Films for Advanced Technological Applications C. Popov and W. Kulisch ...........................................................................215 Diamond like Carbon Films: Growth and Characterization S. Tamuleviþius and Š. Meškinis ................................................................225 Fundamentals of Laser-Assisted Fabrication of Inorganic and Organic Films J. Schou ......................................................................................................241 Nanoparticles of Semiconductors in Sol Gel Glasses R. Reisfeld ..................................................................................................257 Electrochemical Sensor Technology Based on Nanomaterials for Biomolecular Recognitions A. Erdem ....................................................................................................273 The Effects of Doping with Elements from the IIA Group on the Thermal and Electronic Properties of Amorphous Selenium G. Belev, D. Tonchev, S.O. Kasap, and H. Mani .......................................279 Nanoscale Materials for Hydrogen and Fuel Cell Systems M. Suha Yazici ...........................................................................................283 Application of Fe-Nanoscale Materials Useful in the Removal of Arsenic from Waters M. Vaclavikova, K. Stefusova, S. Jakabsky, S. Hredzak, and G. Gallios ...291 Nanopatterning Using the Bioforce Nanoenabler K. Arshak, O. Korostynska, and C. Cunniffe .............................................299
TABLE OF CONTENTS
vii
Electrocatalysts and Electrode Design for Bifunctional Oxygen/Air Electrodes V. Nikolova, P. Iliev, K. Petrov, T. Vitanov, E. Zhecheva, R. Stoyanova, I. Valov, and D. Stoychev...........................................................................305 Part II. General Contributions Preparation of Magnetic Chitosan Nanoparticles for Diverse Biomedical Applications D. Kavaz, T. Çirak, E. Öztürk, C. Bayram, and E.B. Denkbaú ..................313 Anomalous Behavior of Carbon Filled Polymer Composites Based Chemical and Biological Sensors K. Arshak, C. Cunniffe, E. Moore, and A. Vaseashta ................................321 Poly(n-isopropylacrylamide) (PNIPAM) Based Nanoparticles for In Vitro Plasmid DNA Delivery N. Ozdemir, A. Tuncel, M. Duman, D. Engin, and E.B. Denkbas..............325 Rapid, Contactless and Non-Destructive Testing of Chemical Composition of Samples O. Ivanov, L. Stoychev, and A. Vaseashta..................................................331 Synthesis and Application of Metal-Containing Silicas K. Katok, V. Tertykh, and V. Yanishpolskii................................................335 Semiconducting Gas Sensors, Remote Sensing Technique and Internet GIS for Air Pollution Monitoring in Residential and Industrial Areas O. Pummakarnchana, V. Phonekeo, and A. Vaseashta..............................339 Self-Assembled System of Semiconductor and Virus like Nanoparticles Yu. Dekhtyar, A. Kachanovska, G. Mezinskis, A. Patmalnieks, P. Pumpens, and R. Renhofa......................................................................347 Thermal Stability and Optical Activity of Erbium Doped Chalcogenide Glasses for Photonics D. Tonchev, K. Koughia, S.O. Kasap, K. Maeda, T. Sakai, J. Ikuta, and Z.G. Ivanova........................................................................................351
viii
TABLE OF CONTENTS
XRD Study of Pulsed Laser Deposited AlN Films with Nanosized Crystallites S. Bakalova, A. Szekeres, A. Cziraki, E. Gyorgy, S. Grigorescu, G. Socol,and I.N. Mihailescu .....................................................................357 Functionalization of Multi-Walled Carbon Nanotubes (MWCNTs) M. Mohl, Z. Kónya, Á. Kukovecz, and I. Kiricsi ........................................365 Sonochemical Synthesis of Inorganic Nanoparticles J. Kis-Csitári, Z. Kónya, and I. Kiricsi ......................................................369 Novel Transparent Molecular Crystals of Carbon G. Kharlamova, N. Kirillova, A. Kharlamov, and A. Skripnichenko .........373 Hydrogen Microsensor Based on Nio Thin Films I. Fasaki, M. Antoniadou, A. Giannoudakos, M. Stamataki, M. Kompitsas, F. Roubani-Kalantzopoulou, I. Hotovy, and V. Rehacek...379 Design and Characterization of Styrene-Based Proton Exchange Membranes D. Ebrasu, I. Petreanu, L. Patularu, I. Stefanescu, and M. Valeanu .........383 Strontium-Substituted Hydroxyapatite Thin Films Grown by Pulsed Laser Deposition C. Capuccini, E. Boanini, A. Bigi, M. Gazzano, F. Sima, E. Axente, and I.N. Mihailescu....................................................................................389 Growing Thin Films of Charge Density Wave System Rb0.3MoO3 by Pulsed Laser Deposition D. Dominko, D. Starešiniü, K. Biljakoviü, K. Salamon, O. Milat, A. Tomeljak, D. Mihailoviü, J. Demšar, G. Socol, C. Ristoscu, I.N. Mihailescu, and J. Marcus..................................................................399 Single Cell Detection with Driven Magnetic Beads B. McNaughton, R.R. Agayan, V.A. Stoica, R. Clarke, and R. Kopelman .403 Antimicrobial Properties of Titanium Nanoparticles B.K. Erdural, A. Yurum, U. Bakir, and G. Karakas...................................409 CsHSO4/Nanooxide Polymer Membranes for Fuel Cell A. Andronie, A. Morozan, C. Nastase, F. Nastase, A. Dumitru, S. Vulpe, I. Stamatin, and A. Vaseashta ....................................................................415
TABLE OF CONTENTS
ix
IV and CV Characteristics of Multifunctional Ilmenite-Hematite 0.67FeTiO3-0.33Fe2O3 C. Lohn, W.J. Geerts, C.B. O’Brien, J. Dou, P. Padmini, R. Schad, and R.K. Pandey, .......................................................................................419 Electrodeposition of BI1-XSBX Nanowires as an Advanced Material for Thermoelectric Applications J.E. Weber, A. Kumar, and W.G. Yelton ....................................................425 A Solid State Nano-Generator: Concept, Design and Theoretical Estimations M. Vopsaroiu, M.G. Cain, V. Kuncser, and J. Blackburn..........................431
Applications of Statistical Physics to Mixing in Microchannels: Entropy and Multifractals M. Kaufman, M. Camesasca, I. Manas-Zloczower, L.A. Dudik, and C. Liu...................................................................................................437 Synthesis and Characterization of Carbon Supported Pd and PtPd Catalysts for DMFCs A. Morozan, A. Dumitru, C. Mirea, I. Stamatin, F. Nastase, A. Andronie, S. Vulpe, C. Nastase, and A. Vaseashta ... .................................................445 Theoretical Study of the Adsorbed Small Molecule on Twisted Nanotubes by Atomic Scale Simulations V. Chihaia, A. Ghita, B.-S. Seong, and S.-H. Suh .. ...................................449 The Defect Structure of Copper Indium Disulfide D. Perniu, A. Duta, and J. Schoonman .....................................................457 ASI Group Photograph....................................................................................465 ASI Group Photograph Legend.......................................................................467 Selected photographs Taken During the ASI..................................................469 List of Participants...........................................................................................475 Author Index....................................................................................................483 Keyword Index................................................................................................487
PREFACE The primary objective of the NATO Advanced Study Institute (ASI) titled “Functionalized Nanoscale Materials, Devices, and Systems for Chem.-Bio Sensors, Photonics, and Energy Generation and Storage” was to present a contemporary and comprehensive overview of the field of nanostructured materials and devices and its applications in chem.-bio sensors, nanophotonics, and energy generation and storage devices. The study has become one of the most promising disciplines in science and technology, as it aims at the fundamental understanding of new physical, chemical, and biological properties of systems and the technological advances arising from their exploration. Such systems are intermediate in size, between the isolated atoms and molecules and bulk material, where the unique transitional characteristics between the two can be understood, controlled, and manipulated. Nanotechnologies refer to the creation and utilization of functional materials, devices, and systems with novel properties and functions that are achieved through the control of matter, atom-by-atom, molecule-by-molecule, or at a micro-molecular level. Advances made over the last few years provide new opportunities for scientific and technological developments in nanostructures and nanosystems with new architectures with improved functionality. The field is very actively and rapidly evolving and covers a wide range of disciplines. Recently, various nanoscale materials, devices, and systems with remarkable properties have been developed, with numerous unique applications in chemical and biological sensors, nanophotonics, nano-biotechnology, and in-vivo analysis of cellular processes at the nanoscale. On a scientifically related note, the potential and risk for inadvertent or deliberate contamination of the environment, food and agricultural products has recently increased due to the global threats of terrorism. As a result, decentralized sensing has emerged as an important issue for several agencies. To detect the contaminants, the current trend is to make laboratory facilities more mobile and conduct clinical trials employing direct reading, portable, labon-chip systems. A nanotechnology-based sensor platform provides platform for direct electrical detection of biological and chemical agents in a label-free, highly multiplexed format over a broad dynamic range during clinical testing. This platform uses functionalized nanotubes, nanoparticles, and nanowires to detect molecular binding with high sensitivity and selectivity. The platform is capable of detecting a broad range of molecules, viz., DNA, RNA, proteins, ions, small molecules, cells, and even pH values. Detection is possible in both liquid and gas phase and is highly multiplexable, enabling the parallel detection of multiple agents. Recent progress in nanostructured materials and its potential applications in chemical and biological sensors are likely to have a significant impact on efficient data collection, processing, and recognition with minimum false positive xi
xii
PREFACE
counts. Furthermore, nucleic acid layers combined with nanomaterials-based electrochemical or optical transducers produce a new kind of affinity biosensors such as the “DNA Biosensor” or “Genosensor” for small molecular weight molecules. Genosensors are attractive devices for converting the hybridization event into an analytical signal for obtaining sequence-specific information in connection with clinical, environmental, or forensic investigations. The continued development through combined efforts in microelectronics, surface/interface chemistry, molecular biology, and analytical chemistry is expected to lead to the establishment of Genosensor technology as a major component of analytical biochemistry. The design and fabrication of DNA-modified surfaces and materials which are reproducible, stable, and selective to complementary DNA sequences are crucial in the development of emerging analytical tools such as DNA chips or simple diagnostic devices for detecting DNA sequences. These devices have been used extensively not only for the rapid, cost-effective, simple diagnosis of inherited and infectious diseases, but also for the early detection of infectious agents in various environments. In general, the scientific community is confident that nanoscience and nanotechnology will revolutionize research on, and applications in the areas of biology, medicine, and human health. The research will also provide unprecedented means to forewarn and/or protect against the potential and risk for inadvertent or deliberate contamination of the environment and food and agricultural products. Many technologists develop new tools, yet they often have limited understanding of the restrictions that biology places on the proper design of nanotools and nanosystems. Hence, we sought to adopt with this ASI an interdisciplinary approach, bringing together recognized experts in various fields while retaining a level of treatment accessible to those active in specific individual areas of research and development. The NATO ASI titled “Functionalized Nanoscale Materials, Devices, and Systems for Chem.-Bio Sensors, Photonics, and Energy Generation and Storage” was very well planned, organized, and received by the participants. Both directors spent an enormous amount of time in carefully planning the logistics of the event from both international and local organization perspectives. To maximize global participation, the organizing committee focused on a promotional strategy that garnered a tremendous response not only from NATO and partner countries but also from the Asian-Pacific rim. Such an unexpected and substantial response presented a welcome but arduous task for the organizing committee charged with selecting the most deserving participants. Realizing that a productive ASI would result from a large number of participants from multiple scientific disciplines, the organizers accepted a wide range of promising candidates who could be funded through NATO.
PREFACE
xiii
Lectures covering the basic principles and state-of-the-art applications of nanostructured and advanced materials for sensor, optoelectronic, and photovoltaic devices were conducted by 13 experts recognized for advances in nanotechnology. Focused seminar sessions, poster sessions, and interactive feedback sessions stimulated extended interactions between participants and subject matter experts. As a venue for collaborative learning, the interactive lectures and sessions drew enthusiastic response and sharing of information and ideas from all participants. The ASI was held at the Hotel Sinaia, in the eponymous resort town of Romania, which offered a tranquil, congenial setting for the participants. The facility supported formal and informal settings for structured and spontaneous learning and sharing of ideas. The meeting lasted 10 days, with a day free to visit nearby towns, shop, or simply relax. The downtown shops and pubs provided a much-needed break after lectures and in-depth discussions at the ASI. The participants also shared a sightseeing excursion to Bran, Brasov, and Poiana Brasov and a traditional Romanian dinner complete with folk dancing. The unique balance of technical and social interactions materialized in alliances between participants, which have been evidenced by continued correspondence in the months following the ASI. The co-directors interpret the ongoing interaction and positive feedback from participants as an affirmation of a successful ASI. Such a constructive ASI is the outcome of efforts by participants, lecturers, presenters, and co-directors in addition to a host of caring individuals who supported their work. Much appreciation is extended to Mr. Marian Swartz, the Manager of the Hotel Sinaia. We would like to acknowledge editorial assistance from Adina Morozan, Rodica Cristescu, and Silvia Bakalova. Our local organizers, Adina Morozan, Adrian Ghita, Rodica Cristescu, Carmen Ristoscu, Felix Sima, and Ioana Vasiliu handled our everyday logistics with the utmost consideration and efficiency. Our gratitude goes to Dr. F. Pedrazzini, the director of the NATO Scientific Affairs Division for his encouragement, expertise, and financial support. Ms. Annelies Kersbergen with the NATO Publishing Unit of the Springer Academic Publishers has provided us with much appreciated expertise in publishing our proceedings. Similarly, successful proceedings are a result of meticulous preparation of manuscripts by participants to whom thanks are extended. Several funding agencies, such as NSF and TUBITAK, provided travel support for some of the participants and are acknowledged immensely for their generous help. The organizers also acknowledge support from the Romanian Ministry of Education and Research, which provided the conference materials. Thanks are due to the Vaseashta Foundation for poster awards. The co-directors hopefully anticipate that this ASI provides continued success for all participants as they extend collaboration in the pursuit of nanotechnology advancement.
xiv
PREFACE
We would like to close our preface with the following two quotes. “Over the next ten years, the fields of chemistry, physics, material sciences, biology, and computational sciences will converge in a way that will define nanotechnology and impact almost every industry, including computers, semiconductors, pharmaceuticals, defense, health care, communications, transportation, energy, environmental sciences, entertainment, chemicals, and manufacturing. Previously distinct disciplines will also combine: medicine and engineering, law and science, art and physics, etc. This merging will result in developments that are not simply evolutionary; they will be revolutionary.” ............Jack Uldrich and Deb Newberry “The revolutionary promise of molecular nanotechnology (MNT) has become a part of society’s expectations for the future. This technology will provide nanomedicine breakthroughs that could cure cancer and extend lifespace, bring abundance without environmental harm and provide clean sources of energy. These ideas are part of the vision that launched the field of nanotechnology.” ............K. Eric Drexler
Directors Ashok Vaseashta (Washington, DC) Ion N. Mihailescu (Bucharest) April 2008 Organizing Committee Arzum Erdem Ion N. Mihailescu Joop Schoonman Ioan Stamatin Sigitas Tamelevicius Ashok Vaseashta
NANOSCALE MATERIALS, DEVICES, AND SYSTEMS FOR CHEM.-BIO SENSORS, PHOTONICS, AND ENERGY GENERATION AND STORAGE A. VASEASHTA* On detail in Washington DC, USA from Nanomaterials Processing & Characterization Labs, Graduate Program in Physical Sciences, Marshall University, Huntington WV, USA
Abstract – A comprehensive overview of ongoing research efforts and future scientific directions in nanotechnology to develop materials, devices, and systems for potential use in environmental pollution monitoring and mitigation; energy generation and storage; and chemical-biological-radiological-nuclear sensing is presented. Applications of nanomaterials in development of biodegradable, high performance yet light weight and eco-friendly materials are presented to minimize power consumption, green-house gas emissions, and land-fill volume. Societal implications and concerns associated with nanotechnology are addressed by studying fate and transport and development of guidelines for a risk-assessment model. A roadmap of the future of nanomaterials, in-terms of complexity, nexus of disciplines, and emerging green nanotechnologies is presented.
Keywords: Chem.-bio sensors, pollution, satellite, water, energy, storage.
1. Introduction Three most imminent challenges of the 21st century include abundant clean energy, a pollution free environment, and international security and safety. It is imperative that the interdisciplinary scientific community exploit knowledgedriven transformations across scientific fields to develop materials, devices, and systems to address the aforementioned challenges. The scientific study
______ *
To whom correspondence should be addressed: Professor Dr. Ir. A. Vaseashta, email:
[email protected] A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
3
4
A. VASEASHTA
of nanoscale materials and systems is a promising field. Fundamental understanding and technological advances arise from the potential of nanoscale materials to exhibit unique properties that are attributable to their small size such as surface structure, physical characteristics, and chemical composition.1 These properties drive research that serves as a catalyst for new scientific and technological innovations. Furthermore, the geometrical dimensions span those comparable to the smallest engineered entity, the largest molecules of the living systems, and fundamental physical quantities render their study quite captivating. Materials approaching nanoscale dimensions exhibit atypical characteristics with numerous unique and hitherto unexploited applications. Advances in material synthesis, device fabrication and characterization techniques have provided the means to study, understand, control, or even manipulate the transitional characteristics between isolated atoms and molecules, and bulk materials. The unique characteristics and functionalities of nanomaterials have already been utilized in cosmetic, apparel, and sports industries, while proof-of concept electronic and optical devices have been demonstrated and are largely in the developmental stages. Recently, various nanoscale materials with new architectures and improved functionality have been developed with applications in chemical and biological sensors,2 environmental pollution sensing,3 monitoring,3,4 mitigation and remediation,3 next-generation energy generation and storage devices, nanobiotechnology, nanophotonics, in-vivo analysis of cellular processes and futuristic platforms in health and clinical medicine.5 The potential and risk for inadvertent or deliberate contamination of the environment, food and agricultural products, due to global threats of terrorism make decentralized chemical and biological sensing an important research area for academic institutions and governmental agencies. A nanotechnology based sensor platform enables direct electrical detection of biological and chemical agents in a label-free, highly multiplexed format over a broad dynamic range. Nucleic acid layers combined with nanomaterials-based electrochemical or optical transducers produce affinity biosensors such as the “DNA Biosensor” or “Genosensor” that are attractive devices for converting the hybridization event into an analytical signal for obtaining sequence-specific information in connection with clinical, environmental, or forensic investigations.6,7 A perpetual increase in population and consumption of fossil fuels to meet the current energy demand has led to increased pollution worldwide. Pollution in large cities has reached an alarming level and is widely perceived to be a leading contributor to chronic and deadly health disorders and diseases affecting millions of people each year. In a recent study, the World Health Organization (WHO) reported that over 3 million people suffer from the effects of air-borne pollution. Furthermore, reports from the World Energy Congress
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
5
(WEC) suggest that continued fuel consumption at its current rate will result in pollution creating irreversible environmental damage by 2025. Clinical studies show that inhaling particulate matter (PM) is associated with increased mortality rates that are further magnified for people suffering from diabetes, chronic pulmonary and inflammatory diseases. Of general pollutants that contaminate the urban environment, fine suspended PM, Nitrous Oxide (NOx), sulphur dioxide (SO2), volatile organic compounds (VOCs), and ozone (O3) pose the most widespread and acute risks; thus driving studies and measures, such as cap-and-trade, carbon credit, pollution credit etc., to limit emission of greenhouse gases (GHG). Figure 1 illustrates a pollutions footprint of various sources of energy and a proposed “future distributed-source energy solution”. Accordingly, preliminary results of joint investigations to monitor and mitigate envienvironmental pollution are presented. In addition to air pollution, results from joint investigations into the efficacy of nanostructured materials in the detection and remediation of water pollution are presented.
Figure 1. Sources of energy, pollution level, and distributed model of energy.8
Resistance to new technology and the fact that advances in nanotechnology have progressed faster than development of standards, “voluntary code of conduct” by industry, and implementation of regulations by policymakers, has led to the studies of societal implications and concerns associated with the production and use of nanomaterials. Studies of fate and transport of nanomaterials in air, water, and soil, plume modeling and in-silico risk-assessment models using expert elicitation and statistics based decision analysis are under active investigation and are briefly presented. Roadmap of time vs. complexity,
6
A. VASEASHTA
nexus of disciplines, and emerging green nanotechnologies is conceptualized to attain a technological singularity in future. 2. Size Effects in Reduced Dimensions Reduced dimensional systems, in which one or more dimensions are reduced such that material begins to display novel quantifiable quantities hold tremendous potential. For solids, typically reduced dimensions amount to reduction of the coordination number; hence the electrons have less opportunity to hop from site to site; thus reducing kinetic energy of electrons or their bandwidth. A higher Coulomb interaction/bandwidth ratio at a site enhances electron correlation and Mott-transition i.e. the tendencies towards the appearance of magnetism. Furthermore, the symmetries of the system are lowered and the appearance of new boundary conditions lead to surface states and interface states. A change of the quantization conditions alters the eigen value spectrum and transport properties of the solid. A high surface area/volume ratio in nanoscale materials alters mechanical and other physical properties. One critical impact is that surface stresses existing in nanomaterials have a different bonding configuration as compared to bulk atoms. As an example, surface elasticity is an effect that occurs due to the lack of bonding neighbors for surface atoms. The effects of the difference between surface and bulk elastic properties become magnified as the surface area/volume ratio increases with decreasing structural dimension. Studies to calculate surface elastic constants using MD simulations, curvature effect using the Cauchy-Born rule, and electronic effects via effective nuclei-nuclei interaction using DFT calculations provide better understanding of surface and interface effects in reduced dimensions. Despite many investigations on surface elastic effects, many questions remain unanswered. Recently, sensing by nanoscale materials has been utilized for chemicalbiological investigations; primarily because of increased surface area, reactivity, and ability of selective functionalization. For instance, nanoscale resonators9 have allowed various research groups to detect mass of molecules by detecting change in resonance frequency with mass.10 Surface effect may play a role in resonant frequency shift if the thickness of the biomolecular layer becomes comparable to the resonator’s thickness; although, the surface stress effect may affect the resonant frequency shift. We have employed the surface plasmon resonance (SPR) on the optical response2,11 as a function of biological molecule interactions resulting from adsorbate-substrate bonding. As the size of nanomaterials approach fundamental physical quantities and biological molecules (shown in Figure 2(l)), many questions remain unanswered to put the broad topic of emergence of size-effect in different materials into perspective. As an
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
7
Figure 2. (l): Dimensional compatibility, (r): roadmap of nanomaterials.
example, when does the surface effect become important? Is there a critical dimension for materials? Do some processes take place only in nanoscale dimensions and do other processes occur even in micro-scale dimensions? What do we know about the interactions of nanomaterials with biological molecules vis-à-vis “spes altera-vitae”? 2.1. NANOMATERIALS-ENVIRONMENT INTERFACE
Size and surface collectively control characteristics of nanoscale materials due to the existence of large boundaries adjoining its surrounding medium and interplay of physicochemical interactions. The surface free-energy is sizedependent and hence increases almost inversely with the decreasing feature sizes of the material. Collective response of a nanomaterial-medium system that is attributable to reduced dimensions, viz. size (area and distribution), surface structure (groups, functionalization), physical (electronic, optical, photoactivation, luminescence), and chemical (crystallinity, purity, solubility) is vital to developing a scientific model that predicts its response as a sensor, bioadverse response pathways for toxicity, adsorption pathways for materials for storage, and interaction with light for optical response. It is an exceedingly complex task due to a large matrix of parameters consisting of nanomaterials (metals, metal oxides, semiconductors, macromolecules and self-assembled), the surrounding environment, and influencing mechanisms of interactions.
8
A. VASEASHTA
Further complexities arise in nanomaterials production due to variations in, e.g. fabrication with precisely-controlled surface properties, uniform dispersion of nanoparticles in aqueous or organic mediums, linkage of nanoparticles to a polymeric material, and reproducible surface modification of nanoparticles. Hence, as the production and diversity of nanoparticles increases, it will become increasingly important to understand how engineered nanoparticles and biological systems interact in terms of bio-physico-chemical properties of engineered nanoparticles (Figure 2(r)). An extensive investigation is underway in the context of specific functions ranging from bio-nanomaterials interface to toxic potential of industrial pollution. We conducted study of functionalized SAM as biosensors that use biological molecules – usually enzymes, antibodies, or nucleic acids, to recognize simple molecules of interest via hydrogen bonding, charge-charge interactions, and other biochemical interactions to provide molecular information.7 Recent work on affinity biosensors to deliver real-time information about the antibodies to antigens, cell receptors to their glands, and DNA and RNA to nucleic acid with a complimentary sequence,12 provided a multitude of applications. As an example, they can be used to measure blood glucose levels, to detect pollutants and pesticides in the environment, to monitor food-borne pathogens in the food supply, to work as chemical and biological warfare agents, and detect the presence of micro-organisms in foods. A response of a nanomaterials based gas sensor is based on reactions replacement of atoms at the sensing surface of these materials which relies on a change of the resistance of the oxide. Depending on the free electron density in the space charge layer, the depletion region is increased. Since electric properties are influenced by the depletion layer, variation in electrical conductivity indicates sensor response. Similarly, moisture can influence the resistance or conductivity of oxide materials via two pathways: first, the adsorption of monolayer/s of water molecules at the surface; and second, the process of formation of a parallel resistance path by capillary condensation of water via adsorption of the water molecules as protons and hydroxyl groups within pores. The sensitivity and response of nanomaterials of metal-oxide sensors is highly dependent on the roughness of the substrate, which is caused by the increasing surface area and porosity of the film surface modifications in the film surface morphology. The sensitivity (and selectivity) of a can be improved by parameters such as decreasing the crystallite size, the valence control, and using noble metal catalysts. The characteristics which provide beneficial aspects are also believed to be responsible for toxicity of nanomaterials. Consequently, nanoparticle toxicity is studied in context of its ability to induce tissue damage through the generation of oxygen radicals, electron-hole pairs, and oxidant stress by abiotic and cellular
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
9
responses resulting in pro-inflammatory, mitochondrial injury and pro-apoptotic cellular effects in the lung, cardiovascular system and brain.13 It is further believed that nanoparticles absorb cellular proteins which could induce protein unfolding, fibrillation, and thiol cross-linking; leading to neuro-toxicity and reduced enzymatic activity. Nanoparticles which are cationic are also believed to induce toxicity via acidifying endosomes that lead to cellular toxicity and apoptosis in epithelial lung through endosomal rupture through proton sponge mechanism (PSM), mitochondrial targeting, and cytosolic deposition. Nanomaterials composed of redox-active elements are particularly reactive and can possibly provoke potentially damaging chemical transformations. Furthermore, even chemically benign nanoparticles may become activate by light absorption. Hence, fundamental understanding of a nanomaterial-surrounding medium is vital to sustaining technological advances of nanoscale materials as catalyst for new scientific and technological avenues. 2.2. NANOPHOTONICS AND SURFACE PLASMON RESONANCE
Two theoretical models, viz.: the classical electrodynamics for the propagation of light and the solid (or liquid) state theory for the interaction of light with the particle expressed by the complex, frequency dependent dielectric function, DF, of the particle material, which can mathematically be described as,14 according to Mie’s theory: İ(Ȧ) = İ1 + i İ2 = İDrude + Ȥ interband = 1 – (n e2/ İ0 me )/(Ȧ2 + i Ȗ Ȧ) + Ȥ interband
(1)
with n, e, İ0, me, Ȗ are respectively the electron density, elementary charge, field constant, effective mass and relaxation frequency of the conduction electrons of the Drude-Lorentz-Sommerfeld theory, Ȗ = 1/IJ with IJ the Drude relaxation time and (ne2/İ0me )1/2 = Ȧp the “Drude plasma frequency”. The interband transitions exist in most metals thus causing nanomaterials to demonstrate hybrid surface plasmon polariton (SPP) excitations. They consist of collective electronic excitations in the conduction band and (complex) polarization term for deep bands. Most nanomaterials exhibit optical absorption and scattering spectra with complex multi-peak features, characterized by quantities viz., peak height, spectral peak position ȦMax and band width ī. The dielectric function (DF) includes electronic size and surface/interface effects present in a nanoparticle. Although a detailed description is described elsewhere, it suffices to state here that Mie’s theory is used analogously to Fresnels formulae which was derived for samples with planar geometry. Extension of the theory yield effective
10
A. VASEASHTA
permeability and refractive index to produce negative refractive index materials (NRIM) by plasmonic materials to image sub-O features for producing near field super lens.15 The inverted Mie theory yields the realistic DF including all electronic and optical size and surface/interface effects and resulting evanescent field strongly interacts with adsorbed molecules, thus influencing the resulting spectra. The application of inverse Mie theory is essential for structural, electronic and optical effects for selective detection substances in aqueous/biosurroundings using noble metals as plasmonic nanostructures. Integrating plasmonic elements with existing silicon based technology will likely result in plasmonic nanophotonic arrays with multifunctional capability. 3. Chem.-Bio Sensors and Biomedical Platforms International security threats such as escalation in terrorist activities and asymmetric warfare by adversaries drive needs for novel materials and innovative platforms to detect, interdict, and counter threats due to chem.-biologicalnuclear-radiological-improvised explosive (CBRNE) agents, synthetic DNA, contact-poison, and improvised explosive devices (IEDs) in a time-efficient and reliable manner at the site of event. Additional potential risks include inadvertent or deliberate contamination of the environment, food and agricultural products, or even naturally occurring threats such as avian bird flu. In a possible weapons of mass destruction and terrorism (WMDT) scenario, rapid identification of CBRNE events will allow first responders and emergency personnel to implement critical decisions concerning barricading, evacuating, or efficient decontamination, saving hundreds of lives and preventing responders from becoming victims themselves. Conventional detection methods require time-consuming steps to arrive at meaningful data. As an example, methods such as enzyme linked immuno-sorbent assay (ELISA) and polymerase chain reactions (PCR) have been employed for pathogen detection. ELISA and PCR methods require enrichment, isolation, morphological examination, biochemical, and serological testing to positively identify pathogens. In clinical medicine, decentralized laboratory facilities allow mobile facilities for clinical analysis through direct-reading, portable, lab-on-chip (LOC) systems with wireless communications capabilities to centralized servers, command and control units. Recent progress in nanostructured materials based sensor platforms will significantly impact data collection, processing, and recognition to enable the direct detection of biological and chemical agents in a label-free, parallel, multiplexed, broadly dynamic range. This platform utilizes functionalized nanotubes, nanowires, and nanoparticles to detect a broad range of molecules including DNA, RNA, proteins, ions, cells, pH values, and molecular binding with high sensitivity and selectivity.
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
11
Biosensors are intrinsically simple and inexpensive systems that use molecules – usually enzymes, antibodies, or nucleic acids – to recognize simple molecules of interest via hydrogen bonding, charge-charge interactions and other biochemical interactions to provide pertinent molecular information (Figure 3(l)). Progress in nanomaterials and advances in fabrication processes provide opportunity to modify, embed in host matrix, or even customize the nanoparticles for use as highly sensitive and selective sensing materials. Nanomaterials based sensors are used in several configurations such as SPR, electrochemical, optical, electrical transduction, and as shock-wave generators for applications ranging from homeland security to studying the environment pollution. Nanostructures will improve our capability to detect CBRNE events with sensitivity and selectivity by several orders of magnitude, protect through filtration, adsorption, mitigation or neutralization of agents, and provide site-specific in-vivo prophylaxis. Joint investigative efforts are focused towards chemical and biological agents, and radiation detection due to a radiological-dispersal device (RDD) or “dirty bomb”.
Figure 3. (l): Electrochemical sensor basics, (r): SPR based sensor.
3.1. NANOMATERIALS BASED CHEMICAL-BIOLOGICAL SENSORS
CNTs are conducting, can act as electrodes, generate electro-chemiluminescence (ECL) in aqueous solutions, and can be derivatized with a functional group that allows immobilization of biomolecules. CNTs have high surface/volume ratios for adsorption, and have surface/weight ratios ~300 m2/g. The uniform chemical functionalization of CNTs is key to the formation of biosensors. Oxidation of nanotubes with HNO3-H2SO4 leads to high concentrations of carboxylic, carbonyl, and hydroxyl groups on the surface, and removal of the tip to expose the tube interior. Carboxyl groups can readily be derivatized by a variety of reactions allowing linking of biomolecules such as proteins, enzymes, DNA, or even metal nanoparticles. The covalent modification of nanotubes facilitates the creation of well-defined probes, which are sensitive to specific intermolecular interactions of many chemical and biological systems. Integration of the
12
A. VASEASHTA
transducer transducer and probe enables quick, accurate, and reversible measurement of target analytes without the use of reagents. Using sequence-specific attachment, NT-based electronic devices with specific molecular-recognition features of DNA have been reported.16 Covalent modification of single wall CNTs (SWNTs) offers mapping of functional groups at a molecular resolution. Furthermore, chemical processes to link catalysts, such as transition-metal complexes, to the ends of CNTs are useful in creating or modifying the structures at a molecular scale, creating interconnections for electronic devices, and even developing new classes of materials. Covalent functionalization of the sidewalls of SWNTs provides stability and best accessibility; but at the expense of damaging the sidewalls; thereby diminishing the mechanical and electronic properties. However, noncovalent routes to CNTs functionalization offer ease of synthesis and minimum disruption of the tubular structure. During interaction with the polymer coatings, the electrical properties of the nanotubes are altered, enabling detection of the molecules leading to a very sensitive sensing mechanism. In addition to nanotubes, novel materials such as porous silicon17 and porous carbon,18 with porosities of comparable dimensions to those of the biomolecules have been used for biosensor applications. The mesoporous carbon matrix is used for stable immobilization of the biological molecule, and C60 serves as an electron mediator. Both C60 and NTs are good electron mediators when used with a mesoporous carbon matrix or modified metal electrodes. CNT-based transducers, however, show a significant advantage over porous silicon due to the well defined, defect free structures, and also because the NTs promote homogenous electron transfer reactions. Efforts to sort batches of CNTs by length using high-speed centrifuges and functionalization to develop sensors for the food and agriculture industry, genetic analysis, proteomics, drug screening, clinical diagnostics and bio-warfare agent detection are underway. Furthermore, investigations using SAM based SPR and Atomic Force Microscopy (AFM) techniques are in progress to detect several pathogens. The SPR detection technique is rapid, real-time, and requires no labeling, and involves immobilizing antibodies by a coupling matrix on the surface of a thin film of precious metal, such as nanoparticles of gold deposited on the reflecting surface of an optically transparent wave-guide. The precise angle at which SPR occurs depends on several factors (Figure 3(r)). A main response is the refractive index of the metal film, to which target molecules are immobilized using specific capture molecules or receptors along the surface, that cause a change in SPR angle. This can be monitored in real-time by detecting changes in the intensity of the reflected light, producing a sensorgram. The rates of change of the SPR signal can be analyzed to yield apparent rate constants for the association and dissociation phases of the reaction. When the antigens
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
13
interact with antibodies, the refractive index of the medium surrounding the sensor changes producing a shift in the angle of resonance proportional to the change in the concentration of antigens bound to the surface. Various other sensor platforms, viz.: (a) chemo-mechanical micro-cantilever array to provide a quantitative and label-free platform for high-throughput multiplexed biomolecular analysis for detection of various biomolecules based on binding, (b) microarrays chips fabricated using technique matrix assisted pulsed laser evaporation (MAPLE) and laser induced forward transfer (LIFT) for deposition of biopolymers and a variety of biomolecules to detect dangerous gases, aerosols and micro-organisms, and (c) aligned-CNT probes for biomolecular recognition based on charge transport at the CNT transducer with the accuracy down to molecular level for quantitative and selective detection of a range of metabolites including cholesterol, ascorbic acid and uric acid, in buffer solution as well as in human plasma and blood are under investigation. Several other concepts being considered are resonance based9 shock-wave generators that only could detonate military explosives but will also detect biological and chemical weapons. The shock wave can be used for targeted delivery of drugs, chemotherapy, and cure cells without affecting the whole body. Likewise, nanothermite – a composite of fuel and an oxidizer, which in turn generate combustion waves that can hit velocities ranging from mach 4–7, can safely be used for killing cancer cells. Using electrospinning, we have investigated nanofibers prepared from high performance polymer composites embedded with metal oxides, glasses coated with rare earth metals, and biocompatible compounds for sensing, and even controlling electrical, optical, and chemical and biological response (Figure 4 – both panes). Nucleic acids offer analytical chemists a powerful tool for recognition and monitoring of many important compounds. Recent advances in molecular
Figure 4. Use of electrospinning for chem.-bio sensors, health and medicine.11
14
A. VASEASHTA
biology are used to study the effects of proteins and drugs on gene expression, viz. gel mobility shift, filter binding, DNA foot-printing and fluorescence-based assays. Most of these methods; however, are indirect and require various labeling strategies. Electrochemical DNA biosensors play an important role for clinical, pharmaceutical, environmental and forensic applications, because they provide rapid, simple and low-cost point-of-care detection of specific nucleic acid sequences. In recent years, there has been a growing interest towards design of electrochemical DNA biosensors that exploit interactions between surface-confined DNA and target drugs/biological molecules for rapid screening.16,19 Binding of small molecules to DNA primarily occurs in three modes: electrostatic interactions with the negative-charged nucleic sugar-phosphate structure, binding interactions with two grooves of DNA double helix, and intercalation between the stacked base pairs of native DNA. Most electrochemical sensors use different chemistries; and employ interactions between the target, the recognition layer and an electrode. We have followed numerous approaches to electrochemical detection including direct electrochemistry of DNA and devices based on DNA-mediated charge transport chemistry. In direct electrochemical DNA sensors, the analysis is based on a guanine signal where a basepairing interaction recruits a target molecule to the sensor, allowing monitoring of drug/biological molecule-DNA interactions, which are related to the differences in the electrochemical signals of DNA binding molecules for DNA barcoding. It is vital to develop sensing strategies to maintain critical dynamics of target capture to generate a sufficient recognition signal. Standard electrochemical techniques, such as differential pulse voltammetry (DPV), potentiometric stripping analysis (PSA), square-wave voltammetry (SWV), etc. are used as genosensors. Since genosensors are compatible with existing micro and nanofabrication technologies, they enable design of low-cost, devices that offer potential for detection and diagnosis of inherited diseases and clinical potential for detecting pathogenic bacteria, tumors, genetic disease, and forensics via credit card-sized sensor arrays. In the context of security, prevention against agroterrorism is by detecting bovine spongiform encephalopathy (BSE). A possible link between BSE and a disease of humans, a variant form of Creutzfeldt-Jakob Disease (vCJD) using nano-technology based platforms is important to provide clues about the disease. 3.2. USE OF NANOMATERIALS IN HEALTH AND MEDICINE
Sensing strategies can be modified towards applications of nanomaterials in targeted drug delivery, diagnostic, and therapeutic actions,5,20 thus creating pathways for use in health and medicine. By incorporating a drug into bio-degradable polymers, a simple and convenient drug delivery system allows time-controlled
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
15
drug regimen release. We have studied nanofibers for promising medical applications including treatment of primary pulmonary hypertension (PPH), and pulmonary arterial hypertension (PAH) by time-controlled release templates impregnated with anticoagulants and calcium channel blockers, a bioscaffold that mimics extracellular matrix (ECM) topology, polyesters com-bined with phosphatidyl choline for biomimetic applications, intravascular stents from a blend of polyactide and trimethylene carbonate, cellulose based scaffold for cartilage tissue engineering, and esophageal tissue engineering. We studied dystrophin gene immobilized on nanostructured templates as a drug carrier that initiates regeneration and boosted satellite cells mediated repair mechanisms for affected muscles for treating Duchene muscular dystrophy (DMD).21 Likewise, a nanoparticles–cinobufagin-bovine serum albumin based drug delivery mechanism holds promise in treating the hepatocellular carcinoma (HCC) malignancy. Nanotechnology based platforms offer novel opportunities to sense clinical biomarkers by imaging for therapeutic intervention. The use of iron oxide nanoparticles in an MRI system provides 3D image and normal and cancerous hepatocyte cells information.22 Furthermore, nanosized constructs such as dendrimers, liposoms, nanoshells, nanotubes, nanoemulsions, quantum dots (QDs), and even viruses offer use as imaging agents intended as non-invasive probes or targeted disease biomarkers. Gold nanoparticles are used to identify the pathogenic bacteria in a DNA microarray technique,23 and QDs are used to detect the human Y chromosome,24 and for locating cancer markers in cellular imaging.25 One to ten nanometer sized QDs with unique photochemical, photophysical, and physiological properties are proposed for drug delivery or contrast agents in MRI. For medical diagnostics, QDs of varying diameter are embedded into polymeric microbeads to achieve biological assays by multi-color optical coding. Tissue engineering is a relatively new field that seeks to regenerate human tissues through the use of some combination of cells, bioactive molecules such as drugs and mechano growth factors (MGF), and a biomaterial support system or scaffold.26 Hence, biomimetically driven studies are exploring how the topography of a surface can be used to control cell behavior.27 By changing the biomaterial surfaces at nano level, we can observe different types of cell behavior, change in cell adhesion properties – especially for fat-free mass, cell orientation and motility, cytoskeletal condensation, and elastomers for vascular tissue engineering28 to monitor advanced chronic pulmonary diseases (COPD). Recent applications of nanotechnology in dental care (nanodentistry) will permit maintaining near perfect oral health via nanomaterials, biotechnology, and nanorobotics.29 In fact, nanofillers are common nano-dentistry. Additional procedures include renaturalization, permanent hyper-sensitivity cures, and orthodontic realignments. Nanomaterials provide higher mechanical strength, enhanced bioactivity, and restorability for debilitating bone fractures. Nanomaterials-based
16
A. VASEASHTA
orthopedic applications include nanopar-ticles, nanofibres, nanoscaffolds, nanotubes and nano-composites as orthopedic implants. Developing intelligent biomaterials that closely mimic the molecular composition, mechanical responsiveness, and nanoscale organizations of the natural extracellular matrices (ECM) is vital. Recent reports indicate that adding gold nanoparticles to a modified version of an HIV combating drug, TAK-779 creates a compound that prevents the virus from gaining a cellular foothold. These novel and smart biomaterials, combined with defined biophysical cues and biological factors are essential for functional tissue regeneration. 4. Energy Generation and Storage Driven by population growth and the wave of industrialization in developing countries, energy consumption worldwide is increasing relentlessly. Moving forward, an alternate means to generate and store energy is imperative to ease environmental resource constraints, and to push the economy and society to sustainable development. Reports from the European Union’s (EU) recent research framework program (FP) and the United States’ National Nanotechnology Initiative (NNI) have listed several strategies that identify nanotechnology’s role in power generation and storage. Figure 1 further suggests that future power sources will be distributed allowing reduced dependence on fossil fuels leading to pollution reductions and GHG; that are consistent with Jeremy Rifkin’s model of peer-to-peer, shared, decentralized, and distributed generation using hydrogen energy. Significant nanotechnology contributions will influence research in fuel cells, hydrogen generation and storage, improvement in photovoltaic conversion efficiency, super-strong light-weight materials to reduce power consumption, electrical and optical devices requiring less power yet providing higher lumens, and many others. 4.1. HYDROGEN GENRATION AND STORAGE
Hydrogen produces more energy/volume than any other known sources of fuels. The future of a “hydrogen economy” depends on developing clean, safe and efficient methods for producing and storing hydrogen. Hydrogen can be produced by several methods, such as thermolysis, electrolysis, photocatalysis, hybrid processes, and gasification and CO2 sequestration. For sustainable energy, fuel cells (FC) are an important enabling technology to produce clean and renewable hydrogen energy. There are several different types of fuel cells, e.g. alkali, molten carbonate (MCFC), phosphoric acid (PAFC), proton exchange membrane (PEMFC), and solid oxide fuel (SOFC). We focus on PEM based fuel cells. Technically, water splitting can be achieved with coupled solar cell,
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
17
i.e. photo-electrolysis systems. A photo-electrochemical (PEC) cell comprises a semi-conducting photo-anode, an aqueous electrolyte, and an inert counterelectrode. The main advantage of the PEC cell is that the evolving hydrogen and oxygen gasses can be collected in separate volumes. The PEC cells are elegant, however are limited in their applicability due to several practical considerations such as hydrogen and oxygen forming in the same volume; limited lifetime of photo catalysts after which the efficiency drops below acceptable levels; and slow charge transfer across the semiconductor/electrolyte interface. Most research efforts on solar water splitting are focussed on systems with photo-catalyst powder suspended in an aqueous solution. In both photoelectrodes and powder-based photo-catalysts, surface reaction kinetics plays a key role. The PECCS cells combine the advantages of a two-electrode system with simultaneous storage of hydrogen in a metal hydride (MH). Once the MH electrode is fully loaded with hydrogen, it can be withdrawn from the cell and placed in a fuel cell system. Upon heating the electrode, the hydrogen is released and can be used to generate power. A Nernst-type sensor measures the concentration of hydrogen formed in the MH. Using a second MH reference electrode with a fixed hydrogen activity, open-circuit voltage across both MHs, EOC is given by the Nernst equation:30
EOC
RTt H F
§ a ( H ) sample · ln ¨ ref ¸ © a( H ) ¹
(2)
R is the gas constant (8.314 JK–1), T is the temperature (in Kelvin), F is Faraday’s constant (96,485 Cmol–1), tH+ is the ionic transport coefficient for protons, and a(H+)sample and a(H+)ref may be substituted with the hydrogen activities in the MH counter- and reference-electrodes, respectively. The current state-of-the-art membranes are fabricated from perfluorinated sulfonic acid (PFSA) polymers, such as Nafion® and Hydrogen uranyl phosphate tetrahydrate, HUO2PO44H2O (HUP) which can be easily synthesized and show high proton conductivity at room temperature.31 These materials have useful, high proton conductivities when fully hydrated limiting their usefulness to temperatures <100°C. Understanding of parameters such as proton transport, water diffusion, and electro-osmotic drag; ion transport typically described by Grotthus mechanism and Vehicle mechanism; and water management issues are crucial to a successful fuel cell design; as water is being constantly transported from the anode to the cathode leading to flooding. We are studying a new series of polymers based on pyridine cycles originated from poly-acrylonitrile (PAN) in a thermo-oxidative process. The thermo-oxidation in air and at moderate temperature (200–400°C) converts PAN into ladder polymers made of fused pyridine cycles. The membranes are obtained by PAN-SiO2 nanoparticles
18
A. VASEASHTA
solution casting in thin films with simultaneous thermo-oxidation process.31 These PEMs have ability to operate at moderately elevated temperatures, with little to no external humidification, and to be manufactured inexpensively. Suitable photo-anode materials for efficient solar hydrogen generation should have high visible light absorption; high photochemical stability; suitable bandgap (preferably graded); efficient charge transport in the semiconductor; and the conduction and valence band edges should “straddle” the reduction and oxidation potentials of water. Virtually all photo-electrodes display a trade-off between photochemical stability and visible-light absorption. TiO2 has been the most extensively investigated material for photo-electrochemical and photo-catalytic applications. An alternative material that has an intrinsically small bandgap is D-Fe2O3 (hematite). We studied feasibility of solar photo-electrolysis with a system in which an n-TiO2 semiconductor electrode was connected though an electrical load to a platinum back counter electrode placed under near-UV light. When the surface of TiO2 electrode was irradiated with UV light of wavelengths <415 nm, photocurrent flows from the platinum counter electrode to the TiO2 electrode indicating that the oxidation reaction (oxygen evolution) occurs at the TiO2 electrode and the reduction reaction (hydrogen evolution) at the Pt electrode. Hence water splits into O2 and H2 with UV light, without any external voltage, and is described as follows: Excitation of TiO2 by light At the TiO2 electrode At the Pt electrode The overall reaction is
=> => => =>
TiO2 + 2hȞ ĺ 2e– + h+ H2O + 2h+ ĺ ½ O2 + 2H+ 2H+ + 2e– ĺ H2 H2O + 2hȞ ĺ ½ O2 +H2
Recently, hydrogen storage in MHs has been integrated into the photoelectrolysis cell, thus leading to a new type of devices termed as photo-electro chemical conversion and storage (PECCS) cells; in which the cathode is MH allowing in-situ storage of the generated hydrogen. Our efforts are focused on preparation of new materials for PEM fuel cells. Activated carbons (AC), SWNTs and metal–organic frameworks (MOFs) are being investigated as hydrogen adsorption and storage materials.32 The hydrogen storage capacity of nanofibers ranged from |15 > Ȥ > 1 wt % at moderate pressures and temperatures.33 Experiments on SWNTs yielded adsorbed amounts <10 wt % due to narrow pore size distribution of SWNTs. US Department of Energy (US DoE) has recently set performance targets by 2010 for on-board vehicular hydrogen storage systems, e.g. system gravimetric capacity (2 kWh/kg), volumetric capacity (1.5 kWh/L), durability/operability parameters, and charging/ discharge rates. Approaches such as compressed hydrogen gas, cryogenic liquid/ gas, chemical hydrogen storage, high-surface area adsorbents, and MHs are considered at the present time for on-board storage and regeneration and
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
19
off-board regeneration and refilling in terms of parameters such as thermodynamics, kinetics, potential durability/cyclability, and hydrogen discharge kinetics. In general, the materials for hydrogen storage should possess high hydrogen content, low heats of dehydrogenation, and fast desorption kinetics at FC temperature of operation. MHs in particular received a lot of attention recently due in part to inherently safe operation, low operation pressure, low cost, and high storage capacities. Early work on intermetallic hydrides showed good sorption/desorption kinetics but extremely low storage capacities. Since then, many complex MHs have successfully been used for reversible hydrogen storage. Despite several drawbacks, complex metal borohydrides M(BH4)n prepared from direct or indirect (more practical) synthesis are potential contenders for future hydrogen storage. For conditions where hydrogen release and uptake is controlled via temperature and pressure, Ammonia Borane (AB = NH3BH3) has received significant attention in light of its stability and commercial availability. 4.2. ENERGY GENERATION AND RESOURCE CONSERVATION
Nanotechnologies have opened promising new routes for making inexpensive solar cells such as 3D-nanostructured solar cells based on flexible substrates. To develop contacts, nanowires are grown directly onto contact pads to increase carrier collection efficiency. Enhancing the efficiency of solar cells by using electro-active and conjugated poly-(3-hexylthiophene) polymers, composites doped with rare earth metals to shift the incident spectrum to optimize photovoltaic response, active interfaces to reduce the distance of the photo-produced charge carriers, and increasing anti-reflection effects are under active investigation. Also, experimental efforts include nanomaterials in beta-voltaics as a source of power for satellite sub-modules drawing energy from ionizing radiation in space to convert in electrical energy. The need for high-performance and light-weight materials that reduce product weight without compromising cost, performance, or safety is growing. New or improved materials to reduce fuel-consumption, meet consumer expectations, military needs, and regulatory mandates are needed for transport industries. In addition, lightweight materials and structures for body armor and protection, building and construction, sports and leisure goods, and packaging are in demand. Nanomaterial characteristics lead research developments of lightweight materials and structures that withstand arduous operating conditions from high temperatures, dynamic loading, thermal and mechanical stresses, enhance resistance to blast and shock, and enhance ballistic performance without structural degradation. Lightweight components in transport and aviation industries will reduce fuel consumption and environmental pollution. Investigated materials include
20
A. VASEASHTA
carbon fibers, protective coatings, honeycomb, or bonded structures. Using electrospinning, we developed nanofibers using high performance polymers impregnated with CNTs. 4.2.1. Greenhouse gases sequestration The U.S. EIA predicts a 33% increase in CO2 emission by 2030 unless appropriate sequestration and mitigation technologies are instituted. The CO2 sequestration program contributes to the President’s initiative for developing technologies for reducing GHG emissions, lowering the cost and energy penalty associated with CO2 sequestration from large point sources and understanding factors affecting CO2 storage durability, capacity, and safety for geologic formations and ecosystems. Nanoporous carbon and liquid amine absorption are used to adsorb and sequester CO2; however large scale capture requires significant improvements to overcome drawbacks arising from large amounts of energy needed for regeneration, corrosiveness, and solvent degradation. Advanced membrane reactors34 use fossil fuel conversion reactions such as the steamreforming and water-gas-shift with in-situ separation of a reaction product. The fuel chemical equilibrium is hence shifted resulting in reduced thermodynamic losses since conversion is increased significantly as compared to conventional unit operations. A nanostructured ceramic membrane reactor based on alumina allows in-situ separation of H2 from CO, CO2, and possibly H2O gas mixtures. Furthermore, our efforts include sequestering CO2 using electrospun metal oxide nanofibers and propylene oxidation to propylene-epoxide (PO) using a variety of Au/TiO2 catalysts. 5. Environmental Pollution Monitoring, Detection, and Remediation In our technologically dominated society, the environment has been stressed from increased consumption and ever-rising energy demands for better quality of life. Scientific debate supporting and contradicting evidence relating pollution to the earth’s climate change are ongoing, but its impact on health is irrefutable. Nanotechnologies may improve the quality of life by reducing pollution at its origin by use of alternate sources of energy, mitigating pollutants, and curing symptoms from pollution. We have employed principles of green chemistry for producing nanomaterials – “green nanotechnology,”8 and devices that offset fuel consumption and in turn environmental pollution. Satellites play a major role in communication, navigation, climatology, surveillance, and environmental monitoring, and support many applications. Nanotechnology and micro/nano electromechanical systems (MEMS/NEMS) hold the potential to revolutionize the field of satellite design for environmental
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
21
pollution monitoring. Using image processing of the satellite data, we can extract pollution and optical data for any given city. We have conducted experiments with nanoscale metal-oxides sensors to detect pollution. The sensors are connected to a personal digital assistant (PDA) with GPS and wireless communication capability and web-server. Using data from this proptotype device, we have superimposed data from satellite to cross-reference the pollution coordinates. We expanded the investigation to mitigate pollution in air and water from nanoscale materials. The use of such interdisciplinary and transformational technologies has initiated a widespread perception-based public debate with scope ranging from favoring further research to its moratorium in advance of scientific risk assessment. This section provides a brief discussion encompassing ethical and legal considerations, fate and transport, and in-silico riskassessment. 5.1. POLLUTION MONITORING, DETECTION AND REMEDIATION
Satellite image data has traditionally been unexploited for atmospheric pollution studies. Satellite image data consists of earth radiances influenced by the temperature, emissivity of the ground surface, and the atmospheric column above, and its surroundings as observed by its sensors in different bands. Satellite image data can aid in detection, tracking, and understanding of pollutant sources and transport by providing observations over large spatial domains with 3D Models. The pollution assessment of optical atmospheric effects can also be quantified by aerosol optical thickness (AOT). The Differential Optical Absorption Spectroscopy-Method (DOAS) algorithm can analyze the data of the Global Ozone Monitoring Experiment (GOME) on ERS-2 and the SCIAMACHY on ENVISAT to retrieve NO2 columns in the 425–450 nm regions with high accuracy. City specific data can be modeled and correlated with data from ground based stations, such as the U.S. environmental protection agency (EPA), the Texas Commission on Environmental Quality (TCEQ), France’s AIRPARIF, the Netherlands’ National Institute for Public Health and Environment (RIVM), British Atmospheric Data Centre (BADC), and the Natural Environmental Research Council (NERC) among others. There is observational evidence that these aerosols can alter cloud properties. Scattering albedo increases with pollution, thus decreasing backscatter fraction. Aerosol radiative forcing depends on hygroscopicity, which in turn depends on aerosol photo-chemistry. IR absorption spectroscopy effectively identifies trace pollutants in both ambient air and smog systems. Employing Advanced Space borne Thermal Emission and Reflection Radiometer (ASTER) data, we have investigated air pollutants using satellite images 3 comprised of high spatial and spectral resolution images from the TIR and SWIR bands; and
22
A. VASEASHTA
compared the satellite data with EPA air quality and spectroscopic study data of atmospheric pollution. A conceptual model of the interplay of data from satellite and ground based stations is shown in Figure 5(l).3 ASTER is an imaging instrument flying on the Terra satellite as part of NASA’s Earth Observing System (EOS). It is the only high-spatial resolution instrument on the satellite that has 14 bands from the visible to the thermal infrared region. In the visible green and nearinfrared (V-NIR) range between 0.52–0.86 μm, there are three bands with 15 m resolution. In the SWIR range between 1.6–2.43 μm, there are six bands with 30 m resolution. In the third range between 8.125–11.65 μm, there are five bands with a resolution of 90 m. ASTER acquires data over a 60 km swath of which center is pointable. Satellite observed data is processed employing the Multivariate techniques, which are based on groupings in a multivariate data set. Remote sensing software ER-MapperTM (v.7.x) is employed to accurately model the data through feature extraction processes for pattern recognition. Several digital imaging processes such as geometric image registration, radiometric normalization, principal component analysis, and data fusion process the image for accurate feature extraction. Multidate correlations and regression analysis methods are used for individual gases with satellite-recorded reflectance of bands. Nanomaterials satellite components as visualized in Figure 5(r) will improve the resolution, composition, and library of information.
Figure 5. (l): Model of data from satellite and ground based stations, (r): nano-satellite.
A variety of nanomaterials have been studied for their abilities to detect and mitigate pollutants in air,3 water, and soil. We have investigated various ways in which nanomaterials can successfully be used to reduce/isolate atmospheric pollution, such as use of CNT filled in a polymer composite matrix to create a static discharge to remove PM from incoming air, and chemical protective clothing, and breathing filtration masks. Employing immobilized
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
23
TiO2 films, produced by sol-gel process, we explored removal of indoor odors under weak UV illumination. At 10 ppmv weak UV light of 1 μW/cm2 was sufficient to decompose such compounds in the presence of TiO2 photocatalysts. In an experiment, 150 ȝl of an E. coli suspension, containing 30,000 × cells was placed under weak UV (1 mW/cm2) illuminated TiO2-coated glass plate showing antibacterial effect. We are focusing on mitigating water contaminants such as Arsenic (As), which is found in the natural environment, and also from anthropogenic activities. Several studies have linked long-term exposure to small concentrations of As with cancer and cardiovascular, pulmonary, immunological, neurological and endocrine effects; thus necessitating studies for developing efficient methods for removing As from drinking water. We have employed a method based sorption on iron (III) oxides, such as amorphous hydrous ferric oxide (FeOOH), poorly crystalline hydrous ferric oxide (ferrihydride) and goethite (Į-FeOOH) which have been found to be effective in removing both As (V) and As (III) from aqueous solutions. Preliminary results with synthetic akaganeite (ȕ-FeOOH) have shown that the maximum capacity of the sorbents is around 75 mg of As per gm of akaganeite. Also, quite good results were obtained with synthetic magnetite/maghemite, with a maximum capacity of ca. 40 mg of As per gm of sorbent. Incorporating iron oxides/ oxyhydroxides based nanostructures in zeolites and nanoscale zero-valent iron (nZVI) technology provides a method for high capacity As removal. 5.2. FATE AND TRANSPORT OF NANOMATERIALS
The mode of transport, interaction and chemical affects on the human body of atmospheric pollutants is not well understood. Despite beneficial effects and major developments in the field of nanomaterials, there is a significant gap in our knowledge of the environmental, health, and ecological impacts associated with nanostructured materials. A comprehensive and fundamental investigation into the dynamic transport of nanomaterials in the environment and its impact on human health and ecology is needed to guard public welfare. The complex nature of naturally occurring and engineered nanomaterials and transport, either in the environment or via different exposure routes with human body necessitates an ontological modality. A matrix of parameters which govern fate and transport modeling of nanomaterials such as exposure routes, chemical composition, surface structure, solubility, size and shape effects, toxicity, absorption, distribution, metabolism, agglomeration, and excretion rate and mechanisms is under investigation via three most common routes – inhalation, ingestion, and dermal exposure. The essential aspect of employing ontology is in defining not only the item, but also its relationship to other items in the lexicon; to enable information
24
A. VASEASHTA
retrieval from various sources, databases, and expert elicitation. Biodegradation and bioaccumulation of nanomaterials should be sufficiently addressed due to impacts on human health. Studies relating to the thermodynamic properties, interfaces, and free energy of nanoparticles as a function of particle size, composition, phase and crystallinity influence particle dissolution in a biological environment. The accumulation, dispersion, and functional surface groups play an important role in cytotoxicity and in evaluating pathways of cellular uptake, subcellular localization, and targeting of subcellular organelles. This investigation in conjunction with plume modeling and predictive science approaches will assist in prioritizing, testing, and correlating with in-vivo exposure models. 5.3. SOCIAL ACCOUNTABILITY AND RISK ASSESSMENT
The understanding of environmental effects and health risks associated with nanotechnology is very limited and sometimes contradictory. At present, a well defined risk assessment modality is needed, as the “voluntary code of conduct” for using nanomaterials is somewhat trepidation based and not entirely on scientific methodology. Voluntary codes cover seven general principles, including taxonomy, significance, implementation, sustainability, precaution, inclusiveness, and accountability. Risk analysis is a process of evaluating critical assets and systems, threats, vulnerabilities, and controls for mitigating threats. Effective risk analysis involves identifying, assessing, and mitigating risk to a satisfactory level. Outcomes of risk assessment form a strategic plan for managing risks. Nanotechnology risk assessment occurs in a climate of uncertainty and change; therefore, effective decision making by participating experts is critical for successful outcomes. According to Howard,35 high quality decisions are characterized by the following elements: appropriate framing of decisions; analysis of creative alternatives; accurate information and sound models; clear preferences for future status; sound logic; and commitment to process and outcomes. Policy and decision makers are increasingly relying on expert-opinion elicitation techniques for forecasting advances, reliability, and risks related to science and technology. Structured expert-opinion elicitation techniques effectively support complex decision-making in the face of risk and uncertainty. Since nanotechnology risk assessment is in its infancy, incorporating formal expertopinion elicitation methods within the risk assessment process may help to prevent cognitive biases and faulty cognitive processes attributing to poor decision quality. The Delphi Technique is a structured interactive group communications technique effective for reaching consensus about judgments, forecasts, or
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
25
decisions from expert panels. The Delphi process occurs as follows: a facilitator distributes survey to the expert panel; the survey is answered anonymously and independently by the expert panel; a facilitator summarizes and distributes results and rationale, and the expert panel anonymously and independently reviews summarized results and rationale, and panelists are allowed to revise individual responses. The process of eliciting, summarizing, and distributing anonymous and independent responses continues until consensus, stable responses, or a given number of rounds is met. The Delphi technique has proven merit in forecasting trends and risks for many scientists, researchers, policy, and decision makers.36 6. Conclusions, Discussion, and Future Directions Nanotechnologies promise a cornucopia of new products with superior performance and characteristics, and offer solutions to some of the greatest challenges of the 21st century. As proposed in Figure 2r, further advances are necessary to entirely understand, characterize, develop, and optimize the properties of these materials, devices, and systems. A key challenge for such applications is precisely controlled growth of these materials at desired sites with a desired structure and orientation. Specific to remote sensing via satellite, advances in nanophotonics will enhance resolution, thus providing enhanced feature extraction capabilities. Future pollution remediation nanomaterials-based techniques may include pollution transport by nanoparticles which will have medical, radiological, and even national defense implications in terms of human health, safety, and environment. Ontological approaches towards studying and understanding the interaction of nanoparticles with the human body, nanoparticles dynamics in air and aquatic systems, composition-dependent disposition and dispersion of nanoparticles, short and long-term effects of nanomaterials with the human body, immunotoxocity and phototoxocity of nanoscale materials will address some of the societal issues of nanotechnology. Furthermore, this investigation aims to reconcile two visions of the future: first, the transcendence of nanomaterials and systems into nature of self-healing, brain-like computing or artificial intelligence, virtual reality, and bio-mimetic applications; and second, the process in nature produced nanomaterials and system via self-assembly, green chemistry synthesis, and other room temperature processes. The path to reach this technological singularity; specifically the technology to mimic nature and that of nature to develop technology is complex, yet attainable. A roadmap of nanomaterials, in-terms of complexity, discipline nexus, and emerging technologies is displayed in Figure 2r envisaging the invisible boundary between vision and reality.
26
A. VASEASHTA
ACNOWLEDGEMENTS
Immense acknowledgements are due to the NATO for funding for this ASI, Ms. Annelies Kersbergen for help, guidance, and patience during the preparation of this proceeding, and my colleagues Profs. Mihailescu, Reisfeld, Denkbas, Erdem, and Stamatin and students for fruitful discussions and contributions.
References 1. Vaseashta, A., Dimova-Malinovska, D., and Marshall J. (2005), Nanostructured and Advanced Materials, Springer, Dordrecht. 2. Vaseashta, A., and Stamatin, I. (2007), JOAM 9(6), 1506–1613. 3. Vaseashta, A., Vaclavikova, M., Vaseashta, S., Gallios, G., Roy, P., and Pummakarnchana, O. (2007), Science and Technology of Advanced Materials 8, 47–59. 4. Pumakaranchana, O., Phonekeo, V., and Vaseashta, A. (2008), this volume. 5. Denkbas, E., and Vaseashta, A. (June 2008), NANO: Brief Reports and Reviews. 6. Erdem, A. (2008), this volume. 7. Erdem, A., Karadeniz, H., Caliskan, A., and Vaseashta, A. (June 2008), NANO: Brief Reports & Reviews. 8. Vaseashta, A., Riesfeld, R., and Mihailescu, I. (2008), MRS 2008 Spring Proceedings. 9. Pokropivny, V., Pokropivny, P., and Vaseashta, A. (2005), In: Nanostructured and Advanced Materials, A. Vaseashta et al., eds., Springer, Dordrecht, pp. 367–370. 10. Davis, Z., Svendsen, W., and Boisen, A. (2006), Proceedings of the 32nd International Conference on Micro and Nano Engineering, MNE 2006, Spain, pp. 475–476. 11. Vaseashta, A., Erdem, A., and Stamatin, I. (2006), MRS 2006 Spring Proceedings 920. 12. Wang, J. (2003), Analytica Chemica Acta 500, 247. 13. Sahm, T., Rong, W., Bârsan, N., Mädler, L., and Weimar, U. (2007), Sensors and Actuators B: Chemical 127(1), 63–68. 14. Wiederrecht, G. (2004), European Physical Journal Applied Physics 28, 3–18. 15. Anantha Ramakrishna, S. (2005), Reports on Progress in Physics 68, 449–521. 16. Ligia, R., Furlan, A., Garrido, L., Brumatti, G., Amarante-Mendes, G., Martins, R., Cândida, M., Facciotti, R., and Padilla, G. (2002), Biotechnology Letters 24. 17. Sailor, M., Schmedake, T., Cunin, F., and Link, J. (2002), Advanced Materials 14, 1270. 18. Baizeng, F., Haoshen, Z., and Honma, T. (2006), Journal of Physical Chemistry B Condensed Matter, Materials, Surfaces, Interfaces & Biophysical 110, 4875–4880. 19. Wang, S., Zhang, Q., Wang, Q., and Yoon, S.F. (2003), Biochemical & Biophysical Research Communication 311, 572. 20. Cui, Z., and Mumper, R. (2003), Critical Reviews in Therapeutic Drug Carrier Systems 20, 103–137. 21. Vaseashta, A., Boskovic, O., Webb, Al., Ozdemir, N., and Ozturk, E. (2006), In: Functional Properties of Nanostructured Materials, R. Kassing et al., eds., Springer, Dordrecht. 22. Smith, A., Gao, X., and Nie, S. (2004), Photochemistry and Photobiology 80, 377–385. 23. Wu, X., Liu, H., Liu, J., Haley, K., Treadway, J., Larson, J., Ge, N., Peale, F., and Bruchez, M. (2002), Nature Biotechnology, 21, 41–46. 24. Jain, K. (2003), Expert Reviews in Molecular Diagnostics 3, 153–161. 25. Tan, Q., Ji, J., Barbosa, M., Fonseca, C., and Shen, J. (2003), Biomaterials 24, 4699–4705. 26. Curtis, A., and Wilkinson, C. (1999), Biochemistry Social Symposium 65, 15.
NANOSCALE CHEM.-BIO SENSORS AND ENERGY DEVICES
27
27. West, J., and Halas, N. (2000), Current Opinion Biotechnology 11, 215–217. 28. Shi, H., Tsai, W., Garrison, M., Ferrari, S., and Ratner, B. (1999), Nature 398, 593–597. 29. Öztürk, E., Denkbaú, E., and Berestein, G. (2006), 7th International Biorelated Polymers Symposium, in ACS Meeting, 10–14. 30. Wang, C., Wang, X., Lei, Y., Chen, C., and Wang, Q. (1997), International Journal of Hydrogen Energy 22(12), 1117–1124. 31. Ebrasu, D., Stamatin, I., and Vaseashta, A. (June 2008), NANO: Brief Reports & Reviews. 32. Becher, M. et al. (2003), Comptes Rendus Physique 4, 1055–1062. 33. Benard, P., and Chahine, R. (2007), Scripta Materialia 56(10), 803–808. 34. Prasad, P., and Elnashaie, S. (2004), Industrial & Engineering Chemistry Research 43(2), 494–501. 35. Howard, R. (2007), In: Advances in Decision Analysis: From Foundations to Applications, W. Edwards, R. Miles, Jr., and D. von Winterfeldt, eds., Cambridge University Press, New York, pp. 32–56. 36. Vaseashta, A. (2008), Presented at the NATO ARW “Risk, Uncertainty and Decision Analysis for Nanomaterials: Environmental Risks and Benefits and Emerging Consumer Products”, Faro, Portugal, 2008.
NANOSTRUCTURED THIN OPTICAL SENSORS FOR TRACE GAS DETECTION C. RISTOSCU1*, I.N. MIHAILESCU1, D. CAITEANU1, C.N. MIHAILESCU1, TH. MAZINGUE2, L. ESCOUBAS2, A. PERRONE3, AND H. DU4 1 Lasers Department, National Institute for Lasers, Plasma and Radiation Physics, PO Box MG-36, Bucharest – Magurele, RO-77125, ROMANIA 2 Université Paul Cézanne Aix-Marseille III Laboratoire TECSEN, Domaine Universitaire de Saint Jérôme, Marseille, FRANCE 3 University of Salento, Physics Department and INFN, Lecce, ITALY 4 Advanced Materials Research Institute, Northumbria University, UK
Abstract – We report on new features of nanostructured coatings used in trace gas detection applications based on the modification of optical parameters. We review the general physical principles underlying the operation of optical detectors and introduce the basic systems used in trace gas recognition. Pd-doped SnO2 thin films were pulsed laser deposited on Si (100) and quartz substrates in 10 Pa O2 at different substrate temperatures. Coatings on Si were investigated by scanning electron microscopy and atomic force microscopy to determine their surface morphology and by grazing incidence X-ray diffraction for films structure. Twin films deposited on quartz were investigated using the m-line technique to monitor the variation of optical properties when exposed to butane diluted in N2 down to 100 ppm concentration. The acquired expertise was used to design a miniaturized system which was tried in real conditions.
Keywords: Metal oxide films, nanostructured coatings, optical gas sensor, m-line technique.
______ *
To whom correspondence should be addressed: C. Ristoscu, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
29
30
C. RISTOSCU ET AL.
1. Introductory Remarks On-line monitoring of emission gases from different combustion and/or chemical processes has become an important issue as developing techniques in emission and environmental control. Given adequate catalysis, the gas under detection initiates electrochemical reactions that alter the electrical response of a suitably set circuit. By measuring conductivity changes and in respect with catalysts used, the concentration of the gas can be established. Several electrochemical cells are usually employed to provide the required response and selectivity. The operation requires high electrical power, elevated temperatures, and long-term calibration procedures. Over the years, solid-state versions of such electrochemical systems have been developed using semiconductor metal oxides. Their conductivity is changed as a result of gas exposure, largely through reduction-oxidization reactions. In a similar approach systems using organic polymers instead of metal oxides have also been devised. A well developed technology currently available integrates the devices electronically into complete sensor chips. The main drawback of such systems is their electrical operation requiring high power and temperature, electrical connections, etc. They are nevertheless the dominant technology in today’s market. Alternative approaches, based on molecular mass detection in surface acoustic wave or quartz microbalance schemes, are also used, but they involve greater complexity, and the difficulties they face are also larger. A challenging alternative consists in optical schemes that use refractive changes and minimal absorption alterations to enable remote operation and circumvent the need for power supply. Such schemes rely on interferometry and diffraction to provide a unique edge over other detection techniques not only in terms of the microscopic nature of the interactions, but also as concerns the final target products. Their key advantages include remote sensing on long interrogation lengths along with zero power consumption at the sensing point location and hence no need for electrical circuitry at that point or in its vicinity. Besides, they usually operate at room temperature, with zero electro-magnetic interference, and minimal power consumption and can be used in protected or physically isolated areas. Multi-parameter sensorial response is another plus. We review in this paper the basic operation principle of nanostructured optical gas sensors focusing on the example of Pd-doped SnO2 thin films obtained by pulsed laser deposition.
NANOSTRUCTURED THIN OPTICAL SENSORS
31
2. Optical Gas Sensors 2.1. BASIC EQUATIONS
As is well known,1 in the case of non-magnetic homogeneous materials, the electromagnetic field in harmonic regime can be described in terms of electric, E Z , and magnetic fields, H Z , respectively. As a result, Maxwell’s equations can be written as follows:
rot E Z P0
rot H Z
w H Z wt
w DZ J Z wt
(1)
(2)
where DZ is the displacement vector and J Z stands for current density. Equations (1) and (2) can be also written:
rot E Z
jZP0 H Z
rot H Z jZ DZ V E Z
(3) (4)
By introducing the expression
DZ H 0H r Z E Z
(5)
(with H 0 being the electrical permittivity of vacuum and H r the material’s one), Eq. (4) becomes:
§ V · ¸ E Z rot H Z jZH 0 ¨¨ H r j ZH 0 ¸¹ ©
(6)
or further:
rot H Z jZH 0 n 2 E Z 2
(7)
Here n is the complex refractive index of the material which can be expressed as:
32
C. RISTOSCU ET AL.
n2
H r'
Hr j
V ZH 0
(8)
Due to the fact than n is complex, it can be found as n
n' jn' ' . This
2 n'2 n' '2 j 2 u n'un' ' , where n” corresponds to the absorption implies n of the material.
In these conditions, Eq. (8) can further be written:
n '2 n ' '2 H r ° V ® °2 u n'un' ' ZH 0 ¯
(9)
The solutions of this system, which have a physical meaning (n’ > 1), are:
° °n ' ° °° ® °n ' ' ° ° ° °¯
1
2 ·º 2 ª § § · H V ¨ r « ¸¸ ¸¸» 1 1 ¨¨ ¨ «2¨ © 2H 0H rZ ¹ ¸¹»» ¬« © ¼ V 2H 0Z 1
2 ·º 2 ª § « H r ¨1 1 §¨ V ·¸ ¸» ¨ 2H H Z ¸ ¸¸» « 2 ¨¨ © 0 r ¹ ¹» «¬ © ¼
(10)
Equation (10) shows that a variation in the surface conductivity of a material leads1 to a change in its refractive index (the real n’ and imaginary n” parts vary 2 as V ). 2.2. MAIN EXPERIMENTAL SETUPS
In optical sensing, gas detection is achieved by optical interrogation. One has a large choice of laser illuminated interferometers to probe for refractive index variation caused by trace gases. Our option for the studies reported in this paper went to an m-line setup illuminated with a stabilized He-Ne laser source at 633 nm. This technique is suitable and frequently applied in order to measure variations of optical parameters in transparent waveguides.2–4 A typical experimental setup is presented in Figure 1.
NANOSTRUCTURED THIN OPTICAL SENSORS
33
Figure 1. Testing facility for waveguide coupling interrogation schemes.
At the center of the test setup were the prism coupler on its rotation plate and the sensing element, i.e. the thin film, which had been deposited on a transparent substrate pressed on the prism base. Physical interactions between the gas and the sensing material leading to variations in the refractive index of the latter took place at the air-sensitive material interface. Calculations based on equations in Section 2.1. showed5 that the electromagnetic field was enhanced at this interface for the transversal magnetic (TM) polarization state of light and a thin film thickness close to the cut-off thickness of the guided mode (i.e. the thickness below which the mode cannot propagate in the thin film any longer). This configuration ensured the highest sensitivity to refractive index variation under gas exposure. 2.2.1. Coupling by a prism The coupling by a prism of an incident laser beam into a planar waveguide is governed by the incident angle Ts between the beam and the prism base (Figure 2). At a particular incidence angle (called synchronism angle), the resonant coupling of the laser beam into the waveguide can be observed by watching a dark line (known as mode line or m-line) that appears in the reflected beam.
34
C. RISTOSCU ET AL.
dark line
laser beam
Ts
prism (nP)
waveguide (nW)
reflected spot in the Fresnel field region
substrate (nS) Figure 2. Prism coupler.
Consequently, the application of the method is mainly based on measuring the angles corresponding to the m-lines observed. From the propagation constants of the guided modes determined from these angles, the refractive index n and thickness h of the waveguide can be inferred. In the case of an angle measurement accuracy of 10–3, which is rather easily achievable, we have an accuracy of 10–4 for the refractive index estimation and r2 nm for thickness evaluation. We measured the refractive index and waveguiding properties of thin films for the two polarization states of light. The resonant coupling of the laser beam into the waveguide was strongly affected by perturbations that were due to the interaction of the material and surrounding medium. The effective index N of the guided mode for the resonant coupling is connected to the incident angle Ts of light by Eq. (11):
N
ª º § sin T s · ¸ Ap » n p sin «arcsin¨ ¨ n ¸ «¬ »¼ © p ¹
(11)
where np is the refractive index of the prism, and Ap its characteristic angle. From Eq. (11), we deduce the following expression of 'N:
'N
ª º § sinT s · ¸ Ap » n p cosT s cos«arcsin¨ ¨ n ¸ «¬ »¼ © p ¹ 'T s 2 n 2p sinT s
(12)
2.2.2. Coupling by a diffractive grating Light coupling can also be obtained using a coupling grating etched in the waveguide. The advantage of this system over the prism one consists in the possibility of visualizing in the transmitted beam the actual displacement of the m-line in the presence of gas, as shown in Figure 3.
NANOSTRUCTURED THIN OPTICAL SENSORS
35
Laser Beam O Ts
/ h
d N: effective index
n
transparent substrate (ns)
e
Figure 3. Light coupling by a diffractive grating engraved in a waveguide.
The angle is measured in both cases in a similar way. The corresponding relation between synchronism angle, Ts, beam wavelength, O, grating period, /, and the effective index for zeroth order, N, is:
N
sin T s
O /
From Eq. (13), we deduced the following expression of 'N:
'N
O· § 'T s 1 ¨ N ¸ /¹ ©
(13)
2
(14) Note that the results of the two methods are comparable in terms of sensitivity. In both of them, device operation is based on the m-line coupling principle, but the use of a grating coupler increases robustness and enables miniaturization (see further Section 4). 2.3. SENSING ELEMENTS
2.3.1. Optical sensors: critical review In the case of electrochemical conductivity sensors (MOS), reversible charge transfer reactions lead to a collective behavior consisting of electrical conductivity changes. Similarly, alterations of the optical properties occur by agent
36
C. RISTOSCU ET AL.
adsorption in suitably designed materials, such as metal oxides. These changes can be detected optically with very high accuracy and allow for a wide dynamic range multidimensional parametric space for work. Nanocomposite materials exploration by a two-step production and encapsulation process has also been initiated recently. The functionality of advanced composites would further be enhanced by the inclusion of special nanoparticles. The current trends encourage efforts towards the production of composites based on metals, semiconductors, metal oxides, metal-coated nanoparticles and nanotubes. Non-linear optical mechanisms are expected to offer additional innovative sensing tools that will complement the linear optical response under chemical exposure. Further innovation is foreseen by the use of organic materials incorporated in inorganic matrices to selectively adsorb specific agents. A novel approach consists in applying polymer-based sensitive receptors and including functionalized acidic pendants for nitroaromatic detection. We have to mention, though, that polymers age rapidly and are highly prone to corrosion by various aerosols. Both factors strongly influence stability and reliability in gas detection. At present, the chief concern in optical gas sensing is the reversibility of the physicochemical optical changes that take place in materials, mainly thinfilm structures, upon exposure to the chemical environment. Research efforts are aimed at the selective detection at room temperature with sensitivity in the pp-million range and in some cases (e.g., nitro-aromatics) pp-billion range. The predicted response time of the sensors is on the order of ms (depending on the sensor head), while the anticipated recovery time determined by the adsorption affinity ranges from 10 ms to 10 s. Based on these considerations, we made our choice for highly transparent thin metal oxide films, e.g., ZnO, SnO2, TiO2, WO3, as the simplest, very reliable solution for developing a cheap, easy to operate, potentially miniaturized detection system. 2.3.2. Thin metal oxide sensors: pure, doped and clustered Materials used for waveguides have to meet the basic prerequisite of providing high transmittance in the visible or infrared region. Metal-oxide optical sensors, due to their simple construction, may offer advantages in terms of resistance to severe conditions such as high temperature and corrosive environment, and the capability of being placed in close vicinity of emissions. This type of sensor may also have a short response time so as to enable the possibility of real-time control. However, modern applications of these sensors still face the problems of high cross-sensitivity to gases, i.e. rather low selectivity; high sensitivity to
NANOSTRUCTURED THIN OPTICAL SENSORS
37
ambient humidity; and long-term drift, related to oxygen bulk diffusion and transformations of the crystal structure of polycrystalline materials.6 A considerable improvement of operating parameters such as gas response selectivity, stability, and rate, can be achieved by optimizing both the bulk and surface structure of the metal oxide films applied. Efforts to optimize sensor output characteristics have so far mainly focused on finding optimal technological methods for doping metal oxides with nanoscale metal catalyst additives. Decreasing crystallites (grains) to nm dimensions has been another major goal. Dopants can boost the catalytic activity of the base oxide, stabilize a particular valence state, favor formation of active phases, stabilize the catalyst against reduction, and/or increase the electron exchange rate. Inserting active additives into base metal oxides can modify their parameters such as charge carrier concentration; chemical and physical properties of the metal oxide matrix; electronic and physicochemical properties of the surface; surface potential and inter-crystallite barriers; phase composition; sizes of crystallites, and so on. Noble metals due to their electronic state and their distribution both on the oxide surface and into the film can strongly promote gas sensor sensitivity and selectivity. Noble metal nanoclusters have been tested as either dopants or surface clusters for boosting gas-sensing efficiency by catalyzing specific reactions with the detected gases. Several catalytic additives/dopants such as Pt,7,8 Pd,7–13 Sb,14 Au,7 and Ru15,16 were chosen for enhancing sensor sensitivity. Different methods have been applied for adding metals to the films. Among them, pulsed laser deposition (PLD) provides some advantages, including reduced contamination due to the use of laser light, controlled composition of deposited structures, and in-situ doping. Moreover, PLD is a versatile, powerful tool for custom designing nanoparticles with the desired size and composition by merely varying the experimental deposition conditions.17–20 This was therefore our choice for obtaining thin metal oxide films doping with noble metal nanoclusters. 3. Optical Gas Sensor with PLD SnO2 Thin Films Doped with Pd Nanoclusters 3.1. MATERIALS AND METHODS
The experiments were performed with a PLD installation schematically depicted in Figure 4. Before each deposition event, the vacuum chamber was evacuated to a residual pressure lower than 10–5 Pa. Meanwhile, the chamber walls were heated to facilitate desorption of water vapors and other contaminants. Vacuum quality was monitored with a quadrupole mass spectrometer.
38
C. RISTOSCU ET AL.
The SnO2 (99.99%) target was ablated using an UV XeCl* excimer laser source (Lambda Physik LPX 315i, O = 308 nm, WFWHM 30 ns) operating at a frequency repetition rate of 10 Hz. To deposit one film, 20,000 subsequent laser pulses were applied. Before each deposition event, target surface was cleaned by applying 5,000 laser pulses. During this time, a shutter was interposed between the target and the collector to avoid deposition of ablated material from the first layers of the target, which usually contain contaminants. An AR-coated MgF2 lens with 300 mm focal length was used to focus the laser beam into a 1 mm2 spot on the target surface. The target was rotated at 3 Hz frequency during multipulse laser irradiation to avoid drilling. In-situ doping with noble metal of the SnO2 films was achieved by placing a thin 1 mm diameter Pd wire across the target surface (see inset in Figure 4). The laser beam was incident on the target at an angle of 45º, while incident laser fluence during deposition was set at 10 J/cm2. Films were deposited in a dynamic flux of 10 Pa O2. The Pd concentration in the SnO2 films was controlled by varying laser spot position on the target surface. The experimental conditions were optimized on the basis of previous parametric studies.13
Pd wire (I 1 ) SnO Figure 4. Experimental PLD/RPLD setup; the inset represents the SnO2 target with a thin 1 mm diameter Pd wire.
Ablated material was collected on Si(100) or quartz substrates maintained at room temperature, 350qC, or 500qC, during the deposition. They were placed 60 mm from the target parallel to it. All substrates were cleaned in an ultrasonic bath prior to deposition. Scanning electron microscopy (SEM) measurements were carried out using an electron microscope JEOL 6320F equipped with facilities for conducting energy dispersive X-ray spectroscopy (EDS) analysis. Grazing incidence X-ray
NANOSTRUCTURED THIN OPTICAL SENSORS
39
diffraction (GIXRD) analysis was carried out using a Siemens D5000 Diffractometer equipped with a Cu KĮ source (O = 0.1541 nm). All measurements were performed at a grazing incidence angle of 3q. Diffractograms were recorded from 10q to 100q at a step of 0.02q. The surface roughness of the deposited films was studied by atomic force microscopy (AFM). X-ray photoelectron spectroscopy (XPS) analysis was carried out using spectrometer system (Kratos Axis Ultra) with a monochromatic AlKD (hQ = 1486.6 eV) photon source. The system was also equipped with a 165 mm hemispherical analyzer for the acquisition of spectra and a concentric spherical mirror analyzer for XP imaging. The m-line technique was used at 633 nm laser wavelength for waveguide coupling interrogation (see also Section 2.2). Light was coupled by a prism to the material in the form of optical resonance (guided mode), and synchronism angle variations following gas introduction were measured. A He-Ne laser source emitting at 633 nm wavelength in transversal electric (TE) mode was used. Transversal magnetic (TM) polarization was obtained using a O/2 plate placed in the exit of the laser beam. The prism used for light coupling was made of TiO2, rutile phase. Its characteristics are given in Table 1. A photo of the prism setup used for light coupling can be seen in Figure 5. TABLE 1. Characteristics of the rutile prism used for light coupling. O (nm) nTE nTM
633 2.8641 2.5821
LASER
SAMPLE HOLDER PRESS
ROTATION STAGE
Figure 5. Photograph of the prism coupler.
40
C. RISTOSCU ET AL.
The sample was pressed on the prism by a polyethylene support, the manual translation of which helped control applied pressure. Different translation and rotation systems may improve the positioning of the prism coupler (eight degrees of freedom). The incident angle of the beam on the prism base can be varied in order to attain synchronism conditions. At that stage, a photodiode measured the position of the reflected beam. The system was proven to be capable of detecting hydrocarbons in the range of 100–1,000 ppm.21–25 3.2. SENSING PERFORMANCES AND EXPLOITATION CHARACTERISTICS
3.2.1. Physicochemical characterization Figure 6 shows the morphologies for the three types of Pd:SnO2 samples deposited in 10 Pa O2 at (a) RT, (b) 350qC, and (c) 500qC, respectively. The sample deposited at RT exhibited a rough surface with cracks, while the two others were rather smooth and densely packed. A quasi-ordered nanostructure can be noticed in the sample deposited at 500qC.
a
b
c
Figure 6. SEM microimages of Pd:SnO2 structures deposited in 10 Pa O2 at (a) RT, (b) 350°C, and (c) 500°C.
AFM studies (Figure 7) sustained these observations. The surface of the sample deposited at RT was porous, with large particulates (>100 nm). By contrast, samples deposited at 350qC and 500qC were denser, and their particulate dimensions were down to ~100 and ~80 nm, respectively. An even slight increase in SnO2 grain size can lead to significant changes in the film structure that is responsible for both the catalytic and gas sensing properties of metal oxides.26 Typical GIXRD spectra of the investigated samples are given in Figure 8. The films are polycrystalline without any preferential orientation. Using the diffraction database,27 the visible peaks were clearly assigned to either SnO2 or Si. As can be seen from the spectra, the film deposited at 500qC (Figure 8c) was better crystallized than those obtained at 350qC and RT.
NANOSTRUCTURED THIN OPTICAL SENSORS
a
b
41
c
Figure 7. AFM microimages of Pd:SnO2 structures deposited in 10 Pa O2 at (a) RT, (b) 350°C, and (c) 500°C.
Note that the films only consisted of SnO2 phase. The most important lines were identified, as e.g. (110) at 26.6q, (101) at 33.9q, (200) at 37.9q, and (211) at 51.7q.27 EDS results confirmed the formation of SnO2 phase including about 2% wt of Pd. 3.2.2. Detection of trace gas (hydrocarbon) All sensors were exposed at room temperature to different concentrations of butane diluted in nitrogen (1,000, 500, 200, or 100 ppm). The applied testing protocol was as follows: (1) the detection cell was evacuated to clear it from impurities (1 min 30 s); (2) the cell was filled with N2 to atmospheric pressure (2 min 30 s); (3) the N2 + butane mixture was let in (3 min); (4) the mixture inflow was stopped, and N2 was introduced to study the return to baseline (3 min); and (5) cycles 1–4 were repeated. The typical optical responses we recorded are given in Figure 9, in which variation due to the presence of butane is evident. An important feature was that the optical response was repeatable for all tested sensors. In vacuum the signal becomes stable at ~400 mV, while the mixture of butane-nitrogen made it stabilize at about 550 mV. This variation was due to butane and could not be attributed to pressure effects. Moreover, the signal always reached the same level regardless of the preceding phase of the protocol, which indicates high reproducibility and reversibility. All films showed similar responses when the concentration of butane decreased. The shape of the signal variation kept unchanged, but the kinetics was slower at lower butane concentrations. The response time was about 1 min for 1,000 ppm, 2 min for 500 ppm, and 3–4 min for 100 ppm. The response time is defined as the time required for stabilizing the signal level from the phase of carrier (nitrogen) gas flow to that of butanenitrogen mixture exposure.
42
C. RISTOSCU ET AL.
a
130 120 110 100
SnO2 (220)
30 20
Si (331)
Si (220)
40
Si (400)
50
SnO2 (211)
SnO2 (101)
60
SnO2 (200)
70
Si (311)
80
SnO2 (110)
Lin (Counts)
90
10 0 12
20
30
40
50
60
70
80
90
10
2-Theta - Scale
b)
SnO2 (211) SnO2 (110)
Si (400)
SnO2 (312)
SnO 2
Sn O2 (20 2)
(310 )
SnO2 (220)
SnO2 (321)
SnO2 (301)
SnO2 (200)
100
SnO2 (101)
Lin (Counts)
200
0 10
20
30
40
50
60
70
80
90
10
80
90
2-Theta - Scale
SnO2 (321)
SnO2 (332)
SnO2 (221)
100
SnO2 (220)
SnO2 (101)
200
SnO2 (200)
SnO2 (110)
Lin (Counts)
300
SnO2 (002)
SnO2 (211)
Si (400)
c)
400
0 10
20
30
40
50
60
70
10
2-Theta - Scale
Figure 8. GIXRD spectra of Pd:SnO2 structures deposited in 10 Pa O2 at (a) RT, (b) 350°C, and (c) 500°C, respectively (grazing incidence angle of 3º).
250
N2
N2 + butane
V (mV)
200
N2
N2
a)
50
43
N2 + butane
N2 + butane
150 100
vacuum
vacuum
300
vacuum
NANOSTRUCTURED THIN OPTICAL SENSORS
1000ppm 200 ppm
0 0
5
10
15
20
t (min) 600
500
200
vacuum
N2
N2 + butane
N2
N2 + butane
vacuum
N2 + butane
N2 300
vacuum
V (mV)
400
1000ppm 100
500ppm
b
100ppm
0 0
5
10
15
20
t (m in) 600
500
N2
N2 N2 + butane
N2
N2 + butane
vaccum
300
vaccum
vaccum
V(mV)
400 N2 + butane
200
100
1000ppm 500ppm 100ppm
c)
0 0
5
10
15
20
t (min)
Figure 9. Sensitivity of Pd:SnO2 sensor to different concentrations of butane diluted in N2. Quartz substrate temperature during PLD was (a) RT, (b) 350°C, and (c) 500°C.
From Figure 10 we notice that the higher the substrate temperature applied during PLD, the higher the gas sensitivity of the sensors. The (110) and (101) planes of the SnO2 crystal are known to be F (flat) faces. According to the periodic bond chains (PBC) theory,28 surface atoms of F-faces are strongly bound to each other in directions parallel to the face. As a result, these faces have a slight tendency to react with the arriving atoms. It has also been established that (110) and (101) surfaces of SnO2 grow via a layer-bylayer mechanism.
44
C. RISTOSCU ET AL. 600
tension (v)
500 400
vacuum
N2
N2 +butane
N2
vacuum
N2
N2 +butane
vacuum
N2 +butane Room temperature 500˚C
300 200 100 0 0
5
10
15
20
t(s)
Figure 10. Effects of Pd:SnO2 film deposition temperature on the sensitivity to 1,000 ppm of butane.
The various crystallographic planes have different distances between the Sn atoms, which can be ranked as follows: d(110) ~ d(100) < d(101) < d(001).29 Sn atoms are centers of oxygen chemisorption. Changing the indicated distance can therefore influence the rate of dissociative oxygen chemisorption, which is in many cases the main process of gas sensing phenomena.19–23 The observed change in SnO2 crystal faceting could actually be one of the reasons behind the considerable modification of gas sensing characteristics in SnO2 films. It was shown in the case of SnO2 sensors made by selective ionic layer deposition (SILD)30 that the small size of the crystallites (grains) was an essential but not sufficient condition for achieving maximum gas sensitivity together with a fast response. When comparing the gas sensing characteristics of metal oxide-based sensors, one has to primarily consider the size of the conglomerates and their porosity, and only after that the size of grains and thickness of the film.30,31 However, we have to note that the above results are only due to part of the metal oxide matrix parameters which can influence gas sensing properties. Along with geometrical parameters, the physicochemical characteristics of the gas sensing matrix have to be taken into account. They include the chemical and phase composition; the concentration of bulk and surface oxygen vacancies; the size and density of both metal catalyst particles and single atoms on the surface of metal oxides; and the type and concentration of uncontrolled impurities. 3.2.3. Gas contamination and multiple operations The potential gas contamination of the sensors as an effect of butane exposure was studied by AFM and XPS. We compared twin samples one of which had been repeatedly exposed to gas action, while another had been kept in controlled, non-polluted atmospheres.
NANOSTRUCTURED THIN OPTICAL SENSORS
45
AFM analyses (Figure 11) demonstrated that topographies in the cases of unexposed and exposed samples were similar to each other. Nanoparticulates size was in the (45–55 nm) range. The XPS studies indicate that the near surface region of the unexposed sample comprises a mixture of O, Sn, and Pd. The first peak of the Sn 3d doublet occurs at 485.8 eV BE and is assigned to emission from Sn 3d5/2 corelevel of Sn2+ ion in tin oxide. The second component occurs at 494.2 eV and is assigned to emission from the Sn 3d3/2 core level of Sn2+ ions. The 8.4 eV separation of 3/2 and 5/2 components is consistent with this interpretation of SnO/SnO2. The primary component of O 1s occurs at 529.7 eV and is assigned to emission from O2– ions in SnO/SnO2 and PdO. The second component occurs at 530.9 eV and is assigned to emission from chemisorbed OH groups, based on the 2.2 eV separation of the secondary component from the primary one. The third component occurs at 532.9 eV and is assigned to the presence of chemisorbed H2O molecules. The first peak of Pd 3d doublet occurs at 335.7 eV and is assigned to emission from the Pd 3d5/2 core-level of Pd2+ ion in PdO. The second component at 336.8 eV is assigned to emission from the Pd 3d5/2 core-level of Pd4+ ions in PdO2. The XPS spectrum of the “as received” surface of the exposed sample clearly indicates that the near-surface region comprises a mixture of Sn, O, and C. The first component of O1s occurs at 530.6 eV BE and is assigned to emission from O2– ions. The second component occurs at 532.1 eV and, based on the 1.5 eV BE shift, is assigned to emission from adsorbed OH– groups. The Sn 3d5/2 peak occurs at 486.7 eV BE, whilst the Sn 3d3/2 is centered on 495.2 eV BE. The 8.5 eV separation of the two components is consistent with the presence of SnO/SnO2 in the near surface region of the sample. High-resolution C1s spectra were recorded for the unexposed (Figure 12) and exposed (Figure 13) samples, while the main quantitative information was collected in Tables 2 and 3.
a
b
Figure 11. AFM images of Pd:SnO2 samples (a) unexposed and (b) repeatedly exposed to 1,000 ppm butane diluted in nitrogen.
46
C. RISTOSCU ET AL. C1s ZnPd25Q
x102
C1s C1s
Intensity(CPS)
80
60
40
20 290
292
288
286
284
282
Binding Energy (eV)
Figure 12. High-resolution C1s XP spectrum obtained from the surface of unexposed sample. C 1s SnPd26 x 10
3
30
25
CPS
20
15
10
5 290
288
286 Binding Energy (eV)
284
282
Figure 13. High-resolution C1s XP spectrum obtained from the “as received” surface of exposed sample.
By comparing data in Tables 2 and 3, we see that the CH radical did not increase in the exposed sample; on the contrary, there was a reduction of this chemical species in the sample exposed to the butane. It should be stressed one would expect the CH radical to be the butane related species.
NANOSTRUCTURED THIN OPTICAL SENSORS
47
TABLE 2. Summary of quantitative information obtained by analysis of C1s peak as shown in Figure 12. Peak
Position BE (eV)
FWHM (eV)
Raw area (CPS)
RSF
Atomic mass
Atomic conc (%)
Mass conc (%)
C 1s
285.000
1.056
10,004.99
0.278
12.011
60.06
60.06
285.492
1.051
4,625.14
0.278
12.011
27.76
27.76
286.540 287.005
0.890 0.890
639.30 369.07
0.278 0.278
12.011 12.011
3.838 2.215
3.838 2.215
287.946 289.239
1.310 1.191
466.81 553.62
0.278 0.278
12.011 12.011
2.802 3.323
2.802 3.323
(C-C/C-H)
C 1s (C-CO2)
C 1s(C-O) C 1s (C=O)-O
C 1s(C-S) C 1s (shake up)
TABLE 3. Summary of quantitative information obtained by analysis of C1s peak shown in Figure 13. Peak
Position BE (eV)
FWHM (eV)
Raw area (CPS)
RSF
Atomic mass
Atomic conc (%)
Mass conc (%)
C 1s
285.000
0.830
4,350.3
0.278
12.011
45.12
45.12
285.517
0.800
3,052.8
0.278
12.011
31.66
31.66
C 1s(C-O)
285.906
0.728
1,025.1
0.278
12.011
10.63
10.63
C 1s
286.440
0.834
755.6
0.278
12.011
7.84
7.84
289.359
1.175
458.7
0.278
12.011
4.76
4.76
(C-C/C-H)
C 1s (C-CO2)
(C=O)-O
C 1s (shake up)
The latter observation and the results of morphological investigations suggest that the contamination effects due to exposure of the samples to polluted gases are very weak and hence hard to detect, if any. 4. Miniaturized Portable Prototype for Trace Gas Detection The dimensions of the detection system described in Section 2 can be reduced by using an etched coupling grating on the waveguide instead of the prism. We looked for a grating coupler for which the TM0 coupling should take place
48
C. RISTOSCU ET AL.
at normal incidence, i.e. Ts = 0 (see Figure 3). The new setup includes the lithographically submicron etched active material that is required to achieve waveguide coupling. The other basic components are: (i) enclosure for gas insertion, (ii) laser diode source and optics, (iii) high resolution CCD to detect the m-line, and (iv) data acquisition and image processing system for detecting variations of the dark line and reference them to the gas analysis. The new detection system redesigned and rescaled is presented in Figure 14. As a remarkable result, we emphasize that its dimensions do not exceed 15 cm length and 6 cm diameter. The system was recently tried in real conditions. Gas traces below 800 ppm, which is the maximum daily exposure to butane authorized under U.S. Federal Regulations, could be reliably detected. The use of the grating as sensing element also made it possible to cut down sharply the detection time. The system is still being worked on to improve its selectivity and sensitivity. Highly innovative schemes based on direct writing of waveguides in different materials have been recently developed, viz: in LiNbO3 in order to enable active electro-optic control of Mach-Zehender Interferometer, and in inorganic-organic hybrid materials deposited by sol-gel method. Multimode interference couplers written with a UV laser allow for the detection of very low refractive index variations.32 5. Conclusions We reviewed the basic physical principles and applied optical schemes to develop a new generation of high-performance optical gas detectors. General physical equations were introduced and the main solutions for optical sensing elements were critically analyzed. We based our decision to select nanostructured metal oxide thin films as sensing elements on their high sensitivity to gas traces and relative robustness and stability for multiple operations. As an application, we analyzed the case of Pd-doped-SnO2 thin films grown on Si (100) and quartz substrates by pulsed laser deposition in 10 Pa O2 at different substrate temperatures (RT, 350, and 500qC). Films deposited on Si were morphologically and structurally investigated. The results showed that 10 Pa O2 background pressure enabled the formation of nanoparticulates by condensation in the gas phase. Twin films deposited on quartz were investigated by the m-line technique to study the variation of optical properties when exposed to test gases (butane). Our results demonstrated that the gas sensing properties were mainly function of the structure and surface morphology of the samples, which in turn depended on substrate temperature. Moreover, it was evident that the higher the substrate temperature during PLD, the higher the gas sensitivity of the obtained SnO2 sensors. Our investigations did not evidence within measurement limits any remanent contamination of the sensing elements.
NANOSTRUCTURED THIN OPTICAL SENSORS
49
Figure 14. Photo of the miniaturized optical gas sensor prototype.
A miniaturized prototype with a sub-micrometric grating etched in the waveguide film as coupling element was successfully tried for operation in ambient atmosphere. The system is now under development to improve its selectivity and sensitivity to a wide range of detectable gases. ACKNOWLEDGMENTS
The financial support of the EU under the contract NANOPHOS IST-200139112 is acknowledged with thanks. CR also acknowledges the CNR – NATO fellowship Pos. 216.2169, Prot. N. 0015503 and NATO CBP.RIG.982424 contract. The Romanian and French authors acknowledge with thanks the support they received under the 2006–2007 Collaboration Agreement no. 19611 between CNRS and the Romanian Academy.
References 1. 2. 3. 4. 5. 6. 7. 8. 9.
Born, M., and Wolf, E. (1975) Principles of Optics, Pergamon, Oxford, Chapter II. Agan, S., Ay, F., Kocabas, A., and Aydinli, A. (2003) Applied Physics A 80(2), 341. Shi, J., Cao, A., Zhu, J., and Shen, Q. (2004) Applied Physics Letters 84(17), 3253. Luo, Y., Hall, D., Kou, L., Blum, O., and Hou, H. (1999) Applied Physics Letters 75(20), 3078. Mazingue, T. (2005) Ph.D. thesis, Universite Paul Cezane Aix Marseille II, Marseille, France. Korotcenkov, G. (2005) Sensors and Actuators B 107, 209. Wurzinger, O., and Reinhardt, G. (2004) Sensors and Actuators B 103, 104. Gaidi, M., Chenevier, B., and Labeau, M. (2000) Sensors and Actuators B 62, 43. Chatterjee, K., Chatterjee, A., Banerjee, A., Raut, M., Pal, N., Sen, A., and Maiti, H. (2003) Materials Chemistry and Physics 81, 33.
50
C. RISTOSCU ET AL.
10. Korotcenkov, G., Brinzari, V., Boris, Y., Ivanov, M., Schwank, J., and Morante, J. (2003) Thin Solid Films 436, 119. 11. Suda, Y., Kawasaki, H., Namba, J., Iwatsuji, K., Doi, K., and Wada, K. (2003) Surface and Coatings Technology 174–175, 1293. 12. Shatokhin, A., Putilin, F., Safonova, O., Rumyantseva, M., and Gas’kov, A. (2002) Inorganic Materials 38(4), 374. Translated from (2002) Neorganicheskie Materialy 38(4), 462–467. 13. Pererira, A., Cultrera, L., Dima, A., Susu, M., Perrone, A., Du, H., Volkov, A., Cutting, R., and Datta, P. (2006) Thin Solid Films, 497(1–2), 142. 14. Leite, E., Bernardi, M., Longo, E., Varela, J., and Paskocimas, D. (2004) Thin Solid Films 449, 67. 15. Niranjan, R., and Mulla, I. (2003) Materials and Engineering B 103, 103. 16. Rumyantseva, M., Safonova, O., Boulova, M., Ryabova, L., and Gas´kov, A. (2003) Russian Chemical Bulletin International Edition 52(6), 1217 17. Lowndes, D., Geohegan, D., Puretzky, A., Norton, D., and Rouleau, C. (1996) Science 273, 898 18. Chrisey, D., and Hubler G. (Eds.) (1994) Pulsed Laser Deposition of Thin Films, Wiley, New York. 19. Marine, W., Patrone, L., Luk’yanchuk, B., and Sentis, M. (2000) Applied Surface Science 154–155, 345. 20. György, E., Santiso, J., Figueras, A., Giannoudakos, A., Kompitsas, M., and Mihailescu, I. (2005) Journal of Applied Physics 98, 1. 21. Mazingue, T., Escoubas, L., Spalluto, L., Flory, F., Socol, G., Ristoscu, C., Axente, E., Grigorescu, S., Mihailescu, I., and Vainos, N. (2005) Journal of Applied Physics 98(7), 074312 22. Socol, G., Axente, E., Ristoscu, C., Sima, F., Popescu, A., Stefan, N., Mihailescu, I., Escoubas, L., Ferreira, J., Szekeres, A., and Bakalova, S. (2007) Journal of Applied Physics 102, 083103-1-6. 23. Mazingue, T., Escoubas, L., Flory, F., Jacquouton, P., Perrone, A., Kaminska, E., Piotrowska, A., Mihailescu, I., and Atanasov, P. (2006) Applied Optics 45, 1425. 24. Gyorgy, E., Socol, G., Axente, E., Mihailescu, I., Ducu, C., and Ciuca, S. (2005) Applied Surface Science 247, 429–433. 25. György, E., Socol, G., Mihailescu, I., Ducu, C., and Ciuca, S. (2005) Journal of Applied Physics 97, 093527-1_4 26. Golovanov, V., Korotcenkov, G., Brinzari, V., Cornet, A., Morante, J., Arbiol, J., and Rossyniol, E. (2002) CO–water interaction with SnO2 gas sensors: role of orientation effects, in: Proceeding of the 16th International Conference on Transducers, EUROSENSORS-XVI, Prague, Czech Republic, 926–929 (CD). 27. ASTM 21-1250. 28. Hartman P. (Ed.) (1993) Crystal Growth – An Introduction, North-Holland, Amsterdam, The Netherlands. 29. Brinzari, V., Korotcenkov, G., Golovanov, V., Schwank, J., Lantto, V., and Saukko, S. (2002) Thin Solid Films 408(1/2), 51. 30. Korotcenkov, G., Macsanov, V., Tolstoy, V., Brinzari, V., Schwank, J., and Faglia, G. (2003) Sensors and Actuators B 96(3), 602. 31. Korotcenkov, G., Brinzari, V., Cerneavschi, A., Ivanov, M., Golovanov, V., Cornet, A., Morante, J., Cabot, A., and Arbiol, J. (2004) Thin Solid Films 460(1/2), 315. 32. Mazingue, T., Kribich, R., Etienne, P., and Moreau, Y. (2007) Optics Communications 278, 312–316.
X-RAY PHOTOELECTRON SPECTROSCOPY AND TRIBOLOGY STUDIES OF ANNEALED FULLERENE-LIKE WS2 NANOPARTICLES F. KOPNOV1, R. TENNE1*, B. SPÄTH2, W. JÄGERMANN2, H. COHEN3, Y. FELDMAN3, A. ZAK4, A. MOSHKOVICH5, AND L. RAPOPORT5 1 Department of Materials and Interfaces, Weizmann Institute, Rehovot 76100, ISRAEL 2 Fachgebiet Oberflaechenforschung, Fachbereich Materialwissenschaften, Technische Universität Darmstadt, Petersenstrasse 23, 64287 Darmstadt, GERMANY 3 Department of Chemical Research Support, Weizmann Institute, Rehovot 76100, ISRAEL 4 “NanoMaterials” Ltd., Weizmann Science Park, Bldg. 18, 18 Einstein St., P.O. Box 4088, Nes Ziona 74140, ISRAEL 5 Department of Science, Holon Institute of Technology, Golomb St. 52, P.O. B 305, Holon 58102, ISRAEL
Abstract – The temporal chemical changes occurring at the surface of fullerene-like (IF) nanoparticles of WS2 were investigated using X-ray photoelectron spectroscopy (XPS) and compared to those of bulk powder (2H) of the same material. It is possible to follow the long term (surface oxidation and carbonization) occurring at defects on the outermost surface (0001) layer of the fullerene-like nanoparticles. Similar but perhaps more distinctive changes are observed on the prismatic (hk0) surfaces of the 2H powder. Vacuum annealing is shown to remove most of these changes and bring the surface close to its stoichiometric composition. In accordance with previous measurements, further evidence is obtained for the existence of water molecules which are entrapped in the hollow core and interstitial defects of the fullerene-like nanoparticles during the synthesis. They are also shown to be removed by the vacuum annealing process. Chemically resolved electrical measurements (CREM) in the XPS show that the vacuum annealed IF samples become more intrinsic. Finally, ___________ *To whom correspondence should be addressed: R. Tenne, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
51
52
F. KOPNOV ET AL.
tribological measurements show that the vacuum annealed IF samples perform better as an additive to oil than the non-annealed IF samples and the bulk (2H) platelets powder.
Keywords: IF-WS2, nanoparticles, XPS (X-ray photoelectron spectroscopy), tribology, annealing.
1. Introduction The extensive research of the structural, physical and chemical properties of layered transitional metal dichalcogenides, MX2 (M = metal, X = chalcogenide) have been going on for the past 5 decades.1 This is due to their unique structural properties, where MX2 layers are bonded by the weak van der Waals forces through relatively large (several Angstrom) van der Walls gap. The van der Waals gap allows accommodation of foreign atoms and molecules leading to the modifications of the physical properties, such as the electrical conductivity and the magnetic susceptibility. Two of these MX2 materials that have been broadly studied are 2H-MoS2 and 2H-WS2 (2H stands for hexagonal unit cell that consists of two layers); they have been shown to be good solid lubricants2 and catalysts.3 Since it is believed that 2H-WS2 can be a promising material for solar cells, much effort has been paid to the synthesis of WS2 thin films and studying their electronic and transport properties by electrical measurements and X-ray photoelectron spectroscopy (XPS).4 It was found that under special conditions layered transition metal dichalcogenides could be produced as closedcage (fullerene-like) nested nanoparticles (IF) and nanotubes (INT).5 Since the first inorganic fullerenes-like structures and nanotubes of MX2 compounds (M = Mo,W; X = Se,S) were synthesized much work has been done in developing new synthetic routs and elucidating their growth mechanisms.6 Moreover, the structural,7 optical,8 electronic,9 and mechanical10 properties of these materials have been extensively studied. In particular, the IF-WS2 phase revealed intriguing tribological11 performance that offers many potential applications where reduced friction and wear are desired. A 1H NMR study12 of the IF-WS2 nanoparticles showed that the nanoparticles entrap water molecules, or OH moieties, which are produced during the conversion of the oxide nanoparticles into the metal sulfide (IF) and perhaps also some residual hydrogen molecules. The water molecules are believed to be present in defects arising due to folding of the IF-WS2 layers. They may also adsorb in the nanopores between the agglomerated IF-WS2 onions; at the nanoparticles surfaces and occupy voids in the central part (core) of the IF nanoparticle as well. A dedicated
ANNEALED FULLERENE-LIKE WS2 NANOPARTICLES
53
vacuum annealing set-up was assembled in order to extract the remnant water molecules produced during the synthesis or adsorbed onto the nanoparticles surface afterwards. Sequential van der Pauw transport measurements13 of the IF-WS2 pellets showed that the vacuum annealed samples demonstrated higher resistivity than the non-annealed ones. However, apart from a preliminary study14 no systematic X-ray photoelectron spectroscopy (XPS) study of the IF-MX2 structures was carried out, so far. Such a study could reveal the electronic properties of the IF material, i.e. the Fermi level position, the position of the tungsten and sulfur bands, which could also help to elucidate the role of the intercalated water molecules within the IF nanostructures. The following work presents XPS studies conducted on the 2H and IF-WS2 before and after vacuum annealing. Two different types of IF powders were studied, i.e. freshly synthesized powder and another powder produced 3 years before the actual measurements. Our measurements showed that the IF material was indeed a p-type semiconductor, but to a lesser degree than bulk 2H-WS2. The main difference between the old and the fresh powders was that the later exhibited a better stoichiometric ratio. These characteristics support our previous results of transport measurements, i.e. the resistivity of annealed IF samples is higher than that of the pristine ones. Complementarily, the XPS-derived binding energies of W and S in pristine (non-annealed) and annealed IF-WS2 demonstrated systematic shifts, as compared to 2H platelets. Moreover, the most salient chemical change under annealing the IF-WS2 was the loss of oxygen, which was indeed indicative of water degassing. 2. Experimental 2.1. SAMPLE PREPARATION AND HANDLING
The synthesis of the IF-WS2 samples was described previously.14 The 2H-WS2 powder came from a commercial source (Alfa Aesar). The XPS measurements were carried out either on pellets or loose powder. 2.2. XPS
The powder from the 3 years old batch was analyzed by X-ray photoelectron spectroscopy (XPS) measurements using an Escalab 250 (by Thermo-VG) setup at a resolution of 0.3 eV. The protocol used for these measurements was as follows: first, survey spectrum was taken with the following parameters: step size 0.5 eV, pass energy 50 eV, dwell time 50 ms. Subsequently, detailed spectra were taken: step size 0.05 eV; pass energy 10 eV, dwell time 50 ms. The X-ray was running with a power of 150 W at 15 kV with a 500 Pm spot size. All the
F. KOPNOV ET AL.
54
data presented as figures and as series I (pellets) in Table 2 were done with this setup. All spectra were referred to the Fermi level of a freshly sputtered metallic sample (EF = EB = 0eV). Complementary XPS measurements on the freshly prepared samples were performed (series II in Table 1) with Kratos AXIS-HS analytical system using a monochromatized Al KD (1486.6 eV) source at relatively low power (75 W), with a hybrid magnetic and electro-static lens mode and detection pass energy of 20 eV. Control over sample charging was achieved by an electron flood gun (eFG) with typical acceleration of 2–3 eV. The annealed IF samples were analyzed using this setup 2–5 min after exposure to the ambient (see series II (loose powder) in Table 1). TABLE 1. Summary of the results (series II) of the XPS analysis of IF-WS2 (annealed and nonannealed) powders. The numbers (in bold) represent atomic percentage of the constituent elements. The numbers in parenthesis stand for the energy shifts in mV of the corresponding bands: E (annealed)-E (non-annealed) (the certainty of the measurements is §30 mV according to the shifts under electron flux from the eFG).
IF-WS2 pristine IF-WS2 annealed
W(4f)
S(2p)
28.5
57.9
31.1 62.7 (289 meV) (285 meV)
O(1s) C(1s)
8.1
5.5
1.7
4.5
VB (valence band)
(294 meV)
TABLE 2. Summary of the results (series I) of the XPS analysis of 2H-WS2 and IF-WS2 (annealed and non-annealed). The numbers (in bold) represent atomic percentage of the constituent elements. The numbers in the parenthesis stand for the energy shifts in meV of the corresponding bands: E (annealed)-E (non-annealed).
W(4f)
S(2p)
O(1s)
N(1s)
C(1s)
SOx
2H-WS2
17
38
36
4
5
5.7
IF-WS2 pristine IF-WS2 annealed
18
40
31
5
6
4.7
30 (230 ± 50 meV)
61 (240 ± 50 meV)
4
3
2
0
VB (valence band)
(230 ± 50 meV)
ANNEALED FULLERENE-LIKE WS2 NANOPARTICLES
55
2.3. VACUUM ANNEALING
A special setup13 was built for degassing the entrapped gas molecules. The system is evacuated by a turbomolecular pump up to 10–9 Torr (Leybold Turbovac 361). The sample (250–400 mg) is placed in the chamber and heated by a heating element which is controlled by a thermo controller (Eurotherm 2216e). The sample was heated up to 450°C. The usual heating treatment continued for 2 days. The outgoing gases are analyzed by a residual gas analyzer (RGA-Inficon model Transpector 2). 2.4. TRIBOLOGICAL MEASUREMENTS
Friction and wear experiments were performed using a ball-on disk set-up with sliding velocity of 0.4 ms–1 and load of 300 N. A bearing ball with a diameter of 10 mm moved against a steel disk quenched and tempered up to 44–46 HRc. Therefore the harder disk remained intact while the softer ball suffered the wear during the tribological measurements. The surfaces of the ball and the disk were polished (Ra = 0.02 Pm) prior to the experiment. 1 wt % of 2H-WS2 or IF-WS2 powder (pristine and vacuum annealed) were added to a paraffin oil with viscosity of 60 cSt at 20°C. The solid powder was mixed carefully with the oil for 1 h before the test. Few drops of the lubricant were fed to the contact area every minute during the entire duration of the experiment. The friction coefficient and the size of a wear spot on the surface of the ball were measured at a given periods of time after onset of the test. The studied surfaces were analyzed before and after the tribological tests by optical microscopy, electron scanning microscopy (SEM) and Raman spectroscopy. The Raman spectra were measured by a Renishaw micro-Raman microscope 2000 excited with a He/Ne laser (6,328 ǖ). 3. Results and Discussion The most striking observation of the XPS measurements is the shift of the S(2p) and W(4f) bands of the IF samples after annealing. Tables 1 and 2 summarize the results of the XPS analysis. The annealed samples in series II (measured with Kratos AXIS-HS) demonstrated the same band shifts (Table 1) as compared to series I (Table 2) (measured with Escalab 250). The line shifts of the IF powder after annealing are believed to reflect an upwards shift in the Fermi level – EF, i.e. the annealed material becomes less p-doped than the non-annealed one. These results are in line with the previous electrical measurements, where the resistivity of the pristine IF samples was found to be
F. KOPNOV ET AL.
56
higher than that of the 2H-WS2 ones, and furthermore the annealed IF samples had higher resistivity than the non-annealed ones. Another interesting observation is that the annealing process of the 3 years old IF powder led to an improved stoichiometric ratio between S and W (Table 2). The vacuum annealing caused also the disappearance of the SOx moieties. However, the magnitude of the sulfur peaks did not alter suggesting that oxygen atoms had evaporated during the annealing. Additionally, the carbon contamination was drastically reduced during annealing, and the C(1s) peak (before annealing) shifted from the binding energy that fits bonded carbon (probably as WC bonding (283.5 eV)) to the value typical for CH species and graphite (284.4 eV) (after annealing) in both sets of powders (I and II). Furthermore, smaller amounts of the contaminant atoms, like C, O, and N, as well as SOx moiety (less than 1% of the sulfur content ) were found in series II (fresh batch) measurements than in series I (3 years old batch). The larger amounts of SOx moieties and other contaminants found in series I may suggest that the surfaces of the pristine IF powder is not chemically homogeneous, or that some oxidation of the nanoparticle surfaces in air must have happened within 3 years. It should be born in mind, however, that transmission electron microscopy (TEM) and x-ray diffraction analyses (XRD) of the 3-years old sample (series I) did not reveal any noticeable differences compared to the fresh sample (series II). Moreover, the annealing process did not have any influence on the outcome of the TEM and XRD analysis. These observations suggest that the IF nanoparticles do not go through structural changes during the annealing process, which affect their outermost surfaces, only. The tribological measurements revealed that the annealed IF powder demonstrated lower friction coefficient than the non-annealed one (Table 3). The results of the XPS analysis can be interpreted along the following scheme: during the synthesis some oxygen from the oxide core or from the entrapped water molecules can react with sulfur atoms of WS2, consequently carbon atoms that enter the reaction as a contaminant bind to the tungsten atoms. These new chemical moieties provide additional acceptor states. Indeed, Table 2 shows that the quantity of SOx and C at the nanoparticles surface TABLE 3. The results of the tribological tests; each figure represents an average of three trials. Material
Friction coefficient
Wear spot (ȝm)
2H-WS2
0.09–0.11
512–544
IF-WS2 (pristine)
0.09–0.1
480–490
IF-WS2 (vacuum annealed)
0.07–0.08
310–312
ANNEALED FULLERENE-LIKE WS2 NANOPARTICLES
57
decreased nearly by the same amount after vacuum annealing. Vacuum annealing leads to removal of the WC and SOx species. While some of the reactions are restricted to the nanoparticle surface, others may occur throughout the entire volume of nanoparticles (120 nm in diameter), in the central void of a particle or at the defects arising due to folding of the IF-WS2 layers (Figure 1). The products of such a reaction, i.e. CO2, H2O, SO2 are pumped out during the vacuum annealing. Thus the acceptor states disappear and the material becomes more intrinsic and less conductive. Disappearance of the WC peak indicates that chemically a more complex process occurred during the vacuum annealing in the presented case. One of the possible mechanisms is that WC moiety reacts with the extracted water, and that leads to the formation of a volatile tungsten oxihydrate. Such volatile moieties have been documented in the literature.15 In addition, the XPS measurements showed that the surface of the IF nanoparticles underwent oxidation upon exposure to the ambient for a prolong period of time. Since, there are not any structural differences, according to the TEM and XRD analyses, between the (3 years) old and fresh batch of the IF powder, the oxidation can take place at the defects existing within a nanoparticle or at the outermost layer of the nanoparticle (Figure 1 (inset 3)). Noticeably, when the molecular layers fold sharply, a kink is often formed (see Figure 1) leaving a void, which can be filled with water molecules. These separated layers expose prismatic (hk0) faces, which are more prone to a chemical degradation by an oxidation process. It is a well known fact that the
1
2
3
1
3
2
(a)
(b)
Figure 1. (a) TEM image of an individual IF-WS2 nanoparticle shows clearly the hollow core in the centre of the closed nested WS2 layers. (b) Blown up zones 1, 2 and 3 of the nanoparticle exhibiting defects and voids wherever the folding of the layers induces imperfections in the lattice.
58
F. KOPNOV ET AL.
sulfur terminated (basal) (0001) surface of tungsten disulfide layer in the 2H-WS2 crystal is inert towards oxidation. In contrast to that, the prismatic (hk0) faces of the lattice are chemically reactive due to abundance of dangling bonds. The saturation of these dangling bonds with contaminating atoms leads to the appearance of impurity states that impair the performance of these crystals as solar cells and also adversely affects their tribological performance. 6 4. Conclusions XPS investigation of the pristine and vacuum annealed IF-WS2 nanoparticles was undertaken. Furthermore in the course of the measurements 3-years old and freshly prepared powders were studied. It was found that the freshly synthesized pristine powder had less contamination in the form of SOx, C and O moieties and better stoichiometric ratio than the powder of the 3 years old batch. The annealing improved the stoichiometric ratio in the case of the old batch powder and resulted in a reduced level of contamination for both types of powders. In addition, the annealing process led to a shift of the W(4f), S(2p) atomic bands and the valence band towards higher energies. This shift reflects an upward shift of the Fermi energy such that the annealed material becomes more intrinsic than the non-annealed one. The annealing also resulted in a higher energy shift of the carbon (1s) line. The changes in the XPS spectra of the IF powder and the energy shifts of the corresponding moieties indicate that the annealing process resulted in a more stoichiometric surfaces of the nanoparticles. The chemical reactions that occurred resulted in the disappearance of the acceptor states and thus made the material more intrinsic and less conductive.
References 1. Wilson, J., and Yoffe, A. (1969) Transition metal dichalcogenides discussion and interpretation of observed optical, electrical and structural properties, Adv. Phys. 18(73), 193. 2. Watanabe, S., Noshiro, J., and Miyake, S. (2004) Tribological characteristics of WS2/MoS2 solid lubricating multilayer films, Surf. Coat. Technol. 183(2–3), 347–351. 3. Alonso, G., Del Valle, M., Cruz, J., Licea-Claverie, A., Petranovskii, V., and Fuentes, S. (1998) Preparation of MoS2 and WS2 catalysts by in situ decomposition of ammonium thiosalts, Catal. Lett. 52(1–2), 55–61. 4. Gourmelon, E., Lignier, O., Hadouda, H., Couturier, G., Bernede, J., Tedd, J., Pouzet, J., and Salardenne, J. (1997) MS2 (M=W, Mo) Photosensitive thin films for solar cells, Sol. Energy Mater. Sol. Cells 46(2), 115–121.
ANNEALED FULLERENE-LIKE WS2 NANOPARTICLES
59
5. Remskar, M. (2004) Inorganic nanotubes, Adv. Mater. 16(17), 1497–1504. 6. Therese, H., Zink, N., Kolb, U., and Tremel, W. (2006) Synthesis of MoO3 nanostructures and their facile conversion to MoS2 fullerenes and nanotubes, Solid State Sci. 8(10), 1133–1137. 7. Hassanien, A., Tokumoto, M., Mrzel, A., Mihailovic, D., and Kataura, H. (2005) Structural and mechanical properties of MoS2-I-x nanotubes and Mo6SxIy nanowires, Physica E 29 (3–4), 684–688. 8. Loh, k., Zhang, h., Chen, W., and Ji, W. (2006) Templated deposition of MoS2 nanotubules using single source precursor and studies of their optical limiting properties, J. Phys. Chem. B 110(3), 1235–1239. 9. Milosevic, I., Vukovic, T., Damnjanovic, M., and Nikolic, B. (2000) Symmetry based properties of the transition metal dichalcogenide nanotubes, Eur. Phys. J. B 17(4), 707–712. 10. Schwarz, U., Komura, S., and Safran, S. (2000) Deformation and tribology of multi-walled hollow nanoparticles, Europhys. Lett. 50(6), 762–768. 11. Hu, J., Bultman, J., and Zabinski, J. (2004) Inorganic fullerene-like nanoparticles produced by arc discharge in water with potential lubricating ability, Tribol. Lett. 17(3), 543–546. 12. Panich, A., Kopnov, F., and Tenne, R. (2006) Nuclear magnetic resonance study of fullerenelike WS2, J. Nanosci. Nanotechnol. 6(6), 1678–1683. 13. Kopnov, F., Yoffe, A., Leitus, G., and Tenne, R. (2006) Transport properties of fullerene-like WS2 nanoparticles, Phys. Stat. Sol. (b) 243(6), 1229–1240. 14. Feldman, Y., Frey, G., Homyonfer, M., Lyakhovitskaya, V., Margulis, L., Cohen, H., Hodes, G., Hutchison, J., and Tenne, R. (1996) Bulk synthesis of inorganic fullerene-like MS2 (M=Mo, W) from the respective trioxides and the reaction mechanism, J. Am. Chem. Soc. 118(23), 5362–5367. 15. Sarin, V. (1975) Morphological changes occurring during reduction of WO3, J. Mater. Sci. 10(4), 593–598.
THE DEVELOPMENT AND APPLICATION OF UV EXCIMER LAMPS IN NANOFABRICATION I.I. LIAW1 AND I.W. BOYD2* Department of Electronic & Electrical Engineering and London Centre for Nanotechnology, 17–19 Gordon Street, London WC1H 0AH, UK 2 Department of Electronic & Electrical Engineering and London Centre for Nanotechnology, 17–19 Gordon Street, London WC1H 0AH, UK 1
Abstract – Photon-induced processes hold many unique advantages for thin film processing and surface modification. These include low thermal budgets, lack of ionisation, chemical selectivity, and high cleanliness levels resulting in new chemical pathways to facilitate processing at reduced geometries. For such applications, a variety of sources are available including a wide range of lasers and incoherent lamp systems. In this presentation, the principles and properties of ultraviolet (UV) and vacuum ultraviolet (VUV) radiation generated by decaying excimer complexes will be described. Excimer lamps based on this principle provide highly efficient, high intensity, narrow-band radiation at various distinct wavelengths, with no self adsorption, which can, through insightful variations of the geometric configurations of these dielectric-barrier discharges, make large-scale applications possible. For deep UV applications, sources emitting at wavelengths as low as 126 nm have been developed. The high photon energy levels generated from these sources have been demonstrated to directly photodissociate nitrogen molecules, enabling direct nitridation of surfaces. By surveying a selection of publications in the field, a range of applications of applications for these novel sources in the field of nano-fabrication is discussed. These include the photo-deposition of low- and high-dielectric constant layers, low-temperature oxidation of Si, SiGe and Ge, photo-etching and micro-structuring of polymer surfaces, photo-induced metallization, cleaning of surfaces and
______ *
To whom correspondence should be addressed: Ian Boyd, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
61
62
I.I. LIAW AND I.W. BOYD
UV-curing. From these examples, these relatively low cost lamp systems are demonstrated to be capable of providing low temperature alternatives for largescale materials processing in a wider range of nano-scale applications.
Keywords: Excimer lamps, dielectric barrier discharges, photo-chemistry, ultraviolet, low temperature.
1. Introduction Laboratory UV lasers have widely demonstrated the use of monochromatic photons in the ultraviolet spectral range to selectively initiate photochemical reactions, to dissociate gases or liquids, and on substrate surfaces. For industrial deployment, the need for wavelength selective large throughput systems at low cost is essential. Relatively cheap lamps that possess an extended operating lifetime and are easily fabricated in a variety of geometrical shapes are an intuitive choice. The high energy photons (5–12 eV) generated from these dielectric barrier discharge lamps can readily initiate a range of biological, chemical and physical processes. Initial applications to the generation of ozone for water disinfection, have now extended to photochemical degradation and synthesis, polymerization, and materials deposition at the boundaries of microand nano-science. These sources are also finding possible applications in the microtechnology industry and with the reduction in device geometries towards the nanoscale, their ability to induce specific changes in nanostructures will further highlight their potential towards new generations of nano-fabrication processes requiring low-temperature multi-layered materials. 2. Excimer Radiation from Gas Discharges Dense rare gases at pressures around or above atmospheric pressure have a special property that enables the efficient conversion of electron kinetic energy to electronic excitation energy initially stored in excited atomic and ionic states. In this pressure regime, this excitation is funneled rapidly to a few low-lying atomic and excimer energy levels.1 Excimers (excited dimmers or trimers) are unstable excited molecular complexes, which under normal conditions do not possess a stable ground state. Under the influence of short-pulsed particle bombardment these may disintegrate giving off their binding energy in the form of UV or VUV radiation. As an example, the major reactions for the formation of Xe2* excimers are:
NANOFABRICATION BY UV EXCIMER LAMPS
3
Xe*
Xe*
and
3
P1 , 3 P0 2 Xe o Xe2*
P1 , 3 P2 2 Xe o Xe2*
1
63
¦u Xe
(1)
(2)
1
¦ u Xe
Using Figure 1 to illustrate this further, Xe2* excimer formation only occurs at pressures above 0.1 bar. A three body reaction converts the excited Xe atom, Xe* to the exciplex Xe2* whose radiative decay is faster than that of the Xe* itself, which is responsible for a resonance emission at 147 nm (see Figure 2). The emission bands from the bottom two excimer states dominate and these states decay only by radiation and do not interact with the ground state. The fluorescence radiation of these decaying excimer complexes is normally
Figure 1. Photon energy generated by various excimer sources correlated to common bond dissociation energies (eV).
Figure 2. Scheme showing the generation of VUV line radiation at low pressure and the generation of VUV excimer radiation at 172 nm at about atmospheric pressure. 2
64
I.I. LIAW AND I.W. BOYD
restricted to a narrow band of wavelengths (10–15 nm) with a theoretical efficiency between 45–80%3–7 and particle densities around 1025 m-3. The second excimer continua from pure rare gas dimers are shown in Table 1. Using this principle, simple and efficient excimer lamps may be constructed utilizing different rare gases to induce different UV wavelengths. The use of vacuum ultraviolet Xe2* excimer sources has found widespread application due to the availability of high purity silica tubes from which they are constructed which transmit VUV radiation down to 170 nm with little loss. For shorter wavelengths, special window materials such as LiF, MgF2 or CaF2 must be used. At wavelengths <120 nm no known transparent window materials are available and windowless arrangements have to be deployed.8,9 Emission from rare gas mixtures can spread over a wider spectral region due to the formation of heteronuclear diatomic molecules.10,11 Mixtures of rare-gas halide exciplex molecules of the type RgX* (Rg: He, Ne, Ar, Kr or Xe; X: F, Cl, Br or I)3–7 are used to form bound-free BoX transitions which result in a spontaneous high efficiency VUV/UV emission and even narrower spectral bandwidths, typically 2–4 nm.12,13 A summary of several exciplexes and their respective peak wavelengths is shown in Table 2, while Figure 3 shows a selection of characteristic emission spectra produced by dielectric barrier discharge lamps. TABLE 1. Rare gas dimers with corresponding peak wavelengths in nm (commercially available lamps in thick print). He2* 74
Ne2* 83
Ar2* 126
Kr2* 146
Xe2* 172
XeCI* KrCI* Xe2* Kr2*
Intensity
Ar2*
308
222
172
146
126
Wavelength (nm)
Figure 3. Characteristic emission spectra of dielectric barrier discharge lamps.
NANOFABRICATION BY UV EXCIMER LAMPS
65
TABLE 2. Halogen and rare gas halide excimers with corresponding peak wavelengths in nm (commercially available lamps in bold). F2*
ArCl*
ArF*
KrCl*
KrF*
XeI*
Cl2*
XeBr*
Br2*
XeCl*
I2*
XeF*
157
175
193
222
248
253
259
282
289
308
342
354
2.1. SILENT DISCHARGE DRIVEN EXCIMER LAMPS
Excimer lamp systems (also known as Dielectric barrier discharge (DBD) lamps) based on (non-equilibrium) silent discharges can be operated at pressures up to several bar and enable gas excitation without the need for metal electrodes to be in direct contact with the aggressive plasma. For practical applications, this has proved to be much easier to implement compared to high-energy electron beam generators or pulsed capacitor discharges that also tend to be expensive and operate at low repetition rates. These lamps may be constructed from VUV transparent high purity silica (Suprasil) that can have internal protective coatings applied to resist fluorine attack or functional coating formation that would otherwise reduce the ignition voltage. Since Suprasil is an insulator, it cannot pass dc current, thus alternating voltages are required for their use in DBD’s. The dielectric constant and thickness in combination with the time derivative of the applied voltage determine the amount of displacement current that can pass through the dielectric. Additionally, the electric field must be sufficiently high to initiate gas breakdown. Switched mode power supplies operating between 20 kHz–15 MHz and voltages stepped up using transformers or resonant circuits from a few 100 V to several kilovolts (peak-topeak) provide an efficient and reliable driver of these discharges. Sealed lamps of different planar and cylindrical geometries can be designed as shown in Figure 4. Planar designs use electrodes either on both sides of the discharge space applied by screen printing technologies or embedded within the dielectric on one side of the discharge space (co-planar configuration) resulting in a surface Discharge Gap
Perforated or Mesh Electrodes
UV Transparent Dielectric (Silica)
Cooling Duct Mirror HV AC generator Electrodes
Co-Planar Electrode Configuration 13
Figure 4. Sealed cylindrical and planar dielectric-barrier discharge excimer lamp configurations.
66
I.I. LIAW AND I.W. BOYD
discharge being established between adjacent electrodes. In this way, internal phosphor layers are used to convert VUV excimer radiation to visible light within sealed glass cells, which form the basis of mercury-free fluorescent lamps and of flat plasma display panels. High power lamps often utilise the cylindrical form shown in Figure 5 with a cooling channel through the middle, these are the most common form of excimer UV sources. The width of the discharge gap ranges from 0.1 to several millimeters with filling pressures ranging from 0.1 to 50 bar. A third (buffer) gas such as He or Ne is added to the binary excimer-forming mixture to facilitate ignition and provide additional control over the electron energy distribution. For configurations with discharge gaps of a few millimeters and a gas fill of around 1bar, operating frequencies range from 50 Hz and 14 MHz with applied voltage amplitudes of a few kilovolts being required for ignition. At very high frequencies, the current limitation of the dielectric is reduced and for this reason DBDs are operated between line frequency and about 10 MHz. In the case of cylindrical designs, additional photon flux is usually obtained by using a specially mirror polished aluminum (6061-grade) reflector on the side opposite to the irradiation direction in order to reflect photons being emitted in the opposite direction. Commercial excimer lamps are available at 126 nm (Ar2*), 172 nm (Xe2*), 222 nm (KrCl*) and 308 nm (XeCl*) with typical efficiencies from 5–40%. XeCl* lamps of up to 2 m in length are available for UV curing applications. XeI* and XeBr* lamps radiating at 25315 and 282 nm16 respectively have also been reported. Room temperature operation is achieved through the use of cooling water allowing operation at high electrical input powers.
Figure 5. Cylindrical incoherent excimer UV source. 14
NANOFABRICATION BY UV EXCIMER LAMPS
67
As previously mentioned, open or windowless configurations have been used for lower wavelength systems. A good illustration of this concept is presented by Lenk and Menhert8 and shown in Figure 6. The idea is to initiate the discharge within a large chamber containing the both the excimer-forming gas and as well as the sample to be irradiated.
Figure 6. Windowless Ar2* excimer VUV irradiation system. (After M. Lenk and R. Mehnert.8)
3. Applications of Excimer Lamps 3.1. VUV CLEANING OF SURFACES
The cleaning of glass substrates in LCD manufacturing, of reticules for VUV lithography and in wafer processing is a standard industrial process, involving the removal of residual organic surface contaminants, such as aliphatic and aromatic hydrocarbons, which helps promote adhesion and improve wettability. The photon emission of Xe lamps at 7.25 eV can break most molecular bonds in organic compounds but is not strong enough to split the resulting product molecule CO2 and is used for this purpose. This cleaning process is performed in an Ar/O2 mixture at pressures in the 10–100 mbar range. Ar, which is used as a carrier gas, is transparent to the radiation whilst the O2, which absorbs the UV strongly, splits into the very reactive constituents O(3P) and O(1D). Ozone (O3) results from the recombination of these O atoms with O2 molecules and itself absorbs the radiation to photo-dissociate yielding more O(1D) atoms. The cleavage of organic compounds of the general form CxHyOz produces H2O molecules that also absorb this radiation strongly and following excitation, homolyse with the formation of hydroxyl radicals and hydrogen atoms. The result is a highly reactive gas mix containing O3, O(3P), O(1D) and OH radicals,
68
I.I. LIAW AND I.W. BOYD
which produce a powerful cleaning action on the surface contaminants. The intermediate molecular fragments are transformed into stable end-products of H2, H2O, CO and CO2, which are carried away in the gas stream. This produces a cleaner surface with a wettability contact angle of 5–10q compared with 30– 60q prior to the VUV cleaning process, depending on the nature of the substrate. Present technology uses parallel arrays of linear cylindrical Xe lamps that provide up to 100 mW/cm2 of irradiance. The chemical-free nature of such a cleaning process is advantageous when considering the reduced geometrical dimensions of nano-devices and the challenging juxtaposition of materials within the device matrix. 3.2. PHOTO-INDUCED DEPOSITION OF INSULATING LAYERS OF HIGH OR LOW PERMITTIVITY
Many candidate materials have been suggested to replace the ubiquitous SiO2 or more recent SiOxNy and SiN variations as the gate dielectric in future MOS devices, in order to avoid undesirable quantum tunneling in ever thinning gate structures.17 One of these was Ta2O5 but due to non-optimum band-offset characteristics, others are presently preferred. However, due to its compatibility with ULSI processing and its chemical and thermal stability, there are many other possible applications.18,19 Figure 7 schemes an excimer UV lamp system used for the growth of thin Ta2O5 layers on Si. It comprises two stainless steel vacuum chambers (lamp and reactor sections) separated by an MgF2 window, which is transparent to the VUV irradiation generated by the lamp source. Cylindrical excimer UV lamps of Xe2* (O = 172 nm) or KrCl* (O = 222 nm) were used to irradiate <100> orientation c-Si samples (n-type, 2–4 : cm resistivity), which were placed atop a heated (100–400qC) substrate stand, through a low pressure gas mixture with an output intensity of about 30 mW/cm2.20 The gas used was formed by first externally vaporizing a Ta Metalorganic precursor, which was then introduced into the chamber using an Ar carrier gas, and mixed with N2O at a fixed flow rate. A more complete description of this reactor is published elsewhere.21 The basic photochemistry inducing the CVD growth of Ta2O5 from tantalum ethoxide (Ta(OC2H5)5) precursor and N2O follows the reaction:
N 2O hQ o O1D N 2 1 ¦ g
(3)
where the resultant O(1D) oxygen species reacts with the tantalum ethoxide causing its dissociation through a series of reactions leading to deposition. Figure 8 shows the Arrhenius plots of the deposition rate of photolytic versus
NANOFABRICATION BY UV EXCIMER LAMPS
69
Figure 7. Schematic diagram of photo-CVD system incorporating excimer UV lamps.
Figure 8. Arrhenius plot of the growth rate of Ta2O5 deposited by photo- and thermal-CVD on Si.
pyrolytic deposited Ta2O5. Clearly the growth rate of the thermally deposited films strongly depends on substrate temperature. It is also particularly slow and essentially negligible under 400qC with a high activation energy of 1.97 eV. The photolytic film growth is only slightly dependent upon temperature and exhibits a significantly lower activation energy of 0.078 eV, indicating a quite different set of reaction mechanisms. This reduced activation energy enables substrate temperatures as low as 100qC to be used. Following deposition, a subsequent UV annealing step has been shown to lead to the reduction of leakage currents in the as-deposited Ta2O5 layers.
70
I.I. LIAW AND I.W. BOYD
The proposed mechanism for the UV annealing effect is attributed to the active oxygen species (O(3P) and O(1D) and O3, as discussed in (Section 3.1), which are adsorbed onto the Ta2O5 surface, diffuse, accept electrons, and fill any voids, as shown schematically in Figure 9. The subsequent decrease in the number of voids leads to a reduction of leakage current. However, these active oxygen species can also react with Si either at the Ta2O5/Si interface or with any Si species that may be present at the surface, leading to the additional formation of a thin Si oxide layer. The likely contributions to the reduced leakage current provided by the UV annealing step are summarized as: x
Active oxygen species react with, and reduce or remove any suboxides present leading to improved stoichiometry.
x
The active oxygen species decrease the density of defects and oxygen vacancies.
x
A native oxide layer formed at the Ta2O5 and Si interface and/or on the Ta2O5 surface by the reaction between the active oxygen species and Si lead to improved interfacial quality.22
x
Densification of the deposited layers and removal of any impurities present in the as-deposited films.
x
Improvement of surface morphology and reduction of the as-deposited particle size.
Figure 9. Model of the mechanism for reduction of leakage current of Ta2O5 films by UV annealing.
NANOFABRICATION BY UV EXCIMER LAMPS
71
Figure 10 shows the (a) I-V and (b) C-V characteristics of Al/Ta2O5/Si MOS capacitors fabricated using the deposited films. The I-V plot shows that following UV annealing the leakage current is dramatically reduced (up to three orders of magnitude). Table 3 summarises this for various annealing times. The C-V plot of Figure 10b shows that the fixed oxide charge density also decreases with increased annealing times. The flat-band voltages (VFB) are negative indicating the presence of positive fixed charges near the Ta2O5/Si interface. Without annealing the VFB was –1.8 V. This reduced to –0.8 and –0.1 V after annealing for 0.5 and 1 h respectively indicating a reduction in the C-V hysteresis effect and positive surface charge. The dielectric constant of the films was determined from C-V measurements using:
C
H0 kA
(4)
d
10 3 10 4
1.2 250˚C
10 5
350˚C
10 6
400˚C
10 7
0.8 0.6 0.4
8
10
0
0h 0.5 h 1h
1
C /Co
Leakage current density (A/cm)2
where C is the measured capacitance, H0 the electric constant, k the dielectric constant, d the thickness and A the area. The dielectric constant value was found to be 24 (as-deposited).
1.2 0.8 0.4 1.6 Electric field (MV/cm)
2
0.2
8
6
4
2 0 2 Voltage (V)
4
6
8
Figure 10. (a) I-V and (b) C-V characteristics of MOS capacitors fabricated from the deposited Ta2O5 films.
TABLE 3. Comparison of electrical properties of photo-CVD Ta2O5 films at various annealing times. Annealing time (h) 0 0.5 1
Leakage current density at 1MV/cm (A/cm2)
VFB(V)
2.19 u 10– 5 2.55 u 10– 6 1.63 u 10– 8
– 1.8 – 0.8 – 0.1
Fixed charge density (cm–2) 2.9 u 1011 2.3 u 1011 3.7 u 1010
72
I.I. LIAW AND I.W. BOYD
Photolytic processing induced at low temperatures minimizes problems with atom diffusion, dopant redistribution and defect generation otherwise caused at higher temperatures. Further, the surface is not subjected to the damaging ionic bombardment that can occur in other physical deposition methods. The ability of UV light to induce specific gas phase reactions (e.g. the generation active oxygen species) on or at the surface considerably reduces any temperature dependence and allows deposition to occur at much lower temperatures. Reduced device dimensions make the performance at the metal interconnect level crucial. Propagation delays, cross talk, noise and power dissipation at this level becomes significant due to resistance-capacitance (RC) coupling. This delay may be reduced through the incorporation of low permittivity dielectric materials and /or high conductivity metals. Polymeric films show great promise for this purpose with their low dielectric constant and the ease with which they may be applied through patterning and their high thermal stability.23,24 UV photons can play a significant role in the curing of polyamic acid films with results demonstrating the complete transformation of the acid layer to polyimide at temperatures as low as 150qC. The UV processing at lower temperatures and shorter times compared to those for conventional furnace methods resulted in leakage currents an order of magnitude smaller.25,26 Dielectric constants between 1.7 and 3.6 were readily obtained and have clear potential applications as interlayer dielectrics for advanced ULSI devices. 3.3. LOW TEMPERATURE UV ASSISTED OXIDATION OF SI, GE AND SIGE
Continuous downscaling of linear devices in recent years has imposed a severe limitation on the thermal budget associated with the several process steps employed in fabrication. With the approaching nano-scale levels requiring even higher degrees of integration, low temperature alternatives to traditional thermal oxidation methods around 1,000qC are greatly desirable. Excimer lamp radiation can considerably enhance oxidation rates at temperatures up to half that value. Figure 11 shows the oxide thickness of oxides grown at 5 mbar as a function of exposure time compared to previously published data27 for layers grown using a low pressure mercury lamp, visible radiation, ozone and conventional furnace oxidation. The observed oxidation rate using the excimer lamp is by far the highest being about 90 times greater than thermal oxidation at 612qC. Thermal oxidation at 450qC is shown to be negligible even after oxidation times in excess of 5 h. When compared to the same process using the low pressure Hg lamp at 350qC, the excimer lamp is three times more effective. We already know that the 172 nm radiation easily dissociates O2 and that the atoms released can produce ozone which can also be decomposed by the
NANOFABRICATION BY UV EXCIMER LAMPS
73
Figure 11. Oxide thickness grown as a function of time for different photon, particle and thermal sources.
VUV light and that all these species can subsequently react with the Si surface forming SiO2. The photo-oxidation process may be explained by: (1) transport of the oxygen species to the oxide surface, (2) diffusion through the oxide, and (3) reaction at the Si-SiO2 interface. With the overall reaction rate being governed by the slowest reaction. Simple MOS capacitors with 11 nm thick UV-grown oxide layers with an evaporated Al contact of area 3.2 u 10-3 cm2 have been fabricated. Fixed charge densities of the as-grown films were shown to be 2.4 u 1011 cm-2, which are nowhere comparable to the best high temperature thermal oxide films grown in strict clean room conditions. Rapid thermal oxides grown at 950qC also contain high fixed oxide charge density in the 1011cm–2 range.28 However, an additional UV annealing step (up to 2 h in 1 atm of O2 at 400qC) significantly reduced the fixed charge density to 4.5 u 1010 cm-2, which is comparable to that for thermally grown SiO2 at 1,030qC. The stretching peak of Si-O-Si observed from FT-IR spectroscopy was also seen to move to higher wave numbers following annealing, indicating a possible increase in film density.29,30 Direct photo-oxidation of Ge at temperatures d 400qC using a Xe2* lamp shows oxidation rates of 0.1 nm/min which are significantly faster than thermal oxidation methods. According to FTIR and XPS analysis, the formed layers have an identical composition to that of thermally grown GeO2.33 SiGe layers were also shown to oxidise rapidly using 172 nm radiation in comparison to the thermal reaction. An enhanced interdiffusion of Si and Ge atoms and low temperature formation of Ge nanoclusters were observed.34,35 Through photoluminescence measurements, an increase in the size of these nanoparticles resulting from the extended UV oxidation process is observed.36 More recently, the use
74
I.I. LIAW AND I.W. BOYD
of deeper UV sources at 126 nm has been demonstrated to achieve higher deposition rates with up to 8 nm/min being achieved. Reduced leakage current densities as low as 5 u 10-8A/cm2 for 17 nm thin layers oxide at a field of 1 MV/cm were observed.31,32 3.4. PHOTO-INDUCED METAL DEPOSITION
The first report of structured deposition of metallic thin films by VUV driven CVD was reported by Calloway et al.37 Utilising a microwave discharge excited rare gas excimer continuum, trimethylaluminum (TMA) vapor was photodissociated resulting in the deposition of aluminum thin films. Since then, structured metal films obtained through the photo-decomposition of spin-on metallo-organic and organometallic precursors has been demonstrated. The nanoscale application here is for high density multichip interconnects packaged to high densities and multilayer stacks. Much of the concept of nanoscaled interconnects may be inferred from the low-temperature decomposition of palladium acetate films.38,39 A flow diagram of the process steps followed is shown in Figure 12.
precursor
1.deposition
substrate
UV 2.exposure
mask activator
3.stripping
4.electroless metal deposition
Cu,Ni,Cr,Ag,Au
Figure 12. UV-induced metal deposition process.
Electroless plating of metal films has the potential for low-cost, highvolume, high-selectivity and low thermal budget metallization suitable for large area nano-scaled processes. It is well known that the activity of catalysts for electroless metal deposition depends on the thickness, distribution and purity of the metal nuclei on the non-catalytic surface. The precursor material can be synthesized to incorporate a wide variety of metals. The different excimer lamp
NANOFABRICATION BY UV EXCIMER LAMPS
75
parameters (wavelength, UV intensity, exposure time), may be correlated to the different absorption coefficients of wavelengths of UV radiation used. The quality of the resulting coating based on the adhesion of the metal layers on the substrate depends decisively on the activator. By exposing through metallic contact masks, film patterning may also be achieved. This method of deposition offers several advantages over other metallization routes. Other than its low temperature, it is insensitive to substrate properties, and providing large-area and low cost layers. 4. Summary We have shown the simplicity, efficiency and ease of deployment of UV excimer lamps for the production of powerful narrow band incoherent radiation. These sealed sources have an extensive range of applications, more than we have presented here, but great potential for relatively cheap large area processes. Progress in the development of shorter wavelength sources or pulsed excimer lamps could lead to even more efficient and wide-ranging uses.40–42
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
12. 13. 14. 15. 16.
Lorentz, D. C. (1976) Phys. C 82 19. Kogelschatz, U. (2002) Plasma Src. Sci. Technol. 11, A1–A6. Rhodes, Ch. K., Ed., Excimer Lasers (New York: Springer, 1979 and 1984). Eliasson, B., and U. Kogelschatz, U. (1998) Appl. Phys. B 46, 229. Ivanov, V. V., Klopovskii, K. S., Mankelevich, Yu. A., Rakhimov, A.T., Rakhimova, T.V., Rulev, G. B., and Saenko, V. B. (1996) Laser Phys. 6, 564. Oda, A., Sugarawa, H., Sakai Y., and Akashi H. (2000). J. Phys. D: Appl. Phys. 33, 1507. Carman, R., Ward, B., and Mildren, R. (2001) Proceedings of the XXV International Conference on Phenomena in Ionized Gases (ICIPG-XXV), Nagoya, 4, 331–332. Lenk, M., and Mehnert, R. (2001) Proceedings of the RadTech Europe, Basle, 153–158. Lomaev, M., Skakun, V., Tarasenko, V., Shitts, D., and Lisenko, A. (2006) Tech. Phys. Letts. 32, 590–592. Kubodera, S., Honda, M., Kitahara, M., Kawanaka, J., Sasaki, W., and Kurosawa, K. (1995) Jpn. J. Appl. Phys. 34, L618. Gerasimov, G., Volkova, A., Zvereva, G., and Babucke G. (Eds.) (1998) In: Proceedings of the 8th International Symposium on Science Technology Light sources (LS-8), Griefswald, Germany, 248–249. Shuaibov, A., Shimon, L., and Shervera, I. (1998) Instr. Exp. Technol. 41, 427. Shuaibov, A., Shimon, L., Dashchenko, A., and Shervera, I. (2001) Tech. Phys. 46, 207. Kogelschatz, U., Esrom, H., Zhang, J., and Boyd, I. (2000) Appl. Surf. Sci. 168, 29. Zhang, J., and Boyd, I. (1996) J. Appl. Phys. 80, 663. Zhang, J., and Boyd, I. (1996) J. Appl. Phys. 84, 1174.
76
I.I. LIAW AND I.W. BOYD
17. Mueller, D., Sorsch, T., Moccio, S., Baumann, F., Evans-Lutterodt, K., and Timp, G. (1999) Nature 399, 758. 18. Zhang, J., Lim, B., Dusastre, V., and Boyd, I. (1998) Appl. Phys. Lett. 73, 2299. 19. Chaneliere, C., Autran, J., Balland, B., and Devine, R. (1998) Mater. Sci. Eng. Rep, 22(6), 269–332. 20. Zhang, J., Lim, B., and Boyd, I. (1998) Thin Solid Films 336, 340. 21. Zhang, J., Boyd, I., Mooney, M., Hurley, P., O’Sullivan, B., Beechinor, T., Kelly, P., Crean, G., and Senateur, J. (1999) Mat. Res. Soc. Proc. 567, 397. 22. Gellert, B., and Kogelschatz, U. (1991) Appl. Phys. B 52, 14. 23. Singer, P. (1994) Semiconductor Int’l Oct, 34. 24. Murarka, S. (1996) Solid State Technol. 39, 83. 25. Zhang, J., and Boyd, I. (1998) Opt. Mater. 9, 251. 26. Zhang, J., and Boyd, I. (1999) Appl. Surf. Sci. 24, 352. 27. Boyd, I., Craciun, V., and Kazor, A. (1993) Jpn. J. Appl. Phys. 32, 6141. 28. Eftekhari, G. (1993) J. Electochem. Soc. 140, 787. 29. Pliskin, W. (1968) Thin Solid Films 2, 1. 30. Boyd, I., and Wilson, J. (1987) J. Appl. Phys. 62, 3195. 31. Zhang, J., and Boyd, I. (2002) Appl. Surf. Sci. 186, 64. 32. Fang, Q., Zhang, J., and Boyd, I. (2003) Appl. Surf. Sci., 369, 208–209. 33. Craciun, V., Hutton, B., Williams, D., and Boyd, I. (1998) Electron. Lett. 34, 71. 34. Craciun, V., Boyd, I., Reader, A., Kersten, W., Hakkens, F., Oosting, P., and Vandenhouldt, D. (1994) J. Appl. Phys. 75, 1972. 35. Craciun, V., Boyd, I., Hutton, B., and Williams, D. (1999) Appl. Phys. Lett. 73, 1261. 36. Zhang, J., Fang, Q., Kenyon, A., and Boyd, I. (2003) Appl. Surf. Sci. 364, 208–209. 37. Calloway. A., Galantowicz, T., and Ferner, W. (1983) J. Vac. Sci. Technol. A1, 534. 38. Boyd, I., and Zhang, J. (1997) Nucl. Inst. Methods Phys. B 121, 349. 39. Zhang, J. (1993) Thesis, Karlsruhe University Germany. 40. Hilbert, C., Gaurand, I., Motret, O., and Pouvesle, J. (1999) J. Appl. Phys. 85, 7070. 41. Motret, O., Hibert, C., Pellerin, S., and Pouvesle, J. (2000) J. Phys. D: Appl. Phys. 33, 1483. 42. Okazaki, K., and Nozaki, T. (2002) Pure Appl. Chem. 74, 447–452.
FUNCTIONALIZATION OF SEMICONDUCTOR NANOPARTICLES M.-I. BARATON* SPCTS-UMR CNRS 6638, University of Limoges 123 Avenue Albert Thomas, 87060-Limoges, FRANCE
Abstract – Functionalization of nanoparticles surface by attachment of organic entities is used to achieve and tailor many new properties, such as lubrication, optical response, chemical sensing, or biocompatibility. But because at the nanometer scale the surface properties significantly contribute to the overall properties, the consequences of the surface modifications have to be thoroughly evaluated. This paper demonstrates the relevance of Fourier transform infrared spectroscopy to the study of the surface reactions leading to the functionalization, and of the stability of the functionalized surface under the expected working conditions. In the case of semiconductor nanoparticles, this technique additionally allows the analysis of the impact of the functionalization on the electrical properties. This will be illustrated by the case study of tin oxide nanoparticles for chemical gas sensors. The correlation between surface chemistry and electrical properties is critical to optimize the nanoparticles functionalization for the targeted properties.
Keywords: Surface chemistry, surface chemical species, FTIR spectroscopy, gas-surface interactions, semiconductor metal oxides, chemical gas sensors.
1. Introduction An important characteristic of nanoparticles is their enormous specific surface area, that is their very large surface to bulk ratio. Obvious consequences are (i) a high surface reactivity of nanoparticles, leading to fast contamination by any surrounding atmosphere, and (ii) a significant contribution of specific surface
______ *
To whom correspondence should be addressed: M. Baraton, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
77
78
M.-I. BARATON
properties to the overall properties. Therefore, at the nanometer scale, material properties reflect the nature of unavoidable adsorbates and it becomes extremely difficult to define intrinsic properties independent of environmental factors. Attachment of organic entities to a semiconductor surface has been investigated for years in order to achieve and tailor many new properties, including lubrication, optical response, chemical sensing, or biocompatibility. Surface protection or passivation is essential for high-added value applications, and surface phenomena have always been a cornerstone of the microelectronic industry.1 In the case of semiconductor nanoparticles, it is therefore critical to control the surface chemistry and to carefully monitor the surface functionalization. Indeed, electrical and optical properties of the nanoparticles may be adversely affected by the formation of new surface chemical groups. For example, surface chemistry has been shown to impact optical properties of nanoparticles, and computationally it has been predicted that surface chemistry has a great influence on the lowest unoccupied molecular orbital (LUMO).2 The characterization of the surface layers formed during the functionalization remains a challenge, but it is not the only difficulty to overcome. Indeed, when surface functionalization of semiconductor nanoparticles becomes necessary to optimize the fabrication of a device (e.g. homogeneous dispersion of nanoparticles in an organic matrix), to modify the surface chemistry (e.g. control of the surface reactions at the interface), to tailor the electrical or optical properties (e.g. enhancement of gas sensor selectivity), the following steps have to be equally and simultaneously mastered3: 1. Chemical characterization of the first atomic layer of the nanoparticles surface before and after surface functionalization. 2. Selection of the most appropriate molecules to be used for surface functionalization. 3. Determination of the conditions under which the nanoparticles surface should be pretreated to optimize the functionalization. 4. Monitoring of the surface reactions during the functionalization. 5. Stability of the surface functionalization under the final working condi-tions of the functionalized semiconductor. 6. Analysis of the consequences on the surface reactivity and on the elec-trical properties of the semiconducting nanoparticles. The chemical composition of the surface layers on functionalized nanoparticles can be studied using Fourier transform infrared (FTIR) spectroscopy,4–6 X-ray photoelectron spectroscopy (XPS)7,8 or nuclear magnetic resonance (NMR).7,9 Surface layer thicknesses are usually evaluated by transmission electronic microscopy (TEM) but monolayers are fragile to high-energy X-ray and electron beams.10
SEMICONDUCTOR NANOPARTICLES
79
As it will be seen in the following section, the advantage of FTIR spectroscopy over other characterization methods is the possibility to not only identify the surface chemical species on nanoparticles and to investigate their surface chemistry but also to monitor in situ the surface functionalization while simultaneously evaluating the consequences on the electrical properties of semiconductor nanoparticles. Moreover, FTIR spectroscopy is a non-destructive analytical method. 2. FTIR Spectra and Electronic Properties Fundamentals of the theory of infrared spectroscopy can be easily found in many references.11–13 It is however less known that in addition to the standard bulk analysis of inorganic materials, FTIR spectroscopy allows both the characterization of the chemical entities adsorbed on nanoparticles surface and the analysis of the nanoparticle surface reactivity. The detailed surface characterization procedure of nanosized particles can be found elsewhere.4,14–16 A brief summary can be given as follows. The first and obvious step consists in observing the transmission spectrum of the nanoparticles and identifying the absorption bands which cannot be assigned to the bulk modes due to the frequency range where the absorption maxima fall. In general, these absorption bands originate from organic entities such as undissociated precursors or/and contamination by the environment during or after the synthesis. It has to be noted that hydrolysis is the most probable contamination to be found on a material surface resulting in hydroxyl OH surface groups. To check whether identified foreign chemical species are located on the surface or trapped inside the bulk, the sample is desorbed at increasing temperatures (activation). Usually, surface chemical species are either released (desorbed) or transformed by heating under vacuum. It should be noted that after activation, a surface is no longer in equilibrium (activated surface) and will therefore easily react as soon as it is in contact with surrounding molecules. The second step of the surface analysis consists in characterizing the surface reactive sites. This is achieved by adsorbing selected molecules (referred to as probe-molecules) onto the activated surface and by comparing the infrared spectrum of the adsorbed probe-molecule with that of its gas phase.4,17 Once the surface is characterized, it is possible to envisage chemical modifications by reaction of appropriate molecules with the surface reactive sites. An activation of the surface by heating under dynamic vacuum is usually necessary to clear the nanoparticle surface from the various impurities and to free surface reactive sites for further grafting of new molecules.
80
M.-I. BARATON
Even though interatomic vibrations are the most studied features in an infrared absorption spectrum, absorption due to surface electronic phenomena represents an intrinsic part of the infrared spectrum of semiconductors. Indeed, the semiconducting property originates from the mobility of free carriers which are electrons (n-type semiconductors) or holes (p-type semiconductors) and, according to the Drude-Zener theory,18–21 these free carriers contribute to the absorption by the material over the whole infrared range. In 1962, Harrick22 observed the modulation of the free carrier density in a silicon wafer by an external electric field. A broad absorption band was identified as the contribution of the free carriers to the infrared absorption spectrum. Most theories predict a On dependence (O being the infrared wavelength) where n generally equals 2 although variations often occur. The infrared absorption changes due to the free carriers are positive or negative depending on whether carriers are added to or subtracted from the surface. For example, electrons subtracted from the space-charge region of an n-type semiconductor surface lead to a decrease of the IR absorption, whereas electrons subtracted from the space-charge region of a p-type semiconductor surface lead to an increase of the IR absorption22,23 (see examples in Section 3 below). Even though in this pioneered experiment, the free carrier density was modulated by an electric field, modulation can also be obtained by adsorbing molecules on the semiconductor surface.23,24 Practically speaking, a decrease of the free carrier density, that is a decrease of the electrical conductivity, leads to a decrease of the IR absorption, whereas an increase of the free carrier density corresponding to an increase of the electrical conductivity leads to an increase of the IR absorption. Therefore, FTIR spectroscopy allows the in situ monitoring of the reactions taking place at the gas-nanoparticle interface simultaneously with the analysis of variations of the background infrared absorption which are related to the variations of the electrical conductivity. It must be stressed that FTIR spectrometry can combine the investigation of two related phenomena (surface reactions and electrical conductivity) involving different thicknesses of the material. Indeed, the surface reactivity concerns only the first atomic layer whereas the thickness variation of the depletion layer may concern up to several tens of nanometers. 3. Case Study of Tin Oxide Nanoparticles Most of the work dealing with the functionalization of semiconductors refers to silicon. But differently, we present here the example of tin oxide nanoparticles. Tin oxide (n-type semiconductor) is one of the most popular materials for the fabrication of semiconductor chemical sensors because of the large variations
SEMICONDUCTOR NANOPARTICLES
81
of its electrical conductivity occurring under adsorption of reducing gases. Moreover, the use of nanoparticles in the fabrication process of gas sensors by screen-printing technology has been demonstrated to largely increase the sensor sensitivity.25,26 However, an important drawback of such sensors is their crosssensitivity to humidity, precluding their operation in outdoor environment. In earlier work,27 we demonstrated that surface functionalization of the titanium oxide nanoparticles with selected organic molecules may be an asset to reduce the cross-sensitivity to humidity. In particular, the surface modification by grafting hexamethyldisilazane (HMDS) proved to be promising even though the sensor sensitivity may be slightly affected by new surface groups. To optimize the surface functionalization, the influence of HMDS on the tin oxide conductivity has to be studied in more details. The tin oxide nanoparticles used in this study were synthesized by Chemical Vapor Synthesis (CVS) whose advantage is the production of nanoparticles with a narrow size distribution and low degree of agglomeration.28 The produced tin oxide nanoparticles have the cassiterite structure with an average primary particle size of 11 nm and they are substoichiometric in oxygen (unpublished results). In situ FTIR experimental conditions are described in details elsewhere.4 3.1. ACTIVATION AND OXIDATION OF TIN OXIDE NANOPARTICLES
Figure 1 shows the infrared surface spectra of the tin oxide nanoparticles after desorption at different temperatures and subsequent oxidations at each desorption step. Indeed, a heat treatment up to 400°C under dynamic vacuum causes oxygen loss, thus leading to a reduced material close to the metallic state. Due to the high density of free electrons, the sample strongly absorbs the infrared radiation, as explained in the previous section. Therefore, an oxidation has to be performed in order to retrieve the tin oxide material. The observed behavior of tin oxide nanoparticles under desorption and oxidation can be related to the high mobility of oxygen in the material. It should be noted that the experiment corresponding to Figure 1 has been performed in situ and that the spectra have not been translated. We can observe that at 200°C (Figure 1d) and 400°C (Figure 1e), adsorption of oxygen leads to a shift of the spectrum baseline toward lower absorbance, thus indicating a decrease of free carrier concentration. In other words, after oxygen adsorption at high temperatures, the sample is more oxidized than the raw material. This is in agreement with the observed sub-stoichiometry of the as-synthesized tin oxide nanoparticles. The absorption bands in the highest wavenumber region of the spectra (4,000–3,000 cm–1) are due to the stretching vibrations of hydroxyl groups bonded to surface tin atoms in different coordination states.29
82
M.-I. BARATON
Figure 1. FTIR spectra of the tin oxide nanoparticles at room temperature (a), after desorption at room temperature (b), after desorption and subsequent O2 adsorption at 100°C (c), 200°C (d), 400°C and cooled down at room temperature (e).
3.2. ADSORPTION OF HMDS AND HMDSO ON TIN OXIDE NANOPARTICLES
The adsorption of HMDS has been performed at room temperature on the tin oxide nanoparticles previously activated (that is desorbed and oxidized) at 400°C. For comparison purposes, the adsorption of hexamethyldisiloxane (HMDSO) has been studied under similar conditions. HMDS and HMDSO usually react through the surface hydroxyl groups to form trimethylsilyl groups according to the reactions: (CH3)3Si-(NH)-Si(CH3)3 + 2 Sn-OH o 2 Sn-OSi(CH3)3 + NH3 (CH3)3Si-O-Si(CH3)3 + 2 Sn-OH o 2 Sn-OSi(CH3)3 + H2O Both HMDS and HMDSO adsorption experiments were conducted in the same way. The two reactants were adsorbed at room temperature followed by a quick evacuation, thus allowing the elimination of the gas phase and of the physisorbed species. Then, the grafted samples were desorbed at 400°C and, in both cases, the strong reduction of the samples was proved by their complete opacity to the infrared radiation. To oxidize the grafted samples, oxygen was adsorbed at 400°C and the spectra were then recorded after cooling the samples down to room temperature under oxygen. Figures 2 and 3 give the spectra at different steps of the HMDS and HMDSO adsorption experiments, respectively. In these two cases, the experiments have
SEMICONDUCTOR NANOPARTICLES
83
been made in situ and comparison of the absorbance values can be achieved within each of the experiments. In Figure 2, it can be seen that after evacuation at room temperature, the Q(CH3) and G(CH3) absorption bands of the trimethylsilyl groups are still very visible proving that HMDS is linked to the tin oxide surface.3 Moreover, one can note a slight shift of the spectrum baseline towards higher absorbance value, which indicates a higher density of free electrons. After desorption and oxidation at 400°C, the shift of the baseline toward higher absorbance is stronger. Differently from the non-grafted material, the HMDS-grafted tin oxide nanoparticles are not totally re-oxidized once they have been reduced by the thermal desorption, as indicated by the spectrum baseline which does not go back to its original position on the absorbance scale. In addition, the Q(CH3) and G(CH3) absorption bands of the trimethylsilyl groups have almost disappeared whereas a band appears at 3,730 cm–1 assigned to the Q(OH) stretching vibration of newly formed Si-OH silanol groups.3 These silanol groups originate from the oxidation of the trimethylsilyl groups and are bonded to the surface oxygen as proved by the band at 990 cm–1 assigned to the Q(O-Si) vibration in Sn-O-Si groups. Figure 3 depicts the effects of HMDSO adsorption. Formation of trimethylsilyl groups bonded to the tin oxide surface is again observed and these groups resist a desorption at room temperature. The formed H2O molecules, as mentioned in the above reaction, are responsible for a broader and more intense absorption band than in the case of HMDS (centered around 3,200 cm–1). This
Figure 2. Infrared spectra of the tin oxide nanoparticles recorded at room temperature: after desorption and O2 adsorption at 400°C (a); after HMDS adsorption and subsequent evacuation at room temperature (b); after desorption and O2 adsorption at 400°C (c).
84
M.-I. BARATON
Figure 3. Infrared spectra of the tin oxide nanoparticles recorded at room temperature: after desorption and O2 adsorption at 400°C (a); after HMDSO adsorption and subsequent evacuation at room temperature (b); after desorption and O2 adsorption at 400°C (c).
band is assigned to the stretching vibration of H-bonded OH groups. Differently from HMDS, HMDSO adsorption significantly reduces the tin oxide nanoparticles. However, a heat treatment at 400°C followed by an oxidation leads to an almost complete recovery of the oxidation state while silanol groups are formed by oxidation of the trimethylsilyl groups. At the end of the two experiments, the surface spectra of the HMDS-grafted (Figure 2c) and HMDSO-grafted (Figure 3c) tin oxide nanoparticles are similar proving that both functionalizations similarly modify the surface chemistry but, in the case of HMDS, the sample is significantly reduced. The difference in the tin oxide oxidation states at the end of the functionalization experiments can be explained by the different gaseous molecules produced during the grafting reactions. While H2O is able to reduce tin oxide nanoparticles at room temperature without preventing an oxidation at higher temperature, NH3 molecules first coordinate on the surface at room temperature (bands at 3,380 and 1,609 cm-1 respectively corresponding to stretching and bending vibrations of coordinated NH3 species).30 Then, as the desorption temperature is increased, the NH3coordinated species are released from the surface. Because NH3 is a strong reducing gas, the free carrier concentration is drastically and irreversibly affected. It has to be noted that this heat treatment under oxygen (or air) of the modified nanoparticles corresponds to the pretreatment undergone by nanoparticlesbased chemical sensors before operation. Therefore, this heat treatment gives an
SEMICONDUCTOR NANOPARTICLES
85
excellent picture of the evolution and subsequent stability of the functionalization under working conditions of the sensors. In our previous works, it has been proved that surface modification by HMDS reduces the cross-sensitivity to humidity.27 Although the surface chemistry of the tin oxide nanoparticles appears to be identical after modification with HMDSO, further experiments currently are in progress to investigate the consequences of both types of tin oxide modification on the sensitivity of the gas sensors based on these modified nanoparticles. 4. Conclusion and Outlook To summarize, surface functionalization is a powerful method which can simultaneously allow modification of the surface chemistry, tailoring of the electrical properties, and optimization of the fabrication of a device. However, surface functionalization of semiconductor nanoparticles by attaching chemical groups requires a chemical investigation of the surface before, during and after modification, and an analysis of the functionalization consequences on the electrical properties. It is known for example that, in the development of molecular electronics, the electrical properties will depend upon the structural properties of the surface. Depending on the envisaged applications, molecular entities to be attached must be adequately selected, their stability must be checked under the expected working conditions, and post-treatment must be determined to properly tune the final characteristics while retaining the benefit from the functionalization. ACKNOWLEDGEMENT
The author is grateful to Prof. Horst Hahn’s group (FZK, Karlsruhe, Germany) for providing her with tin oxide nanopowder.
References 1. Bent, S. (2002) Surf. Sci. 500, 879–903. 2. Veinot, J. (2006) Chem. Commun. 4160–4168. 3. Baraton, M. (2004) Surface functionalization of semiconducting nanoparticles, in H.S. Nalwa (ed.), Encyclopedia of Nanoscience and Nanotechnology, American Scientific, Stevenson Ranch, CA, Vol. 10, pp. 267–281.
86
M.-I. BARATON
4. Baraton, M. (1999) FT-IR surface spectrometry of nanosized particles, in H.S. Nalwa (ed.), Handbook of Nanostructured Materials and Nanotechnology, Academic, San Diego, CA, pp. 89–153. 5. Buriak, J. (2002) Chem. Rev. 102(5), 1271–1308. 6. Rogozhina, E., Belomoin, G., Smith, A., Abuhassan, L., Barry, N., Akcakir, O., Braun, P., and Nayfeh, M. (2001) Appl. Phys. Lett. 78, 3711–3713. 7. Zou, J., Baldwin, R., Pettigrew, K., and Kauzlarich, S. (2004) Nano Lett. 4, 1181–1186. 8. Legrand, J., Taleb, A., Gota, S., Guittet, M., and Petit, C. (2002) Langmuir 18(10), 4131– 4137. 9. Hua, F., Swihart, M., and Ruckenstein, E. (2005) Langmuir 21(13), 6054–6062. 10. Liao, Y., and Roberts, J. (2006) J. Amer. Chem. Soc. 128(28), 9061–9065. 11. Herzberg, G. (1962) Molecular Spectra and Molecular Structure, Van Nostrand, Princeton, NJ. 12. Wilson, Jr., E., Decius, J., and Cross, P. (1955) Molecular Vibrations. The Theory of Infrared and Raman Vibrational Spectra, Dover, NewYork. 13. Griffiths, P., and de Haseth, J. (1986) Fourier Transform Infrared Spectrometry, Wiley, Chichester. 14. Hair, M. (1967) Infrared Spectroscopy in Surface Chemistry, Marcel Dekker, New York. 15. Boehm, H., and Knözinger, H. (1983) Nature and estimation of functional groups on solid surfaces, in J.R. Anderson and M. Boudart (eds.), Catalysis, Springer, Berlin, Vol. 4, 39–207. 16. Davydov, A. (1984) Infrared Spectroscopy of Adsorbed Species on the Surface of Transition Metal Oxides, Wiley, New York. 17. Lavalley, J. (1996) Catal. Today 27, 377–401. 18. Harrick, N. (1960) J. Phys. Chem. Solids 14, 60–71. 19. Harrick, N. (1963) Ann. N.Y. Acad. Sci. 101, 928–959. 20. Harrick, N. (1967) Internal Reflection Spectroscopy, Interscience, Wiley, New York. 21. Gibson, A. (1958) J. Sci. Instrum. 35, 273–278. 22. Harrick, N. (1962) Phys. Rev. 125(4), 1165–1170. 23. Baraton, M., and Merhari, L. (2005) Synth. React. Inorg., Metal-Org. Nano-Metal Chem. 3, 733–742. 24. Chabal, Y. (1988) Surf. Sci. Rep. 8, 211–357. 25. Baraton, M. and Merhari, L. (2001) Mater. Trans. 42(8), 1616–1622. 26. Williams, G., and Coles, G. (1998) J. Mater. Chem. 8(7), 1657–1664. 27. Baraton, M., and Merhari, L. (2004) J. Europ. Ceram. Soc. 24, 1399–1404. 28. Hahn, H., Winterer, M., Seifried, S., and Srdic, V. (2000) in G.M. Chow, I.A. Ovid’ko, and T. Tsakalakos (eds.), Nanostructured Films and Coatings, NATO Science Series, Kluwer, Dordrecht, Vol. 78, pp. 1–10. 29. Tribout, J., Chancel, F., Baraton, M., Ferkel, H., and Riehemann, W. (1997) in Key Engineering Materials, Euro Ceramics V, Trans Tech, Zurich, Vol. 132–136, pp. 1341–1344. 30. Chancel, F., Tribout, J., and Baraton, M. (1997) in Key Engineering Materials, Euro Ceramics V, Trans Tech Publications, Zurich, Vol. 132–136, pp. 1341–1344.
FLEXOELECTRICITY: A UNIVERSAL SENSORIC MECHANISM IN BIOMEMBRANES AND IN CHEM.-BIOSENSORS A.G. PETROV* Institute of Solid State Physics, Bulgarian Academy of Sciences, 72 Tzarigradsko chaussee, 1784 Sofia, BULGARIA
Abstract – Flexoelectricity provides a reciprocal relationship between electricity and mechanics in membranes, i.e., between membrane curvature and polarization. Experimental evidence of biomembrane flexoelectricity (including direct and converse flexoelectric effect) is reviewed. Biological implications of flexoelectricity in mechanosensitivity, electromotility and hearing is underlined. Flexoelectricity enables membrane structures to function like soft micro- and nano-machines, sensors and actuators, thus giving important input to sensoric applications.
Keywords: Biomembrane, flexoelectricity, mechanosensitivity, biosensors.
1. Introduction This article presents an account on some basic mechanoelectric properties of membranes: flexoelectricity, mechanocapacitance and mechanoconductivity. Their investigation became possible thanks to the BLM model, namely to the fact that a BLM represent a freestanding bilayer that is supported along its periphery only. Experimentally, the way to probe BLM mechanoelectricity is to subject a membrane to a pressure differential, either static or dynamic, or both. BLM models used by us comprise the classic, celebrated Mueller-Rudin-TienWescott discovery of BLM painting a over a hydrophobic hole,1 as well as the tip-dip method where a microsize BLM is patch-clamped inside a hydrophilic borosilicate glass micropipette.2 In the present review we concentrate on the first system.
______ *
To whom the correspondence should be addressed: A. Petrov, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
87
88
A.G. PETROV
Flexoelectricity3 stands for curvature-induced membrane polarization (direct flexoeffect) or voltage-induced membrane curvature (converse flexoeffect), mechanocapacitance is lateral stress-induced capacitance variation, while mechanoconductivity means lateral stress-sensitive ion conduction of a membrane. Stress sensitivity is usually mediated by the presence of ion channels in the lipid bilayer, synthetic or native ones. On the other hand, flexoelectricity and mechanocapacitance can be observed even in unmodified lipid bilayers. All these properties are closely related to some vital function of native membranes, mechanosensitivity and electromotility. They involve two generalized degrees of the membrane, mechanical and electrical ones. Strictly speaking, mechanical degrees of freedom involved are two: curvature change and lateral stretching. Furthermore, in photoactive BLM flexoelectricity becomes light-dependent, i.e. a photoflexoelectricity phenomenon can be demonstrated. This effect involves a further degree of freedom of the membrane, the optical one. In the experiments it is easy to apply controlled amounts of pressure difference, either static or dynamic ones, and measure the current or voltage response of BLM. Quite often, though, these properties are superimposed in the resulting response. In the experimental arrangement of a BLM (and especially in a patch clamped membrane) the two mechanical degrees of freedom are coupled: inducing a curvature of the membrane by its blowing is inevitably related to a lateral stress variation. Nevertheless, since both degrees of freedom are of different symmetry (curvature variation has a polarity, while lateral stretch has not), usually by applying a phase-sensitive analyser it is possible to discriminate between the two contributions to the electric response (see below). 1.1. THEORETICAL REMARKS
The first experimental hints about the generation of a.c. currents by BLMs subjected to oscillating gradients of pressure were obtained in the early 1970s. When a BLM was clamped to certain voltage, these currents were related to mechanoconductivity4 or mechanocapacitance,5 and were found to be 2nd harmonics of the driving pressure’s frequency. Subsequently, we pointed out that the membrane curvature is also a variable in these experiments. Consequently, we related the accidental appearance of “subharmonic” current under zero voltage clamp to the oscillating flexoelectric polarization of the curved membrane3,6: PS = f (c1+c2),
(1)
(c1+c2) is the total membrane curvature, f is flexoelectric coefficient, measured in Coulombs [C] and PS is flexopolarization per unit area.6–8
FLEXOELECTRICITY IN BIOMEMBRANES
89
Phenomenologically, the generation of a “subharmonic noise”, i.e. a 1st harmonic a.c. current by a zero voltage-clamped BLM, follows from the appearance of flexopolarization. Assuming (for simplicity) oscillating spherical curvature and representing it in the form
c1 c 2 ct 2c m sinZ t ,
(2)
Z being angular frequency of oscillations and cm
1 Rm the maximal curvature, we will obtain time-dependent flexopolarization as well. Let us assume that polarization variations instantaneously follow curvature variations3 (for the case of non-zero relaxation time see10): PS t f ct
f 2c m sinZ t ,
(3)
Surface polarization leads to a transmembrane voltage difference (Helmholtz equation, Eq. (4)) that can be measured in an external circuit of two electrodes immersed in the bathing electrolyte on each side of the membrane, connected to a very high impedance electrometer (open circuit conditions, zero current clamp):
U f t
PS
H0
f
H0
2 cm sin Zt
U Z sin Z t .
(4)
The flexoelectric voltage is a first harmonic with respect to curvature oscillation and its amplitude is UZ f H 0 2 cm . When the two electrodes are effectively shorted out via low impedance ammeter (shorted circuit conditions, zero voltage clamp), the displacement current through the meter can be calculated by adopting an equivalent circuit (Figure 1), containing an a.c. voltage generator U f (describing the oscillating flexoelectric voltage according to Eq. (4)) and a capacitor C (where C is the membrane capacitance):
I f t
d (CU f ) dt
dU f . dC Uf C dt dt
(5)
Since PS is by definition the t o t a l membrane polarization due to both permanent and induced dipoles, the relevant membrane capacitance is simply C0 H 0 S0 d , where S0 is the flat membrane area and d is the capacitive thickness of the membrane. When the membrane curvature oscillates, the membrane area will oscillate as well7:
S S 0 ' S sin 2 Z t , where 'S
S0 r 2 cm2 4
(6)
for spherical curvature, with r being the membrane radius. The membrane capacitance will then also oscillate like
90
A.G. PETROV Measuring system
Membrane
RL E out
flexo
Rm $R
Vhold
E
a
in
E
Cm$C In flexo
out
E
in flexo
out flexo
b Figure 1. (a) Equivalent circuit of an oscillating membrane connected to a measuring system; (b) graph of the potential distribution across a planar and a curved membrane. E out and E in are two e.m.f. generators that modulate the surface potentials of the two surfaces of a curved membrane for the case of a positive flexoelectric coefficient, negative surface charge and zero intramembrane field, i.e., zero current clamp. Since the two surfaces’ generators operate in counterphase they can be combined in one flexoelectric generator Uf, as in text. (From Petrov3, with permission from the Publisher.)
C
C0 'C sin 2 Zt .
(7)
The relationship between ' C C0 and ' S S0 is, in general, frequencydependent.9,11,12 With flexoelectricity we are mainly interested in the first harmonic of the response. Capacitance oscillations will contribute to higher harmonics. Therefore, from (4) and (5) we have for the 1st harmonic amplitude of the membrane flexoelectric current IZ f C0 H 0 2cm Z . In this way, measuring U Z and cm , or IZ , C0 and cm , we can determine experimentally the value of the phenomenologically introduced flexocoefficient f in Eq. (1). We have shown3 that several electric multipoles (charges, dipoles, quadrupoles) of membrane components (lipids and proteins) contribute to the flexoelectric coefficient. Experimental elucidation of the importance of different multipoles by their independent variation (especially the partial electric charge per lipid head, E, expressed in percent of proton charge) is of substantial
FLEXOELECTRICITY IN BIOMEMBRANES
91
interest. Another important subject is the influence of surface adsorbed modifiers with electric multipoles that differ from these of the lipid matrix. The variations of the flexoelectric response by such modifiers could then be utilized for sensoric purposes. 2. Flexoelectric Sensors Using Current Response of a BLM
Our first experiments, aimed specifically at registering the flexoelectricity of BLMs from biological lipids (egg yolk lecithin (EYPC),8 bacterial phosphatidylethanolamine (PE)9) were performed in the current registration regime by using a low impedance current-to-voltage converter. The influence of some modifiers of the surface charge and surface dipole, as well as of the membrane conductivity, upon the value of the effect was also studied. 2.1. LIPIDS SURFACE CHARGE AND DIPOLE MODIFIERS
BLMs were formed by the classical method 1 on a 1 mm conical hole in a Teflon cup, immersed in a spectrophotometric glass cuvette, at room temperature. Bacterial phosphatidylethanolamine (PE) ex. E. coli (Koch Light) dissolved in n-decane (20 mg/ml) was used as a membrane-forming solution. As a bathing electrolyte either unbuffered KCl, 10 mM to 1 M (pH 6.2) or buffered KCl with higher or lower pH values was used (KH2PO4 pH 3.0; MES pH 6.0; TRIS pH 8.5; CHES pH 9.3; KH2PO4 pH 10.4). The surface charge of the BLM was varied either by pH variation or by adsorption of the ionic detergent cetyltrimethylammonium bromide (CTAB, Chemapol) added to the electrolyte as a 1 mM water solution. The surface dipole was varied by adsorption of the dipole modifier phloretin (Serva), added as a 10 mM solution in ethyl alcohol. 2.2. METHOD FOR EXCITATION OF MEMBRANE CURVATURE
Membrane curvature oscillations were excited by the method first employed in Ref. 4. The experimental set-up is shown in Figure 2. The Teflon cup with a very small internal volume (0.25 cm2) was closed by a Teflon cap. Two reversible Ag/AgCl electrodes were used for current measurements. One of them was mounted in the cup; another one was immersed in the glass cuvette. Through an opening in the cap connected to a flexible pneumatic pipeline, oscillating air pressure was applied to the electrolyte surface and was transferred in this way to the membrane. An electrically shielded earphone generated
92
A.G. PETROV
Figure 2. Scheme of the experimental set-up for current registration: (1) glass cuvette; (2) Teflon cup; (3) Teflon cap; (4) electrolyte: 10 mM to 1 M KCl, pH 6.2; (5) reversible electrodes, Ag/AgCl; (6) plastic pipeline; (7) ear-phone; (8) function generator, Textronics FG504; (9) currentto-voltage converter, Keithley 427; (10) oscilloscope, C1-78; (11) phase sensitive analyzer, PAR 5204; (12) switch; (13) d.c. voltage source; (14) Boxcar averager, PAR Model 162; (15) X-Y recorder. (From Petrov and Sokolov,9 with permission from the Publisher.)
air pressure. A function generator fed the earphone. The applied frequency was in the range of 10–1,000 Hz. Using a T-pipe, an open branch filled with cotton wool was introduced into the pipeline in order to equilibrate any static pressure difference between the two membrane sides while transferring the dynamic air pressure. 2.3. EVALUATION OF MEMBRANE CURVATURE BY MECHANO-CAPACITANCE
After spreading the membrane, its planarity was adjusted by varying the solution level in the cuvette, and membrane capacitance was measured. The first important point in the study of curvature-electric effects is to evaluate membrane curvature and to assure its constancy at each frequency applied, in order to study the frequency dependence of the effects. To this aim, we made use of the “capacitance microphone effect”.5 Briefly, a small voltage difference 'M (100 mV) was applied to the membrane and the displacement current due to the oscillations of the membrane capacitance was recorded. From the amplitude of the capacitance oscillation the curvature was evaluated. The radius of curvature, Rm 1 cm , was about 2 mm. This is a quite substantial curvature, if compared to the flat membrane radius.
FLEXOELECTRICITY IN BIOMEMBRANES
93
2.4. FLEXOELECTRIC RESPONSE OF BACTERIAL ETHANOLAMINE MEMBRANES
Boxcar-averaged traces of the curvature-electric currents and “capacitance microphone” currents of oscillating membranes demonstrate the doubled frequency of the capacitance response (recorded with 100 mV membrane voltage), the fundamental frequency of the flexoelectric response (recorded with zero membrane voltage) and the phase shift between them. The frequency characteristics of the flexoelectric effect for several membranes in 50 mM KCl are shown in Figure 3. Experimentally, the existence of two frequency regions (corresponding to free and blocked lipid exchange) is evident. The median frequency between them is about 150 Hz. IO (mV) 15
C˚ =2,2 nF PE 50mM KCI pH 6,22 C˚ =2,4 nF
10
5
PE + phloretin 0 0
200
400
600
800
1000 O(Hz)
Figure 3. Frequency dependence of the amplitude of flexoelectric effect. I Q represents the RMS value of the flexoelectric current, voltage-converted with 107 V/A amplification factor. BLM is formed from E. coli PE/n-decane. The electrolyte is unbuffered KCl, 50 mM, pH 6.22. Experimental points: (x) flat BLM capacitance C 0 = 2.4 nF; ({) C 0 = 2.2 nF; u, C 0 = 2.0 nF; (') phloretin-modified BLM: 40 nM phloretin in the bathing electrolyte. (From Petrov and Sokolov,9 with permission from the Publisher.)
In the low frequency region a slight increase of the response from 10 to 50 Hz is followed by a plateau up to 120 Hz and by a marked drop of the response with some membranes around 150 Hz. In the high frequency region a more or less linear growth of the response is observed. Extrapolation of the high-frequency straight line goes approximately to zero. Qualitatively, the shapes of the characteristics correspond to the theory. Phase measurements of capacitance currents from pre-curved membranes were performed with the aim of determining the sign of the flexoelectric coefficient in the two frequency regions (at 70 and 400 Hz). Both capacitance and
94
A.G. PETROV
flexoelectric currents should be in phase when curvature-induced polarization is directed the same way as the transmembrane electric field. For both frequencies this took place when the electric field pointed towards the centre of the membrane curvature. This means that the flexoelectric coefficient was negative in both frequency regions.9 In order to study the influence of the ionic strength, electrolyte solutions of non-buffered KCl 10 mM, 50 mM and 1 M were employed. In the low frequency range the only noticeable influence was upon the median frequency, which seemed to increase with decreasing ionic strength. In the high frequency region a marked correlation between the slope of the frequency characteristic (i.e., the value of the flexocoefficient) and the inverse ionic strength was found. The frequency dependence with 10 mM electrolyte was actually nonlinear, with gradually increasing slope. For comparison, the slope of 1 M curve was the lowest and the generated responses were strictly sinusoidal. The dipolar modifier phloretin is known to decrease the surface potential of lipid head dipoles by about 100 mV because of its oppositely directed dipole moment. With the symmetric addition of 40 Pl of 10 mM ethanol solution to 10 ml bathing membrane electrolyte before BLM formation, its amplitudefrequency characteristic demonstrated a drastic drop of the flexoresponse in the high frequency region and reduction of the overall slope by more than seven times (Figure 3). Unilateral addition to a pre-formed membrane did not produce any definite effect on the first harmonic, but led to the appearance of the second harmonic; this was clearly a capacitance microphone effect from the surface potential difference of the asymmetric membrane. Thus, the ability of this set-up to serve as a flexoelectric sensor for dipolar surface modifiers was demonstrated. 3. Flexoelectric Sensors Using Voltage Response of a BLM
Further experiments on membrane flexoelectricity were performed in the regime of voltage registration (open external circuit) by using a high impedance selective nanovoltmeter. These experiments provided a check of surface charge contribution in the presence of univalent or divalent ions.13 The approach was further elaborated by using stroboscopic interferometry for precise determination the of dynamically excited membrane curvature.14–17 3.1. LIPIDS AND MODIFIERS
BLMs were formed from chromatographically pure lecithin extracted from egg yolk (EYPC). The membrane-forming solution was made by dissolving lecithin in n-decane, in concentration 40 mg/ml. A drop of this solution was applied by
FLEXOELECTRICITY IN BIOMEMBRANES
95
a glass pipette to a 1 mm pretreated orifice in a Teflon plate, partitioning a Teflon chamber into two compartments. In some experiments mixed BLMs of lecithin with negatively charged phosphatidyl serine (PS, Serva, 20 mg/ml chloroform solution) were formed. The necessary amount of PS solution was dried on a piece of glass coverslip and then dipped in the membrane-forming solution EYPC/n-decane until completely dissolved. EYPC BLMs modified by uranyl acetate (UA, Merck) were prepared by adding different amounts of concentrated UA solution in water to the bathing electrolyte. As a bathing solution either unbuffered saline (0.1 M NaCl in distilled water) or a saline with phosphate buffer (pH 6.0) was used. Strong adsorption of UO22 ions on the bilayer surface is expected, thus resulting in a large positive surface charge depending on the UA concentration. The purity of EYPC was controlled by thin layer chromatography on Merck plates. Surface charge was estimated by measurements of the electrophoretic mobility of large multilamellar liposomes prepared after Bangham from EYPC in 0.1 M NaCl. Measurements were performed using Mark II (Rank Brothers) electrophoretic equipment. 3.2. EXPERIMENTAL SET-UP
The scheme of the experimental set-up is shown in Figure 4. A two-compartment Teflon chamber (after Ref.,5 with some modifications) was used. The front compartment was open to the air and had a glass window on its front face, enabling the observation of membrane formation and thinning in reflected light. The rear compartment was closed by a Teflon cap carrying the two reversible Ag/AgCl electrodes and a pipeline transmitting oscillating air pressure generated by a compression loudspeaker. A Teflon plate with a 1 mm cone-shaped orifice was pressed between the two compartments and used as a BLM support. Membrane planarity was controlled by Teflon leveling piston immersed in the front compartment and attached to a micrometer screw. It was also used to deliberately impose a membrane curvature for sign determination of the flexocoefficient. Membrane thinning was followed electrically by the membrane capacitance increase over time. Since the dynamics of membrane oscillations depends on D, the ratio of the torus elasticity to the sum of both BLM and torus elasticities the aim was to study BLMs with possible smaller tori. Such membranes displayed capacitances of more than 2.5 nF; therefore, BLMs of lesser capacitance were discarded. Flexoelectric potentials generated by the oscillating BLMs were registered by a lock-in nanovoltmeter (Figure 4).
96
A.G. PETROV
12 K1.1
11
5
6
7
1 K1.2
4
8
9 10
3
Figure 4. Scheme of the experimental set-up for voltage registration: (1) d.c. voltage source; (2) coupled switches (K1.1 and K1.2); (3) Teflon measuring cell; (4) levelling piston (5) plastic pipeline; (6) amplifier; (7) function generator; (8) selective nanovoltmeter Unitra 237; (9) lock-in nanovoltmeter Unitra 232; (10) oscilloscope, Textronics 2246A; (11) compression loudspeaker; (12) valve. (From Derzhanski et al.,13 with permission from the Publisher.)
3.3. FLEXOELECTRIC RESPONSE OF EGG YOLK LECITHIN MEMBRANES
From electrophoretic measurements of the surface charge of EYPC bilayers yield a negative sign of the surface charge and a rather small value of the electrophoretic charge density. We could estimate a value of partial charge per lipid head E = 0.4% for a freshly opened lecithin ampoule. A partial charge E = 4.6% was found for an older sample, three months post opening. The frequency dependence of the amplitude of the flexoelectric response of pure EYPC membranes is shown in Figure 5. The graph represents the data average over 21 membranes. One can see that above 160 Hz the response is essentially frequency-independent. From the plateau value of the response of each individual membrane we calculated the blocked flexocoefficient. The mean value from the 21 measurements is f = (26.5 r 5.5)10–19 C. The sign of this coefficient turned out to be positive (see below). The data from mixed EYPC + 2 mol% PS BLMs also are given in Figure 6. The points are averaged over four measurements. These measurements were complicated by the shorter lifetime of the BLM in the presence of PS, so a high frequency plateau was not revealed. However, the increased response by a charge modifier at higher frequencies is evident. The data for UA-modified EYPC membranes are shown in Figure 6 for various UA concentrations in the bathing electrolyte, ranging from 1 to 10 mM. In the upper part of this range BLMs were markedly stabilized and measurements could be extended up to higher frequencies. The results with 1 and 3 mM UA differ from those with 5 and 10 mM. In any case, modified BLMs give a higher
FLEXOELECTRICITY IN BIOMEMBRANES
97
UN / MV 2
60 50 40 1
30 20 10 0 40
120
200
280
360
440
N/Hz Figure 5. Frequency dependence of the flexoelectric voltage of BLMs. Curve 1: egg yolk lecithin/n-decane BLM (averaged data from 21 BLMs). Curve 2: egg yolk lecithin + 2% mol phosphatidyl serine/n-decane BLM (averaged data from 4 BLMs). Electrolyte: unbuffered 0.1 M NaCl. (From Derzhanski et al.,13 with permission from the Publisher.)
Figure 6. Frequency dependence of the flexoelectric voltage of BLMs modified by various concentrations of uranyl acetate (UA) in the electrolyte (0.1 M NaCl). Curve 1: no UA (same as curve 1, Figure 5). Curve 2: 1 mM UA. Curve 3: 3 mM UA. Curve 4: 5 mM UA. Curve 5: 10 mM. (From Derzhanski et al.,13 with permission from the Publisher.)
response than unmodified ones for all frequencies. From the plateau values above 300 Hz flexocoefficients f in the range of 12–20010–19 C were calculated, with a negative sign (see below); the minimum value was at 1 mM UA, the maximum at 3 mM. Finally, we describe the results for the sign of the flexocoefficient. Measurements were performed at 300 Hz and ' M = 40 mV (rear electrode positive).
98
A.G. PETROV
With EYPC membranes the comparison of the flexoelectric amplitude to that of the sum of the flexoelectric and capacitance amplitudes revealed an increase at switching ' M > 0 on and a positive (outward) initial curvature, or a decrease at switching ' M > 0 on and a negative (inward) initial curvature. This means that at positive static curvature both currents have the same phase, while at negative static curvature the phases are opposite. Since in the first case ' M ' C ! 0 , while in the second case ' M ' C 0 , from both experiments it follows that ' P ! 0 , i.e., f B ! 0 . This positive sign is opposite to the sign of the surface charge established electrophoretically. Repeating the experiment with the UA-modified BLM where recharging of the surface takes place, we see just the opposite behavior, showing that f B 0 in the whole UA concentration range studied. This negative sign is also opposite the positive surface charge due to the strong UO 22 ion adsorption. Another example of a flexoelectric sensor employing patch-clamped membranes of soy lecithin is provided.21 The effect of heavy metal ions Cd2+ and Hg2+ was investigated. While Cd2+ ions resulted in a biphasic effect, firstly increasing and then decreasing the flexoresponse, the effect of Hg2+ (Figure 7) amounted to a monotonic decrease of the flexoresponse in the millimolar range of concentrations.
Figure 7. Flexoelectric sensor effect of heavy metal ions. Hg2+ concentration dependence of the normalized flexocurrents of 4 soy lecithin membranes in patch pipettes excited at 410 Hz. (From Zheliaskova et al.,19 with permission from the Publisher.)
4. Conclusion
In conclusion, possible applications of flexoelectric BLMs as sensors for ion and dipolar species follows from the great sensitivity of flexoresponse to such adsorbed molecules (e.g. Figure 3). First prototypes of such sensors using BLM
FLEXOELECTRICITY IN BIOMEMBRANES
99
containing lipophilic ions as sensititvity amplifiers18 or patch-clamped bilayers for microscopic usage19 have already been demonstrated. The first observation of the converse flexoeffect in BLMs15,16 permits the potential use of stabilized BLM systems as microtransducers, micro sound wave generators, and microactuators in molecular electronics. Indeed, flexoelectrically induced displacements of a membrane surface that is only nano-meters thick represent (on a molecular scale) a huge, coordinated motion in space of a whole molecular assembly. This effect may, then, find some interesting applications. ACKNOWLEDGEMENTS
This work has been supported by a grant from the National Science Fund of Bulgaria (NP1-03/2004).
References 1. Mueller, P., Rudin, D., Ti Tien, H., and Wescott, W. (1962) Reconstitution of cell membrane structure in vitro and its transformation into an excitable system, Nature 194, 979–980. 2. Coronado, R., and Latorre, R. (1983) Phospholipid bilayers made from monolayers on patchclamp pipettes, Biophys. J. 43, 231–236. 3. Petrov, A. (1999) The Lyotropic State of Matter. Molecular Physics and Living Matter Physics, Gordon & Breach, New York. 4. Passechnik, V., and Sokolov, V. (1973) Permeability change of modified bimolecular phospholipid membranes accompanying periodical expansion, Biofizika 18, 655–660. 5. Ochs, A., and Burton, R. (1974) Electrical response to vibration of a lipid bilayer membrane, Biophys. J. 14, 473–489. 6. Petrov, A. (1975) Flexoelectric model of active transport, in J. Vassileva (ed.), Physical and Chemical Bases of Biological Information Transfer, Plenum, New York, pp. 111–125. 7. Petrov, A., and Derzhanski, A. (1976) On some problems in the theory of elastic and flexoelectric effects in bilayer lipid membranes and biomembranes, J. Phys. Suppl. 37, C3-155–C3-160. 8. Derzhanski, A., Petrov, A., and Pavloff, Y. (1981) Curvature induced conductive and displacement currents through lipid bilayers, J. Phys. Lett. 42, L-119–L-122. 9. Petrov, A., and Sokolov, V. (1986) Curvature-electric effect in black lipid membranes, Eur. Biophys. J. 13, 139–155. 10. Hristova, K., Bivas, I., and Derzhanski, A. (1992) Frequency dependence of the membrane flexoelectric voltage response. Adsorption of multivalent counterions on the surface of curved lipid bilayer, Mol. Cryst. Liq. Cryst. 215, 237–244. 11. Szekely, J., and Morash, B. (1980) The effect of temperature on capacitance changes in an oscillating model membrane, Biochim. Biophys. Acta 599, 73–80. 12. Wobschall, D. (1971) Bilayer membrane elasticity and dynamic response, J. Colloid Interface Sci. 36, 385–396.
100
A.G. PETROV
13. Derzhanski, A., Petrov, A., Todorov, A., and Hristova, K. (1990) Flexoelectricity of lipid bilayers, Liq. Cryst. 7, 439–449. 14. Todorov, A., Petrov, A., Brandt, M., and Fendler, J. (1991) Electrical and real-time stroboscopic interferometric measurements of bilayer lipid membrane flexoelectricity, Langmuir 7, 3127–3137. 15. Todorov, A., Petrov, A., and Fendler, J. (1994) Flexoelectricity of charged and dipolar BLM studied by stroboscopic interferometry, Langmuir 10, 2344–2350. 16. Todorov, A. (1993) Experimental investigations of direct and converse flexoelectric effect in bilayer lipid membranes, Ph.D. thesis, Syracuse University, NewYork. 17. Todorov, A., Petrov, A., and Fendler, J. (1994) First observation of the converse flexoelectric effect in bilayer lipid membranes, J. Phys. Chem. 98, 3076–3079. 18. Sun, K. (1997) Toward molecular mechanoelectric sensors: flexoelectric sensitivity of lipid bilayers to structure, location and orientation of bound amphiphilic ions, J. Phys. Chem. 101, 6327. 19. Zheliaskova, A., Naidenova, S., Marinov, Y., Mellor, I., Usherwood, P., and Petrov, A. (2001) Detection of heavy metal ions (Cd2+ and Hg2+) by their influence on flexoelectricity of patch clamped membranes, C.R. Acad. Bulg. Sci. 51(12), 53–56.
CARBON NANOTUBES: FROM FUNDAMENTAL NANOSCALE OBJECTS TOWARDS FUNCTIONAL NANOCOMPOSITES AND APPLICATIONS W. MASER*, A.M. BENITO, E. MUÑOZ, AND M. TERESA MARTÍNEZ Department of Nanotechnology, Instituto de Carboquímica (C.S.I.C.), C/Miguel Luesma Castán 4,E-50018 Zaragoza, SPAIN
Abstract – In this article we give a general introduction into the field of carbon nanotubes. On one side we describe carbon nanotubes as fundamental nanoscale objects and explain their high application potential. On the other side, we focus on carbon nanotubes as non-homogeneous materials as obtained from different sources. Methods for production, characterization, purification and functionalization and dispersions are presented explaining chances, challenges and limitations. Finally, we deal with the broad field of applications for carbon nanotubes with special emphasis onto high performance carbon nanotube composite materials.
Keywords: Carbon nanotubes, production, characterization, purification, functionalization, dispersions, composites, applications.
1. Introduction Carbon nanotubes1 were first described and brought into context by S. Iijima in 1991. It was quickly realized by the scientific community that his findings had wide ranging implications for science and technology. Driven by early theoretical studies, the very first experiments already confirmed the unique structure-property relationship. Since then, these findings boosted the development of a new and highly exciting field of research. Today, more than 15 years after Iijima’s observation, novel and often highly surprising results on carbon nanotubes are still continuously reported in the literature and in international ______ *
To whom correspondence should be addressed: W. Maser, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
101
102
W. MASER ET AL.
patent databases. The large number of highly promising results, quickly understood as real-world business opportunities, boosted by now the creation of numerous spin-off companies and also contributed to broaden-up core sectors of already well-established companies. With this article we provide the reader with the necessary background knowledge on carbon nanotubes to understand on one hand, their unique properties and opportunities, and on the other hand, the various challenges to be overcome in this broad and fascinating field of research. 2. Carbon Nanotubes: Fundamental Nanoscale Objects 2.1. STRUCTURE
A carbon nanotube (CNT) can be described as a seamlessly rolled-up sheet of graphene, resulting in an open tubular structure composed of carbon atoms arranged in a hexagonal network. The tubular structure is closed by the inclusion of six pentagonal defects into the hexagonal network at each end of the open cylinder to form the corresponding semi-fullerene2 caps. Figure 1 explains the basic structures of carbon nanotubes.
a1 Q
a2
tube axis
armchair A Q C B
zig-zag
A
B
chiral
C Figure 1. The structures of a graphene sheet (A), the family of single-wall carbon nanotubes (B), and a multi-wall carbon nanotube (C).
CARBON NANOTUBES
103
There are three different ways in which a graphene sheet can be rolled up onto itself thus leading to the three basic forms of CNTs, i.e. the family of single-wall carbon nanotubes (SWNTs): (i) Armchair tubes: These are obtained when rolling-up the sheet in such a way that carbon – carbon bounds are perpendicular to the cylinder axis. Looking at the cross-section of such a tube one observes an “armchair” – like structure of the carbon atoms. (ii) Zig-zag tubes: Here, carbon – carbon bounds are parallel to the tube axis and the cross-section results in a “zig-zag”-like pattern of carbon atoms. (iii) Chiral nanotubes: Here, the hexagons of the original graphene sheet are winding-up around tube axis in a helical way letting form the carbon atoms a helical (chiral) pattern along the length of the tube. In any case, the structure of the nanotubes easily can be described by so-called chiral vector C. It is composed of a two-dimensional pair of integer numbers (n, m) corresponding to the numbers of the hexagonal unit cell vectors a1 and a2, respectively, needed to roll-up the graphene sheet onto itself from on point to another one. Therefore, the (n, m) pair directly reveals the diameter as well as the chiral character for every type of tube, e.g. all armchair tubes are characterized by (n, n) indices. Typically, SWNTs have diameters of about 1 nm and lengths of about 1 Pm to several micrometers, according to experimental observations. Often, due to their small diameters, SWNTs like to arrange themselves into bundles which can adapt diameters even up to 100 nm. Additionally, there exists the family of multi-wall carbon nanotubes (MWNTs): They are composed of individual SWNTs concentrically placed inside each other and separated by an interlayer distance of about 0.34 nm slightly larger than in graphite due to curvature. While length and diameter of the most inner tubes are similar to SWNTs, the outer diameters, depending on the number of individual nanotubes easily can reach values up to 20 or 30 nm. Among the MWNTs, double-wall carbon nanotubes (DWNTs) may play an important intermediate role between SWNTs and MWNTs. 2.2. PROPERTIES
Carbon nanotubes are characterized by a very close and unique structure – property relationship. Composed out of only carbon atoms, carbon nanotubes are very light objects with a very low density of around 1.5 g/cm3. Having all atoms at the surface confers CNTs very high specific surface areas of about 1,400 m2/g (outer surface) or even 2,800 m2/g (if inner surface is considered as well), comparable or even better than highly activated carbons. Furthermore, carbon – carbon bounds are one of the strongest in nature. Therefore, very small individual SWNTs may have Young moduli up to 1,800 GPa (100 times stronger than steel), asymptotically reaching values of graphite for very large diameter
104
W. MASER ET AL.
tubes.3 For SWNTs arranged into bundles, due to sliding effects between individual SWNTs, the Young moduli may lower down to values of about 200 GPa. The maximum tensile strength can reach values of up to 30 GPa. Apart of the high stiffness, individual SWNTs show surprisingly high mechanically flexibility.4 Upon bending to very large angles, SWNTs will not break. Instead, they show a reversible buckling behaviour conferring SWNTs an extremely high toughness. Having micrometer dimensions along their axis and nanometersized diameters, CNTs are nano-scale objects with aspect ratios (ratio length to diameter) as high as 1,000. This is an important property whenever it comes to applications where percolation issues play a critical role. On the other side, depending of the point of view, individual SWNTs might be considered as a large molecule or already as a one-dimensional solid state material. The electronic properties5 easily can be derived applying density functional theory for graphene and zone-folding concepts taking into account additionally periodic boundary conditions for the wave function along the nanometer-sized circumference. These boundary conditions reduce the number of allowed states in the two-dimensional Brillouin zone and introduce one-dimensional van-Hove singularities in the density of states. It can be shown that one third of all SWNTs behave as metals, i.e. all (n, m) tubes with n-m = 3i (I = interger), e.g. all armchair tubes, while two thirds of all SWNTs are semiconductors, i.e. all (n, m) tubes with n-m 3i. It is really remarkable that the structural arrangement of carbon atoms in SWNTs so clearly defines the electronic properties and that these are precisely defined for each for each type of nanotube with a specific (n, m) index. In this sense it has been shown that for semi-conducting SWNTs of a given chirality, the energy gap is inversely proportional to the tube diameter.6–8 Additionally, the one-dimensional electronic character of SWNTs, as expressed by the one-dimensional van-Hove singularities in the density of states, provides these nano-scale objects fundamental electronic, optical and vibrational resonance properties of great importance to spectroscopic characterization, e.g. Raman,9 NIR-absorption,10 photoluminescence,11 and scanning tunnelling spectroscopy,12 as well for applications in the fields of quality control and sensors. Furthermore, experimentally it has been found that SWNTs show ballistic transport13 even at room temperature, have current densities14 as high as 1010 A/cm2 (copper and aluminium show values between 107 and 1010 A/cm2), and behave as excellent electron emitters15,16 with low turn-on fields of 1.5–5 V/ȝm at 1 mA/cm2 and low energy spread of 0.25 eV. Finally, it is worthwhile mentioning that CNTs possess an excellent thermal conductivity17 of up to 3,000 W/mK, superior of diamond, one of the best thermal heat conductors. With this unusual close structure-property relationship and a whole bunch of fascinating combined properties, – light, strong, flexible,
CARBON NANOTUBES
105
metal, semiconductor, heat-conductor, electron emitter –, it becomes clear that CNTs bear a great potential for applications in various sectors of technological interest as shown in Figure 2. The fields range from nanoelectronics (CNT based electronic device structures), flexible plastic electronics (photovoltaic and organic light emitting devices), functional composite materials (reinforced structures, tough materials, thermal conductivity, charge dissipation, electromagnetic shielding, intelligent textiles, and coatings, functional adhesives), energy (electrochemical energy storage devices, fuel-cell membranes), nanobio (sensors, drug delivery, cell proliferation, tissue healing), and catalysis (nanocatalyst dispersions). For a more detailed overview on structure and general properties we refer to the book of Dresselhaus.18
Figure 2. An overview of carbon nanotube’s broad application potential.
3. Carbon Nanotubes: Challenging Materials Before profiting from carbon nanotube’s unique properties and using them for various purposes, one has to be aware that carbon nanotubes are not produced as the individual objects described above. Instead, they are obtained as soot materials which are composed of CNTs of all different kinds of characteristics as well as of additional undesired carbonaceous and inorganic byproducts. This is quite challenging since no CNT soot material is like the other and no standard CNT materials exist today. The characteristics of CNT materials largely depends on the production technique being employed: Materials may contain different types of CNTs (SWNTs, MWNTs, DWNTs, …) with different
106
W. MASER ET AL.
Figure 3. Different types of CNT materials: (a) MWNTs from arc, (b) SWNTs from arc/laser, (c) thin MWNT bundles from supported CVD, (d) aligned MWNTs from floating CVD, (e) filled MWNTs from floating CVD, (f) DWNTs from supported CVD, (g) helical tubes from supported CVD.
diameter and chirality distributions, various defect structures, bundle agglomerations, and impurities such as amorphous carbon, graphite particles and attached catalytic nanoparticles. Figure 3 puts in evidence the different CNT materials characteristics. 3.1. PRODUCTION METHODS
To understand better the obtained materials a brief introduction into the most common production methods are given. Basically, one distinguishes between high-temperature and low-temperature processes.19 In high temperature processes, a solid carbon feedstock (e.g. graphite and eventually catalytic particles) is evaporated at temperatures above 3,000ºC by resistive heating or highly concentrated light. In an adequate inert atmosphere carbon species in the gas phase self-assemble to form carbon nanotubes. Electric arc discharge processes20 and laser evaporation methods21,22 are representative for this process. The soot materials obtained are: (i) straight and highly graphitic MWNTs (Øi = 1–3 nm, Øo = 2–35 nm, l = 1 ȝm) with additional polyhedral graphitic nanoparticles but without any catalytic particle (electric arc), (ii) entangled bundles of SWNTs (Øbundle = 20–100 nm, ØSWNT = 1–2 nm, l = 1 ȝm) with amorphous carbon and embedded spherical catalytic nanoparticle, typical Ni, Co, Ni/Y (electric arc, laser evaporation). Low temperature processes are based on the decomposition of a hydrocarbon feedstock at temperatures between 500ºC and 1,000ºC over catalytic particles. Carbon diffuses into the catalytic particle and at lower temperatures precipitates, under the right experimental conditions, in the form of carbon
CARBON NANOTUBES
107
nanotubes. Chemical vapor deposition (CVD) techniques are typical methods being used.19,23–29 Depending on the production parameters (type of hydrocarbon gas, flow conditions, temperature, catalysts, catalyst introduction, support materials, additives etc.) usually various different types of MWNTs are obtained in bended form, as a direct consequence of low temperature processes leading to the inclusion of structural defects. More or less homogeneous materials are composed of thick and thin MWNTs, different numbers of layers, more or less bended, organised in bundles, or as individual objects, in various yields, with various amounts of inorganic impurities (between 2 and 40 wt %) from catalyst (even as filling inside CNTs) and catalyst support as well as low or high amounts of amorphous carbon. Sometimes, soot materials with enhanced fractions of SWNTs25 and DWNTs26 are obtained. CNTs also can be grown in aligned or in patterned form.27–29 A variation is the so-called HiPCO method (high pressure disproportionation of carbon monoxide) developed by Rice University30 and commercialized by CNI company which results in SWNT materials (ØSWNT = 2 nm, l = several micrometers) without any amount of amorphous carbon but with considerable amounts of catalytic iron nanoparticles (up to 40 wt % in as produced materials). Producing an almost “monolithic” CNT material composed only of CNTs of well defined diameters, chiralities, number of walls, without any by-products and in large quantities remains the greatest challenge in this field. Today, as a general rule of thumb one can say that today’s low volume production methods (few gram-scale) result in relatively high quality materials in what concerns CNT characteristics while the more recent large scale – low price production methods (tons-scale, 100 Euro/kg) do not aim yet at welldefined materials characteristics but on the availability of CNTs on industrial scale and on applications based on the overall CNT materials characteristics rather than on the individual CNT properties. 3.2. CHARACTERIZATION METHODS
Depending on the further use and application in mind, one type of CNT material might be more suitable than another one. Therefore it is highly recommendable to collect as much information as possible on structure and properties of the used CNT materials. In the following useful standard lab-techniques and their information content respective to bulk CNT materials, as produced, are briefly described. Scanning electron microscopy (SEM): A few milligrams of powder CNT material is deposited on the SEM sample holder. Eventually, depending on the conductivity degree, a very thin conducting metal layer is sputtered on. This technique generally gives highly valuable structural information on a scale of 100 nm to a few millimeters. Although CNTs can not be directly observed, it
108
W. MASER ET AL.
reveals CNT materials morphology, the presence and organization of the different sample products, density, homogeneity, fibre lengths. It also gives hints on the conductivity state (charging and contrast effects). SEM is a rapid, efficient and relatively economic technique for analysing CNT bulk material. Transmission electron microscopy (TEM): CNT materials are suspended in ethanol and ultrasonicated. A drop is deposited on a TEM grid. Solvent is evaporated and the sample can be analysed whenever the deposited material density is transparent for the electron beam. This technique gives structural information from the sub-nanometer to the 100 μm range. Diameter, layer structure, degree of graphitization, defects, bending and inorganic impurities can be observed. Combined with element mapping, selected area electron diffraction and EELS information on type, crystallization and distribution of catalytic nanoparticles as well as CNT chirality can be obtained. TEM gives valuable information on the level of individual CNTs, however, since only a very small (transparent) volume is probed good statistics is needed to get representative results. Induced coupled plasma spectroscopy (ICPS) and elemental analysis: A few tens of milligrams of powder CNT material is dissolved in acids and the precipitate analysed by flame-spectroscopy. It is a completely quantitative technique and reveals the content of metals, as well of other elements such as oxygen, sulphur and halogens in the bulk sample. ICPS allows the calculation of the composition of a sample and also might reveal ideas on the production yield. Thermogravimetric Analysis (TGA): A few tens of milligrams of CNT powders (or pellets) are deposited in a suspended crucible whose weight-loss is measured as a function of temperature increase. TGA gives information on the oxidation temperatures and residual mass. It may reveal CNT materials composition (presence of different carbon materials) and the amount of catalyst material in the sample. TGA is a valuable qualitative and comparative technique. Very easily different types of CNT materials can be distinguished from each other and information on further treatment temperatures can be obtained. If used for qualitative analysis, a strict protocol has to be employed (oxidation temperatures and metal rest content strongly depend on various experimental parameters. Slow temperature ramps between 3–5ºC/min, low gas flows are recommended). Furthermore, various error sources have to be taken into account (buoyancy effects, sample weight, sample morphology, “fluffy”sample or pellet effect, spontaneous combustion). Powder X-ray diffraction (XRD. About 100 mg of sample material is deposited on a flat-bed sample holder. XRD reveals the sample content, i.e. the type of carbon CNT, Camorph, Graphite, the degree of graphitization, of bundling
CARBON NANOTUBES
109
and alignment, as well as the presence of catalysts or other elements. Easily different CNT samples can be distinguished from each other. Raman spectroscopy: a few milligrams of CNT powder material is placed on a sample holder. Four characteristic features can be obtained: low frequency CNT radial breathing modes (RBM) around 100–200 cm–1 (frequencies are inversely proportional to SWNT diameters), tangential shear mode (G-line) at about 1,580–1,590 cm–1, defect mode (D-line) at about 1,350–1,450 cm–1 and double resonant G’ mode at around 2,700–2,800 cm–1. Therefore, Raman gives information on the carbon structures and the quality of the CNT material. It easily allows to distinguish between different types of CNTs (SWNTs, MWNTs, DWNTs) and reveals the presence of carbonaceous defects and of graphite. Diameters of SWNTs and their distribution as well as the degree of bundling and alignment can be calculated. The conducting character as well as doping and charge transfer effects may be revealed. Due to the one-dimensional electronic structure and the presence of van-Hove singularities, resonance effects upon tuning the laser excitation frequency are observed. Resonance Raman spectroscopy is used to reveal the (n, m) structure of individual SWNTs. Raman spectroscopy is one of the most powerful standard characterization technique for CNTs with a high information content on the sample characteristics and on individual objects.9 The homogeneity degree of the samples has to be probed by mapping different sample areas. Only limitation might be eventual sample degradation effects when using elevated powers which, on the other hand, reveal information about CNT’s thermal stability. Four-point conductivity: Pellets or films are contacted using pressure contacts (most easy approach). This technique establishes the conducting behaviour of bulk CNT networks. It is a good comparative approach to distinguish between MWNTs, SWNTs or modified CNTs. However, absolute values for conductivity strongly depend on the exact calculation of the geometry factor as well as the protocol for pellet pressing and contacting the sample. Nitrogen adsorption isotherms (BET): Bulk technique using a few tens of milligrams of CNT powder (or pellets or films) deposited in the corresponding sample holder. BET gives information on the gas-adsorption process (physisorption or chemisorption), gas kinetics, CNT pore size and specific pore-size area. Different types of CNTs easily can be distinguished. The above mentioned set of techniques to characterize bulk CNT materials reveals sufficient information for further materials processing and also may give first hints on the suitability of CNT materials for certain kinds of applications. Most of them are use in a qualitative and comparative way. It still remains critically distinguishing between different carbon components and quantifying their respective content in the CNT material. An approach to
110
W. MASER ET AL.
solve this problem may by the use of NIR spectroscopy11 carried out in diluted CNT suspensions. 3.3. PURIFICATION AND TREATMENT
As discussed above, CNT materials are not pure and thus there is great interest in improving the samples’ quality for many purposes ranging from the development of proper characterization standards to further use and applications. Improvements refer to: (i) removal of metal particles, (ii) removal of carbonaceous non-CNT contributions, (iii) separation of bundle structures, (iv) removal defective CNTs (v) separation according to length, diameter and chirality. Main difficulty here is that CNTs (and the whole CNT materials) are chemically quite inert and non-soluble. Therefore, purification and treatment methods, in a first step, focus onto the removal of metals and Camorph, and the improvement of CNT structure by oxidation approaches, as shown in Figure 4.
Figure 4. A typical purification scheme for CNT materials.
Typically two types of treatments can be distinguished: On one hand, chemical oxidation in liquid phase using acids (e.g. HNO3,31,32 HCl,31,33 HCl/HNO3,31 HF/HNO3,34 H2SO4/HNO3), under reflux and, on the other hand, thermal treatments, such as air-oxidation processes at temperatures between 300–450ºC. Chemical oxidation focuses onto the removal of metal nanoparticles in the CNT materials. The efficiency strongly depends on the concentration and concentration ratio of the used acids, the reaction time, and the amount of CNTs. Thermal treatments are directed towards the removal of amorphous carbon as well as the creation of pore-structures (activation process). The efficiency depends on the temperature applied (determined beforehand by TGA), the gas-flow, the amount of CNT materials and the macroscopic sample distribution, i.e. the exposed materials surface to the heat and gas-flow. If necessary, both processes can complement each other, even in several repetitive cycles.
CARBON NANOTUBES
111
Typically, a final annealing step is employed in order to heal structural CNT defects, and improve the bundle recrystallization in case of SWNTs. There exists a wide range of recipes which strongly depend on the CNT materials used and on the further use of the material. The following general observations can be made32: Chemical treatments lead to a significant reduction of metal particles up to 95 wt %. This is accompanied by significant material losses (20–60 wt %) including non-metal material losses (5–10 wt%) due to attacked carbon materials. Additionally, basic functional groups, especially carboxylic groups are attached to defect sites in attacked CNTs and also intercalation effects may be observed. The obtained material is highly compacted with drastically reduced specific surface areas of about 2 m2/g and nearly untreatable. Airoxidation does not affect the metal content, but removes defective carbon material, resulting in further material losses and in a relative increase of the metal content. On the other hand, it also results in highly activated materials with drastically increased specific surface areas as high as 750 m2/g. It is to mention that also this process creates defects at the tips and sidewall of CNTs to which carboxylic groups may be attached. The final annealing step contributes to heal defect structures, to recover the bundle organization, to remove created carboxylic groups and intercalated molecules. As one can see, purifycation is a multi-step approach which can be efficiently controlled using the above mentioned bulk techniques in order to judge the materials quality after each step. However, due to serious material losses, one has to find a compromise between resulting CNT quality, process yields, and final materials needs. 3.4. FUNCTIONALIZATION AND DISPERSIONS
Due to the high chemical inertness of CNTs and their insoluble character, chemistry on carbon nanotubes came later into the game and only created increased interest when it became clear that simple pre-oxidation steps as taken out in purification treatments lead to the creation of structural defects in CNTs (tip-opening and side-wall) to which various oxygen containing groups (mainly carboxyl groups) are attached as a function of the strength of the oxidation process (acid strength, and oxidation time) as illustrated in Figure 5. The oxidatively introduced carboxyl groups represent useful sites for further modifications, as they enable the covalent coupling of molecules through the creation of amide and ester bonds. By this method the nanotubes can be provided with a wide range of functional moieties such as dendrimers, nucleic acids, enzymes, metal complexes, nanoparticles etc. Furthermore, the presence of (modified) carboxyl groups leads to a reduction of van der Waals interactions between the CNTs, which strongly facilitates the separation of nanotube bundles
112
W. MASER ET AL.
Figure 5. A nanotube after oxidative treatment with defect sites at tip (open) and at side-walls to which oxidative functional groups (mainly carboxyl groups) are attached.
into individual tubes. Additionally, the attachment of suitable groups renders CNTs soluble in aqueous or organic solvents, opening the possibility of further modifications through subsequent solution-based chemistry. In addition to this “defect” functionalization, direct side-wall addition reactions through thermally activated reactions (preferentially on more reactive small diameter tubes) are documented in the literature. The functionalization degree is low and may vary between 3% and 15%. Fluorination35 and cyclo-addition36 are one of the most prominent side-wall reaction which further can be used for additional substitution reactions with alcohols, amines, Grignard reagents, alkyl lithium compounds and other specific linker molecules.37 Beside covalent functionalization (defect and side-wall addition), noncovalent functionalization is becoming of increased importance. It is based on the highly delocalized ʌ-electron system of CNTs to which many long-chain molecules and polymers have a strong affinity and get non-covalently linked to the CNT surface via van der Waals forces or ʌ-ʌ interactions. Surfactants such as triton X-100 or sodium dodecyl sulfate (SDS) lead to the formation of micelles and help to separate and disperse CNTs.38 Long-chain molecules/ polymers, such as polyelectrolytes,39 peptides,40 lipids,41 DNA,42 ʌ-conjugated polymers43–45 may even fully wrap around the CNTs by undergoing corresponding conformational changes to adapt to the underlying CNT network. These noncovalent approaches bear several advantages: (i) CNT structures are not disturbed and thus the original CNT properties are maintained, (ii) they render CNTs soluble in various solvents and thus is the base for the development of stable homogeneous CNT dispersions and for spectroscopic characterization, (iii) they facilitate selective SWNT separation according to diameter, chirality, length, and conductivity via adequately designed molecules, (iv) they add
CARBON NANOTUBES
113
further functionality to CNTs, (v) they improve interface interactions with matrix materials for the development of high-performance composite materials. Both, covalent functionalization and, even more promising, non-covalent supramolecular functionalization are the base for the development of stable and homogeneous dispersions. This is a very important step towards CNT homogenization, separation, improved solution characterization (NIR,10 photoluminescence,11 dynamic light scattering46 and optical spectroscopy), solution processing of CNTs as a whole material (the deposition and incorporation techniques, transformation technology (buckypaper and fibre fabrication)) and achieving compatibility and synergetic interactions with many molecules and polymer systems. Therefore, stable and homogeneous dispersions are considered as a key-stone to transfer carbon nanotube’s high application potential into realworld applications. 4. Carbon Nanotube Applications With their bunch of fascinating properties carbon nanotubes are of great interest to technological applications in several fields, as already indicated in Section 2.2. In the following, we will not group applications into different sectors, but instead will classify the various applications with respect to the appearance of CNTs in a certain form. This means we distinguish between applications based on individual CNTs, on assemblies of CNTs and on CNT-hybrid materials, such as CNT based composites. This type of classification will allow to better understand the chances and challenges common to CNTs in a certain form of appearance. Hereby, special emphasis will be put onto the field high performance CNT composite materials. 4.1. INDIVIDUAL CARBON NANOTUBES
An individual carbon nanotube with a well defined structure and thus, electronic properties is a highly promising object for the development of nanoelectronic devices. A metal type SWNT can be imagined as a metal wire of great use as vertical interconnector in classical semi-conducting devices. On the other hand, a semi-conducting SWNT can be seen as important logic element in field effect transistor device (FET). It has been demonstrated by several groups around the world that semi-conducting SWNTs can be incorporated successfully as active element into classical silicon-based FET devices. Here they have proven (superior) operational functionality as logic gate,47,48 as gas or bio sensing devices,49,50 or as nanoelectromechanical device.51 Key issues towards comercialization of these products mainly relate to time-costly assembly technologies, which directly is related to the lack of pure SWNT materials composed of only
114
W. MASER ET AL.
one type of SWNTs. Today, three assembling approaches are followed: (i) selection of an individual SWNT of a specific electronic characteristics out of a bulk SWNT material (usually requires coupled AFM, Raman, SEM technology steps), placing the selected SWNT onto a pre-fabricated electronic structure using a nanomanipulator (AFM-based technology), contacting (usually lithographic steps). (ii) dropping a diluted SWNT dispersion onto a prefabricated electronic structure, checking if SWNTs got deposited onto (between) the right contact structures, probing the electronic characteristics of these SWNTs (requires combined AFM (SPM) and Raman techniques), contacting (lithography steps). The limitations for both techniques refer to exhaustive selection and characterization processes for having the right SWNT at the right place. Alternatively, a third approach (iii) is to grow individual SWNTs in CVD processes directly in prefabricated channels of semi-conducting device structures into which catalyst seed particles are incorporated.52 Since no control of CNT characteristics during growth can be achieved yet, these device structures show only functionality at random places. Finally, it remains to mention the dream for purely nanotube-based electronic devices achieved when connecting SWNTs of different specific characteristics. While theoretically easily possible, again, limitations refer to efficient SWNT selection and assembly technologies. Further progress strongly relates to improved SWNT production technology as well as to efficient chemical separation and reliable assembly techniques. 4.2. ASSEMBLIES OF CARBON NANOTUBES
There are several electronic applications which are based on a more statistical approach using as grown assemblies of CNTs. Most prominent example are flat panel devices making use of the electron emission effect of CNTs. As mentioned in Section 3.2, CNTs can be grown in a pixel-like pattern on a conducting substrate.27–29 The micrometer-sized pixel areas contain millions of parallel aligned MWNTs. Placing a top-electrode and applying an electric field, statistically sufficient electron emission can be obtained (although some individual CNTs may fail) to fully satisfy conditions for building commercial flatscreen devices,53 and other devices based on field emission, such as X-Ray cathodes,16 and SEM/TEM tips.54 Aligned grown assemblies of CNTs also have been used for the development of various kinds of filter membranes55 for hydrocarbon separation and waste-water treatment. As a next step, CNT dispersions can be used to obtain statistical networks of CNTs. This can result in micrometer thick CNT networks (buckypapers) or even in ultra thin and even transparent networks of either conducting or semiconducting CNTs which can be coated onto flexible transparent substrates
CARBON NANOTUBES
115
resulting in transparent (vis-IR) and conducting CNT film coatings.56 These further can be assembled into flexible and transparent thin film transistors57 maintaining their electrical performance even upon strong bending. CNT networks (properly functionalized) also serve as transducing material for gas58,59 and bio60 sensing applications as well as biological scaffold materials for cell proliferation.61 CNT dispersions also are used to produce CNT fibre materials by spinning technologies as base for the development of intelligent textiles.62,63 Incorporated into textiles they operate as electrochemical element using the proper sweat of the human body. They may act as supercapacitors (storage of electrochemical energy), detect molecules (sensors), change color (electrochromic device), actuate (close and open pores) and absorb shock-energy. 4.3. CARBON NANOTUBE BASED COMPOSITE MATERIALS
The development of high performance nanocomposite materials is one of the most promising fields for CNT applications impacting on a broad range of technological sectors. The basic idea is to incorporate nanotubes into a matrix material and transfer their unique properties to the host system resulting in a material with enhanced functional, structural and processing properties. Essentially two types of matrix materials are considered: Polymeric and ceramic materials. Since only few works have been reported in the field of ceramics, in the following we will focus only on the class of polymer-CNT composites. Here, most research has been performed in the fields of thermoplastics and thermosets (both of great technological relevance in daily applications) as well in the area of electroactive polymers (of increasing importance for the development of plastic electronics). Independent of the matrix system, there are two key issues of paramount importance to obtain highly functional CNT-based composite materials: (i) Homogeneous CNT dispersion in the host system contributing to achieve maximum stable separation of CNTs and to avoid CNT agglomerations thus resulting in a homogeneous composite material. (ii) Efficient transfer of CNT properties to the host matrix. The development of nanocomposites clearly is an issue at the interface where the formation of a proper interface layer acting as link between CNT and the rest of the host material determines the degree of CNT property transfer and thus the final characteristics of the whole composite material. The interface layer simply may lead to a reorganized polymer structure in the vicinity of the CNTs,64,65 enough to observe already some first significant changes in the overall materials characteristics, or, in the best cases, even allow highly favorable CNT property transfer leading to completely new materials characteristics.45 The development of a favorable interface layer impacting on the whole materials characteristics
116
W. MASER ET AL.
directly is linked to a high degree of CNT dispersion. Both, CNT dispersion and load-transfer are highly complex issues strongly depending on the type of CNT material and the host matrix system, both defining the way towards successful compounding. Latest at this point, it should become clear why a thorough knowledge on all aspects around CNTs, as mentioned in the former chapters is critical towards the development of functional CNT-based nanocomposites. Incorporation of nanotubes into plastics can potentially provide light structural materials with dramatically increased modulus and strength. Thermoplastics, such as polypropylene,66 polystyrene,67 as well as thermosets,68 such as epoxies are widely used for this purpose. While for low CNT loadings, typically in the region between 0.1 and 2 wt %, moderate improvements, due to the development of a proper interface layer, are observed (elastic moduli improvements in the range of a factor or 3–4, increase of thermal conductivity by a factor of 1%), at higher loading rates, agglomeration takes place and the developed materials get highly inhomogeneous. This indicates that far more work is needed to achieve mechanical properties closer to the ones of the proper CNTs. On the other hand, due to CNTs high aspect ratio and conducting (electrical and thermal) properties, conducting filler networks can be obtained already at very low percolation values. Ultra-low percolation thresholds in the range 0.001 and 0.01 wt % of CNT filler ratio (a factor of 1,000 loser than for spherical carbon black particles) have been observed to reach conductivity values around 0.1–1 S/cm.69 Therefore, the formation of percolated CNT networks in different kinds of polymer matrices for the development of lightweight materials able to dissipate electrostatic charge, to shield electromagnetic radiation, and to dissipate heat, is of increased technological interest for commercial applications. Finally, it has been shown that CNTs may interact in a very favorable way with electroactive polymer matrix systems, such as polyaniline, polythiophene, polypyrrole and respective derivatives. Their highly conjugated backbone structure with a highly delocalized ʌ-electron system is highly compatible with the CNTs extended ʌ-electron system leading to composite materials with drastically improved characteristics of special interest to optoelectronic applications. Composites of CNT/poly(p-phenylenevinylene) (CNT/PPV),70–72 CNT/poly(3,4-ethylenedioxythiophene) (CNT-PEDOT)73 have proven their functionality in light-emitting diodes and photovoltaic cells, CNT/polyaniline (CNT(PANI) have been used as printable conductors for thermal-imaging techniques.74 Especially when in-situ polymerization processes are carried out, i.e. polymerization in the presence of CNTs, as shown for CNT/PANI,45,75 highly synergetic effects can be obtained ranging from increased conductivity, thermal stability, deprotonation stability, optical activity, combined with improved processing characteristics. All these are highly promising results towards the development of future plastic electronic devices including improved flexible
CARBON NANOTUBES
117
organic light emitting devices (OLEDs), photovoltaic cells, circuits, and sensing devices. Additionally, these electroactive polymer/CNT composites are of great interest to the development of energy storage devices, such as supercapacitors.76 5. Conclusions and Outlook In this article we have explained the unique properties of carbon nanotubes and shown that these are fundamental nanoscale objects of interdisciplinary, intersectorial interest and of great educative value. They are impacting on broad fields of science and technology. We have shown that carbon nanotubes as a material is highly complex and bears several challenging key issues ranging from proper characterization and purification to functionalization and the development of homogeneous stable dispersions and highly functional CNT-composite materials. While carbon nanotubes certainly will proof advantageous for certain applications, other ones probably will not develop at all. In any case, more than 15 years-time of research in this field has taught us that carbon nantoubes are highly enabling materials and good for many further surprises. ACKNOWLEDGEMENTS
The authors gratefully acknowledge funding from Regional Government of Aragon (DGA) under its “Group of Excellence” programme (DGA-T66). W.M. gratefully acknowledges funding from Spanish Ministry of Education and Science (MEC) and European Regional Development Fund (ERDF) under project NANOPOLICOND (MAT 2006-13167-C02-02). E.M. and A.M.B gratefully acknowledge support by MEC and ERDF under project SENAGAS (TEC200405098-C02-02). A.M.B. and M.T.M. gratefully acknowledge funding from DGA under Priority Research Line Projects (PIPO21/2005 and PIPO15/2005).
References 1. Iijima, S. (1991), Nature 354, 56–58. 2. Kroto, H., Heath, J., O’Brien S., et al. (1985), Nature 318, 162–163. 3. Mintmire, J., Robertson, D., and White, C. (1993), Phys. Chem. Solids 54(12), 1835– 1840. 4. Yakobson, B., Brabec, C., and Bernholc, J. (1996), Phys. Rev. Lett. 76, 2511–2514. 5. Charlier, J., Blasé, X., and Roche, S. (2007), Rev. Mod. Phys. 79, 677–731. 6. Mintmire, J., and White, C. (1995), Carbon 33, 89–902. 7. Hamada, N., and Sawada, S. (1992), Phys. Rev. Lett. 68, 1579–1581.
118
W. MASER ET AL.
8. Saito, R., Fujita, M., Dresselhaus, G., and Dresselhaus, M. (1992), Appl. Phys. Lett. 60, 2204–2206. 9. Dresselhaus, M., Dresselhaus, G., Saito, R., et al. (2005), Phys. Rep. Rev. Sect. Phys. Lett. 409, 47–99. 10. Itkis, M., Perea, D., Niyogi, S., et al. (2003), Nano Lett. 3, 309–314. 11. O’Connell, M., Bachilo, S., Huffman, C., et al. (2002), Science 293, 593–597. 12. Wildoer, J., Venema, L., Rinzler, A., Smalley, R., and Dekker, C. (1998), Nature 391, 59–62. 13. White, C., and Todorov, T. (2001), Nature 411, 649–651. 14. Frank, S., Poncharal, P., Wang, Z., and de Heer, W. (1998), Science 280, 1744–1746. 15. Zhu, Z., Bower, C., Zhou, O., Kochanski, G., and Jin, S. (1999), Appl. Phys. Lett. 75, 873–875. 16. Yue, G., Qiu, Q., Gao, B., et al. (2002), Appl. Phys. Lett. 81, 355–357. 17. Ruoff, R., and Lorents, D. (1995), Carbon 33, 925–930. 18. Dresselhaus, M., Dresselhaus, G., and Avouris, P. (Eds.) (2001), Topics in Applied Physics, vol. 80, Springer, Berlin, Germany. 19. Benito, A., Maser, W., and Martínez, M. (2005), Int. J. Nanotechnol. 2, 71–89. 20. Journet, C., Maser, W., Bernier, P., et al. (1997), Nature 388, 756–758. 21. Thess, A., Lee, R., Nikolaev, P., et al. (1997), Science 273, 483–487. 22. Maser, W., Muñoz, E., Benito, A., et al. (1998), Chem. Phys. Lett. 292, 587–593. 23. Pérez-Mendoza, M., Vallés, C., Maser, W., Martínez, M., and Benito, A. (2005), Nanotechnology 16, S224–S229. 24. Vallés, C., Pérez-Mendoza, M., Castell, P., et al. (2006), Nanotechnology 17, 4292– 4299. 25. Colomer, J., Stephan, C., Lefrant, S., et al. (2000), Chem. Phys. Lett. 317, 83–89. 26. Vallés, C., Pérez-Mendoza, M., Martínez, M., Maser, W., and Benito, A. (2007), Diamond Relat. Mater. C 16, 1087–1090. 27. Fan, S., Chapline, M., Franklin, N., et al. (1999), Science 283, 512–514. 28. Hata, K., Futaba, D., Mizuno, K., et al. (2004), Science 306, 1362–1364. 29. Terrado, E., Redrado, M., Muñoz, E., Maser, W., Benito, A., and Martinez, M. (2006), Diamond Relat. Mater. 15, 1059–1063. 30. Nikolaev, P. (2004), J. Nanosci. Nanotechnol. 4, 307–316. 31. Chiang, I., Brinson, B., Smalley, R., et al. (2001), Phys. Chem. B 105, 1157–1161. 32. Martínez, M., Callejas, M., Benito, A., et al. (2003), Carbon 41, 2247–2256. 33. Jeong, T., Kim, W., and Hahn, Y. (2001), Chem. Phys. Lett. 344, 18–22. 34. Chattopadhyay, D., Galeska, I., and Papadimitrakopoulos, F. (2002), Carbon 40, 985–988. 35. Bettinger, H. (2003), Chem. Phys. Chem. 4, 1283–1289. 36 Tagmatarchis, N., and Prato, M. (2004), J. Mater. Chem. 14, 437–439. 37. Khabashesku, V., Billuops, W., and Margrave, J. (2002), Acc. Chem. Res. 35, 1087– 1095. 38. Islam, M., Rojas, E., Bergey, D., et al. (2003), Nano Lett. 3, 269–273. 39. Correa-Duarte, M., and Liz-Marzán, L. (2006), J. Mater. Chem. 16, 22–25. 40. Dieckmann, G., Dalton, A., Johnson, P., et al. (2003), J. Am. Chem. Soc. 125, 1770.
CARBON NANOTUBES
119
41. Qiao, R., and Ke, P. (2006), J. Am. Chem. Soc. 128, 13656–13657. 42. Heller, D., Jeng, E., Yeung, T., et al. (2006), Science 311, 508–511. 43. Panhuis, M., Maiti, A., Dalton, A., et al. (2003), J. Phys. Chem. B 107, 478–482. 44. Star, A., Stoddart, J., Steuermann, D., et al. (2001), Angew. Chem., Int. Ed. 40, 1721–1725. 45. Sainz, R., Benito, A., Martínez, M., et al. (2005), Adv. Mater. 17, 278–281. 46. Badaire, S., Poulin, P., Maugey, M., and Zakri, C. (2004), Langmuir 20, 10367– 10370. 47. Tans, S., Verschueren, A., and Dekker, C. (1998), Nature 393(6680), 49–52. 48. Avouris, P. (2004), MRS Bull. 29, 403–410. 49. Kong, J., Franklin, N., Zhou, C., et al. (2000), Science 287, 622–625. 50. Star, A., Gabriel, J., Bradley, K., et al. (2003), Nano Lett. 3, 459–463. 51. Bourlon, B., Glattli, D., Miko, C., Forró, L., and Bachtold, A. (2004), Nano Lett. 4, 709–712. 52. Duesberg, G., Graham, A., Liebau, M., et al. (2003), Nano Lett 3, 257–259. 53. Choi, Y., Cho, Y., Kang, J., et al. (2003), Appl. Phys. Lett. 82, 3565–3567. 54. de Jonge, N., and van Druten, N. (2003), Ultramicroscopy 95, 85–91. 55. Srivastava, A., et al. (2004), Nat. Mater. 3, 610. 56. Wu, X., Chen, Z., Du, X., et al. (2004), Science 305, 1273–1276. 57. Cao, Q., Hur, S.-H., Zhu, Z.-T., et al. (2006), Adv. Mater 18, 304–309. 58. Sippel-Oakley, J., Wang, H., Kang, B., et al. (2005), Nanotechnology 16, 2218–2221. 59. Sayago, I., Terrado, E., Lafuente, E., et al. (2005), Synth. Met. 148, 15–19. 60. Liu, G., and Lin, Y. (2006), J. Nanosci. Nanotechnol. 6, 948–953. 61. Zanello, L., Zhao, B., Hu, H., and Haddon, R. (2006), Nano Lett. 6, 562–567. 62. Dalton, A., Collins, S., Muñoz, E., et al. (2003), Nature 423, 703. 63. Bauhgman, R., Zakhidov, A., and de Heer, W. (2002), Science 297, 787–792. 64. Sandler, J., Broza, G., Nolte, M., et al. (2003), J. Macromol. Sci.-Phys. B 42, 479–488. 65. Chen, W., and Tao, X. (2005), Macromol. Rapid Commun. 26, 1763–1767. 66. Castell, P., Maser, W., Benito, A., et al. (2007), Polymer (submitted). 67. Qian, D., Dickey, E., Andrews, R., et al. (2000), Appl. Phys. Lett. 76, 2868–2870. 68. Biercuk, M., Llaguno, M., Radosavljecvic, M., et al. (2002), Appl. Phys. Lett. 80, 2767–2769. 69. Sandler, J., Kirk, J., Kinloch, I., et al. (2003), Polymer 44, 5893–5899. 70. Curran, S., Ajayan, P., Blau, W., et al. (1998), Adv. Mater. 10, 1091–1093. 71. Coleman, J., Curran, S., Dalton, A., et al. (1998), Phys. Rev. B 58, R7492–R7495. 72. Star, A., Stoddart, J., Steurman, D., et al. (2001), Angew. Chem. Int. Ed. 40, 1721–1725. 73. Woo, H., Czerw, R., Webster, S., et al. (2001), Synth. Met. 116, 369–372. 74. Blanchet, G., Fincher, C., and Gao, F. (2003), Appl. Phys. Lett. 82, 1290–1292. 75. Sainz, R., Small, W., Young, N., et al. (2006), Macromolecules 39, 7324–7332. 76. Picó, F., Rojo, J., Sanjuán, M., et al. (2004), J. Electrochem. Soc. 151, A831–A837.
ULTRASHORT PULSE PLD: A TECHNIQUE FOR NANOFILM FABRICATION T. SZÖRÉNYI1* AND Zs. GERETOVSZKY2 1 Research Group on Laser Physics, Hungarian Academy of Sciences, University of Szeged, PO Box 406, H-6701 Szeged, HUNGARY 2 Department of Optics and Quantum Electronics, University of Szeged, PO Box 406, H-6701 Szeged, HUNGARY
Abstract – In the present contribution the peculiarities of laser ablation are discussed with special emphasis on the differences in the mechanisms of nanoparticle formation when ablating materials with pulses of nanosecond vs. femtosecond duration. In the case of ablation using nanosecond pulses the dominating species leaving the target surface are principally atoms and ions. Cluster formation and growth mainly take place, via nucleation and condensation, from the plasma plume within the surroundings. The principal control parameter is the ambient pressure. When the major goal is not the production of colloids (either in form of an aerosol or a sol) but layer growth instead, nanostructured films can be made at pressures higher than a few pascals. On the other hand, ablation with ultrashort pulses produces a plasma plume of biphasic character: its leading edge, consisting of ionic and atomic components is followed by a spatially and temporally well separated cloud of nanoparticles. In this case nanoparticle formation is a direct consequence of the interaction of the ultrashort laser pulse with the target material. This process even works in high vacuum, which provides an additional proof for that here, contrary to nanosecond-ablation, those are the laser parameters that control the characteristics of the nanoparticles produced.
Keywords: Ablation, laser processing, clusters, nanoparticles, thin films
______ *
To whom correspondence should be addressed: T. Szörényi, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
121
122
T. SZÖRÉNYI AND Z. GERETOVSZKY
1. Pulsed Laser Ablation: General Characteristics In laser ablation the energy of the laser pulse is absorbed within a small volume of the surface layer of the target material and converted into heat almost instantaneously. The fast and confined heating builds up extreme pressures, induces phase transitions and leads to material ejection from the heated volume. While the technique is conceptually simple, the strongly nonlinear and far-fromequilibrium nature of the involved processes make it rather complex. Though being clearly an over-simplification, ablation is often envisaged as a sequence of steps, starting with the interaction of laser radiation with the target, followed by the absorption of energy and localised heating within a thin surface layer, phase transitions, and subsequent (very much forward directed) material ejection in form of a plume, which either expands into vacuum or gas, or may even interact with a liquid environment. Depending on the duration of the laser pulse and the nature of the environment, the properties of the components of the ablation plume may further change, due to collisions within the plume and/or between the constituents of both the plume and the ambient, in certain cases even through laser-plume interactions.1,2 Historically, production of nanoparticles by laser ablation dates back to the mid-1980s, when the first papers on fullerenes were published.3 For about 2 decades nanosecond lasers remained the only tools of nanoparticle production. In the third Millennium ultrashort pulse lasers moved to the forefront of laser ablation research, as well. Nevertheless, it became evident only very recently, that nanosecond and femtosecond ablation proceed along fundamentally different routes.4 In this contribution the peculiarities of nanoparticle production by laser ablation will be reviewed by following the life story of species ejected from the target while focusing on the differences between nanosecond vs. femtosecond processing. Special emphasis will be given to that particular setup in which the species are collected on a substrate, i.e. when we grow a layer. First a brief overview of the techniques available for in- and ex-situ analysis of the properties of the ablated species and their interaction within and outside the plume will be given. Then the characteristics of nanoparticle formation will be revealed by analyzing the results of selected case studies. 2. Analytical Toolbox Techniques – The Life Story of the Ablated Species The range of applicable analytical techniques is determined by the ambient the ablated species expand into. In vacuum or in a low pressure gas ambient in-situ plasma spectroscopies offer the most direct way to identify and characterize the just born species and their interaction with the surroundings during plume
ULTRASHORT PULSE PLD
123
expansion. Gated intensified CCD photography is routinely used to record the time evolution of the spatial distribution of the plume. Differentiation between various plume components is possible by working in carefully chosen spectral domains. Once the spatial and temporal evolutions of the plume have been identified via imaging, optical emission spectroscopy can be used to identify and quantify species in their excited states in different locations within the plume. On the other hand, laser-induced fluorescence or optical absorption spectroscopy can probe the “dark” part of the plume, i.e. measure unexited atoms and molecules, along with small clusters. Scattering can also be used to reveal the presence of NPs. Rayleigh scattering imaging is effective for particles larger than just 2–10 nm, so it can be used to study nanoparticle growth. Blackbody radiation can also be imaged, or spectroscopically analysed to reveal the presence and temperature of NPs. Time-of-flight techniques are also routinely applied for spaceand time-resolved analysis of the dynamics of plasma species. In vacuum or relatively low pressures (up to several hundred pascals) the plume components can be collected for (in- and ex-situ) analyses and/or layers can be grown by immersing a substrate into the expanding plume. By following the process of film growth in-situ or by the ex-situ characterization of film properties, information on the characteristics of the ablation products can be derived. It is worth noting, that this technique offers a relatively simple, yet extremely versatile method of thin film production, also known as Pulsed Laser Deposition, PLD. Due to the popularity of PLD, gaining insight into the behaviour of plasma species by means of indirect film-based techniques has a well studied basis. Contrary to the plethora of techniques available in low pressure environments the analytical toolbox is rather empty in denser media. The main reason of that stems from the reduced mean free path at high pressures or in liquid ambients. In certain cases characterization of detonation waves, cavitation bubble formation by shadowgraphy does work, but in the general case the range of analytical techniques reduces to ex-situ characterization of the ablation products. Colloidal systems formed in liquids or in gas atmospheres at pressures above several hundred pascals can be analyzed using the standard techniques of nanoscience/nanotechnology. While stabilizing a nano-dispersed phase in an aerosol is rather problematic, gases offer the cleanest environment and therefore allow for the fabrication of nanoparticles of the highest purity. On the other hand, ablation into a liquid has the advantage of providing an easy means to hinder agglomeration and preserving the original size distribution of the individual nanoparticles via protecting their surface.
124
T. SZÖRÉNYI AND Z. GERETOVSZKY
3. Ablation with Nanosecond Pulses 3.1. SINGLE-WALL CARBON NANOTUBE SYNTHESIS BY LASER VAPORIZATION
Synthesis of carbon nanotubes by laser ablation is worth mentioning not only because it was the first method used to produce fullerenes in the gas phase,3 but also because this technique remained one of the best methods to grow high-quality, high-purity single wall carbon nanotubes, SWNT. A single laser pulse (of typically nanosecond duration) is sufficient to produce a vapor cloud of carbon and catalyst atoms, from which, inside a hot oven and in inert gas atmosphere, the species self-assemble to form SWNTs. The yield is high and the process is relatively insensitive to the diameter of the metal catalyst nanoparticle.5–7 SWNT can also be grown at process temperatures well below 1,000–1,100ºC, even at RT, where high-repetition-rate8 or long-pulse9 lasers provide the sufficient heating of the target and the plasma species. A necessary prerequisite of exploiting the full potential of carbon nanotube growth by laser evaporation is the understanding of the process of SWNT formation. The potential of spectroscopies in following the time evolution of the plasma processes has convincingly been demonstrated by the spatial and temporal mapping of SWNT synthesis in a series of papers from Dave Geohegan’s group at ORNL.10–12 From the results of these studies the timeline of the process has been constructed which depicts, in Figure 1, the birth of laser generated SWNTs. The plume initially consists of atomic and molecular species. Condensation of carbon starts within 0.2 ms after ablation. The small C clusters begin to aggregate and form swirling vortex rings appearing as smoke in the gas flow. As time passes the plasma temperature continuously decreases and the excited metal atoms relax into their ground states. In the next millisecond, also the ground state metal atoms condense, and so by t | 2 ms the plume becomes almost entirely composed of clusters, trapped within swirling smoke rings. This mixture of carbon clusters and metal catalyst nanoparticles forms the feedstock for nanotube formation. The growth is however a relatively long process. After 20 ms, only SWNTs of about 200 nm length are found. Growth up to 10 μm lengths occurs over much longer times – from 100 ms to s – indicating growth rates of 0.5–5 Pm/s. While the conclusion of general interest of these time-resolved growth studies is that SWNTs grow through the conversion of condensed-phase carbon clusters and metal catalyst nanoparticles, in the context of our analysis which aims to compare nanosecond ablation to femtosecond one, the main message is that the clusters, acting as the feedstock of SWNT growth, are formed from atomic species of the plasma by condensation in an inert ambient.
ULTRASHORT PULSE PLD
125
Figure 1. The stages of SWNT formation from laser generated plasma as a function of time and plume temperature.11
3.2. THE ROLE OF PRESSURE IN DETERMINING GAS PHASE PROCESSES; FORMATION OF CARBON NANOCLUSTER FILMS
When the plume expands into a gas, the fate of its species is determined by the frequency and the nature of collisions with the surrounding atoms/molecules. Up to approximately 1 Pa the behaviour of the plume is essentially vacuumlike. With increasing background pressure the frequency of collisions increases, which reduces the initially high kinetic energy of the species. At sufficiently high pressures the expansion eventually stops, and the species become thermalized.12,13 Fast imaging techniques provide insights into the details of these events.14–16 In Figure 2 the effect of confinement is exemplified by comparing the position and shape of carbon plasma plumes in vacuum and in 20.5 Pa Ar. This latter pressure lies already within the pressure domain where the ablated atoms and ions have slowed down so much that cluster formation becomes possible. Now, the size of the clusters is jointly controlled by the laser parameters and the actual pressure. This pressure control of the kinetic energy of plasma species materializes in profound changes in the morphology of films built by collecting the ablated material on substrates.17–19 As an example, Figure 3 shows the evolution of the surface morphology of films deposited by ablating a graphite target with 6 J cm2 pulses of a KrF excimer laser (15 ns, 248 nm) in argon at different background pressures.18 At 0.67 Pa hard, scratch resistant films of smooth morphology are produced. As the Ar pressure increases to 5.32 Pa the films become much
126
T. SZÖRÉNYI AND Z. GERETOVSZKY
Figure 2. Wavelength filtered ICCD images of a carbon plume expanding into vacuum (a)–(c), and into 20.5 Pa Ar (d)–(f). Twenty nanoseconds gate with a delay after the laser pulse of 120 ns for images in the first column, 250 ns for the second column and 400 ns for the third column. The length of each frame is 30 mm.16
Figure 3. SEM images of carbon films deposited at Ar pressures of (a) 0.67, (b) 5.32, (c) 13.3 and (d) 45.2 Pa. Note the different scale bars.18
rougher due to nanoscale clustering. The cluster size derived from the SEM picture is approximately 15 ± 3 nm. At 13.3 Pa clustering becomes more pronounced, and leads to clusters of §45 nm in size. These layers can be easily scratched off, indicating much reduced hardness. By 45.2 Pa the film morphology changes from nanoscale clusters to more filamentary growth together with a further reduction in film density. The appearance of the material resembles very much that of the low density diamond-like carbon nanofoam produced by high repetition rate Nd:YAG laser in argon atmospheres of pressures exceeding §13 Pa.20 When decorated by surface-enhanced Raman-active components,
ULTRASHORT PULSE PLD
127
e.g. metal nanoparticles, such large surface area carbon nanoscaffolds may serve as efficient, low-cost sensor platforms.21 The evolution of film morphology with pressure is very similar in different materials systems, though with different threshold values depending on the material ablated and the nature of the ambient gas used.22 Since the changes in growth mechanism materialize in the variation of all film properties, from the proper analysis of these, information on the underlying mechanisms can also be derived. The comparative analysis of the growth characteristics of carbon and carbon nitride films fabricated by ablating identical graphite targets with an ArF excimer laser (Ȝ = 193 nm) in the same configuration in Ar and N2 atmospheres, respectively, is another example for such a film-based approach.23 The pressure dependence of the apparent growth rates, defined as measured film thickness per number of pulses, is shown in Figure 4. Up to approximately 0.5 Pa the nature of the atmosphere has no effect on the growth rate. It is still in line with the expectations, that in Ar the growth rate drops between 0.5 and 5 Pa. The unexpected increase between 5 and 100 Pa, however, pinpoints a change in the growth mechanism. When ablating in N2, the domain where the increase of pressure has apparently no influence on the growth rate extends up to approximately 50 Pa, suggesting contribution of more than one process to film growth. In the changes in the apparent growth rate both the variation of the number of the constituting atoms and the changes in film microstructure manifest themselves. The two contributions can be separated by recording the change in the number of constituting atoms deposited over unit film area per pulse, within the same pressure domain (Figure 5). A close-up of the change in carbon growth rate in Ar atmosphere reveals that the decrease in the apparent growth rate between 0.5 and 5 Pa, shown in Figure 4, is a consequence of the decrease in the number of carbon atoms reaching the substrate due to collisions with background gas particles. However, the threefold decrease in the number of film building atoms from 0.6 to 0.2 × 1015 atoms cm–2 pulse–1 is only accompanied by an approximately twofold decrease in the apparent growth rate, suggesting a concomitant decrease in the compactness of the films at higher pressures. In the 5–100 Pa domain the carbon growth rate keeps decreasing, while the apparent growth rate increases, revealing that the decrease in film density accelerates. In nitrogen atmospheres, up to approximately 5 Pa, the incorporation of nitrogen nearly compensates for the loss in carbon deposition, resulting in a practically constant apparent growth rate. When exceeding 5 Pa this trend discontinues. The number of carbon atoms incorporated into the films keeps decreasing while that of nitrogen atoms starts to decrease, as well. While the
T. SZÖRÉNYI AND Z. GERETOVSZKY
128
dependence of carbon and nitrogen deposition rates on N2 pressure accounts for the practically constant N/C ~ 0.35 ratios measured in the 5–50 Pa domain,24 it can not any further account for the unchanged apparent growth rate (cf. Figure 4).
-1
Apparent growth rate [nm pulse ]
0.06
0.05
Figure 4. The apparent growth rate of carbon (Ƒ) and carbon nitride (Ŷ) films as a function of pressure, fabricated by ablating identical graphite targets in Ar and N2 atmospheres, respectively.
0.04
0.03
0.02 -2
-1
0
1
2
0.6 0.5
Figure 5. The deposition rate of carbon (Ƒ and Ŷ) and nitrogen (ż) atoms in carbon (Ƒ) and carbon nitride (Ŷ and ż) films, as a function of pressure.
0.4 0.3
Growth rate [10
15
atoms cm-2 pulse-1]
lg (Pressure [Pa])
0.2 0.1 0.0 -2
-1
0
1
2
lg (Pressure [Pa])
3.0
2.6
Figure 6. The density of carbon (Ƒ) and carbon nitride (Ŷ and ż) films as a function of pressure.
-3
Mass density [gcm ]
2.8
2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 -2
-1
0
lg (Pressure [Pa])
1
2
ULTRASHORT PULSE PLD
129
The comparative analysis reveals that the pressure dependence of the absolute number of film constituting atoms (Figure 5) alone can not explain the variations in the apparent growth rates (Figure 4). The mass density vs. ambient pressure plots, derived by dividing the sum of the mass of all constituting atoms over unit area by the measured thickness, shown in Figure 6, display similar dependences for both carbon and carbon nitride, and indicate profound structural changes, indeed. In particular, these results confirm that the macroscopic density of carbon nitride films starts to decrease when the composition exceeds N/C ~ 0.2.25–28 The sudden decrease in mass density above 5 Pa explains why practically no change has been recorded in the apparent deposition rate in the 5–50 Pa domain (Figure 4) in spite of the continuous decrease in both C and N arrival rates (Figure 5). Atomic force microscope17,29,30 and TEM31 studies (Figure 7) verified, that the decrease in mass density of both DLC and carbon nitride films was a direct consequence of increased surface roughness and porosity.
Figure 7. TEM micrographs taken on films deposited with 10 J cm–2 pulses in 5 Pa N2 (left) and with 7.5 J cm–2 pulses in 50 Pa N2 (right). Each picture is 1 × 1 ȝm2 in size.31
The above results confirm that in the particular case of carbon and carbon nitride film growth the critical pressure is at around 5 Pa, indeed. Cluster growth starts at pressures above 5 Pa due to gas-phase collisions of carbon atoms/ions. The thermalized species build a layer of loose, porous network when reaching a room temperature substrate. The corollary of these and any other studies alike is that the primary products of nanosecond-ablation are atoms and ions. Cluster formation is the consequence of the interaction of the expanding plume and the ambient gas species. In inert atmospheres, the contribution of atomic vs. cluster growth to film formation is controlled by pure mechanics, i.e. by the number of collisions.
T. SZÖRÉNYI AND Z. GERETOVSZKY
130
Increasing pressure promotes cluster formation. When ablating in reactive atmospheres, a chemical effect should also be added to the kinetics: the collisions may result in formation of species composed of components of both the plasma and the atmosphere. 3.3. A UNIQUE APPROACH: BACKWARD DEPOSITION OF LASER GENERATED NANOCLUSTERS
As the mean free path of plasma species at pressures higher than a few pascals falls below a centimetre, in the absence of a forced gas flow the nanoclusters formed remain in the vicinity of the source, i.e. close to the ablation spot. A genuine approach is the collection of backward-propagating clusters on substrates placed in the target plane near to the ablated area.32,33 By tuning the laser parameters and the nature and pressure of the surrounding gas, nanoparticles of a great variety of sizes can be produced. The size distribution of nanoclusters deposited by ablating a Si wafer in He (Figure 8) illustrates well the potential of this geometry.
1
2
3
4
5
6
Number of clusters
1J/cm2
Figure 8. Size distribution of Si nanoclusters deposited in He atmosphere at 532 Pa at various energy densities of the ablating ArF laser pulses.32
1.4 J/cm2
3 J/cm2
3.9 J/cm2
1
2
3
4
5
6
Size (nm)
3.4. NANOPARTICLE PRODUCTION BY ABLATING UNDER LIQUIDS
Nanoparticles intended for chemical- and bio-applications in liquid environments should (i) be directly formed in the particular liquid most appropriate for the planned application, (ii) possess a relatively narrow size distribution, (iii) be free of surface impurities, and (iv) have reactive chemical groups on their
ULTRASHORT PULSE PLD
131
surfaces to simplify further attachment of (bio)molecules, if necessary.34 Backed by 100 years of experience colloid chemistry can produce a wide spectrum of nanoparticles, partly meeting these requirements. The peculiarity of ablation is that performed in purified and filtered deionized water a great variety of ultrapure nanomaterials can be produced, and after being in the solution, surface engineering of the nanoparticles becomes possible. Since systematic studies on the physical aspects of ablation in liquid environment have only been carried out for about 7 years, the full potential and the limitations of this technique is far from being explored. Technical problems like the formation of relatively long-living (300–500 ȝs) cavitation bubbles, absorption of laser light by and reflection on the nanoparticles in the liquid, self-focusing (time dependent) inhomogeneities in the refractive index limit the possibilities of following the events in-situ and make the analysis and description of the complicated processes involved rather hard. The state-of-the-art is the following: When ablating with nanosecond pulses in chemically inert liquids (e.g. in pure water), the size of the nanoparticles produced is relatively large, since coagulation and aggregation can hardly be overcome. In practice strongly dispersed particles of 10–300 nm mean diameter have been obtained with limited control by the laser parameters.34 When ablating in chemically active solutions chemistry helps to stabilize the primary nanoparticles, which leads to smaller mean diameters and narrower size distributions, as compared to the inert liquid case. Significant progress has been achieved by the use of aqueous solutions of surfactants. The surfactant covers the just born NPs and prevents them from agglomeration. Sodium dodecyl sulfate (SDS) proved to be one of the most efficient surfactants to stabilize gold and silver nanoparticles at around 5 and 12 nm mean sizes (Figure 9), respectively, in best cases.35–37 3.5. NANOPARTICLES BY NANOSECOND-ABLATION – SUMMARY
When ablating with pulses of nanosecond duration the material leaves the target in form of atoms, ions, and to a much less extent as small clusters. To generate nanoparticles, an environment of relatively high pressure is a prerequisite in order to create an efficient confinement of the plasma and increase its density. The formation and the evolution of the properties of the nanoparticles are essentially determined by the interaction of ejected (atomic) species and the environment. When ablating in gases, the key parameter is the pressure. Here the material can also be collected on substrates, leading to the formation of nanostructured films. Ablation in a liquid ambient results in the formation of colloidal nanoparticle solutions, but small sizes are available only when further chemical treatments are applied.
132
T. SZÖRÉNYI AND Z. GERETOVSZKY
Figure 9. Electron micrograph and the derived size distribution of silver nanoparticles produced by ablation with the second harmonic (532 nm) of a Nd:YAG laser in 0.05 M SDS aqueous solution.35
4. Ablation with Ultrashort Pulses 4.1. INTRODUCTION
In the past decade, femtosecond lasers emerged as unique tools allowing processing of practically any material with unprecedented precision. The advantages of ultrashort pulse processing over conventional laser machining with pulses of nanosecond duration originate in the 4–6 orders of magnitude shorter timescale of energy deposition. The energy injected into the material within picoseconds is absorbed instantaneously by the electrons of a very thin surface layer (skin depth), while the transfer of energy to the lattice takes longer time, typically tens of picoseconds. Although the details of the interaction are different for metals and dielectrics, the net results are the same: clean removal of a small, but well defined volume of material with minimal thermal load to the surrounding material. Historically, motivated by the interest of industrial applications which required high-precision patterning, most efforts have been devoted to the optimization of material removal with much less attention to the properties of the material removed.38 As a consequence, the potentials of femtosecond ablation in nanoparticle production remained hidden until very recently. Ultrashort pulse lasers produce a hot, highly ionized plasma plume that expands in a well-defined fashion and serves as an excellent source for thin film deposition. The plume does not contain micron-sized particulates and droplets,
ULTRASHORT PULSE PLD
133
which is often the case when ablating with nanosecond lasers, enabling thereby the production of particulate-free films. The high intensity pulses produce plumes with large concentrations of ionic species with high kinetic energies (>100 eV) which has been considered beneficial e.g. in the production of high quality diamond-like carbon films.39 However, from the data available to date, one may conclude that ablation with high intensity femtosecond pulses does not necessarily lead to improvement in film properties Instead of striving to achieve maximum intensities, ablation with pulses of lower intensities, conveniently attainable with commercial femtosecond systems, seems to be a more pragmatic approach.40 Moreover, recent results revealed that plasmas generated in low pressure environments by ultrashort pulses having intensities not far above ablation threshold were ideal sources of nanoparticles.41–46 Following the same outline we used to present the peculiarities of nanosecond ablation, in this second part we will highlight those features of the ultrashort ablation process which differ most from that of the nanosecond case. Our approach will again address both the characteristics of the plume, and the properties of the ablation products. This will be achieved by summarizing the results of selected case studies, rather than trying to give a full coverage of all aspects of ablation with ultrashort pulse lasers. 4.2. ABLATION IN VACUUM: PLUME PROPERTIES
In a series of papers published in the last few years S. Amoruso and coworkers convincingly demonstrated that – contrary to the nanosecond case where the material leaves the target surface in a single package – in the plasma generated by ultrashort pulses of intensities within the 1011–1013 W cm–2 domain three populations could be identified, each of which is characterized by different expansion dynamics.41–43 Snapshots of a silicon plume and especially the intensity profiles derived by integrating the emission along directions parallel to the target surface, and shown on the right of the images (Figure 10), clearly prove the presence of three populations in the expanding plasma. From the distances the two fast components traveled in 30 ns a decent estimate can be given to their expansion velocities, while for the determination of the velocity of the third, slow component, which remains very close to the target surface on this timescale, delays of the order of microseconds are necessary (Figure 11). The expansion velocities of the three populations derived from these images are ~107, ~106 and ~5 × 104 cms–1, respectively. Spectral analysis has shown that the fastest population consists of ions, while the second component of the plasma contains both neutral and ionized species. At time delays below 0.5 ȝs the spectra, shown here for the particular
134
T. SZÖRÉNYI AND Z. GERETOVSZKY
Figure 10. The early stage of expansion of a silicon plume produced by ablation with 0.9 ps @1,055 nm pulses of 8.4 × 1011 W cm–2 intensity into vacuum. Gate time: 10 ns.41
Figure 11. Evolution of the slow component of a silicon plume produced by ablation with 0.9 ps @1,055 nm pulses of 8.4 × 1011 W cm-2 intensity into vacuum. Gate times : 1 ȝs at 5.5 ȝs and 5 ȝs at 22.5 ȝs delays, respectively.41
case of Au (Figure 12), are dominated by emissions from atoms. At longer delays, however, structureless, broad black-body spectra appear, indicating the arrival of hot nanoparticles. Detailed analysis of the spectra yields initial temperatures in the order of few thousands Kelvin, continuously decreasing with time due to radiative cooling.42,44,45 The comparison of results obtained on ablating Si, Ni, Fe, Au, Ag, Ti and TiC with pulses of different durations (80–900 fs) and intensities (1011–2 × 1013 W cm–2) suggests that the behaviour of the plasma, outlined above, represent a common, general feature of femtosecond-ablation, at least within the parameter window examined.41–46
ULTRASHORT PULSE PLD
135
Figure 12. Time evolution of the emission spectra of a gold plume recorded 1 mm above the target surface. Ablation with 120 fs pulses of 5 × 1012 W cm–2 intensity of a Ti:sapphire laser @780 in vacuum.42
4.3. NANOPARTICLE FABRICATION IN VACUUM
Similar to the behaviour of the plasma, the size and the size distribution of the nanoparticles produced from different materials are akin, as well. AFM analysis of nanoparticle films of Ag, Au, Ni, Si and Fe of less than a monolayer coverage yielded mean nanoparticle radii ranging from 8 to 25 nm with standard deviations between 5 and 20 nm, as illustrated for the case of Ag in Figure 13.
Figure 13. Left: AFM image of nanoparticles deposited on mica by ablating a silver target with 120 fs pulses of 5.0 × 1012 W cm–2 intensity in high vacuum. Right: size histogram.42
136
T. SZÖRÉNYI AND Z. GERETOVSZKY
Size control of nanoparticles is possible by tuning the laser parameters. An increase in intensity within the 1012–1013 W cm–2 range leads to a slight increase in the mean radii.42 The relatively broad size distribution, characteristic of ablation in the visible, can be narrowed by using UV wavelengths.47 The increase of pulse duration up until the electron-phonon relaxation time, IJe-ph of the material to be ablated has apparently no significant effect on the ablation products. There was very little difference in the morphology of Ag (IJe-ph ~20 ps) and Ni (IJe-ph ~7 ps) films deposited in high vacuum using 10 ps and 200 fs pulses.48 In both cases the films comprised of almost spherical nanoparticles of 50–300 nm in diameter, with apparently more aggregation when ablating with 10 ps pulses. Nevertheless, more experimental work is required to completely characterise the ablation and deposition processes at different pulse durations. In the light of potential future applications it is important to note, that synthesis of nanocrystals of different materials is also possible with this technique. An example: The analysis of X-ray diffraction patterns of Ni and Fe nanofilms indicated that the films consisted of crystallites with sizes of few tens of nanometers, whereas AFM imaging gave exactly the same nanoparticle sizes, suggesting that each and every particle has the same single-crystalline domain, i.e. the films are composed of randomly oriented, but identical nanocrystals.49 The analysis of the magnetic behaviour of Ni nanoparticle films led to the conclusion that the film behaved as a system of isolated magnetic particles, i.e. the individual nanocrystals of 40 ± 19 nm size approached fairly well the ideal, single-domain behaviour.50 4.4. PLUME EXPANSION IN A REACTIVE GAS. A POSSIBLE ROUTE TO COATED NANOPARTICLE FABRICATION
The differences in the mechanisms and the products of ablation performed with pulses of nanosecond vs. femtosecond duration have, to the first glimpse, surprising consequences when performing the experiments in reactive atomspheres. The synthesis of carbon nitride is a good example. The prediction of the unique properties of the ȕ-C3N4 phase by Liu and Cohen in the late 1980s51 initiated tremendous efforts to synthesize this “magic” compound. By ablating carbonaceous targets in nitrogen containing atmospheres carbon nitrides, CNx with × 0.7, i.e. well below its nominal 57 atm % N content could only be obtained.52,53 There were speculations that when producing more energetic carbon species the efficiency of CN formation could be promoted. The experiments performed using femtosecond lasers had puzzling results: Films with maximum nitrogen contents less than 15 atm %, i.e. much lower than those fabricated using nanosecond pulses, could be produced.54,55
ULTRASHORT PULSE PLD
137
Recalling that compound formation materializes as a result of collisions of the carbon plasma with the N2 molecules, it is plausible that the highest nitrogen content can be reached when single C atoms/ions meet N2 molecules or (excited) nitrogen atoms. The interaction of C clusters with nitrogen results in a compound with smaller N content. Meeting of carbon nanoparticles and nitrogen species is the worst case scenario in terms of carbon nitride chemistry. The differences in the plasmas produced by nanosecond and femtosecond pulses therefore explain the apparently surprising result. While the ablation of carbonaceous targets with femtosecond pulses can not produce species rich in N, it may well be that their nitrogen content originates from the outer surface of the nanoparticles. If that is the case, ablation by femtosecond pulses in reactive atmospheres could be a dedicated technique for the production of core-shell nanoparticles. 4.5. SOLID NANOPARTICLE FILM FROM A LIQUID TARGET
Very recently Szörényi and coworkers fabricated Si-doped amorphous carbon nanofilms by ablating a commercial silicone oil with extremely clean 700 fs @248 nm pulses of 4–5 × 1011 W cm–2 intensity in high vacuum.56 AFM and SEM unambiguously proved that the surface of the films, deposited between RT and 250ºC, consisted of nanoparticles (Figure 14). The formation of nanoparticles suggests that the ablation mechanism of a liquid target might be very similar to that of a solid one. The results of a parallel TEM study confirmed that on >100 nm-scale the films were morphologically inhomogeneous, indeed (left panel of Figure 15).
Figure 14. a-C:Si nanoparticle films deposited onto silicon substrate at RT (left) and 200ºC (right) by ablating silicone oil with 700 fs @248 nm pulses of 12.2 and 14.7 mJ total energies, respectively, focused onto 4.1 mm2 areas.
138
T. SZÖRÉNYI AND Z. GERETOVSZKY
Figure 15. Top-view HRTEM images of a free standing a-C:Si film. The blurred contrast is due to defocus and noise filtering.
Even at the highest magnification used (right panel of Figure 15) no ordering could be detected, revealing that at the nanoscale the material is homogeneous and isotropic, possessing a perfectly amorphous structure. 4.6. ABLATION IN VACUUM. EXPERIMENT VERSUS THEORY
The results analysed above suggest that ablation with ultrashort laser pulses is characterized by a number of general features, independently of the nature of the target material. In particular, ablation at intensities not much higher than the plasma formation threshold, more specifically within the 3 × 1011 – 2 × 1013 W cm–2 domain, inevitably leads to the generation of nanoparticles of the target material. This characteristic seems to be true not only for elemental, but also for multicomponent (e.g. binary or even more complex) materials. These general features are consistent with the physical picture of the process, as outlined by the predictions of recent theoretical analyses.57,58 At laser intensities in the range of 1012–1013 W cm–2, corresponding to initial temperatures of a few electron volts, the (nearly) adiabatic cooling drives the material into a metastable phase, and results in the production of a relatively large fraction of nanoparticles through phase decomposition processes. On the other hand, at larger intensities (>1014 W cm–2) the material can never reach the metastable phase, resulting in an almost fully atomized plume. Thus the most promising laser intensity range in the context of nanoparticle production is theoretically predicted to be 1012–1013 W cm–2. Though further comparative studies are inevitably necessary for judging how general this statement could
ULTRASHORT PULSE PLD
139
be, the intensity values predicted by the theoretical approaches turn out to be in very good quantitative agreement with the limited experimental works’ that led to an efficient generation of nanoparticles. 4.7. NANOPARTICLE PRODUCTION IN LIQUID ENVIRONMENT
As it was shown in the case of nanosecond ablation, ablation in liquids offers a clean and efficient way for colloidal nanoparticle production. Now, the limited number of studies and the immature state of femtosecond ablation in liquid ambient do not allow us to draw solid and general conclusions.34,59 One can state though, that with femtosecond pulses – as compared to the nanosecond case – nanoparticles with smaller mean sizes and lower dispersion can be produced, but at much lower production yield.34,59–61 It sounds feasible that in liquid environment two ablation mechanisms compete, depending on the intensity.59,61 The first mechanism manifests itself at relatively low intensities (e.g. at I < 3.6 × 1015 W/cm–2 in the case of ablation of a gold target in pure deionized water) and produces almost monodisperse gold colloids of very small (~3–10 nm) sizes, as shown in the inset of Figure 16. The second process becomes active at high intensities and produces particles of much larger size and broader size distribution.
Figure 16. Size distribution of gold nanoparticles prepared by ablating a gold target under deionized water with 110 fs pulses of 5.5 × 1014 W cm–2 intensity.61 Inset shows a respective TEM micrograph.
140
T. SZÖRÉNYI AND Z. GERETOVSZKY
Similarly to the nanosecond case, in chemically active liquids enhanced stabilization can be established and leads to more efficient control of nanoparticle size.62,63 The first results are quite impressive and support the assumption that ablation with ultrashort pulses in both low pressure and liquid ambients could be a competitive approach to nanoparticle production. 5. Conclusions-Outlook Laser ablation is a general and practical route for the fabrication of nanoparticles and nanoparticle films of both elemental and multicomponent materials in vacuum, gas or liquid environments. From the point of view of the ablation products the principal difference between ablation with traditional, nanosecond and ultrashort, femtosecond lasers is that while in the former case a condensed ambient is indispensable for nanoparticle production (either as a liquid or in the form of a gas atmosphere at pressures above approximately 1 Pa), in the latter one this can even be accomplished under vacuum conditions. In the case of nanosecond ablation the dominating species leaving the target surface are atoms and ions, and cluster formation and growth take place in the plume. When ablating with femtosecond pulses of intensities not far above the threshold of plasma formation, the ablated material leaves the target in form of nanoparticles, i.e. ablation with femtosecond pulses directly produces nanoparticles and therefore independent from the nature and the pressure of the ambient. The properties of nanoparticles can be further tuned by choosing an appropriate environment, making in-situ engineering possible, and thereby opening up the way to produce more exotic (e.g. coated or functionalized) nanoparticles.
References 1. Chrisey, D., and Hubler, G. (1994), Pulsed Laser Deposition of Thin Films, Wiley, New York. 2. Schou, J. (2006), Laser beam-solid interactions: fundamental aspects, in Y. Pauleau, Materials Surface Processing by Directed Energy Techniques, Elsevier, Amsterdam. 3. Kroto, H., Heath, J., OcBrien, S., Curl, R., and Smalley, R. (1985), C60: Buckminsterfullerene, Nature 318, 162–163. 4. Bulgakova, N., Burakov, I., Meshcheryakov, Y., Stoian, R., Rosenfeld, A., and Hertel, I. (2007), JLMN-J. Laser Micro/Nanoeng. 2(1), 76–86. 5. Yudasaka, M., Ichihashi, T., Komatsu, T., and Iijima, S. (1999), Chem. Phys. Lett. 299, 91–96. 6. Thess, A., Lee, R., Nikolaev, P., Dai, H., Petit, P., Robert, J., Xu, C., Lee, Y., Kim, S., Rinzler, A., Colbert, D., Scuseria, G., Tomanek, D., Fisher, J., and Smalley, R. (1996), Science 273, 483–487.
ULTRASHORT PULSE PLD
141
7. Rinzler, A., Liu, J., Dai, H., Nikolaev, P., Huffman, C., Rodriguez-Macias, F., Boul, F., Lu, A., Heymann, D., Colbert, D., Lee, R., Fisher, J., Rao, A., Eklund, P., and Smalley, R. (1998), Appl. Phys. A 67, 29–37. 8. Dillon, A., Parilla, P., Jones, K., Riker, G., and Heben, M. (1998), A comparison of singlewall carbon nanotube production using continuous wave and pulsed laser vaporization, in R. K. Singh, D. H. Lowndes, D. B. Chrisey, E. Fogarassy, and J. Narayan (eds.), Advances in Laser Ablation of Materials. Mater. Res. Soc., Warrendale, PA, USA, pp. 403–408. 9. Kokai, F., Takahashi, K., Yudasaka, M., Yamada, R., Ichihashi, T., and Iijima, S. (1999), J. Phys. Chem. B 103, 4346–4351. 10. Puretzky, A., Geohegan, D., Fan, X., and Pennycook, S. (2000), Appl. Phys. A 70, 153–160. 11. Puretzky, A., Schittenhelm, H., Fan, X., Lance, M., Allard, L., Jr., and Geohegan, D. (2002), Phys. Rev. B 65, 245425. 12. Puretzky, A., Geohegan, D., Fan, X., and Pennycook, S. (2000), Appl. Phys. Lett. 76, 182. 13. Amoruso, S., Toftmann, B., and Schou, J. (2004), Phys. Rev. E 69, 056403. 14. Geohegan, D. (1992), Thin Solid Films 220, 138–145. 15. Amoruso, S., Toftmann, B., Schou, J., Velotta, R., and Wang, X. (2004), Thin Solid Films, 453/454, 562–572. 16. Henley, S., Carey, J., Silva, S., Fuge, G., Ashfold, M., and Anglos, D. (2005), Phys. Rev. B 72, 205413. 17. Geretovszky, Z., Haraszti, T., Szörényi, T., Antoni, F., and Fogarassy, E. (2003), Appl. Surf. Sci. 208–209, 566–574. 18. Henley, S., Carey, J., and Silva, S. (2004), Appl. Phys. Lett. 85(25), 6236–6238. 19. Bolgiaghi, D., Miotello, A., Mosaner, P., Ossi, P., and Radnóczi, G. (2005), Carbon 43, 2122. 20. Rode, A., Gamaly, E., and Luther-Davies, B. (2000), Appl. Phys. A 70, 135–144. 21. Henley, S., Carey, J., and Silva, S. (2006), Appl. Phys. Lett. 89, 183120. 22. Kabashin, A., Sylvestre, J., Patskovsky, S., and Meunier, M. (2002), J. Appl. Phys. 91, 3248. 23. Szörényi, T., and Fogarassy, E. (2003), J. Appl. Phys. 94, 2097–2101. 24. Szörényi, T., Antoni, F., Fogarassy, E., and Bertóti, I. (2000), Appl. Surf. Sci. 168, 248–250. 25. Hu, J., Yang, P., and Lieber, C. (1998), Phys. Rev. B 57, R3185. 26. Alvarez, F., dos Santos, M., and Hammer, P. (1998), Appl. Phys. Lett. 73, 3521–3523. 27. Walters, J., Kühn, M., Spaeth, C., Dooryhee, E., and Newport, R. (1998), J. Appl. Phys. 83, 3529–3534. 28. Spaeth, C., Kühn, M., Richter, R., Falke, U., Hietschold, M., Kilper, R., and Kreissig, U. (1998), Diamond Relat. Mater. 7, 1727–1733. 29. Fogarassy, E., Szörényi, T., Antoni, F., Stoquert, J., Pirio, G., Olivier, J., Legagneux, P., Boher, P., and Pons-Y-Moll, O. (2002), Appl. Surf. Sci. 197–198, 316–320. 30. Fogarassy, E., Szörényi, T., Antoni, F., Pirio, G., Olivier, J., Legagneux, P., and Boher, P. (2003), Appl. Phys. A 76, 15–19. 31. Geszti, O, Radnóczi, G., Bertóti, I., Szörényi, T., Antoni, F., and Fogarassy, E. (2002), Appl. Surf. Sci. 186, 502–506. 32. Marine, W., Patrone, L., Luk'yanchuk, B., and Sentis, M. (2000), Appl. Surf. Sci. 154–155, 345–352. 33. Patrone, L., Nelson, D., Safarov, V., Sentis, M., and Marine, W. (1999), J. Lumin. 80, 217– 221. 34. Kabashin, A., and Meunier, M. (2006), Laser ablation based synthesis of nanomaterials and references therein, in J. Perriere, E. Millon, and E. Fogarassy (eds.), Recent Advances in Laser Processing of Materials, Elsevier, Amsterdam, pp. 1–36.
142
T. SZÖRÉNYI AND Z. GERETOVSZKY
35. Mafuné, F., Kohno, J., Takeda, Y., Kondow, T., and Sawabe, H. (2000), J. Phys. Chem. B 104(35), 9111–9117. 36. Mafuné, F., Kohno, J., Takeda, Y., Kondow, T., and Sawabe, H. (2000), J. Phys. Chem. B 104(35), 8333–8337. 37. Mafuné, F., Kohno, J., Takeda, Y., Kondow, T., and Sawabe, H. (2001), J. Phys. Chem. B, 105(22), 5114–5120. 38. Szorenyi, T., et al., Proceedings of the 4th International Conference on Laser Ablation: Appl. Surf. Sci. 127–129 and (2004), Proceedings of the 7th International Conference on Laser Ablation: Appl. Phys. A 79, 1373–1376. 39. Banks, P., Dinh, L., Stuart, B., Feit, M., Komashko, A., Rubenchik, A., Perry, M., and McLean, W. (1999), Appl. Phys. A 69, S347–S353. 40. Szörényi, T. (2006), Carbon-based materials by pulsed laser deposition: from thin films to nanostructures and references therein, in J. Perriere, E. Millon and E. Fogarassy (eds.), Recent Advances in Laser Processing of Materials, Elsevier, Amsterdam, pp. 76–104. 41. Amoruso, S., Ausanio, G., Barone, A., Bruzzese, R., Gragnaniello, L., Vitiello, M., and Wang, X. (2005), J. Phys. B: At. Mol. Opt. Phys. 38, L329–L338. 42. Amoruso, S., Ausanio, G., Bruzzese, R., Vitiello, M., and Wang, X. (2005), Phys. Rev. B 71, 033406. 43. Amoruso, S., Ausanio, G., Bruzzese, R., Gragnaniello, L., Lanotte, L., Vitiello, M., and Wang, X. (2006), Appl. Surf. Sci. 252, 4863–4870. 44. Albert, O., Roger, S., Glinec, Y., Loulergue, J., Etchepare, J., Boulmer-Leborgne, C., Perrière, J., and Millon, E. (2003), Appl. Phys. A: Mater. Sci. Process. 76, 319–323. 45. Scuderi, D., Benzerga, R., Albert, O., Reynier, B., and Etchepare, J. (2006), Appl. Surf. Sci. 252, 4360–4363. 46. Teghil, R., D`Alessio, L., De Bonis, A., Galasso, A., Villani, P., and Santagata, A. (2006), Thin Solid Films 515, 1411–1418. 47. Amoruso, S., Bruzzese, R., Wang, X., Nedialkov, N., and Atanasov, P. (2007), Nanotechnology 18, 145612. 48. Mannion, P., Favre, S., Ivanov, W., O’Connor, G., and Glynn, T. (2005), Experimental investigation of micromachining on metals with pulse durations in the range of the electronphonon relaxation time (pico to sub-picosecond), Proceedings of the Third International WLT-Conference on Lasers in Manufacturing, Munich. 49. Amoruso, S., Ausanio, G., Bruzzese, R., Lanotte, L., Scardi, P., Vitiello, M., and Wang, X. (2006), J. Phys.: Condens. Matter 18, L49–L53. 50. Amoruso, S., Ausanio, G., de Lisio, C., Iannotti, V., Vitiello, M., Wang, X., and Lanotte, L. (2005), Appl. Surf. Sci. 247, 71–75. 51. Liu, A., and Cohen, M. (1989), Science 245, 841–842. 52. Muhl, S., and Méndez, J. (1999), Diamond Relat. Mater. 8, 1809–1830. 53. Perrone, A. (2002), Jpn. J. Appl. Phys. 41, 2163–2170. 54. Acquaviva, S., Perrone, A., Zocco, A., Klini, A., and Fotakis, C. (2000), Thin Solid Films 373, 266–272. 55. Geretovszky, Z., Kántor, Z., Bertóti, I., and Szörényi, T.(2000), Appl. Phys. A 70, 9–11. 56. T. Csákó et al., Appl. Surf. Sci., to be published. 57. Eliezer, S., Eliaz, N., Grossman, E., Fisher, D., Gouzman, I., Henis, X., Pecker, S., Horovitz, Y., Fraenkel, M., Maman, S., and Lereah, Y. (2004), Phys. Rev. B 69, 144119. 58. Vidal, F., Johnston, T., Laville, S., Barthélemy, O., Chaker, M., Le Drogoff, B., Margot, J., and Sabsabi, M. (2001), Phys. Rev. Lett. 86, 2573. 59. Barcikowski, S., Hahn, A., Kabashin, A., and Chichkov, B. (2007), Appl. Phys. A 87, 47–55.
ULTRASHORT PULSE PLD 60. 61. 62. 63.
143
Tsuji, T., Kakita, T., and Tsuji, M. (2003), Appl. Surf. Sci. 206 314–320. Kabashin, A., and Meunier, M. (2003), J. Appl. Phys. 94, 7941–7943. Kabashin, A., Meunier, M., Kingston, C., and Luong, J. (2003), J. Phys. Chem. B 107, 4527–4531. Sylvestre, J., Poulin, S., Kabashin, A., Sacher, E., Meunier, M., and Luong, J. (2004), J. Phys. Chem. B 108, 16864–16869.
LASER ABLATION AND LASER INDUCED PLASMAS FOR NANOMACHINING AND MATERIAL ANALYSIS D. BATANI* Università di Milano Bicocca, ITALY
Abstract – The goal of this paper is to introduce the basics of laser ablation. After introducing some general concepts for the description of lasers, and their usefulness in laser micromachining, and of the materials with which they interact, we describe in more details the femtosecond and nanosecond pulse duration regimes.
Keywords: Laser ablation, spatial coherence, ablation threshold flux, plasma frequency, Spitzer’s law, hydrodynamics, collisional absorption, crater formation
1. Introduction Soon after the invention of lasers, these systems begun to be used for applications in machining. Initially CO2 lasers were successfully used for efficient laser cutting of metals (and other materials) in various shapes. CO2 lasers are still probably the most used system at the industrial level. Laser cutting is nowadays so used that it deserved a voice in Wikipedia. More recently lasers have been used in applications like micro-hole drilling, creation of surface structures, introduction of calibrated leaks in pharmaceutical packaging (so to have a prolonged and controlled release of medicaments), selective material removal (e.g. polymer jacket from optical fibers), engraving for code marking or decoration, 2.5D micro milling (i.e. creation of structures in depth in materials), 3D engraving of glass and other transparent materials, laser cleaning in art conservation, guaranteeing high precision and high reproducibility in all cases.
______ *
To whom correspondence should be addressed: D. Batani, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
145
146
D. BATANI
Advantages of using lasers include economic advantages (reduction of work times; increase in production quality) but above all technological advantages: laser machining is precise, clean and silent; the beam can be focused on extremely small areas; the area in proximity to the edge may have very low heat alteration; the laser cut has the capacity of operating on complex profiles and with very small rays of curvature. Also, the laser is a non-contact instrument that guarantees absence of mechanical pressure on the piece (unlike water and traditional cutting systems); working capability independent of hardness of the material; capability of cutting coated or surface treated materials; no contact contamination of materials; no wear on the laser (which affects precision) Laser cutting also has a high degree of automation and flexibility able to offer ease of integration with other automated systems; and capability of adapting immediately to changes in production requirements. Finally, in many cases, laser cutting can produce finished pieces that do not require further processing. All these applications of lasers are based on ablation i.e. the capability to remove matter from a substrate by irradiating it with sufficiently high intense radiation. Focusing a high intensity laser on a solid target brings many effects: (1) ablation (removal) of material (hole drilling, micromachining, surface modifications), (2) emission of radiation (this can be used for plasma and material diagnostics but also for developing pulsed radiation sources, in the XUV and soft X-ray range), and (3) redeposition of the ablated material on a substrate (thin film production, PLD, pulsed laser deposition). In this paper, I will present only the first point and describe the basics of laser ablation in the femtosecond and nanosecond pulse duration regime. For PLD please see the contribution by J. Schou in this book. 2. The Laser System Why are lasers good for micromachining? This depends on characteristics of laser vs. “normal” light. Usual light is characterized by: low directionality, low mono-chromaticity, low coherence, low power (thermal source). Laser light is instead characterized by: high directionality, high mono-chromaticity, high degree of spatial coherence, high power (being non-thermal sources, lasers are not constrained by Kirchoff’s law, i.e. the emitted power may exceed the black body limit at a given wavelength). As it is well known, the peculiarities of laser light comes from two “ingredients”, i.e. (i) the different emission mechanism (stimulated vs. spontaneous emission), and (ii) the role of the optical cavity. Lasers appear as an ideal tool for concentrating energy in space and in time thereby realizing huge irradiation intensities with little energy, and higher intensities imply larger effects on materials. The intensity on target is given by
LASER ABLATION
147
I(W/cm2) = E(J) / (S(cm2) W(s))
(1)
where E is the laser pulse energy, S the focal spot area and W the pulse duration. Focusability is related to spatial coherence and hence to beam quality: Lasers are spatially coherent therefore the beam can be focused to a very small spot. This implies a precise cut but also saving of energy because intensity is inversely proportional to focal spot area. Due to diffraction, focal spot diameter decreases with the wavelength O. With perfect beams and aberration-free focusing optics, we get d = 2.44 O f / D = 2.44 O/F#
(2)
where F# = D/f (the constant 2.44 holds is flat-top laser beam profile in the near field, otherwise a different number appears, always of the order of unity). Provided their optical quality is good, short wavelengths offer the possibility of improved resolution and also of allowing higher intensities on target I v (F#/O)2
(3)
The other way of increasing laser intensity is by reducing the pulse duration W. With this respect after the introduction of Q-switching and mode-locking in the 1970s, which allowed pulses in the picosecond-regime to be produced, the recent breakthrough is the well known introduction of the Chirped Pulse Amplification (CPA) technique1 in the 1990s, which has allowed the generation of laser intensities as high as 1021 W/cm2. Such values are of course not directly interesting for laser-ablation and here we will limit ourselves to much lower irradiation values. In any case CPA has made possible to extend laser ablation to the femtosecond-pulse duration regime. TABLE 1. Laser parameters, in red the useful range for laser ablation.
Wavelength Repetition frequency Average power Energy per pulse Pulse duration Peak power Intensity (irradiance)
Visible, near IR, near UV CW, 100 MHz, MHz, kHz , Hz, single shot MW, W, kW pJ, nJ, mJ, J, kJ μs, ns, ps, fs MW, GW, TW mW/cm2, 109, 1016, 1021 W/cm2
O Q P E = P/Q W Po = E/W I = P/S
Several parameters are useful for a discussion of laser performance in ablation, and in all cases the variation extends over several orders of magnitude: see Table 1 (in red the useful range for laser ablation). As for the choice of the laser-type (i.e. the active medium and the pumping tool) the systems, which are more largely used in today’s laser ablation and laser micromachining, are gas
148
D. BATANI
lasers (especially excimers due to their high efficiency and short wavelength), and solid-state lasers (Nd:glass, Nd:YAG and Nd:YLF and Ti:Sa for the femtosecond regime). For such brief discussion, it appears that the “ideal” laser for ablation and micromachining should have the following characteristics: high beam quality; short wavelength; short pulse duration; high energy per pulse (and high absorption); high repetition frequency; and finally high conversion efficiency (conversion from plug-in to laser energy which implies reduced costs). 3. The Material After describing the laser, let’s now present the other principal actor of laser ablation, i.e. the material. First of all, we must notice how ablation is the removal of matter from the target (i.e. of atoms). But atoms are heavy and inertial so that laser light preferentially interacts with the electrons in the material, while the atoms are heated later due to energy exchange processes. Therefore first we need to characterize the electrons in the materials (Figure 1) and their interaction with light, and then the energy transfer to ions.
Figure 1. Electronic structure in the material.
Let’s first consider the case of metals (we will briefly speak about the case of insulators in next section). As it is well known, electrons are packed in the conduction band and only electrons at the top of the band can effectively exchange energy (the others being constrained by Pauli’s exclusion principle). Electrons in the conduction band are characterized by the value of the Fermi Energy. In practical units EF (eV) § 3 10–7 (ne(cm–3))2/3
(4)
where ne is the density of free electrons in the conduction band. Usually EF is a few electron volts, for instance for aluminum it is 11.63 eV, for copper 7 eV.
LASER ABLATION
149
Another important value is the extraction work ), i.e. the energy difference between the most energetic electrons in the conduction band and the continuum. As for the atoms in the material, they are bonded to the lattice with energies which are characteristics of the materials, but typically Ebond § 1 eV to a few electron volts. This defines an ablation (evaporation) heat: :§ (ni Ebond) = U NA Ebond/A
(5)
where ni is the atom (ion) density in the material, U the material density, NA§ 6 1023 the Avogadro number, A the atomic weight of the material. For instance in the case of Al, U/A§0.1 and assuming Ebond§1eV =1.6 10–19 J we get :§9,000 J/cm3. The corresponding value per unit mass is :U§ 3,300 J/g or, in general :U§ NA Ebond/A. Let’s also notice that the density of ions an free electrons in the material are related by ne§Z* ni, where Z* is the ionization degree or the average number of electrons in the conduction band per ion. Finally we will need to describe thermal conduction in the material and its own time scale Wth. Since Wth is finite and, as we will see, of the order of a few 10 ps, we can already infer that a fast deposition of laser energy (short pulse W, femto and picoseconds) will not allow spreading and penetration of energy to large volumes. In this regime therefore high intensity ablation (I > 1013 W/cm2) will be characterized by direct evaporation of the matter and negligible thermal effects to the surrounding material, resulting in high precision. On the contrary, for very long pulses (microseconds or sub-microseconds) ablation will take place at low laser intensities and evaporation will be accompanied by fusion of material resulting in low precision. To this we must add the fact that in general short pulse lasers are also characterized by a better beam quality, as compared to longer pulse ones. 4. Laser Ablation in the Femtosecond Regime Although of course laser ablation in the femtosecond regime is recent, follow-ing the introduction of femtosecond lasers, we start by describing it because it is a clearer process. We will also see how the description also applies to the few picosecond regime. Laser ablation is a multi step process. First, energy is deposited in a region of thickness G (which, as we’ll see soon, can be identified with the skin depth) being absorbed by the free electrons in the material (electrons in the conduction band). Second, energy is transported by electrons to a thickness l (electron thermal conduction, diffusion process). Third, we have the interaction with the atomic lattice, heating of the atoms, breaking of the chemical bonds, and finally ablation of the material and the formation of a plasma plume. In order for this third step to take place, after thermalisation, ion energy must exceed the bond-
150
D. BATANI
ing energy: this condition can be expressed as CiTi > U: where classically Ci = 3/2ni with number of atoms per unit volume (ionic density). The steps are qualitatively illustrated in Figure 2. In the following we will illustrate the basic physical mechanisms determining the depths G and l. Before that, however, let’s notice one important point. Let’s assume that energy is deposited with an exponential profile in the material with a typical scale-length l, i.e. the energy flux (energy per unit surface) is F (J/cm2) = Fo exp (-x/ l)
Figure 2. Schematic of the ablation process. Energy is absorbed to a thickness G and then diffuses to a larger penetration depth l.
Then the energy deposited between xo and xo + dx over a surface S is
dE S
dF dx dx x x 0
S Fo exp(xo / l )
dx l
(6)
For ablation to take place, this deposited energy this must exceed the ablation heat U: in the volume dV = S dx
dE dV
§x · Fo exp¨ ¸t U: l ©l ¹
(7)
The distance x = L at which the two quantities are equal defines the depth of ablated material. By solving for L we get
§F · § F · L l ln¨ o ¸ l ln¨ o ¸ ©U:l ¹ ©Fth ¹
(8)
LASER ABLATION
151
where the quantity Fth (J/cm2) = U:l is the threshold flux for ablation. This must be corrected for the absorption coefficient of laser light A = 1-R, so that Fth = U:l/A. Let’s notice two important consequences of the logarithmic dependence of L over Fo: (1) Below Fth there is no ablation, but L = l when the logarithm is 1, i.e. when Fo = e Fth § 3 Fth. (2) If you are working with a high repetition frequency laser and above the threshold flux, increasing the inward energy flux Fo (therefore the laser inten-sity I) is not advantageous. Indeed with N shots at a given laser intensity I, the total ablation depth is §NL, while if we increase the intensity N times, the ablation depth only increases as ln(N) (Moreover increasing the intensity may increase unwanted “collateral” effects). Let’s finally observe that this model is 1D. However neglecting 2D dynamics and flows and assuming a radial Gaussian profile of the energy source we can also use the above law in order to derive an equation for the radial size of the ablated region:
F(r) Fo exp(r2 / 2wo2 ) t Fth §F · 2wo2 ln¨ o ¸ ©Fth ¹
r2
From which
(9) (10)
4.1. THE OPTICAL PENETRATION DEPTH G
The initial distance G over which the laser energy is absorbed is determined by the penetration of light in the material, i.e. for a metal, metal, the skin depth. The free electrons in the conduction band form a plasma with electronic density ne = Z* ni and the dispersion relation of the electromagnetic wave (laser) in the plasma, in the collisionless approximation is
Z 2 Z 2p c 2 k 2 4Snee 2 me
Zp
(11) (12)
where Zp is the plasma frequency, which defines the plasma refractive index
n
1
Z 2p Z
2
n 1 e nc
(13)
152
D. BATANI
where nc is the critical density at which the (angular) frequency of laser light equals the plasma frequency. In practical units Zp and nc are given by
Z p (Hz) 5.64 10 4 ne (cm3 ) nc (cm3 )
1.1 10 21 ( O(Pm))2
(14) (15)
Usually Zp !!Z and then the refraction index n becomes imaginary. Therefore the wave is evanescent in the material, penetrating only over a skin depth G § c/Zp For instance for Al, Z* § 3 and ni § 6 1022 cm–3. Then Zp § 2.4 1016 Hz and G § 12 nm. In reality, in a real case, the plasma will be collisional (even strongly collisional) and the plasma frequency may not be much higher than the laser frequency. Therefore determining the skin depth may become a complex problem. Finally, from G, we can calculate the average energy in the initial plasma contained in the volume ʌwo2G, which is
Ee
A EL 2 (Swo G ) niZ *
(16)
Now, as said before, electrons in the conduction band are characterized by the value of the Fermi Energy. As a guiding rule, we can assume that if Ee EF then the characteristics of the material are not affected. We call this “cold solid approximation”. On the contrary if Ee > EF, which happens at high laser intensities, then we will need to take into account modifications in the target characteristics and, in particular, to include plasma effects. Finally the case Ee << EF practically corresponds to no ablation (since EF, Ebond, ) are all of the same order, i.e. a few electron volts). 4.2. THE THERMAL PENETRATION DEPTH l
The next step is the description of the penetration of the absorbed energy in the material at a distance l, from the initially heated region of dimension G. This step is classically described by the so-called “two temperature model”.2 There are several quantities involved in the model: the relaxation time for electrons We, the relaxation time for the ionic lattice WI, the laser pulse duration W, the equilibrium temperature Teq i.e. the common temperature finally reached by ions and electrons in the volume of thickness l, the thermal diffusivity which is classically given by D = (1/3)v2We, the electron-lattice coupling constant J, the laser source flux S, and the heat flux Q(z) at a certain depth z inside the material.
LASER ABLATION
153
The equations of the 2-T model are simply the energy conservation equations for ions and electrons
wTe wQ( z) ° °Ce wt wz J (Te Ti ) S ® °C wTi J (T T ) e i ° ¯ i wt
(17)
In the electron equation, energy balance is given by heat propagating inside the material, by the relaxation towards ions (the lattice) and the flux of energy from the initially heated region. The ion equation only contains exchange with electrons (which is now positive acting as a source term). Initially electron temperature will increase due to the term S, then heat will propagate inside; at the same time ions will remain substantially cold. Finally due to relaxation electron temperature decreases as ion temperature increase until they approach the same limiting value:
Teq
§ z · § z ·º Fo 1 ª l exp¨ ¸ G exp¨ ¸» « Ci l 2 G 2 ¬ © l ¹ © G ¹¼
(18)
Finally, the heat penetration length will be calculated as
DW
l
(19)
The square root dependence is not a surprise this being a diffusion process, and also it is maintained by the laser pulse, therefore, unless it is too short, the time that appears here is just the laser duration. Let’s also notice that in the “usual” case l >> G we get the expression which we have written before, which yields the value of the flux ablation threshold Fth
CiTeq
§ z · Fo exp¨ ¸ l © l ¹
(20)
§ z · exp¨ ¸ G © G ¹
(21)
If instead case G >> l then we get
CiTeq
Fo
Penetration of energy is still exponential but it is now governed by optical propagation, yielding a value of ablation threshold Fth = U:G.
154
D. BATANI
4.3. CALCULATION OF THE ATOMIC PARAMETERS
Let’s consider as an example the case of Cu irradiated by laser light at 532 nm (2nd harmonic of Nd:YAG). Cu has a density of atoms ni = 8.5 1022 cm-3 and Z* = 1, therefore ne = ni. This corresponds to Zp= 1.6 1016 Hz, and G = 1.8 10-6 cm = 18 nm. Also such electronic density gives EF = 7 eV and vF = 1.57 108 cm/s. This is the velocity of the electrons at the top of the conduction bands, those with energy EF and the only ones that can exchange energy. Therefore it seems appropriate to substitute v = vF in the expression for the thermal diffusivity D. The calculation of the electronic relaxation time We requires some specifications. Such time is related to the dc electrical conductivity V by the well-known expression
V
nee 2W e me
Z 2pW e 4S
(22)
therefore it is a function of material density and temperature. Figure 3 reports the behavior of V for solid density Al for temperatures between room temperature and 10 keV. Experimental measurements have been made by Milchberg et al.3 by measuring the reflectivity of femtosecond-heated Al and calculating resistivity. These are compared to the classical Drude model and to a simplified quantum calculation (Eidmann-Huller model4). All curves gives the same behavior: first conductivity decreases and then increases again when the material temperature becomes so large that the plasma regime is reached. In this case conductivity scales with temperature according to the Spitzer’s model, i.e. Vv Te3/2. Let’s notice that the simple classical Drude model, although quantitatively wrong, gives the correct qualitative behavior. All curves show a minimum value of conductivity, called “saturated” conductivity. It is possible to show that this minimum corresponds to We= a/vF, where a is the inter-ionic distance in the material (Ioffe-Regel’s limit). In other terms, in this limit, the electrons moving at velocity vF make a collision with every atom they meet on their path. In the case of copper, we get a = ni-1/3 = 2.3 10-8 cm = 2.3 Å and We = a/ vF = 2.2 10-16 s = 0.22 fs. It is easy to see that such is an incredibly fast time, much shorter than even shorter laser pulses used in ablation. The ion relaxation time is much longer because the energy exchange between very light and very heavy particles is proportional to the mass ratio, i.e. is given by:
Wi which yields § 26 ps in our case.
miW e me
(23)
LASER ABLATION
155
109
Electrical conducibilty MKS (:m)
-1
experimental data from Milchberg et al. Quantum model by Eidmann and Huller Semiclassical Model (Drude + Spitzer)
108
Vcold Al = 3.7 107 :-1m -1
107
106
T (eV)
105 0.1
1
10
100
1000
4
10
Figure 3. Behavior of electrical conductivity vs. temperature for solid state Al.
Using these values we get for the thermal diffusivity D = (1/3) vF2We = 1.8 cm /s. Now, assuming as an example W= 40 ps (at the very upper limit of the validity range for the present model of laser ablation in the femtosecond-regime) we get l = (DW § 85 nm. Finally we can calculate the threshold flux for ablation as Fth = U:l/A = 1.2 J/cm2 where we have taken into account that A = 0.3 at 532 nm. Of course this is the value calculated for a polished and clean Cu surface, which will certainly be very different from a real rough and oxidized surface; however this is not too important: A appears inside a logarithmic terms so small differences in A will not change the value of Fth too much. Coming back to the cold solid approximation, we therefore see that in such approximation (Ee§EF) no characteristics of the material is strongly affected: the electronic relaxation time remains the same, there is no ionization above the normal value of Z* therefore the electronic density doesn’t change, the values of G, A, EF, vF don’t change. On the other end, as already said, if Ee<<EF, practically there is no ablation. 2
4.4. AN EXAMPLE OF MEASUREMENTS
As an example, let’s consider the results from an ablation experiment, which was performed at the LOA Laboratory (Palaiseau, France) with the goal of drilling holes in capillary tubes for gas inlet for an electron acceleration experi-ment.5 The used laser was a dye at 620 nm, pulse duration 80 fs (FWHM), repetition rate 10 Hz, energy per pulse <3 mJ. In the experiment §60 μJ of laser energy were delivered on target while the laser focal spot diameter was varied from 13
156
D. BATANI
to 30 μm giving a laser flux F on the target surface between 2 and 40 J/cm2, corresponding to a maximum laser intensity on target 5 1014 W/cm2. The energy reproducibility from shot to shot was ±10%. The targets used in the experiments were stainless steel and fused silica of 120 μm thickness and were irradiated in air at normal incidence. A shutter placed on the beam allowed to choose the number of pulses. The ablation depth was measured by exposing the target with 1,000 pulse. Quantitative measurements of ablation depth were then performed with optical and scanning electron microscopes and the ablation depth per shot was estimated as the total ablation depth (measured value) divided by number of shots (1,000). In this experiment, precise machining was not a requirement as compared to fast operation. Hole morphology is shown in Figure 4. Figure 5 shows instead the ablation depth results obtained by varying the energy flux on target.
Figure 4. Scanning electron microscope (SEM) images. (Left) Front surface of the ablation site after exposing the target surface with 1,000 laser pulses at F = 105 J/cm2. (Right) Rear surface of the ablation site after exposing the front side with a number of pulses sufficient to get complete drilling at F = 33.6 J/cm2.
The ablation results show a reduction of ablation depth due to the presence of air, which induces re-deposition of the ablated material (but the threshold flux remains the same). 4.5. THE CASE OF DIELECTRICS
In the case of dielectrics, the valence band is full and the con-duction band is empty, there are no free electrons. Light penetrates and it is attenuated with an exponential law (Lambert-Beer law) with an attenuation coefficient D, i.e. I(z) = Io exp(-Dz). Therefore we can replace G with 1/D.
LASER ABLATION
157
Ablation Depth per Shot (Nm)
0,2
0,15
Nolte’s model (with Fth = 0.335 J/cm2 and I = 44 nm)
0,1
0,05
0
5
10
20 25 15 Flux (J / cm2)
30
35
40
Figure 5. Ablation depth (μm) per pulse vs. laser flux F on target: experimental results (black points) compared to (a) an interpolation L (nm) = 9 ln (F/Fth), and (b) to the formula L = l ln (F/Fth) where l = 44 nm is calculated with the 2T model. In both cases Fth § 0.335 J/cm2.
Figure 6. Effect of multiphoton absorption in dielectrics.
However due to the high laser intensity photons are absorbed by multiphoton absorption so that electrons cross the gap and move to the conduction band or even by multiphoton ionization and go to the continuum (free electrons), as in Figure 6. However, when the density of created free electrons reach the critical density nc, the laser can no longer propagate in the plasma and the process stops (self-regulation). Therefore with high-intensity short-pulse lasers there is no substantial difference between insulators and conductors with a free electron density nc. 4.6. HIGH INTENSITIES: PLASMA EFFECTS
At higher laser intensities (>1013 W/cm2) and short pulses, we get Ee > EF. This results in strong heating of the material and plasma effects: (1) ionization of
158
D. BATANI
material above the normal value Z: the electronic density changes inducing changes in G, A, EF, vF; (2) high temperature of the initial plasma. Formulas for heat transfer derived from solid state physics are no longer applicable. In particular there is an increase in the velocity of electrons (vF is replaced by electronic thermal velocity vth) and there is an increase in We for which we must now use Spitzer’s law giving We in a plasma
We
10 6 Te3 / 2 s 2.91 ne ln /
(24)
The plasma ionization degree should be calculated in a self consistent way. Figure 7 shows the average ionization degree Z* for gold at density U = 0.1 g/cm3 vs. temperature in various approximation or using different formulas. 80 70 60
Z*
50 40 30 20 10 0 10
1000
100
104
T (eV)
Figure 7. Average ionization degree Z* for gold at density U = 0.1 g/cm3. Squares: coronal limit; circles: Thomas Fermi model; solid lines: analytical formula for coronal equilibrium (CE); dashed line: analytical formula by Colombatnt and Tonon6 (valid in Collisional Radiative Equilibrium, CRE).
In principle, we should also calculate the energy lost to ionization by summing over the different ionization energies for each electron in Cu. (shown in Figure 8). However it is possible to show that usually the total energy lost to ionization Eion is negligible with respect to the thermal energy. The following table shows a comparison of the ablation parameters calculated in the cold solid approximation and in plasma model. The pulse duration is 400 ps. In the plasma case, values are calculated for an intensity I § 1014 W/cm2 giving a temperature Te § 500 eV.
LASER ABLATION
159
104
1000
100
10
1 0
5
10
15
20
25
30
Figure 8. Ionization energy of each electrons for Cu ions. TABLE 2. Comparison of ablation parameters calculated in the cold solid approximation and taking the plasma effects into account.
Z* Free electron density ne Electron velocity ve Electron relaxation time We Penetration depth l Ablation threshold flux Fth = U:l/A
Cold solid approximation 1 8.5 u 1022 cm–3 vF = 1.6 108 cm/s 2.2 u 10–16 s 0.1 μm 1.2 J/cm2
Plasma model 18 – 19 15 u 1023 cm–3 vth = 1.5 109 cm/s We (Spitzer) = 10-15 s 2 μm 70 J/cm2
The transition between the cold solid approximation and the plasma case is qualitatively illustrated in Figure 9. The plasma case onset corresponds to a higher ablation threshold flux, but gives larger ablation depths L. L L
§ F · o ¸ lplasma ln¨ ¸ ¨F © th plasma ¹
L
§ F · lcold ln¨ o ¸ ©Fth cold ¹
Figure 9. The transition between the cold solid approximation and the plasma case The plasma case onset corresponds to a higher ablation threshold flux, but gives larger ablation depths L.
160
D. BATANI
As an illustration of ablation measurements in the plasma case, let’s consider the following measurements performed at Milano Bicocca.7 A Nd:YAG (Quanta System) laser was used with maximum energy 60 mJ per pulse, 10 Hz repetition frequency, working in the green (532 nm) with focal spot radius wo=25 μm and pulse duration W = 40 ps giving a maximum intensity on target I = 1014 W/cm2. Figure 10 shows the depth and diameter of the ablation region on Cu targets in air and vacuum, as a function of energy flux. Figure 11 instead compares results obtained in vacuum for Copper and Lead. In both cases, results were obtained by analyzing the irradiated samples with optical microscopy. First, data show that the presence of air decreases the ablation depth of about 40%: as in the case of the previous experimental results obtained at LOA (Figure 5) this is due to the re-deposition of the ablated material (while the threshold flux is not affected). From available data, we know that for Cu U = 8.96 g/cm3 and : = 4723 J/g, while for Pb U = 11.35 g/cm3 and : = 859 J/g. Then interpolation of data yields the following penetration depths and ablation threshold: Cu, l | 5 Pm and Fth | 70 J/cm2, Pb, l | 4 Pm and Fth | 7 J/cm2. The ablation rate for lead is larger than for copper but this is due to the different in the evaporation rate while the heat penetration rate is practically the same (indeed this is mainly a function of plasma conditions, including electron temperature, which are mainly fixed by the char-acteristics of the laser pulse). Also, we see that in this experiments, the penetra-tion lengths are of the order of a couple of micrometer, as indeed it should happen in the plasma case, and not of several 10 nm, as in the cold solid approximation.
Diameter Depth (Pm)
30
(Pm)
160 140
VACUUM
25 120
VACUUM 20
100 80
15
60 10 40 5
AIR
AIR
20 0
0 0
1000
2000
3000
Flux (J/cm2)
4000
5000
0
1000
2000
3000
4000
5000
Flux (J/cm2)
Figure 10. Depth and diameter of the ablated region on Cu vs. energy flux on target. Data have been interpolated with the formulas introduced in Section 4.
Alternatively, the ablation depth per shot can be measured by counting the number of shots, which is needed to drill a clear hole in a foil with a given
LASER ABLATION
Diameter
Depth (Pm)
161
( Pm)
40
200
LEAD
35 150
LEAD
30 25
100
20 15
COPPER 50
10
COPPER
5 0
0 0
1000
2000
3000
4000
Flux (J/cm2)
5000
0
1000
2000
3000
4000
5000
Flux (J/cm2)
Figure 11. Depth and diameter of the ablated region for Cu and Pb vs. laser flux in vacuum.
Figure 12. AFM images of surface modifications induced on the irradiated Al target.
thickness. For instance, in the same experiments, we measured that 53 shots at the maximum energy flux are needed to drill a hole in a Cu foil 1-mm thick. This implies an ablation depth per shot of 19 μm, in fair agreement with the measurements reported in Figure 11. In the same conditions, the measured hole diameter was 120 μm. Finally, let’s notice that removal of material is of course not the only consequence of laser ablation. Indeed the characteristics of the material surface are also modified, an aspect that is also often used in application8,9 (e.g. surface hardening, surface cleaning, surface polishing). Let’s see for instance the analysis of the irradiated Al surface obtained with the same laser system. The Atomic Force Microscopy (AFM) pictures in Figure 12 (40 u 40 μm and 80 u 80 μm respectively) show two samples irradiated at Milano Bicocca (10 laser shots on an Al surface). The laser induced a change in the surface: the smooth surface was not irradiated, the part with “bubbles” was irradiated). The second picture shows a granular and irregular surface following laser irradiation.
162
D. BATANI
5. Laser Ablation in the Nanosecond Regime The situation for laser ablation in the nanosecond regime is very different: there is indeed one very important point, which is implicit in what we said until now. During the duration of the laser pulse W, matter has no time to move, there-fore the ionic density ni remains constant. Indeed the sound velocity is
cs
JZ * Te mi
9.8 10 5
JZ * Te (eV ) A
cm / s
(25)
Even if Te = 100 eV, we get cs = 7 106 cm/s, and then if W = 100 fs, then the typical extension of the plasma is L= cs W = 0.07 μm. This shows that there is no appreciable movement or expansion of the target material. The situation in the nanosecond regime is very different being completely dominated by hydrodynamics. Because matter has time to move, we have plasma expansion in vacuum. A low density plasma corona is formed in front of the target surface. The laser interacts with the plasma and it is absorbed in the plasma corona at the critical surface. Energy is then transported inwards by electronic thermal conduction (and/or radiation and/or suprathermal electrons). Matter is constantly removed from the solid at the ablation surface (where U = Uo of the cold solid). The plasma expands in vacuum and plasma expansion is fed by laser ablation. Beyond the ablation surface a shock is formed as a consequence of momentum conservation: the material is compressed by such ablation (shock) pressure. Shock front and ablation front moves inward in the material. The situation is evidenced in Figure 13. Ablation region
Plasma Expansion
ne ns
Overcritical region
LASER
Te ^ Kev nc corona
xs xa
xc
Figure 13. Typical density and temperature profiles established during nanosecond-laser interaction.
LASER ABLATION
163
It is possible to accurately describe the 1D hydrodynamics though analytical formulas obtaining scaling laws for the relevant quantities, the ablation pressure and the mass ablation rate.10,11 In practical units: 2/ 3 § 2 · § ·2 / 3§ A ·1 / 3 I (W / cm ) 1 ¸ ¨ P( Mbar ) 12.3¨ ¨ ¸ ¨ ¸ ©O (Pm) ¸ 14 ¹ ©2Z ¹ © 10 ¹ 1/ 3
§ · I (W / cm 2 ) ¸ m (kg / s cm 2 ) 150¨ ¨ ¸ 14 ¹ © 10
4/3
§ 1 · ¸ ¨ ©O (Pm) ¹
(26) 2/ 3
§ A · ¨ ¸ ©2Z ¹
= 32 kg/cm2s. For instance irradiation at 1012 W/cm2 with O = 1 μm yields m For a laser pulse duration W = 1 ns, this gives a mass removed for unit surface m W § 3 10-5 g/cm2. Finally for U = 2.7 g/cm3 (Al), this gives and ablated depth W/U = 0.12 μm. At 1014 W/cm2 it becomes 0.6 μm. d= m The ablation rate is a large number but the ablated mass and the ablated depth become very small because of the small focal spot and short pulse duration. However such ablated depth can be difficult to be measured because in this regime often the shock is also a very effective factor producing further ablation of material, as we will see in the following. Let’s also notice that ablation pressure can very easily reach the several 10 Mbar range. Indeed this kind of experiments are also used to compress matter to high pressures and study physical characteristics and equation of state of compressed matter.12 5.1. HYDRODYNAMICAL CODES
A number of 1D and 2D hydro codes are well adapted to describing the hydrodynamics and the ablation of nanosecond-laser irradiated targets, e.g. the code MULTI developed by R. Ramis et al. at Universidad Politecnica, Madrid.13,14 In Figure 14, we see the interaction of a 6 μm foil Al target with laser radiation with W = 3 ns at 800 nm and 1012 W/cm2. After shock breaks out on target rear side (at about 0.75 ns after the beginning of the laser pulse), the whole target is accelerated by the laser beam. At the same time a relaxation wave is generated and reaches the front side again. Since the laser pulse is still on at this time, then a second shock is generated and reaches the rear side just before 2.5 ns. The ablated mass corresponds to the rear side reaching low densities (green area). Of course if the target is massive, then the shock will never breakout on
164
D. BATANI
Figure 14. Interaction of a 6 μm Al foil with a laser with W = 3 ns at 800 nm and 1012 W/cm2.
rear side and only front side dynamics will be important. In this case relaxation will take place only at the end of the laser pulse, although some residual superficial compression may be maintained in the material (and actually it may be useful for surface hardening of the material). 5.2. THE QUESTION OF LASER ABSORPTION
The question of laser absorption is very important for an efficient laser-target coupling and then for an efficient ablation, as we have noticed in Section 2. While the absorption of femtosecond-lasers is still poorly understood, the situation is clearer for nanosecond-laser pulses In the corona laser energy is absorbed collisionally (i.e. by inverse bremsstrahlung). The absorption coefficient then reads,
D (cm1 ) 3.1 107 Z * ne2Z 2 ln / Te3/ 2 n 1
(27)
On this basis, very often we find in text books that IR lasers are absorbed more because of longer wavelengths (smaller Z). In reality this view is only partial and true only for very low intensities (or rather short pulses) for which an extended plasma corona does not develop. But as soon we get an extended corona (which is the main feature of nanosecond-interactions) this is not true. We must remember the plasma refraction index n becomes 0 as the critical density is approached, and also that the critical density is inversely proportional to the square of wavelength. Therefore, at shorter wavelengths, the laser can penetrate deeper in the plasma and reach higher intensities where absorption is higher.
LASER ABLATION
165
Experimental measurements by Fabre et al.15 performed already in the 1980s, clearly show such behavior (see Figure 15) A = absorption % 2.1010 2.1011 100
2.1012
intensity (W / cm2) 2.1013 2.1014 2.1015
0.53Mm 2ns
90 80 70 60
1.06Mm100 ps
1.06Mm 2.5ns
50 40 30 20 E(J) for 100 ps pulse 1/25E(J) for 2.5 ns pulse
10 0 10 4
10 3
10 2
10 1
1
10
Figure 15. Fraction of laser light absorbed in a plasma as a function of laser intensity on target, laser wavelength and pulse duration.
5.3. SHOCK EFFECTS AND CRATER FORMATION
At high laser irradiances in the nanosecond-regime, the shock wave becomes the dominant aspect of the interaction and leads to the formation of a crater.16 This is shown for instance in the SEM images in Figure 16. In general this regime is not interesting for micromachining, unless precision is not a requirement. However it is still very interesting from the point of view of physics.
Figure 16. SEM images showing the surface morphology of caters created on Al flat targets: (left) diameter D = 30 μm, flux § 2,200 J/cm2; (right) D = 100 μm, flux § 4,000 J/cm cm2.
166
D. BATANI
Shock dynamics is governed by Rankine-Hugoniot relations which represent the conservation laws for mass, momentum, and energy (per unit mass) across the shock front. For strong shocks, these read: Po UoUD
P
Uo D U ( D U ) 'E
E Eo
(28)
1 2 U 2
Here P, U, E, D and U are respectively the shock pressure, material density, internal energy per unit mass, shock and fluid velocity (i.e. the velocity of matter behind the shock front), and the suffix “o” refers to the unperturbed material ahead of the shock front. The condition for ablation requires the energy increase per unit mass to be larger than ablation threshold, i.e. 'E = U2/2 > :. Hydrodynamics is governed by scaling laws i.e. it is universal or scaleinvariant. Crater formation is governed by self-similar laws (i.e. power laws). The same laws concerning crater depth and crater diameter hold from micronsize laser crater to huge craters created by nuclear explosion and meteorite impacts (according to the theory formulated by Raizer in the approximation of a pointlike, in space and in time, energy release17). It is possible to derive the scaling law for crater size in the following way. The crater size Rcrat is determined by the condition 'E(Rcrat) = :. In the case of crater formation we have a spherical shock, which is expanding in the material starting from the laser focal spot, and thereby reduces its pressure according to:
§Ro ·2 Po ¨ ¸ © R ¹
P
(29)
In a material the shock and fluid velocity are usually related by the relation D = cs + SU where cs is the sound velocity in the material. For strong shocks gives D§ SU from which P = Po + UoDU §UoDU §UoSU2. Therefore
'E
U2 2
P 2UoS
2 Po §Ro · ¨ ¸ 2 U o S © R ¹
§Ro ·2 k ¨ ¸ © R ¹
and the crater size is
Rcrat
Ro Po / 2 Uo S:
Ro k / : Ro
Rcrat / k / :
We now integrate to get the total energy spent in creating the crater
(30)
LASER ABLATION
Eabs
³
Rcrat 0
167
§Ro ·2 k ¨ ¸ 2SR 2 dR 2SkRo2 Rcrat © R ¹
(31)
and by substituting the previous relation between Rcrat and Ro we finally get
Eabs
2SkRo2 Rcrat
2Sk Rcrat / k / :
Rcrat 2
3 2SkRcrat / : (32)
or in other words a cubic root dependence 1/ 3
Rcrat v Eabs
(33)
Under such conditions, the crater may (or indeed will) be much larger than the focal spot (i.e. the laser acts as a point source). 6. Conclusions We have revised the physical basics of laser ablation in the femtosecond and nanosecond pulse duration regimes. Laser ablation is a complex physical phenomenon which involves optical penetration of laser in the material and the transport of energy inside the material, a problem at the border line of solid state and plasma physics, which may imply quantum mechanical calculations. In the nanosecond-laser interaction regime, ablation is dominated by hydrodynamics: ablation feeds plasma expansion continuously and the laser interacts within the corona, producing an increased absorption at shorter wavelengths. Finally, shock dynamics and the formation of a laser crater is often a key phenomenon associated to high-intensity nanosecond-laser interactions.
References 1. 2. 3. 4. 5.
Perry, M., and Mourou, G. (1994), Science, 264, 917–924. Nolte, S. et al. (1996), J. Opt. Soc. Am. B, 14(10), 2716. Milchberg, H. et al. (1988), Phys. Rev. Lett., 61(20), 2364. Eidmann, K., Meyer-ter-Vehn, J., Schlegel, T., and Hüller, S. (2000), Phys. Rev. E, 62, 1202. Di Bernardo, A., Batani, D., Desai, T., Courtois, C., Cros, B., and Matthieussent, G. (2003), Laser Part. Beams, 21, 59–64. 6. Tonon, G., and Colombant, D. (1973), J. Appl. Phys., 44, 3524–3537.
168
D. BATANI
7. Desai, T., Batani, D., Rossetti, S., and Lucchini, G. (2005), Radiat. Eff. Defects Solids, 160(10–12), 595–600. 8. Trtica, M., Gakovic, B., Maravi, D., Batani, D., Desai, T., and Redaelli, R. (2006), Mater. Sci. Forum, 518, 167. 9. Trtica, M., Gakovic, B., Batani, D., Desai, T., Panjan, P., and Radak, B. (2006), Appl. Surf. Sci., 253, 2551–2556. 10. Lindl, J. (1995), Phys. Plasmas, 2, 3933. 11. Batani, D., Stabile, H., Ravasio, A., Desai, T., Lucchini, G., Desai, T., Ullschmied, J., Krousky, E., Juha, L., Skala, J., Kralikova, B., Pfeifer, M., Präg, A., Nishimura, H., Ochi, Y., et al. (2003), Phys. Rev. E, 68, 067403. 12. Batani, D., Stabile, H., Tomasini, M., Lucchini, G., Ravasio, A., Koenig, M., BenuzziMounaix, A., Nishimura, H., Ochi, Y., Präg, A., Hall, T., Milani, P., Barborini, E., Piseri, P., et al. (2004), Phys. Rev. Lett., 92, 065503. 13. Ramis, R., Schmalz, R., Meyer-ter-Vehn, J. (1988), Comput. Phys. Commun., 49, 475. 14. Ramis, R., and Meyer-ter-Vehn, J. (1992), MULTI2D-A computer code for two-dimensional radiation hydrodynamics, MPQ-174. 15. Garban-Labaune, C., Fabre, E., Max, C., Fabbro, R., Amiranoff, F., Virmont, J., Weinfeld, M., and Michard, A. (1982), Phys. Rev. Lett., 48, 1018–1021. 16. Bussoli, M., Batani, D., Desai, T., Milani, M., Trtica, M., Gakovic, B., and Krousky, E. (2007), Laser Part. Beams, 25, 121–125. 17. Zeldovich. Y., and Raizer, Y. (1967), Physics of Shock Waves and High Temperature Hydrodynamic Phenomena, Vol. 1, Academic, New York.
PHOTO-, DUAL- AND EXOELECTRON SPECTROSCOPY TO CHARACTERIZE NANOSTRUCTURES
Y. DEKHTYAR* Biomedical Engineering and nanotechnologies Institute, Riga Technical University, Kalku 1, Riga LV-1658, LATVIA
Abstract – Physical basics of prethreshold photo-, dual- and exo-electron spectroscopy to characterize nanostructures, as well as relevant measurement technique are described. Application cases are considered.
Keywords: Prethreshold photo-, dual- and exo-electron spectroscopy, nanostructures.
1. Introduction Advances in nanotechnologies need non-destructive characterization methods. To reach this, electron emission measurements, when energy to escape an electron from a tested object is not enough to destroy its molecular/atomic couples, are employed as such energies have values ~0.1–1 eV for non-destructive characterization. In such a case the excited electron has a mean free path (L) ~10–100 nm,1 correspondingly within the analysed emitter. These values of L fit sizes of nanostructures. This is a key point to apply electron emission for nanomaterials characterization. The paper is targeted to consider a prethreshold photo-, dual- and exo-electron emissions and their capabilities in respect with nanotechnologies. 2. Photoelectron Emission Analysis 2.1. PHYSICAL BASICS
A current (I ) of the prethreshold single photon photoelectron emission (PE) is described by the classical formula
______ *
To whom correspondence should be addressed: Y. Dekhtyar, e-mail,
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
169
Y. DEKHTYAR
170
conts(hQ M ) m
I
(1)
where hQ – energy of exciting photon, Melectron work function, m-power index. For the prethreshold emission mode (hQ t M the ҏvalues of M are equal to several electron volts,2 and therefore a mean pass of the photoelectron within the emitter is around 10–100 nm.1 To emit the electron it is excited from an initial (i) to a final (f) states that correspond to the energies Ei and Ef . The electron that leaves the specimen to the vacuum has the energy
E f Evac
W
(2)
Evac – energy of the vacuum level. Because of the energy conservation law
hQ
E f Ei
following main photoelectron spectroscopy methods could be employed. When hQ = const,
W
Ei hQ Evac ,
dI dW
dI dEi
(3)
the approach on initial states spectroscopy is provided. When Ef = const, hQ = var
dI d (hQ )
dI dE f
(4)
the spectroscopy of the final states is available. If Ei = const, however Ef = var and hQ = var with the condition 'Ef = 'hQ (' – means increment) (1) comes again to (4). When electron emission is delivered from the local states (Ei = const) and hQ = var, W
I ~ ³ N i ( E ) N f ( E , hQ )dE ,3
(5)
0
N i , N f – density of the initial and final electron states, correspondingly. Using
this way
dI ~ N i (W ) N f (W , hQ ) . dW
(6)
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
171
However,
dI dW
dI . dhQ
(7)
Because of (6) estimation of the N i , N f combination could be estimated. The electron work function in a case of emission from metals and semimetals that do not have an energy gap (Eg) is equal to an electron affinity (F). For the materials with Eg (semiconductors, dielectrics)
M Eg + F . The index m (formula (1)) characterises transition channel to excite the electron (see please the Table 1 below): TABLE 1. Index m in for different electron transitions. Material
Transition
Crystal
Direct Nondirect From the local level
Amorphous
Value of m
Source
1 2.5 1.5 >3
5 5 6 7
2.2. INSTRUMENTATION
To get the advantages (estimation of m, M and I(M) of the prethreshold PE measurements the photon energy should be as possible as close to the electron work function: hQ | t M. To supply the single photon electron emission mode and to avoid heating of a tested object, the flux of the photons should be rather weak. This stipulates small values of I. Because of this very sensitive electron detectors must be in use. Typically the secondary electron multipliers that have a noise 0.1–1 electron/second are applied. The value of F could be effected by ions/dipoles sorption/desorption on/from an emitter surface.4 This asks for stable and “passive“ electron emission measurements environment (preferably vacuum; 10–2…–7 Pa8). Typically the values of M are not distributed uniformly over a surface of the emitter. As a result a contact potential difference is induced between the surface arrears and electrical fields are induced. It decreases/increases I from the places with lower/higher M, correspondingly.4 In fact, when characterisation of M distribution is an aim of measurements, their results could become false. To avoid this, an external compensating electrical field ~102…3 V/cm directed from the surface to the detector should be provided.4
Y. DEKHTYAR
172
To measure energy of the escaped electron (for the approach by the formula (3)) an energy analyser should be in use. The approaches by (4)–(7) are available, when a monochromator of the photons is provided. 2.3. APPLICATION CASES
2.3.1.
Employing of the electron work function
2.3.1.1. Mechanical tensions of crystalline silicon surface layer Influence of any mechanical load on a negative increment of the silicon energy gap9 was in use to identify location of mechanical tensions induced within a surface layer because of its local doping with boron.10 A thickness of the doped layer was equal to 100 nm. A distribution of related to the tensions I over the surface is demonstrated in Figure 1. A crown at the boundary of the image corresponds to increased local deformation. Doping (B) 1000 Å
Si (B) |V|
Energy of electron
x Vacuum Cond. band
Cond. band
M2 Valence band
Valence band
Cond. band
M1 Valence band
Figure 1. PE Image of the mechanical tensions distribution over the Si surface.10
2.3.1.2. Thickness of nanofilm An insulator film Si3N4 was deposited with a different thickness (5–100 nm) on a Si substrate. Because of simultaneous electrons emission from the film and the substrate, both of them having differentM, their contribution to the value of I depended on a thickness of the film. Figure 21 provides the PE image of the non-uniform nanofilm. Figure 3 demonstrates a response of I during plasma chemical etching of a multi layer structure. Measurements were supplied in situ using a PE spectrometer incorporated into an etching machine. 2.3.1.3. Size depended electrical potential of nanoparticles It is known that mechanical surface tensions depend on a size of the spheri-cal like particles. This could influence of density of the surface charge, particularly in a case of ion crystals. As the result value of M should be affected.
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
173
Figure 2. PE current distribution over the nanofilm having different thickness.
Figure 3. PE current in dependence on multi layer system etching (in situ).
Figure 4 demonstrates a correlation of M on a size (X) of the hydroxyapatite nanoparticles.11 2.3.2.
Density of electron states
2.3.2.1. Electron states of the nanoparticles Computational simulation (CS) evidenced that the result in Figure 4 was stipulated because of different electron state density induced by protons.11 To verify these approaches (6) and (7) were employed.11 The achieved results demonstrated that increasing of the particle size (X) shifts the detected local
Y. DEKHTYAR
174
0,71
log M>M@ = eV
0,705 0,7 0,695 0,69 0,685 0,68 0,675 0,67 1
1,5
2
2,5
3
3,5
log X, [X]= nm Figure 4. Correlation of M on a size of the nanoparticles.11
dI/dhQ , arb. units
Shift of the local state
Tail of states
1,2 1 0,8 0,6 0,4 0,2 0
X=100 -1000 nm X=20-60 nm
-0,2 -0,4 -0,6 4,9
5,4
5,9
hQ , eV
Figure 5. Distribution of particle size.11
dI / d (hQ ) ~ N i N f
on dependence on the hydroxyapatite nano-
centre in accord with CS the and provides a tailed distribution of the electron states that is similar with disordered systems (Figure 5). The latter was evidenced because of the index m (1) increasing on 50%.11 2.3.2.2. Electron states of the bone The approach (3) was applied jointly with fluorescence and dual emission (see please below) measurements to explore electron states of the bovine bone.12 The result presented in Figure 6 evidenced that the bone has a semiconductor like electron density states distribution.12
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
E eV
vacuum level
175
Density of electron states, arb. units
-1.0 -2.0 -3.0 -4.0 -5.0 -6.0 Figure 6. Electron density of the states in bone.12
3.
Dual Electron Emission Analysis
3.1. BASIC PHYSICS
When the tested object has and energy gap, electrons and hols may be generated because of radiation by photons having an energy hQad t Eg. The excited charge carries are mobile and therefore provide an opportunity to compensate a surface charge that is resulted by the Fermi level pinning at the surface13 and adsorbed ions/dipoles.14 If the photons are switched alongside with PE measurements of I, the latter demonstrates an increment ('I). This is a dual emission mode (DE). The values of 'I may be calibrated in terms of the surface charge density, behaviour of 'I on time reflecting surface electrical potential relaxation induced by photon flux switching on/off.12 3.2. INSTRUMENTATION
The technique on DE measurements is similar that is in use to detect PE, additional light source and monochromator to select suitable hQad should incurporated. To exclude influence of hQad on direct escaping of the photoelectron, the DE is restricted by the condition hQad < hQ, where hQ is the photon energy from (1). On the other hand to minimize an effect of hQ on DE, a ratio of light fluxes must be taken into account:
Y. DEKHTYAR
176
Fad /F >> 1, where Fad and F are the fluxes of light beams that supply hQad and hQ, correspondingly; Fad is not enough for multi photon processes. To select a contribution of hQad , it should be supplied in a pulse mode. 3.3. APPLICATION CASES
3.3.1.
Detection of the energy gap
'I, arb. units
Figure 7 provides an example on Eg detection of the CdTe semiconductor. The value of 'I becomes from zero to a positive value, when hQad = Eg .
1 0.8 0.6 0.4 0.2 0
Eg
1
1.1
1.2
1.3
1.4
1.5
1.6
hQad, eV Figure 7. 'I dependence on hQad in respect to evaluate the energy gap of CdTe surface layer.
3.3.2.
Light induced time depended excitation and relaxation
The flux Fad was switched (+hQad) to radiate CdTe crystal and the value of 'I was increased in time (Figure 8 ). When Fad was switched off (-hQad), 'I had been relaxed. 4. Exoelectron Emission Analysis 4.1. BASICS OF PHYSICS
Electron emission, when its current is induced15 or modulated16 because of relaxation processes within a surface layer of the solid is named as exoelectron
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
+hQ Qad
177
-hQ Qad
1
'I, arb. units
0.95 0.9 0.85 0.8 0.75
time, minutes
0.7 0
2
4
6
8
10
Figure 8. Behavior of 'I because of Fad , in a case of CdTe.
emission (EE). To reach EE two steps should be applied. First, a tested object in advance typically has to be provided with imperfections, their concentration is not in a thermodynamically equilibrium state. After that heat is delivered to the object to induce relaxation of imperfections. As a result EE is supplied. This is a mode of thermostimulated EE (TSE). On the other hand the object tested with PE could be heated alongside to supply EE. In such a case photothermostimulated EE (PTSE) is achieved. EE current (IEE) depends on temperature (T) of the tested specimen, behaviour of IEE having a maximum (Figure 9). The latter is a subject of research and practical applications of EE for analysis. The temperatures to reach EE are not enough to provide significantly detectable thermoelectron emission current. TSE is typically delivered by following mechanisms: (a) Thermoionisation of local states trapping electrons and belonged to the imperfections15; such a mechanism is available from materials with Eg; the ionisation potential should have a little value. (b) Auger transitions of electrons,15 from local electron traps of imper-fections; this channel is available when Eg > F. (c) Field emission from the electron traps by imperfections because of heat induced electrical polarisation/depolarisation of the emitting surface layer.17
Y. DEKHTYAR
178
IEE
Ti
Tmax
Tf
T
~ (+20 … 600) o C Figure 9. Typical spectrum of EE. (Ti, ,Tf , Tmax – initial, final temperatures of the spectrum and temperature of its maximum, correspondingly).
PTSE is provided as a single photon prethreshold PE, when heat delivers: (a) Modulation of the density of the electron local states induced by imperfections and emitting electrons16 (b) Shift of the Fermi level because of (a)16 The above mechanisms fit materials having the energy gap. EE from metals is supplied perhaps from their surface oxides having properties of insulators or semiconductors.15 For the ionisation mechanism Tmax corresponds to the ionisation potential of the electron trap.18 There are experimental evidences15 that
I EE (T ) ~
dC dT
(8)
where C – concentration of relaxing imperfections. In fact the total emitted charge (Q) is directly proportional to C at Ti.15,16 An activation energy (Er) of relaxation for the simplest first order annealing reaction may be estimated from (8)15,16: Er = - kT ln [IEE(T)/N(T))]
(9)
where the k-Boltzmann constant, Tf
N (T )
³I
T
EE
(T )dT
(10)
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
179
In a case of semiconductors the EE spectrum is contributed by annealing of point type complexes of defects supplying single imperfections. Such the reaction is represented by the up-going branch of the spectrum (Ti < T < Tmax ). 16 For the first order reaction its activation energy (Eact) is available from the equation (9)16. However T f should be replaced with Tmax .16 The down-going branch of the EE spectrum (Tmax < T < Tf ) corresponds to migration of the generated single imperfection to the surface of the specimen. An activation energy (Em ) of this process in a case of uniformly distributed single defects within a specimen may be estimated from the equation:16
Em
kT ln
I EE (T max) I EE (T )
d[
I EE (Tmax ) ] I EE (T ) dT
(11)
4.2. INSTRUMENTATION
The above conditions considered to detect PE and DE (when DE is applied jointly with EE) are related to EE too. However, the temperature of the specimen should not exceed the condition, when a significant thermoelectron emission current is provided. 4.3. APPLICATION CASES
4.3.1.
Electron traps at the interface
A thin film of Si3N4 (10 nm) was deposited on a SiO2 substrate. The speci-men was irradiated with weak electrons. The latter were supplied with differ-enced energy to fill in the traps in Si3N4 , Si3N4/SiO2 , consequently. TSE was detected after each step of radiation. The TSE current demonstrated significantly different behaviour, when electron radiation reached the Si3N4/SiO2 interface (Figure 10). This is in favour to apply EE to detect interfaces of the nanostructures. 4.3.2.
Concentration of implanted atoms
A crystalline Si was irradiated by As+ (50 keV) ions. Measured total emitted charge of PTSE was proportional to the delivered fluence of ions (Figure 11).16 The value of Q was sensitive to 10–6 atomic % of As atoms concentration.
Y. DEKHTYAR
180
Si3N4
8 IEE , arb. units
Si3N4/SiO2
4
0 0
300
600
o
T, C Figure 10. TSE of the multi layer system Si3N4/SiO2.19
lg Q, arb. units
2 1.8 1.6 1.4 1.2 1 10
11
12 13 14 15 -2 lg (Flunece), cm
16
Figure 11. Emitted charge of PTSE in dependence on As+ ions fluence supplied to Si.16
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
4.3.3.
181
Activation energy of point type defects annealing and migration
4.3.3.1. Annealing of complex of single defects Tetra vacancies (V4) were generated in a monocrystalline Si because of P+ ions (100 keV) radiation. An activation energy (Er) of complexes annealing was acquired from PTSE measurements (9).16 At a high values of delivered fluence (>5.1014 cm–2) Er decreased because of interactions between V4 (Figure 12).
0.8
Er , eV
0.6 0.4 0.2 0 13.5
14
14.5
15
lg (Fluence), cm-2 Figure 12. Behavior Er of Si tetra vacancies in dependence on P+ ions fluence.16
4.3.3.2. Migration of single defects Single vacancies were generated in a monocrystallne Si owing to dissociation of the vacancy contending defect complexes. The latter were induced due to electron radiation (5 MeV). A activation energy (11) of the liberated vacancy decreased at the fluence level >4.1015 cm–2 (Figure 13), that was stipulated by interference of imperfections.16 From the experimental data, as shown in Sections 4.3.3.1 and 4.3.3.2, it is demonstrated that the photothermostimulated emission instrument can be used to estimate: (a) A thermodynamically non equilibrium concentration of defects (b) A threshold of the concentration of imperfection that provides their interaction (c) Annealing and migration activation energies of imperfections.
Y. DEKHTYAR
182
lgEm, [Em]=eV
14.8
15
15.2 15.4 15.6 15.8 16
16.2
0 -0.5 -1 -1.5 -2 lg (Fluence), cm-2
Figure 13. Em of Si single defects in dependence on electron fluence.16
5. Conclusion The review demonstrates that PE, DE and EE are capable to supply sensitive, accurate, and non-contact technique for the needs of nanotechnologies. The techniques are directed to measure the electron work function (to characterise surface potential), distribution of the electron states density (PE); energy gap and charge relaxation (DE); concentration of the point type imperfections, their annealing and migration (EE). ACKNOWLEDGMENTS
The author graciously acknowledges figures and contributions provided by the Author’s colleagues Mr. V. Noskov, Riga Technical University, Latvia, Dr. A. Balodis, Riga Technical University, LV, and Mr. V. Noskov, Riga Technical University, LV.
PHOTO-DUAL- AND EXO-ELECTRON EMISSION
183
References 1. Froitzheim H. (1977), Electron Spectroscopy for Surface Analysis. Ed. H. Ibach, Springer, New York, 315 p. 2. Fomenko V.S. (1981), Emission Properties of Materials. Naukova Dumka, Kiev, 338 p. (in Russian). 3. Kinding N.B., Spicer W.E. (1963), Band structure of cadmium sulphide photoemission studies, Phys. Rev. 138: A561–A576. 4. Dobrecov L.N., Gomoyunova M.V. (1966), Emission Electronics. Nauka, Moscow 564 p. (In Russian). 5. Kane E.O. (1962), Theory of photoelectric emission from semiconductors, Phys. Rev. 127(1): 131–141. 6. Pihtin A.N. (1983), Physical Basics of Quantum Electronics and Optoelectronics. Vsishaya shkola, Moscow, 304 p. (in Russian). 7. Dekhtyar Yu. D., Vinyarskaya Yu. A. (1994), Exoelectron analysis of amorphous silicon, J. Appl. Phys. 75(8): 4201–4207. 8. Dekhtyar Yu.D., Sagalovich G.L. (1988), Exoelectron emission of monocrystalline silicon and its practical application, Proc. USSR Academy of Sci., Physical Ser. 52: 1611–1613. (in Russian). 9. Polyakova A.L. (1979), Deformation of Semiconductors and Semiconductor Devises. Energiya, Moscow, 168 p. (in Russian). 10. Sagalovich G.L., Balodis A.Ya., Dekhtyar Yu.D. (1984), Emission test of mechanical tensions in surface layers of silicon, Electron. Tech., Ser. 8. 5(110): 25–27. 11. Bystrov V., Bystrova N., Dekhtyar Yu., Filippov S., Karlov A., Katashev A., Meissner C., Paramonova E., Patmalnieks A., Polyaka N., Sapronova A. (2006), Size depended electrical properties of hydroxyapatite nanoparticles, in: IFMBE proceedings. V. 14. CD version. Springer, Seoul, pp. 3149–3150. 12. Arvin H., Bogucharska T., Dekhtyar Yu., Hill R.M., A. Katashev, Pavlenko A., Pavlenko I., Zakaria M. (2000), Electronic transitions and structural changes in bone, Latvian J. Phys. Tech. Sci. 6(S): 50–55. 13. Nesterenko B.A., Cnitko O.V. (1983), Physical Properties of Atomic Clean Surface of Semiconductors. Naukova Dumka, Kiev, 264 p. (in Russian). 14. Volkenstein F.F. (1987), Electron Processes on Semiconductors Surface During Chemosorption. Nauka, Moscow, 430 p. (in Russian). 15. Kortov V.S., Shifrin V.P. Gaprindashvili A.I. (1975), Exoelectron spectroscopy of semiconductors and insulators, Microelectronics 8: 28–49 (in Russian). 16. Dekhtyar Yu.D. (1993), Exoelectron Spectroscopy of Point Type Defects in Semiconductors. Riga Technical University, Latvia, Riga, 59 p. (in Russian). 17. Rosenman G.I., Rez I.S., Chepelev Yu.L., Angert N.B. (1981), Exoemission of defected surface of lithium tantalite, J. Tech. Phys. 51(2): 404–408 (in Russian). 18. Nassenshtein G. (1962), Electron emission from the solid state surface after mechanical treatment, in: Exoelectron Emission, Inostrannaya literature, Moscow, pp. 72–95 (in Russian). 19. Rosenman, M., Naich, M., Molotskii, Dekhyar Yu, Noskov V. (2002), Exoelectron emission spectroscopy of silicon nitride thin films, Appl. Phys. Let. 80(15): 2743–2745.
LASER INTERACTION WITH NANO-SPHERES: APPLICATIONS IN SUB-MICRON PARTICLES REMOVAL AND NANODOT ARRAY FABRICATION M. SENTIS1*, D. GROJO1, PH. DELAPORTE1, AND A. PEREIRA2 1 Lasers Plasmas and Photonic Processing Laboratory, CNRS – Universities of Aix-Marseille, Campus de Luminy – Case 917, 13288 Marseille Cedex 9, FRANCE 2 INRS-EMT, 1650 blvd. Lionel-Boulet, Varennes (Quebec) J3X1S2, CANADA
Abstract – Laser-assisted nanoparticle removal processes like selective particle ablation, mechanical ejection, local substrate ablation, explosive evaporation are considered by experimental and theoretical approaches. Experiments are based on the estimation of particle removal efficiencies and thresholds depending on material and size of particles. Afterwards, a novel and efficient photonicbased method to synthesize porous alumina membranes (PAMs) and subsequent metal nano-dot arrays on various substrates, based on near laser field enhancement by nano spheres, is reported.
Keywords: Laser nano-particle interaction, near-field enhancement, nanodot arrays.
1. Laser-Assisted Nanoparticle Removal Processes 1.1. INTRODUCTION
Due to the emerging nanotechnologies, new cleaning solutions providing the ability to remove nanometer defects from material surfaces are now required to ambition a reliable fabrication of nanoscale devices. In this context, the dry
______ *
To whom correspondence should be addressed: M. Sentis, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
185
186
M. SENTIS ET AL.
laser cleaning (DLC) technique might be consider as a potential approach. In this method, the particles are ejected from the surfaces by the illumination of the polluted materials with nanosecond laser pulses. Some studies have already demonstrated that this technique is efficient for specific applications1 and allows eliminating particles with size below 100 nm.2,3 However, there is still a lack of knowledge of its underlying mechanisms. This may be explained by the complexity of the processes which can be combined in the studied situations. Among them, the adhesion and the laser-matter interactions involved in the study of particles at the nanometer scale exhibit singular characteristics. The literature on DLC, for more than 15 years now, shows that the removal of particles results from a combination of at least three mechanisms: (i) the mechanical force (inertia or elastic) exerted on the particles while they move with the rapid thermal expansion of materials,4,5 (ii) the local substrate ablation due to the enhancement of the laser light which can be observed underneath small particles,6,7 (iii) the explosive evaporation of the adsorbed humidity from the air at the interface between the particles and the surface of the substrate.8,9 On the basis of this latter contribution which has been evidenced recently, we may consequently wonder how dry the so-called dry laser cleaning technique is. Indeed, as we shall see later on, the humidity can play a major role on the physical processes occurring in the experiments. For this reason, we extensively investigated this aspect. The interpretation of the results of these experiments contributes to a better understanding of the humidity-based adhesion and laser-assisted removal processes involved in experiments with submicrometer particles. 1.2. EXPERIMENTAL DETAILS
Laser-induced particle ejection experiments were carried out with an ArF (Olas = 193 nm) excimer laser delivering pulses of 15 ns duration and 300 mJ energy. The samples were irradiated by single laser pulses in residual pressure conditions to avoid particle redeposition. The image relay of a metallic mask (using a lens) permits a near-uniform irradiation (S = 2 u 1.5 mm2) of the target materials. Samples were silicon (Si) (100) substrates on which 250 nm radius SiO2 spheres were spin-coated. A density of #105particles/cm2 was obtained with more than 90% of isolated particles. The samples were placed in a vacuum chamber allowing to reach a residual pressure of 10–4 Pa. Inside the chamber, the stainless steel sample holder was back-side equipped with a halogen lamp to adjust the temperature of the irradiated target up to 750K. Thus, the substrates can be subjected to different degassing step between each of the experiments and the water meniscus inevitably formed in a humid medium at the interface between the particle and surface can be gradually eliminated. In our experiments, the water meniscus may be formed due to the humid particle deposition process
LASER INTERACTION WITH NANO-SPHERES
187
or condensed from ambient (i.e. air) humidity before the sample was entered into the chamber. Software analysis of optical micrographs (dark field) of the laser cleaned substrates was used to determine the Particle Removal Efficiency (PRE ) as a function of the experimental conditions (laser energy, degassing step). The values of PRE correspond to the statistic fraction of particles which was removed by the irradiation i.e. PRE = 1–n/n0 where n and n0 are respectively the densities of particles on the surface before and after the laser irradiation. Fast imaging of the ejected particle clouds was also performed with a fast intensified CCD (ICCD) and a forward scattering-based probing technique. The temporal and spatial resolutions of these measurements were ~1 μs and 5 μm, respectively. A complete description of the experimental apparatus can be found in Ref.10 On the basis of this diagnostic, time-of-flight analyses of the ejected particles allow to determine the detachment velocity of the particles. The measurement of the initial velocity, which can be considered as an experimental signature of the different ejection mechanisms,11 is used to distinguish between them and to identify their respective validity domains. 1.3. RESULTS AND DISCUSSIONS
1.3.1. Time-of-flight (T.O.F.) analyses Detailed information on the particle cloud ejection dynamics is obtained from the distribution of the scattering signal from a CW probe beam as a function of the distance from target I(z) for different delays t between the laser pulse and the ICCD observation gate of 1 μs duration. The distance between the particles (#30 μm) being large compared to the probe wavelength (Oprobe = 532 nm) the measured signal is directly proportional to the particle density. The I(z) curves which were captured for a time delay t = 8 μs and for three different laser fluences above the particle removal threshold fluence (Fth # 130 mJ cm–2), are shown on Figure 1. Each profile was fitted with a Gaussian function G F ,t ( z ) . Consequently, according to the consistency of the fits, the complete velocity distribution f F (v ) of the ejected particles for a given laser fluence F is also well described by a Gaussian function which can be deduce from the profiles that are captured at a time delay t. This distribution is given by the expression: f F ( v ) v tG F ,t ( vt ) . The profiles shows that very different behaviors are observed as a function the laser energy deposited onto the polluted substrates. In particular, Figure 1b shows that under appropriate conditions, the ejected particle cloud is split in two distinct components with different velocities. According to the position of the detected species zmax at about 200 and 650 μm for each of these components, the corresponding propagation velocities of the particles are respectively 25 and
188
M. SENTIS ET AL.
81 m s–1. This observation means that, although we study calibrated systems, the particle detachment results from a competition between at least two different mechanisms in this given situation. For lower laser fluences (Figure1a), only the slow component is observed. For larger laser fluences (Figure1c) only the fast one remains. These observations evidenced the existence of two regimes where only one removal mechanism is responsible for the particle ejection.
Figure 1. Spatial repartition of the ejected particles (given by the forward scattering signal) with respect to the distance from target for a time delay t of 8 μs after the irradiation of the target with the nanosecond laser pulse. The profiles (and their Gaussian fitting curves) are shown for three different laser fluences: (a) Flas = 270 mJ cm–2, (b) Flas = 410 mJ cm–2 and (c) Flas = 550 mJ cm–2.
To analyze the different regimes in more details, Figure 2 shows the determined velocities of the particles as a function of laser fluence. We note that the analyses were possible for laser fluences very close to the removal threshold fluence Fth. Above Fth, particles are ejected with a characteristic velocity which gradually increases from 7.6 to 36.8 m sí1. In the range 300–500 mJ cmí2, while the scattered intensity from the particle cloud decreases a faster component appears. This transitory behavior evidenced that some particles are now removed on the basis of a physical process of different nature to that observed for the low laser energy domain. For larger fluences, only the fast component remains and particles acquire considerably more speed when increasing Flas. Propagation velocity of the particle cloud linearly increases from 66 to 231 m sí1 when fluence is varied from 415 to 680 mJ cmí2. Thus, analyzing the characteristic velocities of the ejected particles, we determined two laser energy domains where two different detachment processes dominates but that still have to be identified.
LASER INTERACTION WITH NANO-SPHERES
189
Figure 2. Velocities of the ejected particles measured by fast imaging of the scattering clouds as a function of the laser fluence. The determined velocities (above the particle removal threshold fluence) are compared with a calculation of the surface expansion dynamics for similar conditions.
Among the possible mechanisms, we first examined the likelihood of a mechanical ejection resulting from the rapid thermal expansion of the materials. For submicron particles irradiated with nanosecond laser pulses, the elasticity of particles can be neglected and the contaminants move as a whole, following the thermally-induced displacement of the surface of the substrate.10 Figure 2 presents the results of a calculation of the maximal surface velocity which is reached in the vicinity of the particles. The model consists of a numerical resolution of the 3D thermo-elasticity problem, taking into account the focusing power of the SiO2 spheres. This aspect is described using the Lorentz-Mie theory. Indeed, we already showed that, the surface of the substrate is locally illuminated by a bright spot with a Gaussian-like shape and a full width at half maximum of 160 nm12 underneath each of the spheres. The subsequent hot spots enhance the expansion dynamics of the materials (in comparison with a uniform irradiation with clean surface). However, as shown on Figure 2, the expansion velocity of the surface remains two orders of magnitude below the experimentally measured particle ejection velocities. This demonstrates that the dominant removal mechanism in these experiments is not the rapid thermal expansion of materials.
190
M. SENTIS ET AL.
1.3.2. Self-limited local ablation of the surface After the irradiation experiments, the cleaned substrates were carefully analyzed by means of SEM. For Flas corresponding to the presence of the fast component (i.e. Flas > 300 mJ cm–2), we successfully observed ablated craters in places where originally located the particles. The hole sizes were in the order of 135 nm in diameter for laser fluences less than 500 mJ cmí2. This size is relatively consistent with the laser energy distribution given by the Lorentz-Mie theory solution.12 It demonstrates that, in this fluence domain, the ejection of the particles results from the local ablation of the substrate which is induced by the optical near-field enhancement underneath the particles. This approach is not relevant in a context of extreme cleaning but could be used as a technique of nanomanufacturing. 1.3.3. Humidity dependence of the laser-induced removal of particles As stated previously, the nature of the interactions between, the laser light, the submicrometer particles and the surface may be strongly modified by the water meniscus at the interface between the particles and surface. The influence of this aspect on the particle removal with moderated laser fluences is analyzed in this section. To better understand the role of the humidity, we repeated PRE measurements as a function of laser energy after successive periods during them the polluted substrates were conserved under vacuum and at relatively high temperature. Consequently, the water meniscus under the particle was progressively reduced with respect to time in our experimental investigations. Figure3 presents the evolution of the particle removal threshold when the humidity is progressively reduced. The last value which corresponds to the driest substrate (as dry as we could get) is close to the local substrate ablation threshold reported in the previous paragraph. It means that the removal of the water meniscus leads to the disappearance of the low fluence regime mechanism which was discriminated in the TOF analyses. Thus, for dry substrates, the only way to eject particles is the substrate ablation. From this result, we propose that the explosive evaporation of the water meniscus which is induced by the laser irradiation8,9 is the particle removal mechanism in this regime. The measured ejection velocities (see Figure 2) strengthen us in our proposition. Indeed, the velocities found in the literature relating the ejection (ablation) of thin films of condensed water on metal surfaces by nanosecond laser pulse irradiations are also in the range 10–40 m s–1 for similar conditions.13 In this approach, the cleaning is still based on the momentum transfer from laser-ablated species (water) but does not require the consumption of the substrate.
LASER INTERACTION WITH NANO-SPHERES
191
Figure 3. Dependence of the particle removal threshold fluence as function of time during an experiment where the humidity trapped on the substrate is gradually reduced.
It would be over-simplified to claim that the presence of humidity transforms only the laser-matter interaction processes in our experiments. Indeed, the nature of the adhesive interaction between the particle and the surface is also strongly modified by the water meniscus which is formed between them. In particular, this aspect needs to be taken into account to explain the nonmonotonous variation of the particle removal threshold shown in Figure 3. The observation of a minimum as a function of time evidences that it exists an amount of water which optimizes the removal process. This behavior is confirmed by the ejected particle velocity measurements performed before and after that the samples was left during 24 h in a 10–2 Pa atmosphere (corresponding to the first and second points in Figure 3). Figure 4 shows that this gentle degassing step leads to a 20% decrease of the particle removal threshold Fth and a significant increase of the ejection velocity for laser fluences close to Fth It reveals that, when the humidity is slightly decreased the particle ejection becomes easier (in terms of laser energy); hence it corresponds to a better compromise between the adhesion and the cleaning force. In principle, the adhesion of submicrometer objects to surfaces is dominated by van der Waals forces. A typical value of these forces for a 250 nm silica particle is #100 nN.14 However, the presence of a water meniscus may provide an additional contribution to the adhesion by the capillary force. The capillary force can be estimated by using the standard approximation Fc = 4SJR where
192
M. SENTIS ET AL.
Figure 4. Measurement of the velocities of the ejected particles as function of the laser fluence. After that the humidity was slightly reduced by a degassing step of 24 h under a pressure P = 10–2 Pa, a decrease of the particle removal threshold fluence (# 20%) and a significant increase of the kinetic energy of the ejected particles are observed.
J is the surface tension of water (7.28 u 10–4N cm–1).15,16 For R = 250 nm, this (#200 nN) is not negligible and can easily dominate the van der Waals force. Surprisingly, on the basis of this often used approximation, the adhesion force is independent on the relative humidity RH, and consequently, on the amount of water at the interface between the particle and the surface. However, recent numerical simulations16,17 and capillary force measurements (based on Atomic Force Microscopy)18 demonstrated that while this approximation is viable for particles above 1 μm, its validity is limited for particles as small as the particles we studied. In particular, O. Pakarinen et al.16 developed a model to numerically calculate the exact (non-circular) meniscus profile from the Kelvin equation and, the subsequent capillary force in real environments. They showed that for the geometry “sphere on flat surface”, the standard approximation remains acceptable for RH # 1. However, the adhesion vanishes gradually with the humidity (with slopes increasing as the size of the particles decreases). Consequently, the non-monotonous variation of the particle removal threshold given in Figure 3 is explained by the competition between two humidity-based forces: the capillary adhesion force and the cleaning explosive evaporation (of the water meniscus) force. In the experiments, at the very beginning, when the humidity is gently reduced, the adhesion force is severely decreased but there is still enough water trapped under the particles to eject them efficiently
LASER INTERACTION WITH NANO-SPHERES
193
by the laser-produced explosive evaporation of this humidity. Thus, the removal threshold is reduced as a result of an optimization of the environment. After this time, while the sample is degassed further, the removal threshold increases progressively as a result of the reduction of the quantity of water possibly ablated and the subsequent dragging force exerted on the particles. This is observed until the quantity of water becomes sufficiently low to make impossible the ejection of the particles by momentum transfer from the evaporated water. Then, the only way to remove the particles is based on the local ablation of the substrate. 1.4. CONCLUSION OF LASER NANOPARTICLE REMOVAL
We demonstrated that the removal of submicrometer particles in nanosecond laser experiments results from a competition between two mechanisms based on the momentum transfer from laser ablated species to the particles. The local ablation of the substrate resulting from the near-field enhancement underneath the particles provides the cleaning force for large fluences. At low laser fluence, we observed that the high efficiency of the direct illumination laser cleaning technique is damage-free. However, the experimental results reveal a large dependence on the amount of humidity inevitably trapped at the interface between the particles and surface. The humidity plays a major role in the cleaning force as well as in the adhesion force exerted on the particles. This humidity dependence may explain the lack of constancy of the results of the numerous experimental studies on DLC performed in non-controlled atmospheres for more than 15 years now. The explosive evaporation of the trapped water is actually the cleaning mechanism in our studied systems. Consequently, the so-called DLC is a humidity-based process. Thus, it is rather similar to the Steam Laser Cleaning (SLC) where a liquid film is deposited on the surface before the laser irradiation. In comparison with SLC, the advantage of our approach is that the water is located only where the water is needed (at the interface between the particle and the surface). However, to optimize the technique, the relative humidity must be controlled and adjusted to lead to the optimum quantity of water trapped under the particles. 2. Preparation of Metal Nanodot Arrays by Near-Field Enhancement In this paragraph we exploit the local ablation of substrates resulting from the near-field enhancement underneath the particles to synthesize Porous Alumina Membranes (PAM) and subsequent metal nanodot arrays on various substrates.
194
M. SENTIS ET AL.
2.1. LASER PROCESS
The fabrication process is outlined in Figure5. A monolayer of self-assembled spheres is formed onto a thin alumina (Al2O3) film, which was previously coated on a silicon substrate by means of Pulsed Laser Deposition (Figure 5a). Then, pores are optically drilled in the Al2O3 film by particle-assisted near field enhancement. This is accomplished through illumination of the spheres with a single nanosecond laser pulse at the wavelength Olas = 193 nm. This leads to the local removal of the 20 nm thick Al2O3 film under each sphere. Since the spheres are arranged in a hexagonal array at the surface of the substrate, the aluminum oxide film is decorated with an ordered arrangement of holes. (Figure 5b). Using this Laser Fabricated PAM (LF-PAM) as a mask for the deposition of metal (Figure 5c), a series of ordered metal nanodots are formed at the surface of the substrate upon dissolution of the alumina layer (Figure 5d). As we shall see later on, we demonstrated the controllability of this schematic approach by forming an ordered array of gold nanodots.
Figure 5. Scheme of the nanodot array fabrication method.
Numerous studies devoted to nanostructuring by particle-assisted near-field enhancement deal with 1-μm PolyStyrene (PS) spheres. However, Piparia et al.19 demonstrated recently that, in the near UV region (<300 nm), the interaction process exhibits a self-limited character. In this spectral region, short wavelengths are strongly absorbed in most polymer materials. To avoid this limitation, we choose to use silica (SiO2) spheres. Also, spheres with radii R = 250 nm were chosen to prepare nanodot array with a larger surface density (#5 u 108dots/cm2) than accessible with 1-μm spheres. The most critical aspect of this approach is the controlled and selective nanodrilling of the alumina film. Al2O3 is a very low absorption material largely used for optical coatings in the UV region. At 193 nm, the reflectance of Al2O3
LASER INTERACTION WITH NANO-SPHERES
195
is R # 0.09 and its optical penetration depth is several orders of magnitude larger than the film thickness used in this study.20 Consequently, the laser light is mainly absorbed by the Si substrate. 2.2. DISCUSSION AND RESULTS
We carefully analyzed, by means of Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM), the effect of the pulse energy on the “spheres-Al2O3-Si” stacks. For laser fluence as low as 200 mJ cm–2, the silica spheres are ejected from the surface. According to the laser cleaning mechanism described in part 1. Also, craters of few nanometers deeps are formed at the surface of the Al2O3 film (AFM observations), showing that this mechanism slightly damages the surface. At laser fluences close to 340 mJ cm–2, the silica spheres and the aluminum oxide film are removed. The typical diameter of the ablated craters is 130 r 15 nm. The bottom of the crater is flat, indicating that the underlying Si substrate is not significantly affected by the process. In this situation, the ablation of the Al2O3 layer is due to the spallation mechanism induced by the confinement of the deposited energy at the alumina-silicon interface.21 For laser fluences above 400 mJ cm–2, the Si substrate is damaged. In that regime, sombrero-like structures having a diameter of about 280 nm are observed. The center of these structures protrudes the original surface of the (undamaged) aluminum oxide film by about #8 nm. Recent studies reported similar phenomenon at larger size scale with pure silicon substrates.22 The formation of such nanostructures relates to inward flows of the melted materials caused by the inhomogeneous temperature distribution in the heated layer.19,23 Figure 6 shows AFM and SEM images of gold nanodot arrays created on Si substrates by performing LF-PAM at intermediate laser fluence (340 mJ cm–2). Software analyses of the images allow us to characterize the size and shape of the gold nanostructures. From the object area Sdot, the average size of the structures Rdot = (Sdot/S)1/2is 48.2 r 2.6 nm (rSD). The circularity, given by the ratio of the square of the perimeter to 4S Sdot, is 1.47 r 0.1. Each dot has a uniform height (20 nm) and spacing (500 nm) to its neighbor. The height determined by the AFM profile (Figure6c) is consistent with the original Al2O3 film thickness. 2.3. CONCLUSION ON LASER PROCESSING TECHNIQUE FOR METAL NANO DOT ARRAY FABRICATION
We demonstrated that the fabrication of metal nanodot arrays is feasible by using exclusively laser processing techniques. The size of the dots thus generated is the same as that of the holes drilled on the PAM. This approach is quite
196
M. SENTIS ET AL.
Figure 6. SEM image (a) AFM image (b) and depth profile (c) of gold nanodots created on silicon substrates by the LF-PAM based process.
general and can be applied to wide range of metals, semiconductors or complex oxides. This technique can be extended to spheres of variable diameter, thus allowing the distance between the nanodots to be varied for a defined application. However, Mie theory calculations show that the improvement in resolution (size of the near-field enhancement region) induced by the reduction of the sphere size will not be dramatic.24 Focusing at a scale of a few nanometers is still a challenge that simple particle-masks will not solve. To improve resolution
LASER INTERACTION WITH NANO-SPHERES
197
and accuracy, one of the possible extensions of the technique will be to exploit the deterministic and non-linear character of the damage processes involved in femtosecond laser machining.25
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.
Grojo, D., Cros, A., Delaporte, P., and Sentis, M. (2006) J. Phys. IV 138, 191–201. Lizotte, T., and O’Keeffe, T. (1996) Proc. SPIE 2703, 279–287. Wu, X., Sacher, E., and Meunier, M. (1999) J. Appl. Phys. 86, 1744–1748 Luk’yanchuck, B., Wang, Z., Song, W., and Hong, M. (2004) Appl. Phys. A 79, 747. Arnold, N. (2003) Appl. Surf. Sci. 15, 208. Mosbacher, M., Munzer, H., Zimmermann, J., Solis, J., Boneberg, J., and Leiderer, P. (2001) Appl. Phys. A 72, 41. Arnold, N., Schrems, G., and Bauerle, D. (2004) Appl. Phys. A 79, 729. Vereecke, G., Rohr, E., and Heyns, M. (1999) J. Appl. Phys. 85, 3837. Mosbacher, M., Bertsch, M., Munzer, H., Dobler, V., Runge, B., Bauerle, D., Boneberg, J., and Leiderer (2002) P. Proc. SPIE 4426, 308. Grojo, D., Cros, A., Delaporte, P., and Sentis, M. (2006) Appl. Phys. B 84, 517–521. Grojo, D., Cros, A., Delaporte, P., and Sentis, M. (2006) Proc. SPIE 6261, OE. Pereira, A., Chaker, M., Guay, D., Grojo, D., Delaporte, P., and Sentis, M. Appl. Phys. Lett., submitted. Zapka, W. (2002) Laser Cleaning, B. Luk’yanchuk (Ed.), World Scientific, New Jersey, 23–48. Grojo, D. (2006) Ph.D. thesis/University Marseille, 17–32. Cappela, B., and Dietler, G. (1999) Surf. Sci. Rep. 34, 1–104. Pakarinen, O., Foster, A., Paajanen, M., Kalinainen, T., Kataienen, J., Makkonen, I., Lahtinen, J., and Nieminen, R. (2005) Model. Simul. Mater. Sci. Eng. 13, 1175–1186. Jang, G., Schatz, C., and Ratner, M. (2004) J. Chem. Phys. 120, 1157. Zitzler, L., Herminghaus, S., and Mugele, F. (2002) Phys. Rev. B 66, 155436. Piparia, R., Rothe, E., and Baird, R. (2006) Appl. Phys. Lett. 89, 223113. French, R., Mullejans, H., and Jones, J. (1998) J. Am. Ceram. Soc. 81, 2549. Xia, Z., Shao, J., Fan, Z., and Wu, S. (2006) Appl. Opt. 45, 8253. Wysocki, G., Denk, R., Piglmayer, K., Arnold, N., and Bauerle, D. (2003) Appl. Phys. Lett. 82, 692. Georgiev, D., Baird, R., Avrustsky, I., Auner, G., and Newaz, G. (2004) Appl. Phys. Lett. 84, 4881. Ahn, C., and Rabe, K. (2004) J. Mater. Sci. 303, 488. Joglekar, A., Liu, H., Meyhofer, E., Mourou, G., and Hunt, A. (2004) PNAS 101, 5856.
CLEAN FOSSIL FUELS: ADVANCED MEMBRANE REACTORS T. TRAN, K. STOITSAS, AND J. SCHOONMAN* Department DelftChemTech-Energy, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands
Abstract – The development of processes with high efficiency and low carbon emissions seems to be more necessary than ever in order to face the temperature rise of our planet. In this paper preliminary results are presented of the development of membrane reactors which will be used in the steam reforming reaction so as to produce pure Hydrogen with higher efficiency than the conventional reactors. The sol-gel technique is used for the development of porous ceramic membranes that will separate Hydrogen from Carbon Monoxide or Dioxide due to the molecular sieving effect. Atomic Layer Deposition will be used for reducing the pore size of the primary membranes made by the sol-gel technique.
Keywords: Greenhouse effect, hydrogen production, carbon membrane reactors, ceramic membranes.
1. Introduction In 2003 the Nobel Laureate Richard Smalley1 presented his view on the humanity’s top-ten problems for the next 50 years. On a prominent first place Energy was presented. Every day the world uses about 85 million barrels of crude oil and nearly about 10 billion cubic meters of natural gas and demand is rising. Energy scenarios anticipate that by 2030 the total global energy demand will grow by close to 50% and by 18% in Europe – EU and non-EU. These numbers do not take into full account the rising economies in India and China. To date, fossil fuels cover about 80% of the present energy demand, with about 7% nuclear, and about 13% renewable energies, like, for example, solar energy, wind energy, biomass, hydrothermal, geothermal, and tidal wave energy. Especially the utilization of fossil fuels causes the emission of the greenhouse
______ *
To whom correspondence should be addressed: J. Schoonman, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
199
200
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
gas (GHG) carbon dioxide, and the emission of this GHC is less, when using natural gas, instead of oil, and, in particular, instead of coal. To date, the total annual CO2 emission to the atmosphere equals 3.5 Gt Carbon2 and global warming is related to the increase of the CO2 concentration. A simplified explanation of the greenhouse effect is as follows. The solar radiation passes through the clear atmosphere and some solar radiation is reflected by the Earth’s surface and the atmosphere, but most radiation is absorbed by the Earth’s surface and warms it. Infrared radiation is emitted from the Earth’s surface and some of the infrared radiation is absorbed and re-emitted by water vapor and greenhouse gases. The effect of this is to warm the surface and the lower atmosphere.3 In fact, the prediction of climate change due to human activities began with a prediction made by the Swedish chemist Svante Arrhenius in 1896. Arrhenius took note of the industrial revolution then getting underway and realized that the amount of carbon dioxide being released into the atmosphere was increasing. He predicted that if atmospheric carbon dioxide doubled, the Earth would become several degrees warmer.3 Improve energy efficiency, switch to low-carbon fuels, switch to renewable energy sources where possible and sequestration of CO2 derived from fossil fuels are options for reducing CO2 emissions into the atmosphere and attract widespread attention. In this respect hydrogen is an attractive energy carrier and in a transition towards the sustainable application of fossil fuels and towards renewable energy sources, clean fossil fuels and renewable hydrogen, i.e., hydrogen prepared using renewable energy sources will play an important role. Clean fossil fuels can be obtained using advanced membrane reactors and in this paper materials and processes for advanced membrane reactors will be presented. 2. Production of Hydrogen Hydrogen is the third most abundant element, but it is only available in bound form. It can be produced from many sources as well as a variety of technologies: natural gas by reforming and partial oxidation, coal by gasification and in-situ gasification with CO2 sequestration, and nuclear energy by electrolysis (water splitting) or for a thermo-catalytic process. Solar energy and wind energy can also be used to electrolyse water and together with gasification of biomas provide renewable hydrogen sources. The photo-electrochemical production of hydrogen has been presented by Schoonman et al.4 Hydrogen from these processes is sustainable and renewable. Hydrogen is not widely used today but it has great potential as an environmentally clean energy carrier for the near future. Energy is generated by burning directly hydrogen with oxygen producing only water as shown in reaction 1. Using hydrogen and oxygen in a fuel cell generates electrical energy.
CLEAN FOSSIL FUELS
H2(g) + ½ O2(g) l 2H2O(l) 286 kJ/mol
201
(1)
A comparison with fuels, like coal, brown coal, wood, gasoline, diesel, methanol, and natural gas reveals, that hydrogen as an energy carrier has a highenergy content per unit mass of 119.9 kJ/g and a low-energy content per unit volume of only 10 kJ/l. Therefore, the efficient storage of hydrogen is an important scientific and engineering challenge. The key reactions for the industrial chemical production of hydrogen from fossil fuels are reactions (2) and (3). 750-1,000 °C
CH4 + H2O l CO + 3H2 catalyst
ǻǾ°298 = 206 KJ/mol
(2)
ǻǾ°298 = – 41 KJ/mol
(3)
300–500°C
CO + H2O l CO2 + H2 catalyst
According to Le Chatelier’s principle, by removing H2 from the gaseous product, the equilibrium can be shifted to the product side. Effectively, this can lower the reaction temperature to 400–500°C and higher efficiencies 85–90 instead of 75% and improves the purity of the hydrogen.5 Other approaches to obtain pure hydrogen are the removal of the formed CO2 and hydrotalcite materials and Ionic Liquids are being studied for CO2-selective membranes. The removal of hydrogen or carbon dioxide is studied within the Stanford Global Climate and Energy Project (GCEP), which focuses on the fundamental and pre-commercial research on technologies that would foster the development of a global energy system with low greenhouse gas emissions. 3. Membrane Reactors McLeary et al.6 have published a recent review paper on zeolite films, membrane and membrane reactors, focussing on the progress in the field and the prospects. These authors have presented the classification of membrane reactor configurations according to membrane function and location, based on the recent literature. The authors describe the catalytic membrane reactor (CMR), the packed-bed membrane reactor (PBMR), the catalytic non-permselective membrane reactor (CNMR), the non-permselective membrane reactor, and the particle-level membrane reactor. The applications of membrane reactor processes, the reactor configuration, the type of membrane, and the operating conditions for the reaction types, i.e., based on dehydrogenation reactions, hydrogenation reactions,
202
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
oxidation reactions, and examples of organic syntheses. In addition, examples of scarce industrial membrane reactor processes are presented.6 Inorganic membranes can be classified into two types, i.e., dense and porous membranes. Dense membranes prepared from Pd-Ag alloys, or oxide-ion conducting, and mixed ionic-electronic conducting perovskites only allow for hydrogen or oxygen, respectively. In the perovskites oxygen permeates via oxideion vacancies. While such membranes exhibit extremely high selectivities, they have limited permeabilities, but recent research has resulted in fluxes within reach of targets. Microporous silica membranes have proven to be promising for molecular sieving applications. Precise control of the pore size allows for separation on the basis of size by molecular filtration, or “sieving”. It is known that the hydrophilic surface of silica membranes leads to pore blocking by H2O and especially at higher temperatures this can seriously degrade the microstructure.7 Therefore, the microporous silica membranes are not stable under the steam reforming reaction of methane to separate hydrogen from CO. Here to, we have designed a reactor based on alumina and silica-alumina composite microporous membranes, which combine the steam reforming reaction, or the water-gas shift reaction with in-situ hydrogen separation in one system as illustrated in Figure 1. This type of membrane reactor will be expected to remove selectively hydrogen from the reactions and to achieve high conversion of methane at lower
Reaction catalyst
Nano-structured Layer Atomic Layer Deposition (ALD) (<0.32 nm) Intermediate Porous Layer Sol-Gel (1-5 nm)
Macroporous Support (D- Al2O3)
Figure 1. Concept of membrane reactor for hydrogen separation.
CLEAN FOSSIL FUELS
203
temperatures. The purpose of the research is to develop a microporous membrane material based on J-alumina prepared by the sol-gel route, followed by Atomic Layer Deposition. This technique allows depositing a monolayer or a multilayer inside the pores, thereby reducing the initial pore diameter to the molecular dimension of hydrogen. In this way, the separation is mainly based on the differences of the molecular diameters of the molecules involved as shown in Figure 2.
Figure 2. Kinetic diameter of H2, CO2, CO, and CH4.
4. Synthesis of Mesoporous Membranes Mesoporous membranes are developed using the sol-gel technique. Sol-gel processing is one of the most interested methods for the preparation of porous ceramics membranes. This method involves the transition of a system from a colloidal or polymeric solution (sol) into a gelatinous substance (gel). There are two kinds of sol-gel routes according to the state of the sol: colloidal route and polymeric route as illustrated in Figure 3. In general, the sol-gel route involves the hydrolysis-condensation reactions of metal alkoxides M-OR (M = Si, Ti, Zr, Al; OR = OCnH2n+1) precursors dissolved in water or an organic solvent, leading the formation of a metal-oxobased macromolecular network. The overall hydrolysis-condensation reactions of the sol-gel route can be presented as follows: Hydrolysis: M-OR + H2O ĺ M-OH + R-OH Condensation: M-OH + M-OX ĺ M-O-M + X-OH Oxolation M-OH + M-XO+H ĺ M-O+H-M + X-OH Olation (X= H or R )
(4) (5) (6)
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
204
Metal salt or metalorganic precursor
Colloidal route
media
polymeric route
Sol (polymeric)
Sol (colloidal)
membrane coating
Polymeric gel
Colloidal gel
Drying (hybrid organic-inorganic membrane)
Sintering (pure inorganic membrane)
Figure 3. Diagram showing two sol-gel routes used in ceramic membrane preparation.8
The synthesis of Ȗ-Al2O3 membranes proceeds through the colloidal route. Thus a colloidal boehmite sol is prepared by hydrolysis of aluminum-tri-secbutylate at a temperature above 80oC. Two litres of water are used per mole of alkoxide. In order to peptize the sol particles 0.07 mol HNO3 per mole alkoxide is added.After the addition of alkoxide, the sol keeps boiling in the open reactor for 30 min to evaporate most of the butanol that is formed. Finally the solution is kept at 90oC under reflux conditions for about 16 h in order to form a stable boehmite sol. The concentration of the final sol is 0.92 M.9–11 For the synthesis of unsupported systems small amounts of boehmite are poured into different petri dishes and subsequently dried for 4–5 days under ambient conditions. After drying, calcination follows at 600oC for 3 h with a heating rate of 60oC/h. Supported membranes are developed on the surface of macroporous Į-alumina supports that NGK INSULATORS LTD provided to us. 12 The support is plate shaped with a diameter of 25 mm and a thickness of 2 mm. It consists of three successive layers. The final layer on the top of which the Ȗ-alumina films are developed has a thickness of 10 ȝm with a pore size of 100 nm. The preparation of the supported membranes proceeds through the slipcasting mechanism by bringing the dry support into contact with the colloidal sol of boehmite (Figure 4). Capillary forces created from the porous support
CLEAN FOSSIL FUELS
205
Support
Capillary pressure
Boehmite sol
Gel formed on the top of the support
Figure 4. Schematic representation of the dipping process.
pull the liquid into the structure. The concentration of the boehmite particles increases on the surface of the support and gelation occurs. After the deposition, drying and calcinations is performed at high temperatures to form the final microstructure. In order to improve the rep reducibility of this method an amount of dissolved poly-vinyl alcohol (PVA) is added to the boehmite sol. The final concentration of PVA in the sol is 2.5 wt%. It is possible that during the dipping procedure pinholes are formed due to the presence of irregularities or defects on the support surface or due to the presence of air bubbles in the sol. Repeating the dipping, drying and calcination steps several times can repair this type of defects. Note that multiple dipping results in an increase in membrane thickness and can be used to modify this parameter. The pore size is the important parameter for the separation performance of the membranes. For this reason unsupported Ȗ-alumina membranes are characterized by nitrogen adsorption-desorption at –196oC to calculate the pore size distribution. The samples are initially out-gassed under vacuum and at a temperature of 150oC for 16 h so as to remove moisture from the structure. The nitrogen adsorption-desorption diagram is type IV, characteristic for mesoporous solids with the typical hysteresis13 (Figure 5). The specific surface area of the membranes is estimated at 235 m2/g using the BET (Brunauer, Emmett and Teller) equation at low relative pressures (P/Po <0.4) whilst the porosity of the membranes is estimated at 48%. The pore size distribution of the Ȗ-alumina membranes is estimated by the nitrogen desorption isotherm applying the Density Functional Theory (DFT) theory, that takes into account the interactions of nitrogen with the specific material.14 The mean pore diameter of the membranes is 5.5 nm with a rather broad pore size distribution of 3–8 nm.
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
206
0.01
160
0.009
140
0.008
120
0.007
Dvc/Ddp (cm 3 /g/A0 )
Volume of adsorbed gas (cm3 STP/gr)
180
100 80 60 40
Ȗ-Al2O3 calcined at 600 C o Ȗ-Alfor 2O33calcined hr in airat 600 C for 3 hr in air
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.005 0.004 0.003 0.002
o
20
0.006
0.001 0 1
Relative pressure P/Po
0
20
40
60
80
100
Pore size dp(A0)
Figure 5. Nitrogen adsorption-desorption isotherm and pore size distribution using DFT model for unsupported Ȗ-Al2O3 thin films.
Characterization of the supported Ȗ-Al2O3 samples was performed with the permporometry technique in order to determine the pore size distribution of these membranes. Permporometry is a characterization method that can provide information of the pore size and the existence of pinholes or cracks in the supported layers. In addition, by using permporometry the discrimination between the open pores, called “active” pores, and the dead-end pores, called “inactive pores”.15 Permporometry is based on the controlled stepwise blocking of pores by condensation of a vapor, present as a component of a gas mixture and the parallel measurement of the gas flux through the membrane (Figure 6). The relative pressure of water is controlled by the flow of Helium at the feed side. By applying a pressure drop of 200–500 mbar along the membrane, a mixture of He/H2O is forced to pass through the membrane. As the relative pressure of the condensable gas increases capillary condensation follows the adsorption mode blocking the pores for the diffusion of the non-condensable gas (Figure 7). The relative pressure at which pore filling starts depends on the radius of the capillary and can be calculated from the Kelvin relation (Eq. (7)).
ln(
P ) P0
2JV cos(T ) r RT
(7)
Where P is the actual vapor pressure, Po is the saturated vapor pressure, Ȗ is the surface tension, V is the molar volume of condensate gas, R is the universal gas constant, T the temperature, r is the radius of the capillary and ș is the contact angle of the liquid on the solid.
CLEAN FOSSIL FUELS
207
Figure 6. Permporometry separation unit.
Prel = 0, All pores are open
0 < Prel < 1, Small pores are filled with condensate
Prel = 1 All pores are blocked Figure 7. Blockage of the pores as a function of the relative pressure (Prel).
The flux of He through the Ȗ-Al2O3 membranes is shown in Figure 8 during the desorption and adsorption of H2O vapor. This typical flux vs. relative pressure plot can be divided into three main areas. At low relative pressures the flux of He decreases slightly. In this region all the pores are empty and available for transport but still the available space for the diffusion of He decreases due to monolayer adsorption of H2O. At pores
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
208
8.E-06 Adsorption
He Permeance (mol/m2.Pa1.S1)
7.E-06
Desorption Multilayer adsorption, capillary condensation
6.E-06 5.E-06 4.E-06 3.E-06 Monolayer adsorption
2.E-06 1.E-06 0.E+00 0
1
2
3
4
5
Kelvin diameter (nm) Figure 8. Permporometry adsorption and desorption curve of Ȗ-alumina membrane with H2O used as adsorbate.
with size 2 nm < dp < 4 nm the flux of He decreases with relative pressure because multilayer diffusion and capillary condensation takes place. For pores larger than 5 nm the flux of He is almost zero indicating that all the pores are blocked and no cracks or pores larger than 5 nm are present in the structure of this type of membranes. The pore size distribution of the Ȗ-Al2O3 membranes is shown in Figure 9 for the adsorption and desorption mode. For the determination of the pore size of the active pores the desorption branch is used due to the fact that the equilibrium of the adsorption process is more difficult to reach. Therefore, the quantitative analysis of the desorption process is preferred.16 The pore size distribution of the active pores of the Ȗ-Al2O3 membranes is rather sharp and narrow with a mean pore size of 2.5 nm according to the desorption branch. It is clear that membranes exhibit a very well defined structure, which is almost in agreement with results found with N2 physisorption. Small differences between the pore size of unsupported and supported samples arise from the fact that during drying and calcination, stresses created by the presence of the support result in a more closed structure for the supported samples.17 Scanning electron microscopy (SEM) and transition electron microscopy (TEM) pictures of the supported Ȗ-alumina membranes confirm the absence of cracks or large pores (Figure 10). The interface between the membrane and the support reveals that the structure of the membrane is composed by primary particles which are stacked in a
CLEAN FOSSIL FUELS
209
dF(cc/bar/m)
3.E+04 Desorption
2.E+04
Adsorption
2.E+04 1.E+04 5.E+03 0.E+00 0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Kelvin diameter (nm) Figure 9. Pore size distribution of supported Ȗ-Al2O3 membranes.
way that the large planes are parallel to each other. The final structure of the membranes is a card-pack structure with slit-shaped pores.10,18 The temperature stability of Ȗ-Al2O3 films was studied having in mind that these membranes are going to be modified with the Atomic Layer Deposition (ALD) technique qt high temperatures. In addition, these membranes are going to be used in steam reforming reaction, which will be exposed to high temperatures necessary for substantial CH4 conversion. In Table 1 the most important structural properties of samples heated in air for the different times are presented. Ȗ-Al2O3
Į-Al2O3
Figure 10. SEM micrograph and TEM micrograph of the interface between the membrane and the macroporous cross section of a supported Ȗ-Al2O3 membrane.
All the samples remain mesoporous with a pore size smaller than 10 nm. Significant differences start appearing when the samples are heated for 1 week at 600ºC. The pore size increases almost 11% in comparison the initial pore
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
210
TABLE 1. Structural properties of the Ȗ-alumina films by N2 physisorption.
Samples Ȗ-Al2O3(3 h/600oC) Ȗ-Al2O3 (1 week/600oC) o
Ȗ-Al2O3(3 weeks/600 C)
Mean pore size (nm)
Pore volume (cm3/g)
Porosity (%)
5.88
0.23
46
BET surface (m2/g) 198
6.56
0.26
49
176
6.56
0.28
51
184
size, while the surface area is increasing due to sintering effects. Longer time (3 weeks) at 600ºC do not seem affect to the structure, as the mean pore size remains constant.
5. Synthesis of Microporous Membranes It is known that the hydrophilic surface of the silica membranes leads to pore blocking by H2O and especially at higher temperatures this can seriously degrade the microstructure.7 For this reason composite SiO2-Al2O3 membranes were synthesized with various silica-alumina ratios, which are mesoporous, microporous or a combination of both. Composite membranes were synthesized by mixing the colloidal boehmite and polymeric silica sols in different proportions. The polymeric silica sol is prepared via the sol-gel technique by hydrolysis tetraethylorthosilicate (TEOS) in ethanol.7 From Figure 11 it is clear that upon reducing the concentration of SiO2 in the composite samples the micro porosity of the materials is reduced. The samples with low SiO2 concentrations (until 20% mol) comprise a mesoporous structure with a mean pore size of 2.5 nm. On the other hand, samples with SiO2 con-centrations between 20% and 50% mol exhibit a bimodal pore size distribution. The micropores have a size of 0.55 nm and a rather broad pore size distribution. At the mesoporous regime the pores have a size of 2.5 nm. The micropores have a size of 0.55 nm and a rather broad pore size distribution. At the mesoporous regime the pores have a size of 2.5 nm. The isotherms of these samples show the characteristic hysteresis of the mesoporous solids. Increasing the concentration of SiO2, the hysteresis starts to vanish and the isotherms tend to become type I. Thin films with concentrations of SiO2 higher than 50% mol have a totally microporous character without mesopores, whilst the micropore size remains 0.55 nm.
CLEAN FOSSIL FUELS
211
200 180
Volume (cm 3 /g)
160 140 120 100
36 mol%SiO2 y-Alumina, 600ºC 36 mol% SiO 2 25.5 mol% SiO2 Ȗ-Al2O3 SiO2 25 mol% SiO2 2 53 SiO mol%SiO2
80 60 40 20 0 0
0.2
0.4
0.6 53 mol% 0.8SiO2
1
Relative pressure P/Po
Figure 11. N2 adsorption-desorption isotherms for Ȗ-Al2O3, SiO2 and composite films calcined at 600ºC for 3 h in air.
The pore size distribution of the thin films is estimated from the nitrogen desorption isotherm applying Density Functional Theory (DFT) model for the mesopores, which takes into account the interactions of nitrogen with the specific material. For the micropores the Horvath Kowazoe model is applied for the determination of its pore size.19 Preliminary EXD experiments reveal that the samples are homogenous without segregation of a separate phase of Al2O3 or SiO2. Supported composite membranes are synthesized by dip coating of Į-Al2O3 supports (NGK INSULATORS LTD), with a mean pore size of 100 nm, in a sol containing 53% mol SiO2. Permporometry measurements for a sample dipped one time with a sol containing polyvinyl alcohol 1.5 wt % are presented in Figure 13. The permeation curve in this case is characteristic for microporous membranes (pore diameter smaller than 2 nm). At small relative pressures a minor reduction of the He permeance is observed due to the blocking of micropores. At higher relative pressure the permeance of He reaches a plateau. The density of the defects of this membrane is high which explains the large final permeability of He around 1·10–5 mol/m2 s·Pa. Mesopores are not present in this membrane, which actually is in agreement with the N2 physisorption results for the unsupported flakes. Another interesting remark is that desorption curve does not follow the adsorption possibly due to strong adsorption of H2O at the hydrophilic surface of SiO2.
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
212
36.3 mol% SiO2 Dv (d) (cm 3/A 0 /g)
0.01
25.5 mol% SiO2
0.008
y-Al2O3
0.006 0.004 0.002 0 0
10 20
30
40
50 60
70
80 90 100
3
Dv (d) (cm /nm/g)
Pore w idth (A0)
1
53 mol% SiO2 36 mol% SiO2 25 mol% SiO2
0.8 0.6 0.4 0.2 0
0. 0. 0. 1 1. 1. 1. 1. 4 6Pore 8 width 2 (nm) 4 6 8 Figure 12. Pore size distributions by applying the DFT model (left) and the HK model.
Permeation He [mol/ m2.Pa1.S1]
2.0E-05 1.8E-05 Adsorption
1.6E-05
Desorption
1.4E-05 1.2E-05 1.0E-05 8.0E-06 6.0E-06 4.0E-06 2.0E-06 0.0E+00 0
1
2
3
4
5
6
7
8
9
10
Kelvin diameter [nm] Figure 13. Permporometry adsorption and desorption curves of an alumina silica composite mem-brane with 53% mol SiO2.
CLEAN FOSSIL FUELS
213
6. Conclusions Mesoporous supported Ȗ-Al2O3 membranes with a sharp and narrow pore size distribution and a mean pore size of 2.5 nm are synthesized via the sol-gel method. The Ȗ-alumina membranes exhibit very good temperature stability. Microporous silica and alumina-silica composite membranes are synthesized via the sol-gel method as well. Microporous pure silica and composite membranes have a pore size of 0.5 nm. ACKNOWLEDGEMENTS
The authors are grateful to Professor F. Kapteijn for providing the set-up for nitrogen adsorption-desorption measurements. Dr. W. Haije is gratefully acknowledged for determining the pore size distribution of the microporous materials. Finally, the authors would like to thank Dr. P. Kooyman for the TEM micrographs and Dr. J. Vente for having contributed to permporometry studies. GCEP sponsors are acknowledged for financial support.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
Smalley, R. (2003) Energy & Nanotechnology Conference, Rice University, Houston, May 3, 2003. http://www.grida.no/climate/vital/index.htm. NASA Facts (1994) Goddard Space Flight Center, March 1994. NF-222. Schoonman, J., Perniu, D., and Van de Krol, R. (2007) Photo-Electrochemical Production of Hydrogen and In-Situ Storage of Hydrogen. NATO-ASI, June 4–15, 2007, Sinaia, Romania. This Proceedings. Bredesen, R., Jordal, K., and Bolland, O. (2004) Chem. Eng. Process. 43, 1129–1158. Mc Leary, E., Jansen, J., and Kapteijn, F. (2006) Micropor. Mesopor. Mat. 90, 198–220. Iler, R. (1979) The Chemistry of Silica, Wiley, New York, Chapter 6. Guizard C. (1996) Sol-gel chemistry and its application to porous membrane processing, Fundamentals of Inorganic Membrane Science and Technology, Vol. 4, Edited by A. J. Burggraaf and L. Cot, Elsevier, Amsterdam, 227–258. Yoldas, B. (1975) Amer. Ceram. Soc. Bull, 54, 289. Leenaars, A., Keizer, K., and Burggraaf, A. (1984) J. Mater. Sci. 19, 1077. Kikkinides, E., Stoitsas, K., and Zaspalis, V. (2003) J. Colloid Interface Sci. 259, 322. http:/ww.ngk.co.jp Greg, S., and Sing, K. (1982) Adsorption Surface Area and Porosity, second edn., Academic, New York, 25. Lastoskie, C., Gubbins, K., and Quirkeft, N. (1993) J. Phys. Chem. 97. Cuperus, F., Bargeman, D., and Smolders, C. (1992) J. Membr. Sci. 57. Sing, K. S. W. (1996) Physical Adsorption: Experiment, Theory and Applications, Kluwer, Dordrecht, The Netherlands.
214
T.TRAN, K. STOITSAS, AND J. SCHOONMAN
17. Kumar, K. (1993) Nanostructured Ceramic Membranes: Layer & Texture Formation Ph.D. thesis. 18. Stoitsas, K., Sklari, S., Kikkinides, S., and Zaspalis, V. (2006) Microporous Aluminophosphate and Silica Ceramic Membranes for Hydrogen and Olefin/Paraffin Separation 9th International Conference on Inorganic Membranes, Norway, 149–152. 19. Horvath, G., and Kawazoe, K. (1983) J. Chem. Eng. Jpn. 16, 470.
NANOCRYSTALLINE DIAMOND FILMS FOR ADVANCED TECHNOLOGICAL APPLICATIONS
C. POPOV1* AND W. KULISCH2 Institute of Nanostructure Technologies and Analytics, University of Kassel, Heinrich-Plett-Str. Kassel, GERMANY 2 Institute for Health and Consumer Protection, Joint Research Centre, Ispra, ITALY 1
Abstract – Nanocrystalline diamond films have been grown by microwave plasma chemical vapor deposition from methane/nitrogen mixtures. The films – composed of diamond nanocrystallites (3–5 nm) embedded in an amorphous carbon matrix with grain boundary widths of 1–1.5 nm – were investigated with respect to their possible structural, tribological, biomedical and biosensoric applications.
Keywords: Nanocrystalline diamond films, biomedical applications, biosensors.
1. Nanocrystalline Diamond Films Nanocrystalline diamond (NCD) films have drawn the attention of the scientists in the last years because of their exceptional properties like high mechanical strength, low friction coefficient, high chemical stability, low electron emission threshold voltage, etc.1 The small grain size (typically 3–5 nm) makes the films valuable for tribological, field emission and structural applications. Using topdown approaches, two- and three-dimensional structures can be fabricated from NCD films. The small grain sizes lead to higher feature resolution compared with polycrystalline diamond (PCD) films, where the grain sizes are on the order of several microns. The submicron resolution, combined with excellent mechanical properties, makes NCD films a successful competitor of silicon for the preparation of micro- and nano-electromechanical systems (MEMS and NEMS).2
______ *
To whom correspondence should be addressed: Cyril Popov, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
215
216
C. POPOV AND W. KULISCH
In addition to all its outstanding mechanical, tribological, thermal, optical and chemical properties, diamond possesses the advantage of being inherently biocompatible.3,4 As a consequence, in recent years many studies have been performed considering the application of diamond thin films in the fields of biomedicine, biotechnology and biosensorics. Among others, diamond films have been proposed as a material for glycose sensors,5 for catheter ablation in the heart,6 as encapsulation for retinal implant microchips,7 and especially as a template for the immobilization of biomolecules for biological investigations or for biosensoric applications.8–10 PCD films suffer usually from a very high surface roughness3 which prevents their applications in many bio-related fields, e.g. as a coating for implants. A solution can be found in NCD films, either in pure form or as a NCD/a-C composite.1,11 Such films are very smooth but still retain most of the extreme properties of PCD; consequently, bio-applications such as BioMEMS sensors,4 coatings for temperomandibular joints,12 and again as a template for immobilized biomolecules13,14 have been reported. For the latter application it is of utmost importance that nanocrystalline diamond surfaces do not interact unspecifically with biomolecules such as RNA or proteins. In the present work, we report on the results from the investigations of thin nanocrystalline diamond films for structural, tribological, biomedical and biosensoric applications. 2. Growth and Basic Properties Nanocrystalline diamond films were prepared by microwave plasma chemical vapor deposition (MWCVD) from CH4/N2 mixtures in a deposition set-up described in detail elsewhere.15 The deposition experiments were performed at substrate temperatures between 520°C and 770°C, a working pressure of 26 mbar, with CH4 concentrations in the gas phase of 9% or 17%, and a MW plasma input power of 800 W. The duration of the deposition processes was 390–420 min. The films were comprehensively characterized with respect to their composition, morphology, topography, crystallinity and bonding nature by a variety of techniques. NCD films deposited under the above conditions consist of diamond nanocrystallites with diameters of 3–5 nm, embedded in an amorphous matrix.16,17 The films are continuous, smooth (rms roughness 12 nm) and almost featureless although top-view SEM images showed the existence of some rounded features at the surface with diameters of 0.5–1 Pm, which in turn seem to possess a kind of substructure.11 The diamond nanocrystals are randomly oriented; they are separated by the matrix forming grain boundaries of 1–1.5 nm width. The volume ratio matrix/crystallites is about unity. The density of the films is about 2.75 g cm–3 as determined by XRR,16 from which a matrix density of ca. 2 g cm–3
NANOCRYSTALLINE DIAMOND FILMS
217
can be calculated. The matrix itself is a mixture of sp2- (20–30%) and sp3-bonded units; the films contain about at.10% hydrogen which is almost exclusively bonded in the matrix to sp3-carbon atoms. 3. Structural Applications 3.1. SELECTED AREA DEPOSITION
NCD films were coated on complex Si/SiO2 structures by selected area deposition (SAD) using silicon oxide masks. The process combines standard lithography, ultrasonic pretreatment in a suspension of diamond powder, and MWCVD. The advantage of this technique is that the random growth of undesirable diamond off-side the selected areas can be reduced to a large extent, and even entirely avoided by etching off the oxide in HF/NH4F in an ultrasonic bath before as well as after the deposition. NCD films were deposited on different Si/SiO2 structures, e.g. trenches and inverted pyramids. The flat surfaces at the top were coated with SiO2, the inclined side walls and the bottom of the structures were bare silicon. As seen from Figure 1, after SAD only isolated nodules of NCD are present on the top SiO2 coated surface, while the inclined walls inside the pyramids are completely covered with a NCD film. The rims of the pyramids served as active nucleation sites; as a result well formed frames built up of NCD nodules can be observed at the rims (Figure 1). In a similar way the side walls of trenches were coated with a continuous NCD film.
Figure 1. Selected area deposition of NCD films inside inverted pyramids (left) and trenches (right).
Closer observation reveals a difference of the thickness of the film at the rims and that of the isolated nodules on the top, on the one hand, and that on the walls, on the other hand. Geometrical reasons could be supposed to be responsible for the different growth rates at different positions; for example, the
218
C. POPOV AND W. KULISCH
space angle, from which the film-forming species can hit the surface of the substrate in the vicinity of a growing nodule, depends on its position. This geometrical factor should be taken into account by deposition of NCD on complex structures for preparation of diamond cantilevers, diamond tips, etc. 3.2. DEPOSITION ON ULTRASHARP SILICON WHISKERS
Ultrasharp silicon whiskers with heights of about 40 μm served as nucleation sites for the growth on NCD. As a result, almost perfect NCD balls were obtained on top of part of the whiskers (Figure 2). In this case, the absence of geometrical constraints promotes spherulitic growth around the nucleation site. This growth mechanism, observed for “ballas”,18 a natural form of nanocrystalline diamond, requires a low density of nucleation sites which is entirely fulfilled in this case, and a high rate of secondary nucleation, a specific condition for the growth of NCD.11 The diameter of the NCD balls can be controlled by the growth rate, i.e. by the process parameters, like the substrate temperature or the CH4 content in the gas phase. For the conditions used during this experiment, the growth rate was 10 nm/min, a little bit higher than that on the top of the Si/SiO2 structures.
Figure 2. NCD ballas on top of a Si whisker.
Figure 3. NCD/PCD layer system.
3.3. DEPOSITION OF LAYER SYSTEMS
For many applications it is important to prepare multiplayer systems. NCD films were successfully deposited onto a number of layers of technological interest, such like SiO2, Si3N4, TiN, c-BN, PCD. The latter system is of great interest for future advanced mechanical and tribological applications, if it turns out possible to combine the advantages of both materials, e.g. the extreme mechanical properties of polycrystalline diamond and the still very good mechanical
NANOCRYSTALLINE DIAMOND FILMS
219
properties19 together with the much improved surface roughness of NCD.1,16 Figure 3 shows a NCD film deposited on a PCD layer. As seen from the SEM image, some rounding and smoothening of the originally very rough PCD layer by the NCD film is achieved; this effect can be increased by increasing the thickness of the NCD film. We have grown also a PCD film on top a NCD layer. In both cases, no substrate pretreatment has been used prior to the deposition of the top layer: both underlayers, either PCD or NCD, provide the required nucleation sites for the subsequent films. As a result, a kind of “epitaxial” growth (with sharp interfaces free of voids) is observed although we cannot speak about real homoepitaxy due to the completely different rates of secondary nucleation: very high for NCD and almost none by PCD. These experiments have demonstrated that such nano/poly diamond multilayers could be prepared in the same reactor by simply switching the gas phase composition. 4. Tribological Applications 4.1. NANOHARDNESS
The hardness of the NCD films on different substrates was studied by nanoindentation measurements with a diamond Berkovich indenter applying a linear loading rate of 8 mN/min up to a maximum load of 4 mN. All measurements showed similar nanoindentation load/displacement curves, which yielded comparable results for the film hardness and Young’s modulus (Figure 4), with the exception of the TiN sample. The average indentation hardness is 34–40 GPa, the indentation modulus 325–390 GPa, and the elastic recovery about 75– 78%.20 The large error observed for the PCD/NCD system can be attributed to the rough surface of this coating (Figure 3). The absolute values of the hardness are considerably lower than those for bulk diamond and other NCD films (100 and 70–90 GPa, respectively; see Ref. 19 and the reference cited therein). A possible reason for the somewhat low values found in our study may be the presence of the amorphous matrix with a volume fraction of 40–50%. On the other hand, the tribological performance of a wear protecting coating is not only determined by its hardness; rather, the toughness is of even higher importance. Here, the presence of the amorphous matrix may even be of advantage as it can help to prevent fatal brittle failure. For the TiN substrate, the hardness is considerably lower; this can be explained by the indentation depth of ca. 300 nm, which means that the substrate, which is softer in the case of TiN than for the superhard materials diamond and c-BN, contributes to these values. On the other hand, for the measurements with the silicon substrate, the coating was thick enough to exclude a substrate contribution.
220
C. POPOV AND W. KULISCH
Figure 4. Hardness (left) and Young’s modulus (right) of NCD films on different substrates.
4.2. WEAR RESISTANCE AND FRICTION COEFFICIENT
The Nano Tribo tests on the NCD/a-C films have been performed in ball-ondisk configuration with an Al2O3 ball as counterpart.19 No delamination (neither inside nor alongside the tracks) has taken place during the tests of the sample on Si even after 10,000 laps (Figure 5). From the width of the visible track of almost 300 Pm a considerable wear of the Al2O3 ball must be concluded, which has very probably produced some amount of debris. The friction coefficient P recorded during the same measurement shows initially a rather high value of about 0.5; it decreases gradually until after about 4,000 laps a value of 0.1 or even lower is reached. During the entire measurement, frequent peaks of the friction coefficient reaching up to 0.6 were observed, presumably caused by wear debris, most probably from the Al2O3 ball. The initial high value of P at the beginning of the measurement is most probably a result of the short and medium range roughness of the film surface, observed by SEM and AFM; these protrudings are worn off at the begin of the test, which may also have contributed to the production of debris. 4.3. ADHESION
For the nanoscratch tests, performed with a progressive loading at a scan speed of 5 mm/min, a Rockwell C diamond indenter was used. The critical loads for the first cracking, the first rupture and the full delamination were determined from the friction force/ penetration depth curves. Very good results for the NCD film adhesion were obtained for Si and TiN substrates20 (Figure 6). Owing to the rough surface of the PCD/NCD system (Figure 3), scratch tests were difficult to perform. The low critical load for delamination of the c-BN/NCD
NANOCRYSTALLINE DIAMOND FILMS
221
system can be explained by the rather poor adhesion of c-BN films on silicon substrates, i.e. the failure is determined by the weakest link, the poor adhesion between Si substrate and c-BN layer. Full delamination had in some cases not taken place at the final load of 275 mN. Images of the scratches show that delamination, once occurred, is not restricted to the width of the scratch but took place in much larger areas.19 In the delaminated areas the underlaying substrate is heavily damaged, which proves the protecting nature of the coating. 300
Critical load (mN)
250 200 150 100 50 0
Si
c-BN
TiN
Substrate
Figure 5. Wear track on NCD film after 10,000 laps.
Figure 6. Adhesion of NCD films on different substrates.
5. Biomedical and Biosensoric Applications 5.1. CELL PROLIFERATION AND BIOACTIVITY
The biocompatibility of NCD films was studied by direct contact tests with osteoblast-like cells, fibroblasts and endothelia cells.21–23 All cells showed good adhesion and spreading on the NCD surfaces following the incubation (Figure 7). After several days of cultivation they formed confluent monolayers; comparisons with cells from control samples showed that the NCD films are not cytotoxic and do not affect the cell viability and proliferation. The NCD coatings are also bioinert as revealed by simulated body fluid (SBF) tests. The exposure to SFB with a composition close to that of blood plasma for 10 days did not result in the formation of hydroxyapatite as shown by analyses of the SBF composition and of the film surface.21 5.2. IMMOBILIZATION OF BIOMOLECULES
We investigated the attachment of double-stranded RNA and proteins onto NCD surfaces by scanning probe microscopy (SPM). The biomolecules did not
222
C. POPOV AND W. KULISCH
stick to the NCD surface, in contrast to glass or mica substrates. To exclude the possibility that proteins had stuck to the NCD surface, but had fortuitously not been imaged, we compared the forces exerted between a bovine serum albumin (BSA) functionalized AFM cantilever and a glass or NCD surface. An interaction between the BSA cantilever and the glass surface was observed in the force curves in 38% of all measurements, but not even once for the NCD surface, thus confirming that the proteins do not interact unspecifically with the NCD surface.22
Figure 7. Fibroblast (left) and endothelial (right) cells grown on NCD films.
In a series of experiments, we have further investigated whether double stranded RNA could be deposited directionally on the functionalized surface of the NCD films.23 For this purpose, the as-grown films were treated in microwave hydrogen plasma to achieve H-termination of the NCD surface. The surface modification was followed by photochemical attachment (O = 254 nm) of NH2containing molecules (in our case 1-amino-3-cyclopentene hydrochloride) and deprotection of the amine group. The resulting amine-terminated surface of the NCD films was then functionalized with sulfosuccinimidyl 4-(N-maleimidomethyl) cyclohexane-1-carboxylate (SSMCC) which serves as a linker for thiol-terminated DNA. These DNA oligonucleotides (30 nucleotide long), attached to the surface of the NCD films as described above, played the role of anchoring points. The 374 base pair long dsRNA transcript used features a single stranded 5ƍ end that can form Watson-Crick base pairs with the anchoring DNA oligonucleotide. When the RNA sample was applied to NCD films subjected to the above functionalization scheme, a number of small objects on the order of 20 nm can be seen in the amplitude AFM image (Figure 8, left), and a fraction of these is also visible in the height image. The size of the objects observed, however, deviates from that of a perfect A-type RNA helix of 374 base pairs, which is expected to be about 100 nm in length. Presumably, this size difference can be attributed to the fact that RNA molecules do not interact unspecifically with the surface of NCD films as discussed above. Rather, if hit by the tapping cantilever
NANOCRYSTALLINE DIAMOND FILMS
223
in the AFM experiments, they might be moved on the NCD surface, which does not keep them in place, unlike the positively charged mica surfaces conventionally used in AFM experiments. This also implies that the RNA molecules, with the exception of their 5' end base-paired to the anchoring DNA oligonucleotides, are freely accessible on the surface. This feature might be beneficial for future force measurements of the interaction of double stranded RNA with protein domains. To prove that the objects in Figure 8 (left) indeed are RNA molecules and to exclude that they are contaminants incidentally present on the surface with the RNA sample, we incubated the surface of functionalized NCD films with the identical buffer that has been used to apply the RNA molecules. The images of this surface reveal the complete absence of any 20 nm structures (Figure 8, right). This indicates that the structures seen in Figure 8 (left) indeed are RNA molecules that have bound to the anchoring DNA oligonucleotides on the functionalized surface of the NCD films.
Figure 8. AFM phase images of functionalized NCD films after interaction with RNA in TE buffer (left) and upon exposure to TE buffer only (right).
6. Summary The results from the investigation of some application relevant properties of NCD films showed that they have a great potential in fields like micro- and nano-machining for the production of MEMS/NEMS, membranes and cantilevers, in tribology as wear resistant coatings and in biomedicine and biosensorics as coatings for implants and heart valves, and as templates for immobilization of biomolecules in biosensors and DNA chips.
224
C. POPOV AND W. KULISCH
References 1. Gruen, D. (1999) Annu. Rev. Mater. Sci. 29, 211. 2. Gruen, D. (2001) In: Properties, Growth and Applications of Diamond, edited by M.H. Nazare and A.J. Neves (INSPEC, London), 313 3. Koidl, P., and Klages, C. (1992) Diamond Relat. Mater. 1, 1065. 4. Carlisle, J., and Auciello, O. (2003) Electrochem. Soc. Interface 12, 28. 5. Troupe, C., Drummond, I., Graham, C., Grice, J., John, P., Wilson, J., Jubber, M., and Morrison, N. (1998) Diamond Relat. Mater. 7, 575. 6. Müller, R., Adamschik, M., Steidl, D., Kohn, E., Thamasett, S., Stiller, S., Hanke, H., and Hombach, V. (2004) Diamond Relat. Mater. 13, 1080. 7. Xiao, X., Wang, J., Liu, C., Carlisle, J., Mech, M., Greenberg, R., Guven, D., Freda, R., Humayun, M., Weiland, J., and Auciello, O. (2006) J. Biomed. Mater. Res. B 77, 273. 8. Hamers, R., Butler, J., Lasseter, T., Nichols, B., Russell Jr., J. Tse, K., and Yang, W. (2005) Diamond Relat. Mater. 14, 661. 9. Strother, T., Knickerbocker, T., Russell Jr., J., Butler, J., Smith, L., and Hamers, R. (2002) Langmuir 18, 1968. 10. Knickerbocker, T., Strother, T., Schwartz, M., Russell Jr., J., Butler, J., Smith, L., and Hamers, R. (2003) Langmuir 19, 1938. 11. Kulisch, W., and Popov, C. (2006) Phys. Stat. Sol. (a) 203, 203. 12. Papo, M., Catlegde, S., Vohra, Y., and Machado, C. (2004) J. Mater. Sci. Mater. Med. 15, 773. 13. Yang, W., Auciello, O., Butler, J., Cai, W., Carlisle, J., Gerbi, J. Gruen, D., Knickerbocker, T., Lasseter, T., Russell Jr, J., Smith, L., and Hamers, R. (2002) Nat. Mater. 1, 294. 14. Wang, J., Firestone, M., Auciello, O., and Carlisle, J. (2004) Langmuir 20, 11450. 15. Popov, C., Novotny, M., Jelinek, M., Boycheva, W., Vorlicek, V., Trchova, M., and Kulisch, W. (2006) Thin Solid Films 506–507, 297. 16. Popov, C., Kulisch, W., Gibson, P., Ceccone, G., and Jelinek, M. (2004) Diamond Relat. Mater. 13, 1371 17. Popov, C., Kulisch, W., Boycheva, S., Yamamoto, K., Ceccone, G., and Koga, Y. (2004) Diam. Relat. Mater. 13, 2071 18. Lux, B., Haubner, R., Holzer, H., and de Vries, R. (1997) Int. J. Refrac. Metals Hard Mater. 15, 263. 19. Kulisch, W., Popov, C., Boycheva, S., Buforn, L., Favaro, G., and Conte, N. (2004) Diamond Relat. Mater. 13, 1997. 20. Kulisch, W., Popov, C., Vorlicek, V., Gibson, P., and Favaro, G. (2006) Thin Solid Films 515, 1005. 21. Popov, C., Kulisch, W., Jelinek, M., Bock, A., and Strnad, J. (2006) Thin Solid Films 494, 94. 22. Popov, C., Kulisch, W., Reithmaier, J., Dostalova, T., Jelinek, M., Anspach, N., and Hammann, C. (2007) Diamond Relat. Mater. 16, 735. 23. Popov, C., Bliznakov, S., Boycheva, S., Milinovik, N., Apostolova, M., Anspach, N., Hammann, C., Nellen, W., Reithmaier, J., and Kulisch, W. (2007) Diamond Relat. Mater. (submitted for publication).
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION S. TAMULEVIýIUS* AND Š. MEŠKINIS Institute of Physical; Electronics of Kaunas University of Technology, Savanoriǐ 271, 50131 Kaunas, LITHUANIA
Abstract – Present study is devoted to the application of the DLC films as wear resistant coatings for protection of surfaces of the steel tools as well as to the investigation of the optical and hydrophobic properties of SiOx and silicon doped DLC films. It was found, that in the case of the deposition of DLC on the steel surface chemical composition of DLC/interlayer interface as well as mechanical stress in interlayers and composition of hydrocarbon gas should be taken into account. Raman scattering spectra of the all synthesized amorphous carbon films were typical for DLC films. In the case of SiOx containing DLC films, Raman scattering spectra additional features typical for trans-polyacetylene-like segments has been observed. Contact angle with water of the all investigated films did not depend on the deposition conditions. Absorption coefficient of HMDSO + C2H2 films was several times larger than absorption coefficient of the HMDSO + H2 films, but substantially lower than absorption coefficient of DLC films deposited from acetylene gas. Additional Ar or N2 gas flow during the deposition resulted in increased optical transparence of SiOx doped DLC films (HMDSO + H2 films). Despite lower absorption coefficient, optical bandgap of HMDSO + C2H2 DLC films was smaller that optical bandgap of “conventional” hydrogenated DLC film.
Keywords: DLC, ion beam synthesis, SiOx containing DLC, XPS study, Raman scattering spectroscopy, optical properties, contact angle with water.
______ *
To whom correspondence should be addressed: S. Tamuleviþius, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
225
226
S. TAMULEVIýIUS AND Š. MEŠKINIS
1. Introduction Diamond like carbon (DLC) and related films received considerable interest due to their outstanding properties. Diamond like carbon it is a metastable form of the amorphous carbon containing significant amount of sp3 bonds.1 Due to that structural feature, properties of the diamond like carbon films are similar to the properties of the diamond. Hydrogenated DLC films are synthesized by different PECVD methods as well as by ion beam synthesis using different ion sources. Ion beam synthesis offer advantages of the increased deposition process control such as independent control of the ion energy and ion current density. Particularly DLC ion beam synthesis by closed drift ion source is already used in industry due to its relatively simple maintenance, because it is griddles and filamentless ion source. Despite numerous research devoted to the use of the diamond like carbon films in medicine, micro- and optoelectronics the main application areas of that films remains applications related to tribological and optical properties of DLC. Present study is devoted to the two applications of the diamond like carbon films. The first one is related to application of the DLC films as wear resistant coatings for protection of surfaces of the tools or biomedical implants.2–4 In this case hydrogenated amorphous carbon (a-C:H) films very often are deposited onto different steels and other ferrous materials. However, adhesion problems often occur in such a case.5–7 Formation of intermediate layers between the ferrous substrate and diamond like carbon film has been the most common approach in order to solve these obstacles for diamond like carbon films deposited on the steel.6 In present research interfaces of the DLC fabricated on different interlayers were investigated. The second application is related to the tuning of optical, hydrophobic and mechanical properties of DLC. Recently nanoimprint lithography received considerable interest as a simple and cost effective alternative to the convenient lithographic techniques such as optical or electron beam lithography.8,9 One of the most important problems to be solved in the nanoimprint lithography is fabrication of the stamps with anti-sticking surface. However, Si, quartz, Ni – the most often used materials for imprint stamp formation – have high surface energy and, as a result, bad antiadhesive properties. Till now low surface energy fluoride films such as Teflon-like coatings10,11 and fluorinated silane layers12,13 were investigated as anti-sticking layers for the nanoimprint stamps. However, protective film degradation due to the fluorine mass transfer at elevated imprint temperatures was observed.10 After certain number of complete imprint cycles this film needs to be replaced. Diamond like carbon (DLC) films can be a good choice for this application due to the combination of the hydrophobicity with outstanding mechanical and tribological properties of these films.1 Thick DLC
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 227
layers were reported as an efficient mould material, where it was used in the imprint technology.14,15 DLC and SiOx doped DLC films are already used as protective coatings for increased life-time of the DVD/CD molds. In present study ion beam deposited DLC films were investigated as possible candidates to increase anti-sticking properties of the imprint lithography stamps. The research was done varying systematically the gas content, investigating doping effect to satisfy criteria for high temperature embossing (nanoimprint lithography) and UV assisted molding (UV nanoimprint lithography). 2. Experimental In present study diamond like carbon films were synthesized on Si(111), steel and quartz substrates by DC ion beam deposition using closed drift ion source.16,17 In all cases ion energy was 800 eV. Acetylene gas has been used as a hydrocarbon source for synthesis of DLC. Nitrogen doped DLC were deposited from the mixture of acetylene and nitrogen. SiOx containing and Si doped DLC films were synthesized using hexamethyldisiloxane vapor and hydrogen gas mixture as well as hexamethyldisiloxane vapor and acetylene gas mixture. Description of the XPS measurement techniques as can be found in Refs.18,19 For Raman scattering spectra measurements please see Ref.20 Wider description of the used optical measurements technique can be found in Refs.20,21 3. Synthesis of Diamond like Carbon Films on Steel: Interfaces, Adhesion, Chemical Composition In present study effects of Ti and a-Si interlayers as well as steel surface nitrogen ion beam treatment and DLC film nitrogen doping on adhesion and interface structure were investigated. For wider information on sample synthesis as well as XPS, AFM and stress measurements techniques please see Refs.18,19 Adhesion of the formed hydrogenated amorphous carbon films was very sensitive to the surface treatment of the steel substrate as well as to the doping. The peeling occurred after exposition of the films to the air if hydrogenated amorphous carbon films were deposited onto the steel substrate without any surface ion beam pretreatment. The 30 min steel pretreatment with N2 ion beam was essential in forming mechanically stable films. However, some partial peeling of these films occurred as well after the more prolonged exposure to air (2–4 weeks). E-beam deposition of the a-Si interlayer onto the steel substrate resulted in substantial improvement of the adhesion between the steel and DLC film. In some cases partial peeling of these coatings occurred only after the
228
S. TAMULEVIýIUS AND Š. MEŠKINIS
more prolonged exposure to air (2–4 weeks). Essential difference of the adhesion properties of a-C:H on Ti sublayer was observed between the hydrogenated amorphous carbon films formed by two different equipments (equipped with turbo-molecular and diffusion pumps). In case of the diffusion pump deposition of the Ti interlayer resulted in bad adhesion with synthesized DLC film when diamond like carbon film was grown immediately after the deposition of the Ti film. No substantial improvement of adhesion due to Ti surface treatment by argon or nitrogen ion beams just before the carbon film deposition was observed. However, prolonged (t2 weeks) aging of the Ti interlayer before deposition of the DLC film resulted in improved adhesion. In this case partial peeling was observed only after the more prolonged exposure to air. While in the vacuum chamber equipped with the turbo-molecular pump adherent DLC films were deposited on both “fresh” and aged Ti interlayer. And finally no peeling was observed for a-CNx:H films deposited onto the nitrogen ion beam treated steel. These films were stable for long time and were suitable for mechanical applications. After the nitrogen ion beam treatment substantial decrease of the metal oxide related O1s, Fe2p and Cr2p XPS peaks has been observed.22,23 The N1s XPS peak (Figure 1) clearly shows presence of the some chromium24 and/or iron25 nitrides as a result of the steel nitridation by ion beam. Measured binding energy of the Cr2p and Fe2p XPS peaks are somewhere between the binding energies typical for chromium and chromium nitride and iron and iron nitride correspondingly.24,25 It seems, that Cr and Fe oxides during ion beam bombardment were replaced by corresponding nitrides. Deposition of the ultrathin a-C:H and a-CNx:H layers resulted in formation of the different carbon and nitrogen bonds that is clearly seen from the shift of the binding energy of N1s XPS peak and fitting of this peak.26 However, some Cr and/or Fe nitrides were still presented.20 However replacement of the lower valence index nitrides to the higher valence index nitrides (e.g., from CrN to Cr2N) possibly takes place. Oxygen in the samples with a-C:H and a-CNx:H overlayers was mainly detected in form of the carbon oxides.22 But some metal oxides (most possibly Cr) still existed.22 It can be mentioned, that during the deposition of the a-C:H and a-CNx:H layers formation of some iron carbides took place.20 However no substantial qualitative differences between the XPS spectra from the ultrathin a-C:H and a-CNx:H layers were found. As was mentioned above, a-C:H films of 500 nm thickness partially peeled from ion beam nitrided steel after the prolonged air exposure. While in the case of the a-CNx:H films no peeling was observed. These differences can be explained by different structural properties of the films such as lower internal stress in a-CNx:H films.18–27 and, possibly, different thickness of the transition a-CNx:H layer.
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 229
Figure 1. The XPS N1s peak evolution due to different treatments of steel surface: (a) ion beam nitridation, (b) ion beam nitridation + deposition of ultrathin a-C:H, (c) ion beam nitridation + deposition of ultrathin a-CNx:H. 530.1
Intensity (a.u.)
O1s
reference
530
nitrogen ion treated a-C:H layer 532.4 531.5 a-CNx:H layer
535
530
Binding energy (eV) Figure 2. O1s x-ray photoelectron spectra: (a) reference sample, (b) ion beam nitridation, (c) ion beam nitridation + deposition of ultrathin a-C:H, (d) ion beam nitridation + deposition of ultrathin a-CNx:H.
Different behavior of hydrogenated amorphous carbon films deposited by ion beam at room temperature on Ti and a-Si interlayers already at the initial growth stages was observed by XPS spectroscopy. After the deposition of ~2.5 nm thickness carbon film (20 s deposition time) on a-Si, substantial suppression
230
S. TAMULEVIýIUS AND Š. MEŠKINIS
reference
576.4
Cr2p
574.9
Intensity (a.u.)
nitrogen ion beam treated
574.7
a-C:H layer
576.6
a-CNx:H layer 590
585
580
575
570
Binding energy (eV) Figure 3. Cr2p x-ray photoelectron spectra: (a) reference sample, (b) ion beam nitridation, (c) ion beam nitridation + deposition of ultrathin a-C:H, (d) ion beam nitridation + deposition of ultrathin a-CNx:H.
of Si2p peak intensity occurred, C1s peak shape is typical for hydrogenated amorphous carbon films – growth of the diamond like carbon film already began with relatively small intermixing. It seems, that carbon film growth without substantial intermixing occurred on the a-Si and no carbide related C1s peak can be seen – SiC compound formation takes place only at the carbon-silicon interface. However, some oxygen atoms present in this layer. On the other hand, carbon layer growth onto the Ti interlayer is more sensitive to the different variations of the deposition conditions (Figure 4). In this case a-C:H film growth depends on whether turbo-molecular or diffusion pump (oil vapor present in vacuum chamber) was used. Analysis of the curvature of substrate-interlayer has shown, that mechanical stress of the Ti films decreased by an order as a result of the aging (Figure 5). Possible presence of the oil vapor results in suppressed growth of the a-C:H film onto the stressed Ti film. While in the case of aged, less stressed Ti interlayer, a-C:H film growth rate is higher and adhesion of the formed film is better.21 However, deposition rate of a-C:H is lower as compared to the a-Si interlayer.21 In all cases strong Ti-O related Ti2p peak was observed even after 60 s deposition when ~5 nm thickness carbon film was already grown on Si(111) (Figure 4). It seems, that in this case strong intermixing occurred
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 231
Figure 4. Ti2p XPS spectra of the Ti interlayer and ultrathin a-C:H films deposited onto it.
Figure 5. Residual stress of the Ti thin films immediately after the deposition and after 2 weeks aging in air (three different samples were measured).
between the growing carbon film and Ti interlayer. The TiC subpeak of C1s peak still can be seen, however Ti atoms are preferentially bound to oxygen, but not to carbon.21 4. Si and SiOx Containing DLC Films: Chemical Composition, Structure, Optical and Hydrophobic Properties SiOx doped DLC films mostly are deposited using hexamethyldisiloxane (or its mixtures with different transporting gases) as a precursor. In present study DLC films were ion beam deposited using hexamethyldisiloxane vapor and hydrogen gas (HMDSO + H2) as well as hexamethyldisiloxane vapor and acetylene gas
232
S. TAMULEVIýIUS AND Š. MEŠKINIS
(HMDSO + C2H2) mixtures. “Conventional” DLC films were synthesized using acetylene as hydrocarbon source for comparison purposes. Ion beam energy in all cases was 800 eV. Ion current density in most experiments was 100 PA/cm2 (otherwise it is stated in text). Thickness of the films in most experiments was ~200 nm (otherwise it is stated in text). 4.1. SI AND SIOX CONTAINING DLC FILMS: CHEMICAL COMPOSITION
Table 1 presents chemical composition of the hydrogenated amorphous carbon films deposited from hexamethyldisiloxane and hydrogen as well as from hexamethyldisiloxane and acetylene gas mixtures. Only 2% atomic fraction of Si in the HMDSO + C2H2 film was found. Amount of the oxygen (9.52%) in the film was comparable with one for undoped DLC films deposited from acetylene without hexamethyldisiloxane or reported for DLC-Si films deposited from tetramethylsilane.27 On the other hand, atomic concentration of O in HMDSO + H2 is higher than that reported in Ref.27 for films deposited from hexamethyldisiloxane by radio-frequency plasma beam, while Si concentration in that films was beyond 30%.27 These differences between present study and previously reported results most probably could be attributed to higher ion energies (energies per atom) used is our case as well as differences between the continuous (DC) and oscillating (radio-frequency) ion bombardment. Position of the Si2p XPS peak of HMDSO + C2H2 film is 100.7 eV – typical for SiC28–31 (Table 2). While binding energy of the Si2p of HMDSO + H2 film is 101.7–101.8 eV – that is typical for SiOx30,31 (Table 2). However, some SiC related Si2p peak shoulder can be seen in spectra of the HMDSO + H2 film as well. O1s peak binding energy was nearly the same for all the analyzed films – 532.4–532.5 eV. This position is closer to the O1s binding energy reported for undoped DLC (532.6 eV) and SiO2 (532.5 eV) than to the SiC (532.8 eV).27 The differences between the chemical composition of the HMDSO + H2 and HMDSO + C2H2 films can be explained in following way. In the case of C2H2 the total amount of the carbon is contributed by two components (both acetylene and hexamethyldisiloxane). As a result atomic concentration of the carbon as well as C:O and C:Si ratios are increased. One can see, that atomic concentrations of the C, Si, O differ only for several percents for HMDSO + H2 films deposited under different synthesis process conditions (Table 1). Si2p XPS peak in all cases can be described as a superposition of the SiC and SiOx (x < 2) and, possibly, SiO2 sub-peaks. Some Si-O-C bonds can be present as well, because position of the summary Si2p peak (101.5–101.8 eV) was described as Si-O-C in Refs.29,32. Si-O-C related29 FTIR absorbance subpeak (shoulder) at 1,100 cm–1 (especially for sample No. 5) is in good accordance with this suggestion. It can
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 233 TABLE 1. Chemical composition of samples. Sample No.
Atomic concentration (%)
1 2 3 4 5 HMDSO + C2H2
Si 20.85 21.64 20.25 21.75 22.36 2.06
O 23.83 27.02 22.55 22.12 21.74 9.52
C 55.32 51.34 57.2 56.12 55.9 88.42
Composition related to Si SiO1.14C2.65 SiO1.25C2.37 SiO1.11C2.82 SiO1.02C2.58 SiO0.97C2.5 SiO4.62C42.92
TABLE 2. Si2p XPS peak binding energy. Sample No.
(HMDSO + H2)/H2 gas flow ratio
Current density (mA/cm2)
Binding energy (eV)
1 2 3 4 5 HMDSO + C2H2
1:1 1:1.5 Only HMDSO + H2 Only HMDSO + H2 Only HMDSO + H2 –
0.1 r 0.01 0.1 r 0.01 0.1 r 0.01 0.05 r 0.01 0.15 r 0.01 0.1 r 0.01
101.7 101.8 101.5 101.5 101.5 100.7
be seen, that additional hydrogen flow results in shift of the resultant Si2p peak to the higher energy range. It possibly can be described by decrease of the SiC related bonding and increase of the SiOx related bonding. 4.2. RAMAN SCATTERING SPECTRA
Raman scattering spectra of the deposited silicon oxide doped hydrogenated amorphous carbon films (HMDSO + H2 and HMDSO + C2H2) can be seen in Figure 6. In both cases spectra are typical for diamond-like carbon.1 The main broad Raman scattering peak at ~1,455 cm-1 (for HMDSO + H2 film) and at ~1,525 cm–1 (for HMDSO + C2H2 film) can be seen. This peak (G-band) is designated as the stretching vibration mode of graphite crystals1 or as a superposition of the G peak and disorder induced (D) peak. In the case of the HMDSO + H2 DLC film deposited by 50 PA/cm2 current density ion beam well pronounced additional shoulder at ~1,150 cm-1 can be seen. It can be seen, that position of the G peak of HMDSO + C2H2 film is very close to the position of the G peak of DLC films deposited under the same technological process conditions (ion energy, ion current density) using acetylene as hydrocarbon source. While in
234
S. TAMULEVIýIUS AND Š. MEŠKINIS
the case of the both HMDSO + H2 films it is shifted towards lower wavenumbers. It should be mentioned, that Raman scattering peaks related to the trans-polyacetylene-like segments are reported at 1,150 cm–1 33 and at ~1,480 cm-1 in nanocrystalline diamond films.34–36 Therefore, in our case peculiarities of the SiOx doped DLC films Raman scattering spectra in comparison with Raman spectra of “conventional” hydrogenated DLC films can be explained by presence of trans-polyacetylene chains. Presence of the chains is more pronounced for films deposited at lower ion current density.
Intensity (a.u.)
HMDSO+H2 (j=50MA/ cm2) HMDSO+H2 (j=100 MA/ cm2)
HMDSO+C2H2 1330 cm 1 1530 cm 1 1150 cm 1 1480 cm 1 1100 1200 1300 1400 1500 1600 1700
Wavenumber (cm 1)
Figure 6. Raman scattering spectra of DLC films.
4.3. WETTABILITY OF DIAMOND LIKE CARBON FILMS
It can be seen in Table 3, that contact angle with water was substantially lower than 100o reported in Ref.37 Additional hydrogen flux only slightly dec-reased contact angle with water of the films. This behavior is different from the case of the undoped a-C:H films: contact angle with water of the films depo-sited using additional hydrogen was 94o38,39 in comparison with 65–82o for films deposited from the pure hydrocarbon gas.37,40 Increase of the ion current density resulted in decreased contact angle with water (to 61o). It should be mentioned, that for undoped a-C:H films, deposited by plasma enhanced CVD, increase of the contact angle with decrease of the plasma power was reported.40 Additional doping by nitrogen decreased contact angle with water of the SiOx doped DLC films as well. However at low N2 flux (2.5% in total gas flow) this decrease was very small (from 70 r 1o to 69 r 1o). Higher additional N2 fluxes (10% and 20% in total gas flow) resulted in more pronounced decrease of the contact angle to the 58 r 1o–60 r 1o respectively. Extra doping by argon (10% in total gas flow) resulted in decrease of the contact angle with water to the 59 r 1o.
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 235 TABLE 3. Thickness, deposition rate and contact angle with water of synthesized films. Sample No.
(HMDSO + H2)/H2 gas flow ratio
Additional doping gas flow (%)
1 2 3 4 5
1:1 1:1.5 Only HMDSO + H2 Only HMDSO + H2 Only HMDSO + H2 Only HMDSO + H2 Only HMDSO + H2 Only HMDSO + H2 Only HMDSO + H2
– – – – – 2.5% N2 10% N2 20% N2 10% Ar
Current density (mA/cm2)
Contact angle with water (o)
0.1 r 0.01 0.1 r 0.01 0.1 r 0.01 0.05 r 0.01 0.15 r 0.01 0.1 r 0.01 0.1 r 0.01 0.1 r 0.01 0.1 r 0.01
67 r 1 67 r 1 70 r 1 65 r 1 61 r 1 69 r 1 58 r 1 60 r 1 59 r 1
The dependence of the contact angle with water of the SiOx containing diamond like carbon film on deposition time is presented in Figure 7. The films in this experiment were deposited at the same conditions, as sample with the highest contact angle with water – sample No. 3. It can be seen, that contact angle with water for of 4 nm thickness films (deposition time 5 and 10 s respectively) has the same value as contact angle for the 180 nm film (deposited for 40 min). Dependence of the contact angle on film thickness is presented for “conventional” DLC film deposited from acetylene for comparison purposes. It can be seen, that in the case of the undoped DLC film higher film thickness is necessary to achieve contact angle with water in the range of 70o. It should be mentioned,
Figure 7. Dependence on film thickness of the contact angle with water and refractive index of the SiOx doped DLC film.
236
S. TAMULEVIýIUS AND Š. MEŠKINIS
that 1 h air annealing did not change contact angle with water of the 4 nm thickness film. While for 180 nm thickness film, decrease of the contact angle with water was observed.21 As it was mentioned in experimental section, it is difficult to estimate thickness of the 5s or 10s deposited SiOx containing diamond like carbon films itself: refractive index of the two layer system native SiO2/ SiOx containing DLC has the same refractive index like the SiO2 film, while the thickness of the native silicon oxide film measured before DLC film deposition is the same (~4 nm). Therefore, it can be supposed from the calculations using deposition rate of the thick film, that thickness of the deposited layer is somewhere in the range of the ellipsometer measurement error – less than 1 nm. 4.4. OPTICAL PROPERTIES OF SI AND SIOX CONTAINING DLC FILMS
Optical transmittance spectra of the DLC films in UV-VIS range were investigated to evaluate it as a possible antisticking layer in UV imprint lithography. In all cases ion beam energy was 800 eV and ion current density – 100 PA/cm2. UV-VIS optical absorption spectra (Tauc plot) of the synthesized DLC films are presented in Figure 8. Optical bandgap of DLC films investigated has been calculated by extrapolation of the linear part of Tauc plot. Slope of the linear part of Tauc plot was calculated as well as. The highest absorption coefficient was observed for “conventional” hydrogenated DLC deposited from the acetylene precursor. Absorption coefficient of HMDSO + C2H2 films was several times larger than absorption coefficient of the HMDSO + H2 films, but substantially lower than absorption coefficient of DLC films deposited from acetylene gas. Additional Ar gas flow during the deposition resulted in increased optical transparence of SiOx doped DLC films (HMDSO + H2 films). N2 doping dec-reased absorption coefficient of HMDSO + H2 DLC films even more. It can be seen in Table 4 and Figure 8, that despite lower absorption coefficient, optical bandgap of HMDSO + C2H2 DLC films was smaller that optical bandgap of “conventional” hydrogenated DLC film. Optical bandgap of SiOx doped DLC films was substantially larger (2.15 eV in comparison with 0.89 and 0.96 eV). Additional Ar flow resulted in increase of the optical bandgap of HMDSO + H2 DLC films. Optical bandgap increased as a result of N2 codoping as well. Optical bandgaps of SiOx containing DLC films deposited under additional argon or nitrogen gas flow in all cases were in 2.65–2.81 eV range. Slope of the Tauc plot of the different DLC films has been calculated to evaluate disorder in investigated films. It was already shown for a-Si:H and a-SiC:H films, that slope of Tauc plot (B parameter) increases with decrease of the disorder.41 It can be seen in Table 4, that the highest slope was observed in the case of the “conventional” hydrogenated DLC film deposited using acetylene gas as a
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 237
sqrt(Energy*Absorbtion coef.)(eV*cm 1)1/2
hydrocarbon source. Slope of HMDSO + C2H2 DLC film was substantially smaller. It was even lower than slope of HMDSO + H2 film. Both additional Ar gas flow and N2 doping resulted in decreased slope of SiOx doped DLC film. Nitrogen co-doping resulted in more pronounced increase of disorder in comparison with effects of additional Ar gas flow during the SiOx containing diamond-like carbon films synthesis process.
400
C2H2 HMDSO HMDSO +C2H2 +H2
300
+10% Ar +2.5%N2
200
+10%N2
100
0
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Energy (eV)
Figure 8. Optical absorption spectra of DLC films. TABLE 4. Optical properties of the DLC films. Gas (gas mixture) used for synthesis C2H2 HMDSO + C2H2 (CH3)3SiOSi(CH3)3 + H2 (CH3)3SiOSi(CH3)3 + H2 (CH3)3SiOSi(CH3)3 + H2 (CH3)3SiOSi(CH3)3 + H2 (CH3)3SiOSi(CH3)3 + H2
B (eVcm)1/2
2.4 2.1 1.8
Optical bandgap Eopt (eV) 0.96 0.89 2.15
10% Ar
1.8
2.65
215
2.5% N2
1.8
2.67
192
10% N2
1.8
2.81
194
20% N2
1.7
2.68
172
Extra co-doping gas – – –
Refractive index (O = 632.8 nm)
380 188 240
238
S. TAMULEVIýIUS AND Š. MEŠKINIS
5. Conclusions Considering adhesion between diamond like carbon film and steel enhancement issues by different interlayers or steel surface ion beam treatment techniques chemical composition of DLC/interlayer interface as well as mechanical stress in interlayers and composition of hydrocarbon gas should be taken into account. SiOx containing DLC films were deposited from hexamethyldisiloxane vapor and hydrogen gas mixture. Raman scattering spectra of the all synthesized films were typical for DLC films. However, in the case of SiOx containing DLC films, Raman scattering spectra features typical for trans-polyacetylene-like segments has been observed. Contact angle with water of SiOx containing DLC films under different synthesis process conditions was very close to the contact angle with water of the “conventional” DLC films deposited from acetylene gas. The highest optical absorption coefficient was observed for “conventional” hydrogenated DLC deposited from the acetylene precursor. Absorption co-efficient of HMDSO + C2H2 films was several times larger than absorption coefficient of the HMDSO + H2 films, but substantially lower than absorption coefficient of DLC films deposited from acetylene gas. Additional Ar or N2 gas flow during the deposition resulted in increased optical transparence of SiOx doped DLC films (HMDSO + H2 films). Despite lower absorption coefficient, optical bandgap of HMDSO + C2H2 DLC films was smaller that optical bandgap of “conventional” hydrogenated DLC film. The most promising combination of the UV absorption and anti-sticking properties had N2 co-doped SiOx doped DLC films deposited by 800 eV energy ion beam. ACKNOWLEDGEMENTS
Support of the Lithuanian Science and Studies Foundation is acknowledged.
References 1. Robertson, J. (2002) Materials Science and Engineering: R: Reports vol.37, p.129–281. 2. Taube, K. (1998) Surface and Coatings Technology vol.98, p.976–984. 3. Grill, A. (2003) Diamond and Related Materials vol.12, p.166–170. 4. Hauert, R. (2003) Diamond and Related Materials vol.12, p.583–589. 5. Chi-Lung Chang, Da-Yung Wang (2001) Diamond and Related Materials vol.10, p.1528–1534. 6. Michlera, T., Grischke, M., Bewilogu, K., Hieke, A. (1999) Surface and Coatings Technology vol.111, p.41–45. 7. Chi-Lung Chang, Da-Yung Wang (2001) Diamond and Related Materials vol.10, p.1528–1534.
DIAMOND LIKE CARBON FILMS: GROWTH AND CHARACTERIZATION 239 8. Podgornik, B., Vizintin, J., Ronkainen, H., Holmberg, K. (2000) Thin Solid Films vol. 377–378, p.254–260. 9. Guo, L. Jay (2004) Journal of Physics D: Applied Physics vol. 37, p.R123–R141. 10. Harriott, L.R. (1998) Materials Science in Semiconductor Processing vol.1, p.93–97. 11. Jaszewski, R.W., Schift, H., Schnyder, B., Schneuwly, A., Groning, P. (1999) Applied Surface Science vol.143, p.301–308. 12. Soo-Beom Jo, Min-Woo Lee, Se-Geun Park, Jung-Keun Suh, Beom-hoan (2004) Surface and Coatings Technology vol.188–189, p.452–458. 13. Schift, H., Saxer, S., Park, S., Padeste, C., Pieles, U., Gobrecht, J. (2005) Nanotechnology vol.16, p.S171–S175. 14. Park, S., Schift, H. H., Padeste, C., Schnyder, B., Kötz, R., Gobrecht, J. (2004) Microelectronic Engineering vol.73–74, p.196–201. 15. Watanabe, K., Morita, T., Kometani, R. et al. (2004) Journal of Vacuum Science & Technology B vol.22, p.22–26. 16. Morita, T., Watanabe, K., Kometani, R. (2003) Japanese Journal of Applied Physics Part 1 vol.42, p.3874–3876. 17. Zhurin, V.V., Kaufman, H.R., Robinson, R.S. (1999) Plasma Sources Science and Technology vol.8, p.R1–R20. 18. Kopustinskas, V., Meškinis, Š., Grigalinjnas, V., Tamuleviþius, S., Pucơta, M., Niaura, G. et al. (2002) Surface and Coatings Technology vol. 151–152, p.180–183. 19. Meškinis, Š., Andruleviþius, M., Kopustinskas, V., Tamuleviþius, S. (2005) Applied Surface Science vol.249, p.295–302. 20. Meškinis, Š., Andruleviþius, M., Tamuleviþius, S., Kopustinskas, V., Šlapikas, K., Jankauskas, J., ýižinjtơ, B. (2006) Vacuum vol. 80, p.1007–1011. 21. Kopustinskas, V., Meškinis, Š., Tamuleviþius, S., Andruleviþius, M., ýižinjte, B., Niaura, G. (2006) Surface and Coatings Technology vol. 200, p.6240–6244. 22. Meškinis, Š., Kopustinskas, V., Šlapikas, K., Tamuleviþius, S., Guobienơ, A., Gudaitis, R., Grigalinjnas, V. Thin Solid Films vol.515, p.636–639. 23. Miola, E.J., de Souza, S.D., Nascente, P.A.P., Olzon-Dionysio, M., Olivieri, C.A., Spinelli, D. (1999) Applied Surface Science vol.144–145, p.272–277. 24. Roosendaal, S. J., van Asselen, B., Elsenaar, J.W., Vredenberg, A.M., Habraken, F.H.P.M. (1999) Surface Science vol. 442, p.329–337. 25. Ku-Ling Chang, Shih-Chun Chung, Shih-Hsiang Lai, Han-C. Shih (2004) Applied Surface Science vol.236, p.406–415. 26. Alphonsa, I., Chainani, A., Raole, P.M., Ganguli, B., John, P.I. (2002) Surface and Coatings Technology vol.150, p.263–268. 27. Tamuleviþius, S. Kopustinskas, V., Meškinis, Š., Augulis, L. (2004) Carbon vol.42, p. 1085–1088. 28. Toth, A., Mohai, M., Ujvari, T., Bertoti, I. (2005) Thin Solid Films vol.482, p.183–187. 29. Santoni, A., Frycek, R., Castrucci, P., Scarselli, M., De Crescenzi, M. (2005) Surface Science vol.582, p.125–136. 30. Veres, M., Koos, M., Toth, S., Fule, M., Pocsik, I., Toth, A., Mohai, M., Bertoti, I. (2005) Diamond and Related Materials vol.14, p.1051–1056. 31. Zhang, P., Tay, B.K., Yu, G.Q., Lau, S.P., Fu, Y.Q. (2004) Diamond and Related Materials vol.13, p.459–464. 32. NIST XPS database http://srdata.nist.gov/xps/ 33. Shou-Yong Jing, Heon-Ju Lee, Chi Kyu Choi (2002) Journal of Korean Physical Society vol.41, p.769. 34. Ferrari, C., Robertson, J. (2001) Physical Review B vol.63, p.121405-1–121405-1-4. 35. Ikeda, T., Teii, K. (2006) Diamond and Related Materials vol.15, p.635–638.
240
S. TAMULEVIýIUS AND Š. MEŠKINIS
36. Kuzmany, H., Pfeiffer, R., Salk, N., Gunther, B. (2004) Carbon vol.42, p.911–917. 37. Pfeiffer, R., Kuzmany, H., Knoll, P., Bokova, S., Salk, N., Gunther, B. (2003) Diamond and Related Materials vol.12, p.268–271. 38. Grischke, M., Hieke, A., Morgenweck, F., Dimigen, H. (1998) Diamond and Related Materials vol.7, p.454–458. 39. Je-deok K., Kyung-hwang L., Kyu-young K., Hiroyuki S., Osamu T., Yunying Wu, and Yasushi I. (2003) Surface and Coatings Technology vol.162, p.135–139. 40. Je-Deok K., Hiroyuki S., Osamu T. (2002) Vacuum vol.66, p.379–383. 41. Swain, B.P., Patil, S.B., Kumbhar, A., Dusane R.O. (2003) Thin Solid Films vol.430, p.186–188.
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF INORGANIC AND ORGANIC FILMS J. SCHOU* Department of Photonics Engineering, Risø Campus, Technical University of Denmark, DK-4000 Roskilde, DENMARK
Abstract – The standard method for producing films by laser-assisted methods, Pulsed Laser Deposition (PLD) will be reviewed. The films considered are usually inorganic films, but also films of organic materials have been produced. Also the deposition of organic films by MAPLE (Matrix Assisted Pulsed Laser Evaporation), in which the target is replaced by a frozen matrix containing a few per cent film material, will be reviewed.
Keywords: Pulsed laser deposition (PLD), matrix-assisted pulsed laser evaporation (MAPLE).
1. Introduction The application of lasers for production of thin films of a broad range of materials has advanced tremendously during the last decade. The increasing demand for new materials has been accompanied by a successful development of reliable lasers in the ultraviolet (UV)-regime with adequate pulse energies. Laser-assisted film fabrication has primarily been made by Pulsed Laser Deposition (PLD), which has become the preferred technique for fast production of thin, high-quality films of complicated stoichiometry, in particular for metal oxide films.1–4 Laser irradiation of a solid is not only a method to transfer material from a target to a substrate, but can also be used to induce chemical changes in the target or to pattern a surface.2,5–6 Recently, also thin films of organic materials such as polymers and biomaterials have been produced by a refinement of the original PLD technique.7–9
______ *
To whom correspondence should be addressed: J. Schou, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
241
J. SCHOU
242
A PLD experiment is surprisingly simple (shown in Figure 1). A UV laser with nanosecond pulses is directed onto a target in a vacuum chamber such that the material is removed from the target. This “explosive removal” of material during high intensity laser irradiation is usually called laser ablation. The ablated material expands in a flow perpendicular to the surface and is collected on a suitable substrate in a holder system. The accumulation of materials from a large number of pulses leads to the growth of films with thickness in the subnanometer range to micrometers. Even though PLD has been extensively used for more than two decades, many features in the process are still not understood (Section 3). For organic materials film production by laser-assisted methods it is rather the rule than the exception that the deposition processes are not known (Sections 4–6).
Vacuum Gas inlet chamber Substrates
Target
Holder with heater UV laser
Aperture
Rotating target holder
Lens UV transparent window
Figure 1. A standard PLD setup with a plasma plume expanding from the target. (From Thestrup et al.)10
The simplifying picture in Figure 1 does not reveal the fact that the absorption of light from a laser with high pulse energies in a solid is a complicated process, in which the absorption changes dynamically during the pulse because of the surface modification and production of free electrons. Once the material leaves the target, the plume dynamics depends strongly on the geometry of the laser beam spot, and the dynamics of the ablation plume also changes strongly in the presence of a background gas. Finally the growth parameters of the films are influenced by the energy of the arriving particles, the temperature of the substrate, and the type of background gas. Some of these processes will be discussed in Sections 2 and 3.
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
243
2. Pulsed Laser Deposition (PLD) PLD is advantageous in many respects (see Eason,1 Schou11): x The stoichiometry is conserved during the material transfer from target to the thin film. x The film deposition can be carried out in vacuum, in a reactive or inactive background gas. x The atoms and ions in the plume have sufficient kinetic and internal excitation energy to increase the sticking and nucleation rate as well as the surface mobility. x “Sandwich” films of different materials can be produced comparatively easily in-situ with a multitarget system in a vacuum chamber. x Highly perfect interfaces in a multilayer system can be achieved by controlling the number of pulses, the laser repetition rate and the pulse energy. PLD has primarily its strength in the growth of crystalline oxides. The growth of complex oxides requires the arrival of flow at the growing film with the correct stoichiometry in an oxidizing ambient gas with a background pressure that is favorable for the desired phase formation. The achievements in the late 1980s with high-temperature superconducting thin films of YBa2Cu3O7-x fabricated by PLD (Cheung and Horwitz12) created world-wide activity in PLD research and applications. During the last decade very complicated metal oxides such as Nd, Cr::Gd3Sc2Ga3O12 (a potential material for integrated semiconductoroptoelectronic applications)13 and epitaxially grown magneto-plumbate Ba2Co2 Fe12O22 (a ferromagnetic oxide with potential applications for magnetic devices) have been produced.14 This ability to grow epitaxial, multication complex thin films has been a substantial feature for the widespread use of PLD. A related direction has been the development of superlattices, especially with perovskite structure, and excellent film flatness and crystallinity have been obtained for these films. PLD also offers the opportunity to generate atomically sharp interfaces, for example monitored with RHEED (Reflection High Energy Electron Diffraction) (Rijnders and Blank15). Another material group, for which PLD has turned out to be superior to most other methods, is quasicrystals, typically formed by a high number of different metal atoms, which may contain atoms with very different evaporation pressure.16 PLD is a technique with a lot of flexibility. In contrast to a most other methods, many of the experimental parameters are not coupled and can be controlled independently, e.g. substrate temperature and laser pulse energy. Most of the experimental parameters can also be changed relatively easily, usually without breaking the vacuum.
244
J. SCHOU
(i) Type of substrate: For many oxides of complicated structure, where a specific epitaxial growth is required, a suitable choice of substrate can induce the desired crystal structure. A typical substrate for films of high temperature superconductors is SrTiO3. However, by choosing a substrate with a structure completely different from that of the film material one may enforce a new structure in the growing film similar to that of the substrate, but different from the equilibrium phase structure of the ablated material. (ii) Temperature of the substrate: The temperature is completely decoupled from the ablation conditions, e.g. energy of the plasma particles and electron temperature. This means that the film growth temperature can be adjusted to the desired phase of the material. (iii) Type of background gas: Growth of oxides and nitrides typically requires an ambient gas which is favorable for the phase formation considered. The background gas is particularly important for high temperatures, such that the chemical equilibrium can be maintained. The oxygen constituents are largely more volatile than the metal atoms, and the stoichiometry can, therefore, be maintained with a flow from the background. In some cases the gas atoms and the ablated metal atoms/ions can form oxides and nitrides directly during the transfer from the target to the substrate.17–18 The background gas also acts as a moderator for the fast ions, which are slowed down and scattered by background gas atoms. As discussed below, this reduces the damage and sputtering in the growing film caused by energetic particles from the plasma plume. In cases, where one needs a moderate oxygen partial pressure, one may use a background gas of a mixture of the chemically active component, oxygen, and an inactive gas such as argon. (iv) Pressure of the background gas: As mentioned in (iii) the magnitude of the pressure can be adjusted to the desired partial pressure for the formation of a particular phase of the growing film. In addition to this, the pressure can be chosen in such a way that the energetic plume particles are slowed down to a few electron volts or below, where damage of the film is largely reduced. (v) Fluence of the laser: The energy of the laser pulse is a key quantity that determines the number of ablated atoms and the initial energy of the atoms in the plume. Since the number of ablated atoms typically also depends on the area of the irradiated surface, one usually considers the fluence, i.e. the total laser pulse energy per area unit, as a general mea-sure of the ablation process. The fluence is controlled either with the optics to change the focus of the beam partly and thereby change the irradiated area, or simply just to reduce the pulse energy directly of the laser (if this is possible). For a standard PLD process one will usually run at a fluence somewhat above the evaporation threshold of the solid such that the flow from the target does not contain the most volatile constituents of a compound alone, i.e. the
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
245
stoichiometry of the flow would differ substantially from that of the target. On the other side, a too high fluence leads to partial or complete shielding of the target by the plume, and the laser energy is not used efficiently for ablation. The flexibility in fluence can also be used for layer-by-layer growth. In that case the fluence is low, such that only a fraction of a monolayer (~0.01) is deposited for each pulse. This gives an unprecedented control of an accurate deposition of a fraction of a monolayer. (vi) Geometry of the laser beam spot: PLD is often used with apertures for the laser beam such that areas with low intensity of a beam spot are removed. This gives a homogeneous fluence, which is necessary for the deposition of high quality films. On the other hand, one will be reluctant to cut away too much of the available laser energy. In addition, for fast film deposition one usually wants to have a large beam spot with the desired fluence so that a large number of ablated atoms arrive at the substrate for each laser pulse. A complicating issue is the plume dynamics from a nonsymmetrical beam spot, for which the dissimilar propagation of the plume front may change the form of the plume. This flip-over process is an important issue for the production of uniform films.19–21 Some other parameters are difficult to change. For nanosecond lasers the pulse length is determined by the construction of the laser and has a duration from 5–10 ns for Nd:YAG lasers and 20–30 ns for excimer lasers. For femtosecond lasers it is in many cases possible to change the pulse length, but it is usually not a trivial task to go below 100 fs. The wavelength of the excimer lasers is determined by the gas mixture in the laser, e.g. the most common excimer laser types are the KrF operating at 248 nm and ArF at 193 nm. For excimer lasers several wavelengths are available (see Chrisey and Hubbler22), but a change of wavelength does not only require exchange of the gas, but also exchange of optical components with a suitable antireflection coating. Many Nd:YAG lasers with the fundamental frequency operating at 1,064 nm are also equipped with doubling or tripling systems for wavelengths at 532 nm and 355 nm. With a system equipped with different exit apertures the wavelength can be changed in a few minutes, but the optical elements along the beam path to the target also need to be replaced with other elements with the correct coating. 3. Fundamental Processes During PLD The interaction between a laser beam and a solid is a complicated sequence of different processes which usually even occur simultaneously. This section is largely based on Schou,11 Willmott and Huber23 and Gorbunoff.3
246
J. SCHOU
1. The laser light strikes the solid and produces electronic excitations in the solid. After a period of tens of picoseconds the atoms and electrons equilibrate which leads to strong heating of the irradiated volume. During this stage (1) the laser light interacts with a solid. 2. As a result of the heating in stage (1) material is ejected. This material is continuously absorbing laser light and expands as a one-dimensional plume of partially ionized plasma. During this stage (2) the material behaves like a gas or plasma. This stage continues until the end of the laser pulse (which for typical PLD lasers is at 5–30 ns) 3. After the termination of the laser beam, the ablation plume expands adiabatically in three dimensions. If the ablation takes place in vacuum, the plume atoms will eventually flow away with a constant velocity. This stage lasts for microseconds in a typical chamber with a target–substrate distance in the range 10–200 mm. 4. If the ablation takes place in a background gas, the plume pressure is initially so high that the plume expands similar to a plume in vacuum. Once the driving pressure of the plume is reduced because of expansion and cooling, the propagation is completely determined by the interaction between the plume atoms and the molecules/atoms of the back-ground gas. This stage may last for milliseconds. A good starting point for the discussion is the total ablation yield Y = FA/ U0 ,
(1)
where the fluence F during a laser pulse with an incident intensity I0 and a duration WL onto an area A is defined as Wp
F=
³ I dt 0
(2)
0
For each material under consideration the total ablation yield depends on the beam area and the cohesive energy U0 per atom (but also on the reflectance of the light at the particular wavelength). U0 ranges from 1 eV for the most volatile metals to 10 eV for the most refractory metals. Unfortunately, U0 is not known well for many materials, in particular for chemical compounds. The formula is qualitative since it does not include reflection of the laser light from the surface or absorption in the plume, both of which may vary strongly during the laser pulse.24 Also conduction of heat away from the surface is not included. For metals the yield from a number of selected materials as a function of the cohesive energy at a fluence of 2 J/cm2 was reported by Schou.11 It was demonstrated that the ablation yield is significantly enhanced for metals with a low cohesive energy (Schou11). A typical yield is Y | 1 u 1015 atoms/pulse at 2
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
247
J/cm2. At higher fluence some of the laser energy will be absorbed by the plume, and with increasing fluence the yield will saturate due to the shielding by the plume.24 There is a tendency in the literature to derive the absolute ablation yield from depth measurements such that the yield is the number of atoms removed from the volume with the depth d and the area A from a single shot. Usually the depth d is not constant over the ablation spot which can exhibit a pronounced surface roughness. The work on silver by Toftmann et al.20 and Amoruso et al.25 provides us with some numbers for the ablation (Table 1). Silver has a cohesive energy U0 that is not too high and a typical attenuation length of 20 nm (which is similar to other metals for UV laser irradiation). The reflection coefficient varies strongly from the initial value of 0.75 because of the roughening of the target surface during the pulse. The table also shows that there are three length scales TABLE 1. Estimated parameters for a PLD process on silver with a fluence of 2 J/cm2 at 355 nm.
Fluence : 2 J/cm2 Wavelength: 355 nm Pulse length: 6 ns 2 15 Ablation yield: 2 u 10 Ag atoms/pulse. Beam spot: 0.04 cm Material: Silver Cohesive energy: 2.9 eV/Ag atom Density of solid silver: 5.86 u 1022 atoms/cm3 Ablation depth d Attenuation depth 1/Į Thermal diffusion length (IJLD)1/2 Pulse energy: Energy required for ablation (cohesive energy) Velocity of plume: Extension of plume at t = IJL : Density of plume at t = IJL : Density of atmospheric air at 20oC: Plume pressure at t = IJL at 20oC: Pressure for particle energy of 0.5 eVa
16 km/s 92 ȝm 8.7 u 1018 atoms/cm3 2.5 u 1019 atoms/cm3 0.2 bar 4.0 bar
Deposit of silver per pulse (80 mm from target)b : Estimated deposit per pulse (60 mm from target) : a
Estimated from Toftmann et al.27 b From Toftmann et al.20
: 8 nm : ~20 nm : ~1,000 nm : 80 mJ : 0.95 mJ
14.7 ng/cm2 26 ng/cm2
248
J. SCHOU
of importance in this system, the ablation depth d, the attenuation length 1/Į and the thermal diffusion length. The ablation depth is usually the smallest quantity for energies considered in PLD. For metals the attenuation length is typically smaller than the thermal diffusion length (IJLD)1/2, where D is the thermal diffusivity. With a typical pulse duration of the order 10 ns the length ranges from 1,000 nm for good conductors such as silver and copper down to 100 nm for badly conducting metals. For the example given in Table 1 it is striking that the pulse energy is much larger than the energy required for ablation calculated from the cohesive energy. It reflects the fact that the laser light is reflected from the metal surface (in stage 1), that the energy is absorbed by the expanding plume (stage 2) and finally that the ablated atoms are not thermal, but may have an energy exceeding several electron volts. If the pulse energy was delivered equally to all ablated atoms without considering other loss channels, the average energy of the atoms would be ~255 eV. This energy is much higher than the experimental values, but one should note that the peak of the energy distribution for angles close to the normal is at ~150 eV.26 In any case, there is large number of high-energy ions with an energy exceeding 100 eV. After the termination of the laser pulse at the time t = WL, i.e. the end of stage 2, the ablated particles form a dense plume which covers the target spot. At this time the front of the plume has reached a distance of the order ~100 Pm corresponding to WLvp, where vp is the front velocity of the plume (see Table 1). The thickness of the plume will increase if the front velocity is high (or the laser pulse is long). Initially a considerable fraction of the energy in the dense plume is thermal, but already at the end of stage 2 the kinetic energy of the plume is comparable to the thermal energy. A simplified view is that the laser energy is stored as potential energy for the expansion in terms of excitations, ionizations and thermal energy of the plume atoms. This energy drives the expansion such that the plume atoms asymptotically will move away from the target with a constant velocity. After the end of the laser pulse, the expansion is adiabatic, since there is no further energy and mass transfer to the plume. The complicated interactions in the plume dynamics in stages 3 and 4 have recently been reviewed by Schou11 and Schou et al.21 The growth of films by PLD has been treated by Willmott.28 4. Organic Materials: From PLD to MAPLE and IR-Irradiation There is an increasing demand for polymers and biomaterials as coatings for pharmaceutical and biotechnological applications (Piqué7). Most of these coatings can also be produced by chemical methods, but there is frequently a
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
249
need for homogeneous films of well-controlled thickness or films which can be deposited in a dry environment. Obviously, irradiation of organic materials by those UV lasers that are most frequently used (O = 193, 248 and 355 nm) leads to damage and fragmentation of organic molecules in the target. Therefore, the quality of organic films deposited by PLD with UV lasers is reduced considerably (Chrisey et al.9and Piqué7) There are two ways to overcome the fragmentation of the organic materials. The first possibility is that the material to be deposited is embedded in a matrix which protects the organic material from damage. This technique is called MAPLE (matrix assisted pulsed laser evaporation). The second possibility is to use lasers with a long wavelength, typically in the infrared (IR) region such that the photon energy is so low that bond breaking of the organic material is avoided. The energy transfer to the material is then optimized by letting the laser light have resonant absorption in the material, Resonant Infrared PLD (RIR-PLD). This technique is described in Bubb and Haglund.8 Another technique for film transfer is LIFT (Laser induced forward transfer) (see Piqué et al.29). In this technique a laser pulse is directed onto the rear side of a transparent substrate. The film material is blown off by the violent energy deposition at the film-substrate interface and transferred to another substrate in the direction of the laser beam. The transfer is typically performed over a fraction of a millimeter, but can be carried out in atmospheric air. Since the geometry is very different from PLD and MAPLE, this technique will not be discussed further. 5. MAPLE (Matrix Assisted Pulsed Laser Evaporation) The MAPLE (Matrix assisted pulsed laser evaporation) technique was first introduced by Piqué et al.30 The authors were inspired by the MALDI (Matrix assisted laser desorption ionization) technique, which has become a standard technique in analytical biochemistry for determining the mass of an analyte fast.31 A schematic overview is sketched in Figure 2. The material to be deposited is dissolved in a solution of a volatile matrix, which is frozen, and subsequently inserted into a vacuum chamber. The idea of the technique is that the matrix absorbs the laser light, so that decomposition of the film material is avoided. During the subsequent ablation the solvent evaporizes, whereas the (film) material is deposited on the substrate. In MALDI ions from the analyte dissolved in the matrix is directed into a mass spectrometer such that the mass of the analyte can be determined. The advantage of MAPLE is primarily the high flexibility:
J. SCHOU
250
1. Films of any organic material which can be dissolved in an appropriate solvent can be deposited. 2. The consumption of solute, i.e. the film material to be deposited, is much less than at the corresponding PLD process. Usually only a minor fraction of the film material is required, e.g. 1–5 wt %. This is particularly important for very expensive biomaterials. 3. Film production by MAPLE can be combined with PLD or other vacuum techniques including deposition through masks.
Figure 2. Schematic drawing of the MAPLE technique. (Courtesy – Andreea Matei.)
The drawbacks of MAPLE are that 1. MAPLE is not a UHV experiment. 2. It is not always possible to find a solvent with a freezing point which is convenient during the deposition, i.e. in the range of –100–0oC. 3. It is not always possible to find a solvent with “matrix absorption” such that the laser light is strongly absorbed by the matrix. 4. Occasionally, the solvent reacts with the film molecules and form undesired radicals which are deposited in the growing film. 5. The surface of the film is often nonuniform, covered with large clusters of micrometer size (see below). As for PLD many parameters can be controlled easily, and some are not different from those used in PLD, e.g. type of substrate, pressure of background gas and target-substrate distance. However, some parameters are decisive for MAPLE, e.g. target temperature and type of matrix.
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
251
(i) Target: type of matrix: Typically, the matrix is a flash-frozen solvent containing 1–5 wt % of the material to be deposited. As discussed above the role of the matrix is to absorb the laser light and transfer the energy to the film molecules. Several organic solvents have been used, e.g. toluene, chloroform, tetrahydrofuran, isopropanol (Chrisey et al.9; Piqué7 and Toftmann et al.32), but also a simple liquid as water can be used. These solvents are absorbing in the UV range, but not always up to 355 nm. Flash frozen water ice has also been used as matrix,33 even though water is not absorbing radiation of wavelengths exceeding 200 nm. However, it was pointed out by Mercado et al.34 and Rodrigo et al.35 that nonlinear absorption processes outside the expected absorption range may occur as well. (ii) Target temperature: The temperature of the target needs to be adjusted in such a way that the target evaporates with a rate which is not too high. In such a case the beam may penetrate fast through the matrix and ablate material from the target holder into the growing film. Usually the target temperature is a compromise between consumption of cooling liquids and an acceptable evaporation rate. Rotation of the target and/or rastering of the beam serve to avoid hole drilling in the ice by the laser beam. In any case, the MAPLE technique requires a solvent with a melting point within a practical cooling range. Therefore, solvents with a freezing point between – 100oC and 0oC are preferentially used for MAPLE. (iii) Substrate temperature: For the MAPLE technique most of the film materials applied can be destroyed partly or completely by extensive heating. Heating of the substrate can therefore only be used moderately. Rodrigo et al.36 tried to achieve a uniform layer of PEG on a heated substrate surface with a large number of particulates of lateral dimension up to micrometers. The substrate was gradually covered by a bottom layer, but the particulate abundance or size were not reduced. Also Mercado et al.34 found a pronounced tendency to particulate formation on top of films of poly-lactide-co-glicolide (PLGA) (iv) Fluence: Most of the depositions have been carried out for a fluence less than or about 1 J/cm2. The reason is primarily to avoid photodamage in the solute. For horseradish peroxidase and insulin Ringeisen et al.37 reported that it was necessary to use a very low fluence, 0.2 J/cm2 for the deposition of these proteins. The dependence of fluence has actually only been studied systematically by Rodrigo et al.35 and by Purice et al.38 for water ice matrices. These two studies demonstrate that also for water ice with PEG or lysozyme the deposition can take place below or around 2 J/cm2. (v) Repetition frequency: Since the pressure in the chamber increases strongly during laser irradiation, many experiments have been performed at a frequency of 1–10 Hz. Deposition measurements of PEG in water ice was not possible for frequencies exceeding 2 Hz (Rodrigo et al.35).
252
J. SCHOU
Some of the other parameters are difficult to change similar to the case of the PLD process. The wavelength is as mentioned above an important parameter which determines the absorption of the laser light in the matrix. No systematic experiments have been carried out to investigate the dependence of the pulse length on the MAPLE results. 6. Fundamentals of MAPLE The MAPLE process is even more complex than PLD because the major part of the laser light is absorbed by a frozen solvent which usually has a complicated chemical composition. Some characteristics of MAPLE are similar to MALDI, but even this well-known technique is not understood in detail (see Dreisewerd39). Since the fluence in a typical MALDI experiment can be considerably below 0.1 J/cm2, the extrapolation to the MAPLE regime has to be carried out with some caution. However, since there are very few data on fundamental interactions in MAPLE, some MALDI results will be included as well. Similar to PLD the full MAPLE process can be described in four stages: 1. As for PLD, the laser light strikes the solid matrix with the solute and is absorbed, primarily in the matrix. The light induces a variety of complicated electronic excitations, the nature of which is strongly dependant on the specific matrix. 2. The excitations drive material out of the irradiated volume of the matrix as a one-dimensional plume in forward direction. During the ejection the molecules and fragments in the plume are irradiated by laser light resulting in secondary reactions in the plume, until the laser pulse is terminated. 3. The plume expands in three dimensions in vacuum, or more correctly, in a background gas of evaporating matrix molecules. If the density of the plume is sufficiently high, plume reactions may still take place. This stage lasts up to microseconds. 4. The plume molecules and fragments arrive at the substrate. The nonvolatile molecules of the film material stick at the surface instantaneously or move around on the substrate surface until they nucleate into larger nanoclusters or thermalize. The solvent molecules are pumped away during the expansion of the plume or from the surface because of the low sticking probability. This stage may last up to milliseconds. The absorption of laser light leads to localized ionizations and excitations in the matrix which eventually are converted to translational energy by recombination and repulsive transitions analogous to processes which happen under electron and ion bombardment of insulating solids (see Schou40 and Baragiola et al.41). From MALDI39 and RIR-MAPLE,8 it is known that the ejection of the
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
253
solute is most efficient at wavelengths with a strong absorption. The absorption is exponentially decreasing as a Lambert``-Beer law at low fluences.31 However, Rodrigo et al.35 and Mercado et al.39 found that the absorption may become highly nonlinear with increasing fluence. Zhigilei and Garrison42 as well as Chen et al.43 have demonstrated that phase explosion in the matrix can be responsible for the material ablation even at moderate fluences, and a plasma plume from ionization breakdown has actually been observed by Rodrigo et al.35 at a fluence above 3.5 J/cm2. Unfortunately, the fluence dependence of the deposition rate for MAPLE has been investigated systematically in few cases. For water ice Rodrigo et al.35 found that the PEG deposition from a water ice matrix reached a constant value at a fluence of 4–5 J/cm2, but also that the deposition yield exhibited large variations from run to run (because of nanoclusters as discussed below). However, chicken egg lysozyme showed a clear maximum in the deposition rate at 2 J/cm2.38 In both cases the mass of the total deposit was measured, and it was not possible to determine the fraction of intact molecules to that of fragments. Since the increasing energy should lead to a larger ablated volume of the matrix, one would naively expect the deposition yield to increase with fluence as well. In MALDI the ion signal decreases with fluence as a result of an enhanced fragmentation of the analyte ions at higher fluence,39 whereas the fragmentation in MAPLE preferentially leads to loss of the light fragments which have a much broader distribution in a plume in vacuum (see Urbassek and Sibold44 and Pureztky et al.45). In a background gas of the evaporating matrix molecules the light fragments will be scattered even more than the intact film molecules. A strongly enhanced fragmentation rate will therefore lead to a reduction of the total deposition rate, i.e. of fragments as well as of intact molecules. The measurements of the velocity of the matrix particles indicate a front velocity of 700–1,000 m/s,45 which means that a “dense” plume of thickness of 4–6 ȝm is developed after a laser pulse of duration 6 ns has terminated (at the end of stage 2). Compared with a PLD plume, this plume is denser, because the front velocity may be slower and the ablated number of molecules from the matrix much larger (see Table 2). Another issue for MAPLE is the abundance of particulates of regular or irregular shape up to micrometer size.7,34,35,46 Computational and experimental studies by Leveugle et al. 47,48 demonstrate that formation of nanoclusters is an inherent issue for MAPLE. The authors suggest that the ejection of large matrixpolymer clusters is responsible for the polymer features on the surface. The evaporation of matrix molecules during the flight or after the deposition on the substrate leads to residues of complicated shape on the substrate. If this is a
J. SCHOU
254
general feature of MAPLE, it may be difficult to produce uniform films with MAPLE. A closer look at Table 2 reveals several important differences between the processes during PLD (Table 1) and MAPLE. There is no visible plume for MAPLE at 2 J/cm2, whereas there is a distinct plume from silver. Similar to other insulating materials, the thermal diffusion length is much smaller in water ice than in metals. The major difference is the volatility of water ice which means that the evaporation yield may be two orders of magnitude larger. Since also the front velocity below the onset of plasma formation is much smaller, the density of the plume will be comparable to that of water ice. This indicates that the collision rate in the initial plume will be high, but also that the translational expansion of the plume will carry the film molecules (and possible fragments) to the substrate. An interesting point is also that the efficiency of transfer (~1 ng/cm2/pulse) is much higher for lysozyme, which is heavier than the PEG molecule (and its fragments). This efficiency is then in turn much smaller than the corresponding one for the solid silver target (Table 1), which, of course, is related to the fact that the MAPLE results originates from a 1 wt % matrix. TABLE 2. Estimated parameters for a MAPLE process at a fluence of 2 J/cm2 at 355 nm.
Fluence : 2 J/cm2 Wavelength: 355 nm Pulse length: 6 ns Beam spot: 0.015 cm2 Material: Water ice matrix with 1 wt % solute Cohesive energy: 0.45 eV/H2O molec.a Density of water ice: 3.01 u 1022 molec./cm3 (estimated for 0.9 g/cm3) Estimated ablation depth d : 7 Pm Estimated ablated water ice : ~9 Pg Attenuation depth 1/Į : ~1.5 mmb 1/2 Thermal diffusion length (IJLD) : ~80 nmc Velocity of plume: 1,000 m/s Extension of plume at t = IJL : 6 ȝm < 1.85 u 1022 molec./cm3 Density of plume at t = IJL : Solutes: polyethylene glycol (PEG): ~1,333 amu, lysozyme: 14,302 amu Deposit of PEG 1,500 per pulse (60 mm from target): 0.1 ng/cm2 b Deposit of lysozyme per pulse (60 mm from target): 1 ng/cm2 a
From Sack and Baragiola.49 Estimated from Rodrigo et al.35 c From Focsa and Destombes.50 b
FUNDAMENTALS OF LASER-ASSISTED FABRICATION OF FILMS
255
References 1. Eason, R. ed. (2007) Pulsed Laser Deposition of Thin Films, Wiley, Hoboken, NJ, 1–682. 2. Phipps, C. ed. (2007) Laser Ablation and Its Applications, Springer, New York, 1–586. 3. Gorbunoff, A. (2002) Laser-Assisted, Fortschritt-Berichte; Reihe 9, VDI 357 ,VDI Verlag, Düsseldorf, 1–229. 4. Miller, J., and Haglund, R., eds. (1998) Laser Ablation and Desorption, Methods of Experimental Physical Sciences, Academic, New York, vol 30, pp. 1–647. 5. Lippert, T. (2004) Adv. Polym. Sci. 168, 51–246. 6. Bäuerle, D. (2000) Laser Processing and Chemistry, 3rd edition, Springer, Berlin, pp. 1–788. 7. Piqué, A. (2007) Deposition of polymers and biomaterials using the matrix-assisted pulsed laser evaporation (MAPLE) process, in: Pulsed Laser Deposition of Thin Films, Eason, R., ed., Wiley, Hoboken, NJ, 63–83. 8. Bubb, D., McA, D., and Haglund, Jr., R. (2007) Resonant infrared pulsed laser ablation and deposition of thin polymer films, in: Pulsed Laser Deposition of Thin Films, Eason, R., ed., Wiley, Hoboken, NJ, 35–61. 9. Chrisey, D., Pique, A., McGill, R., Horwitz, J., Ringeisen, B., Bubb, D., and Wu, P. (2003) Chem. Rev. 103, 553–576. 10. Thestrup, B., Schou, J., Nordskov, A., and Larsen, N. (1999) Appl. Surf. Sci. 142, 248–252. 11. Schou, J. (2006) Laser beam-solid interactions: fundamental aspects, in: Materials Surface Processing by Directed Energy Techniques, Pauleau, Y., ed., Elsevier, Amsterdam, 35–66. 12. Cheung, J., and Horwitz, J. (1992) Mater. Res. Bull. XVII 2, 30–36. 13. Willmott, P., Manoravi, P., and Holliday, K. (2000) Appl. Phys. A 70, 425–429. 14. Ohkubo, I., Matsumoto, Y., Ueno, K. Chikyow, T., Kawasaki, M., and Koinuma, H. (2003) J. Cryst. Growth 247, 105–109. 15. Rijnders, G., and Blank, D. (2007) Pulsed laser deposition of metals on selfassembled monolayers, in: Pulsed Laser Deposition of Thin Films, Eason, R., ed., Wiley, Hoboken, NJ, 177–190. 16. Willmott, P. (2007) Quasicrystalline thin films, in: Pulsed Laser Deposition of Thin Films, Eason, R., ed., Wiley, Hoboken, NJ, 633–648. 17. Lowndes, D. (1998) Growth and doping of compound semiconductor films by pulsed laser ablation, in: Laser Ablation and Desorption, Experimental Methods in the Physical Sciences, Miller, J.C., and Haglund, R.F., eds., Academic, New York, 30, 475–571. 18. O’Mahony, D., and Lunney, J. (2007) Group III nitride growth, in: Pulsed Laser Deposition of Thin Films, Eason, R., ed., Wiley, Hoboken, NJ, 291–312. 19. Anisimov, S., Luk’yanchuk, B., and Luches, A. (1996) Appl. Surf. Sci. 96–98, 24–32. 20. Toftmann, B., Schou, J., and Lunney, J. (2003) Phys. Rev. B 48, 104101-1-5. 21. Schou, J., Amoruso, S., and Lunney, J. (2007) Plume dynamics, in: Laser Ablation and Its Applications, Phipps, C.R., ed., Springer, New York, 67–95. 22. Chrisey; D., and Hubler, G., ed. (1994) Pulsed Laser Deposition, Wiley, New York, 1–613.
256
J. SCHOU
23. Willmott, P., and Huber, J. (2000) Rev. Mod. Phys. 72, 315–328. 24. Amoruso, S. (1999) Appl. Phys. A 69, 323–332. 25. Amoruso, S., Schou, J., and Lunney, J. (2006) Europhys. Lett. 76, 436–442. 26. Hansen, T., Schou, J., and Lunney, J. (1998) Appl. Phys. Lett., 72, 1829–1831. 27. Toftmann, B., Schou, J., Hansen, T., and Lunney, J. (2000) Phys. Rev. Lett. 84, 3998–4001. 28. Willmott, P. (2004) Prog. Surf. Sci. 76, 163–217. 29. Piqué, A., Kim, H., and Arnold, C. (2007) Laser forward transfer of electronic and power generating materials, in: Laser Ablation and Its Applications, Phipps, C.R., ed., Springer, New York, 339–373. 30. Piqué, A, McGill, R., Chrisey, D., Leonhardt, D., Maina, T., Spargo, B., Callahan, J., Vachet, R., Chung, R., and Bucaro, M. (1999) Thin Solid Films 355–356, 536–341. 31. Karas, M. (1993) Mat. Fys. Medd. Dan. Vid. Selsk. 43, 623–641. 32. Toftmann, B., Papantonakis, M.R., Auyeung, RC.Y., Kim, W., O’Malley, S.M., Bubb, D.M., Horwitz, J.S., Schou, J., Johansen, P.M., and Haglund, R.F. (2004) Thin Solid Films 453–454, 177–181. 33. Toftmann, B., Rodrigo, K., Schou. J., and Pedrys, R. (2005) Appl. Surf. Sci. 247, 211–216. 34. Mercado, A., Allmond, C., Hoekstra, J., and Fitz-Gerald, J. (2005) Appl. Phys. A 81, 591–599. 35. Rodrigo, K., Toftmann, B., Schou, J., and Pedrys, R. (2004) Chem. Phys. Lett. 399, 368–372. 36. Rodrigo, K., Czuba, P., Toftmann, B., Schou, J., and Pedrys, R. (2006) Appl. Surf. Sci. 252, 4824–4828. 37. Ringeisen, B., Callahan, J., Wu, P., Piqué, A., Spargo, B., McGill, R., Bucaro, M., Kim, H., Bubb, D., and Chrisey, D. (2001) Langmuir 17, 3472–3479. 38. Purice, A., Schou, J., Kingshott, P., and Dinescu, M. (2007) Chem. Phys. Lett. 435, 350–353. 39. Dreisewerd, K. (2003) Chem. Rev. 103, 395–425. 40. Schou, J. (1987) Nucl. Instrum. Methods B 27, 188–200. 41. Baragiola, R., Vidal, R., Svendsen, W., Schou, J., Shi, M., Bahr, D., and Atteberrry, C. (2003) Nucl. Instrum. Methods B 209, 294–303. 42. Zhigilei, L., and Garrison, B. (2000) J. Appl. Phys. 88, 1281–1298. 43. Chen, Z., Bogaerts, A., and Vertes, A. (2007) Appl. Phys. Lett. 89, 041503-1-3. 44. Urbassek, H. M., and Sibold, D. (1993) Phys. Rev. Lett. 70, 1886–1889. 45. Puretzky, A., Geohegan, D., Hurst, G., Buchanan, M., and Luk’yanchuk, B. (1999) Phys. Rev. Lett. 83, 444–447. 46. Frycek, R., Jelinek, M., Kocourek, F., Vrnata, M., Myslik, V., and Vrbová, M. (2006) Thin Solid Films 495, 308–311. 47. Leveugle. E., Zhigilei, L., Sellinger, A., and Fitz-Gerald, J. (2007) Appl. Surf. Sci. 253, 6456–6460. 48. Leveugle. E., and Zhigilei, L. (2007) J. Appl. Phys. 102, 074914-1-19. 49. Sack, N., and Baragiola, R. (1993) Phys. Rev. B 48, 9977–9978. 50. Focsa, C., and Destombes, J. (2001) Chem. Phys. Lett. 347, 390–396.
NANOPARTICLES OF SEMICONDUCTORS IN SOL GEL GLASSES R. REISFELD* Department of Inorganic Chemistry, The Hebrew University, Jerusalem, ISRAEL
Abstract – Nanoparticles of II-VI and III-V semiconductor materials are a subject of growing theoretical interest arising from their high potential of practical applications. However in order that these materials will be of practical value in lasers, nonlinear optics, biological markers and sensors for biological and environmental impurities the highly reactive species have to be incorporated in an inert, stable, transparent medium. We present the theory of nanoparticles with emphasize of quantum size effect. Preparation of a number of semiconductor particles and rods in different inorganic, organic and inorganicorganic hybrid matrices are presented. The practical results of optical properties as well as electrical conductivity of a number of materials are discussed. The characterization of the materials by absorption and fluorescence spectroscopy, XRD, EDAX, SEM, TEM, and AFM are presented Electrical measurements were performed. A number of applications of the new materials proposed.
Keywords: Semiconductor quantum dots and rods, sol gel glasses, inorganic-organic hybrid matrices, zirconia-silica-polyurethane, optical and electrical properties
1. Introduction Semiconductor quantum dots exhibit a wide range of optical and electrical properties that are of fundamental and technological interest.1 However because their large surface to volume ratio the particles are subjected to strong ambient influence and quick deterioration. It is therefore of utmost importance to incorporate them into transparent inert medium that will not change their basic properties but keep them unaffected by outside atmosphere.
______ *
To whom correspondence should be addressed: R. Riesfeld, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
257
258
R. REISFELD
It is the subject of this paper how nanoparticles of semiconductors can be incorporated in situ by the use of the sol gel technique. The first part will describe the sol gel process. The second part will present the experimental results obtained in our laboratory of optical and electrical measurements. The third part will summarize the important relevant findings obtained by other groups. 1.1. THE DESCRIPTION OF SOL-GEL PROCESS
The sol-gel method is a low-temperature technique for creating solid glass bulks or thin-films. Using this method, coatings on glass, ceramic, metal or other solid substrates are easily fabricated. In addition, the relatively gentle synthetic conditions allow for the addition of dopants such as organic dyes or inorganic ions, which convert the resulting glass/dopant combination into an active material which may be used in optical or sensing applications. The incorporation of organic materials into glasses prepared using sol-gel methods was first described in 19842 followed by a paper describing incorporation of inorganic ions into thin-film sol-gel coatings.3 The precursor solution for sol-gels consists of various alkoxydes or inorganic sols. This solution can be applied by dip coating, spin coating or laminar coating.4 The most common precursors are used to produce silicates, titania, germanates, alumina, zirconia, tungstates, vanadates and ormosils.5,6 In this section, we first provide a description of the “generic” sol-gel process. Then, a number of specific sol-gel or composite sol-gel/polymer systems are described. For each of these systems, a typical synthesis is given, and properties specific to that system are summarized. The sol gel process is based on hydrolysis and polycondensation reactions of metalorganic compounds such as silicon alkoxides. Commonly used silicon alkoxides include the family of tetra alkoxysilanes, which have the general form Si(OR)4, where R is an alkyl group, and therefore OR is an alkoxy group (usually ethoxy or methoxy). Examples include tetraethoxy-silane (TEOS) and tetramethoxy-silane (TMOS). TEOS or TMOS are used as a component in a starting mixture which also contains water as the second reactant. The sol gel process can be realized without using any solvent, but in most cases, it is preferable to use a solvent in the starting mixture in order to control the reaction rate of the process. Alcohol based solvents are commonly used for this, but there are other possibilities (e.g. acetone). The molar ratio between the different components of the starting mixture greatly influences the final product.
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES
259
For a starting mixture which includes tetra alkoxysilane and water, hydrolysis reactions of the form {Si-OR + H2O o {Si-OH + ROH
(1)
occur. This is the first stage of the sol gel process. Complete hydrolysis of a tetra alkoxysilane molecule, produces: Si(OR)4 + 4H2O o Si(OH)4 + 4ROH
(2)
a solution of silanol groups, Si(OH)4, in alcohol, as shown in Eq. (2). The hydrolysis can be catalyzed by acidic catalysts via an electrophilic mechanism. The reaction rate of the hydrolysis is increased with the strength of the acid. Therefore, HCl and HNO3 are commonly used catalysts. On the other hand, the reaction can also be base-catalyzed by a nucleophilic mechanism.5 The reaction rate of the hydrolysis is also influenced by steric considerations and is decreased with the size of the alkoxy group. For this reason, TMOS hydrolysis is faster then TEOS hydrolysis. The second stage of the process consists of polycondensation of the hydrolysis products. Each condensation reaction of two hydrolysis products has the form: {Si-OH + HO-Si{ l {Si-O-Si{ + H2O
(3)
If we consider, for simplicity, an ideal model in which condensation reactions take place only after a complete hydrolysis, so that the reactants are silanols, then a general form which describes the polycondesation stage is given by: nSi(OH)4 o (SiO2)n + 2nH2O
(4)
The result of the polycondensation reactions is the creation of a three dimensional crosslinked polymer network, leading to formation of sol particles. Further condensation reactions link between the sol particles, forming a wet gel. Evaporation of the liquid leads to a dry gel (xerogel), which is porous glass. The polycondensation can be catalyzed by hydrofluoric acid (HF), since F– ions can replace hydroxyl ion in the hydrolysis product Si(OH)4 and being more electronegative than the hydroxyl, they increase the attraction to other silanols, leading to Si-O-Si bonds.7 In a similar way, the polycondensation can be catalyzed by base. The rate of the condensation reaction influences glass porosity. Higher condensation reaction rates result in higher porosity. Therefore, the
R. REISFELD
260
choice of catalyst (acid or alkaline) allows a degree of control of the porosity of the glass. For example, when HF is used as a catalyst, the porosity increases with increasing concentration of HF.8,9 One of the main problems of sol gel glasses is cracking, which is attributed to inhomogeneous drying.10 Furthermore, even uncracked glasses suffers from a tendency to develop cracks when introduced into liquids. Since the reaction, Eq. (3), is reversible, the presence of water in the glass prevents complete condensation, and large amounts of penetrating water can cause hydrolysis, resulting in cracking of the glass. Heating the glass to high temperatures evaporates the water, leading to de-hydroxylation and complete condensation, which stabilizes the glass. A thermal treatment at more than 800°C makes the glass completely impenetrable for water, but many applications require penetration of liquids into the glass. For this purpose it is common to heat the glass up to 500°C. This produces a glass structure which enables penetration of liquids only into the pores, while preventing damage to the glass. The Figure 1 is a schematic representation of the classical sol-gel process for preparation of silica skeleton.
Si
OR
+
HOH
Hydrolysis Reesterification
Si
OH
+
Si
OH
Si
OH
+
ROH
Water Condensation Si
O
1
Si
+
HOH
2a
Si
+
ROH
2b
Hydrolysis
Si
OH
+
Si
OR
Alcohol Condensation
Si
O
Alcoholysis
Figure 1. Classic sol-gel reactions scheme.
1.1.1. Ormosil hosts Organically modified silicates (ORMOSILs) usually exhibit lower porosity and enhanced mechanical properties which allow cutting, grinding and polishing prior to heat treatment. A typical ORMOSIL gel network contains a significant amount of organic functionalities, which offers great flexibility with respect to the chemical compatibility of the gels with the nanoparticles to be incorporated. Sol-gel process, can be improved by modifying the sol
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES
261
process and using a variety of organofunctional silicone alkoxides. Modified alkoxide precursors RSi(OEt)3 where R is an organic group such as methyl, vinyl or amyltriethoxysilane (denoted as MTEOS, VTEOS and ATEOS, respectively) lead to organic-inorganic hybrid matrices. The covalently bonded organic groups decrease the mechanical tensions during the drying process.11 Functionalized alkoxides F-R’-Si(OEt)3, where F is a functional group such as amino or isocyanate and R’ is an alkyl spacer, enable dopants to be covalently grafted onto the xerogel matrix, in order to avoid phase separation. After drying, optically clear and dense inorganic-organic hybrid xerogels were obtained (30 mm diameter and 15 mm thick).12 In our lab, ORMOSIL glass samples were prepared by a one step process at room temperature which led to the formation of hybrid organic/inorganic materials. Zirconia matrix has been found as a very useful material for incorporation of CdS13 and CdSe14 and CdTe.15 Preparation of zirconia matrix is given in scheme 2. 1.1.2. Polyurethane A newly organically modified zirconia silica polyurethane was found to be an excellent host for organically active species.16 Preparation of zirconia polyurethane and its precursor are given in schemes 1–4. The hybrid material Zirconia-Silica-Urethane (ZSUR) was obtained by using three precursors: (a) poly(ethylene)glycol chain covalently linked by urethane bridges with thriethoxysilane groups synthesized separately, (b) epoxy-silica ORMOSIL precursor and (c) a zirconium oxide precursor. The preparation procedure of ZSUR is shown in the schemes 1–4. The final solution of ZSUR was obtained from DURS (di-urethane siloxane), ESOR (epoxy-silica ormosil) and ZrO2 precursor solutions. The ZrO2 matrix solution was obtained from zirconium n-tetrapropoxide in propanol and acetic acid was used as a chelating agent to stabilize the zirconium oxide precursor. The DURS-olygomer precursor was synthesized by reacting of ICTEOS (3-isocyanatopropylthriethoxysilane) and PEG-600 (polyethylene glycol) with the molar ratio (2:1). The reagents were stirred in chloro benzene at boiling temperature under reflux for 3 h. The residual solvent was evaporated and the DURS inorganic-organic hybrid oligomer material was obtained. Isocyanate functionalized siloxane coupled with the polyol to form urethane linkage is known as a system improving bond strength. The epoxy- silica-ormosil (ESOR) precursor was obtained from tetramethoxysilane (TMOS) and 3-glycid oxypropyl trimethoxysilane (GLYMO) with molar ratio 1:1.
R. REISFELD
262
3 ml of CH3 COOH
10 ml of Zr(OPr)4
15 min. stirring, at room temperature
20 ml of propanol
3.3 ml of solution H2 O: CH3COOH = 1:1
15 min. stirring, at room temperature
Filtration
ZrO 2 matrix solution Scheme 1 Preparation of the zirconium oxide matrix solution..
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES
3ml of TMOS Si(OCH3)4
3ml of MTMOS CH3 -Si-(OCH3 )3
0.25 ml Acetic Acid CH3 COOH
15 min. stirring, at room temperature
10 ml Ethanol + 3.6 ml water 15 min. stirring, at room temperature
4.8 ml of GLYMO
2 hour stirring, at room temperature
Filtration
ESOR precursor solution
Scheme 2 Preparation of the epoxy-silica ormosil ESOR precursor solution.
263
R. REISFELD
264
Cl-benzene C6 H5 Cl
Poly(ethylene Glycol), PEG HO-(-CH2 -CH 2-O-)n-H
(C2H 5) 3Si-(C3H 6)-N=C=O, ICTEOS
Stirring under reflux 3 hours at the boiling temperature
Cl-benzene evaporation
Cooling to room temperature
Filtration
DURS precursor solution
Scheme 3 Preparation of di-urethane- silica DURS precursor solution.
First, TMOS was hydrolyzed at room temperature (Tor) for 2 h (TMOS: CH3OH:H2O:CH3COOH =15.0:12.8:7.2:2.4), then GLYMO was added and stirred for 3 h at Tor. In order to obtain ZSUR, these two composites, DURS and ESOR were combined with a zirconium oxide (ZrO2) matrix, which was used as an inorganic hetero network and as a promoter catalyst for the epoxy polymerization. Although the reaction at room temperature between zirconium and epoxy groups is limited (about 27% unreacted epoxy groups) as well as at high temperature (70oC/8 h, about 24% unreacted), in the case of ZSUR unreacted epoxy groups (at least 24%) can be reacted with secondary amino groups in urethane linkage of DURS. The nominal molar ratio in the final sol was: SiO2:ZrO2 = 69:31; and of Urethane to Epoxy was 24:76. Due to strong chemical bonding between the inorganic coupling agent isocyanatotrietoxysilane and the organic polymer (polyethylene glycol), GLYMO and zirconium oxide, it is possible to combine the strength and hardness of solgel matrices with the processibility and ductility of polymers which offer better mechanical properties (elasticity, flexibility) and higher chemical stability.
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES ESOR Epoxy-silica ormosil precursor solution
265
ZrO 2 Zirconium oxide matrix solution
30 min. Stirring at room T oC
DURS Di- urethane- silica precursor solution
1 hour Stirring at room T oC
Dip-coating and treatment of ZSUR Zirconium –Silica-Polyurethane thin films
Scheme 4 Synthesis and preparation of ZSUR zirconium–silica–polyurethane thin film.
2. From Theory to Applications 2.1. NANOPARTICLES IN GLASSES PREPARED BY THE SOL-GEL METHOD
2.1.1. Quantum size effects Semiconductor nanocrystals in transparent media have recently received attention due to their properties and promising applications in the non-linear optics and optical switching. Nanocrystalline semiconductors exhibiting quantum size effects have been prepared and studied in a number of different forms, mainly as dispersed colloidal particles in a liquid or solid matrix but also in thin films. The incorporation of CdS into silica film and into zirconia film prepared by sol-gel method has been reported previously.17,18
R. REISFELD
266
When the diameter of a semiconductor crystallite approaches, or becomes smaller than the bulk exciton diameter, changes occur in the energy band structure. Since the levels are confined to potential wells of small lateral dimensions, quantization occurs, resulting in increased spacing between levels as crystal size becomes smaller. The effect is observed as blue shift in the absorption and emission spectra with decreasing crystal size. These changes in the absorption spectrum of materials with changing particle size have been used to illustrate the quantum size effect (QSE). From theoretical point of view these materials provide models for three dimensional quantum confinement and allow identification of the excited states. The energy states of the quantum dots are positioned between the discrete energy levels of atoms or bands of molecules, and the broad band of the condensed phase. The basic ideas of quantum confinement or quantum wells were introduced in the 1970s. In these nanoclusters, the dimensions of the wave functions of the electron-hole pair (exciton) in the lowest excited state of the nanocluster are comparable to the physical size of the particle. This quantum confinement of the exciton means that the continuum band of energies becomes more molecular in character, with narrow ranges of energy and line structure in the optical spectra. From chemical viewpoint, less delocalization means less energy stabilization; a reflection of this is that the absorption band for direct transmissions of nanosized semiconductor clusters is shifted to higher energies than in the extended bulk parent materials. The “photo charging” of the semiconductor particle, that is, photo inducting high electron concentration into a narrow conducting band with a small effective density of states, also changes the band gap and the resulting absorption edge. Generally, the blue shift of the exciton absorption peaks is discussed in two cases: one is valid for the excitons having an effective Bohr radius R much larger than the crystal size a0, and the other is for exciton having an effective radius much smaller than the crystal size (a0/R > 4). These two regimes are named the “electron-hole confinement” and the “exciton confinement”, for which energies are approximately given by, respectively,
E
E g = 2S 2 / 2 Pa 02
E
Eg = S
2
2
/ 2Ma
(5)
2 0
Where P and M is the reduced mass (P = me × mh) and the translational mass (M = me × mh), respectively, of the exciton. Many theoretical models for the expected variation of effective bandgap have been proposed. The most commonly used, at least until recently, was the effective mass approximation, based on parabolic energy bands, used by Efros
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES
267
and Efros19 and Brus.20 This model treats the particles as a sphere-boy confining the electrons and holes with an infinite potential barrier, including a correction for coulombic interaction between electrons and holes. In general, the lower excited state energies can be calculated within the effective mass approximation using the different mass for the conduction and valence bands. The Coulomb interaction can be included to give an approximate expression for the lowest energy.
E* Eg
=2S 2 ª 1 1 º 1.8e2 « » 2a02 ¬ me mh ¼ H a0
(6)
where H- is the dielectric constant of the semiconductor, Eg- is the bandgap of the bulk material, e- electron charge, me- is the reduced electron mass and mh- is the reduced hole mass. The quantum size effect is dependent on the crystal size a0. The effect increases within the decreasing of the crystal size. Compared to the well-known II-VI and III-V materials, PbS have a larger exciton radius (aB = 18 nm) and a small fraction of atoms at the surface at the same carriers confinement regime. The absorption edges of PbS NCs cover the 0.5–2.5 μm wavelength region, which are important for telecommunication. More recent reports discuss the incorporation of PbS NCs in polymers,20 glasses21 and in sol-gel materials.22–25 The sol-gel process extends the conventional glass melting methods, as it al lows incorporation of semiconductor NCs at low temperatures and predetermined concentration and size. Furthermore, the sol-gel technology has advantages in the formation of films with controllable thickness, three-dimensional protection of the NCs, prevention of NCs growth, aggregation and oxidation. The widely investigated matrix materials are SiO2, ZrO2 SiO2/TiO2 and ZrO2/Ormosil. PbS NCs of sizes 2–4 nm in ZrO2,23 2–3.5 nm in SiO2/TiO2 and 4–8 nm in ZrO2/Ormosil matrix26,27 have also been studied recently. 2.1.2. PbS NCs-polyurethane composite We have reported27 the incorporation of PbS NCs ranging between 4–8 nm into a Zirconium-Silica-Urethane (ZSUR) matrix obtained by the sol-gel method. The sizes of the particles were controlled by temperature treatment and by concentration of PbS in the ZSUR matrix. PbS NCs sizes were determined by transmission electron microscopy (TEM) measurements. The quantum size effect could also be extracted from optical absorption and photoluminescence spectra. The new matrix allows incorporation of 5–50% PbS forming a characteristic structure of PbS NCs with increasing of NCs concentration.
R. REISFELD
268
Table 1 summarizes the energy of the first exciton transition and particles size as a function of annealing temperatures and mole per cent of PbS materials within the different matrices. Table 2 summarizes the band gap energy between uppermost valence band and the conduction band and types of structure of well-known semiconductors. I-V characteristics of the films with PbS NCs in ZSUR were measured for three configurations of the films: one layer uniform PbS NCs in hybrid film (ITO/ PbS-hybrid film/Au) structure exhibiting nearly symmetric nonlinear characteristics; two-layer structure (ITO/PbS-hybrid film/PbS bulk/Au) depending on the DURS concentration in hybrid matrix exhibiting rectifying behavior with rectifying ratio 50–200 at 1 V; and the third uniform PbS thin film (Au/PbS thin film/Au) structure with thickness dependence of specific conductivity. TABLE 1. Optical properties of PbS nanoparticles in ZSUR and ZrO2 sol gel films.
PbS (mol %) in ZSUR (a, b, c) films (a) 20
(b) 30
(c) 40
Tan o
( C) 130 170 200 220 130 170 200 220 130 170 200 220
Absorption band edge (eV) 2.75 2.30 2.00 2.25 2.60 2.05 1.52 1.88 2.15 1.72 1.28 1.85
PbS (mol %) in ZrO2 (d, e) films (d) 20
(e) 30
Tan o
( C)
Absorption band edge (eV) 2.05 1.65 1.35 1.92 1.55 1.35 1.65 1.90
200 250 300 350 200 250 300 350
TABLE 2. Band gap energies Eg, eV between the uppermost valence band and the conduction band, and the shift in nanoparticles.
Semiconductor CdS CdSe
Structure Wurtzite Zinc blende Wurtzite Zinc blende
Band gap energies, Eg (eV) 2.58 2.45 1.84 1.67
The shift of Eg (eV) in nanoparticles26 3.0 2.2
CdTe
Zinc blende Wurtzite
1.45 1.58
2.3
PbS
Cubic
0.41
1.28–2.3
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES
269
2.1.3. PbS nanocrystal assembles PbS nano-wires and nano-rods were synthesized by surfactant assisted solvothermal reaction and was reported recently by us.28 The synthesis is based on decomposition of lead thiocyanate in boiling benzyl alcohol with Cetyl trimethyl ammonium bromide used as a surfactant. Nanowires of PbS (about 2–3 mm with an average diameter of 30–50 nm) and nanorods (200–300 nm in length with an axial ratio of 4–5) were synthesized. The nanostructures were characterized by high resolution transmission electron microscopy (HR-TEM), scanning electron microscopy (SEM), selected area electron diffraction (SAED) and X-ray diffraction analysis. The experimental results indicate that the reaction duration and concentration of surfactant play key roles in determining the final morphologies of PbS blocks building and also in their crystallinity. SEM characteristics of rods and wires are given in the Figures 2 and 3.
1
Figure 2. SEM images of PbS nanorods prepared using 1.7 mMol of CTAB in BA at 200oC with a reaction duration of 6 h, SAED characteristic are shown in the Figure 4.
2
Figure 3. SEM images of PbS nano-wires prepared using 3.7 mMol of CTAB in BA at 200oC with a reaction duration of 6 h.
R. REISFELD
270
3
Figure 4. SAED of PbS nano-rods obtained with a reaction duration of 6 h.
2.1.4. Indium arsenide semiconductor quantum rods and wires Indium phosphate and indium arsenide nanorods were recently reported by Shweky et al.29 The use of a hydrophobic polymer, polystyrene, serves to encapsulate the nanocrystals inside the spheres awhile maintaining many of their original properties. The nanospheres and nanowires were characterized by transmission electron microscopy, scanning electron microscopy, energy dispersive X-ray spectroscopy, and by a single-particle fluorescence spectroscopy. A synthesis of soluble III-V semiconductor quantum rods using gold nanoparticles to direct and catalyze one-dimensional growth was developed. The growths took place via the solution-liquid-solid (SLS) mechanism where proper precursors were injected into a coordinating solvent. The synthesis of InP nanorods using indium acetate and myristic acid with gold nanoparticles as the catalysts in the SLS growth mode. A similar route was successfully developed for growth of InAs nanorods. The Au catalyst in the reaction is as important parameter to achieve shape control. The length of the InP rods could be controlled by changing the ration of the In/P precursors. InP rods show evolution of the band gap from 0D dots ti 1D wires upon elongating the rod axis. Furthermore, they can potentially be integrated in composite nanocrystalspolymer solar cells. 3. Conclusions Nanoparticles of variety of II-VI and III-V semiconductors are a subject of a large number of investigations because of their theoretical interest in quantum theory and the high potential of practical application. However in order that these materials will be of practical value in lasers, nonlinear optics, biological
SEMICONDUCTOR NANOPARTICLES IN SOL GEL GLASSES
271
markers and sensors for biological and environmental impurities the highly reactive species have to be incorporated in an inert stable, transparent medium We discuss here preparation of a number of semiconductor particles and rods. Their incorporation in different inorganic, organic and inorganic-organic matrices, optical properties is combined with electrical studies of the materials. A short theory of quantum effect was presented and the synthetic procedure for preparation of a number of hosts, inorganic and inorgano-organic, the experimental ways of incorporating the quantum dots and rods is presented. The characterization of the materials by absorption and fluorescent spectroscopy, electric measurements, XRD, EDAX, SEM, TEM and AFM is shown. I-V characteristics measured for two-layer structure (ITO/PbS-hybrid film/PbS bulk/Au) depending on the DURS concentration in hybrid matrix exhibiting rectifying behavior with rectifying ratio 50–200 at 1 V; and the third uniform PbS thin film (Au/PbS thin fim/Au) structure with thickness dependence of specific conductivity. A number of applications of the new materials proposed. ACKNOWLEDGEMENTS
The author is deeply grateful to Tsiala Saraidarov for providing many experimental data.
References 1. 2. 3. 4. 5. 6. 7.
8. 9. 10. 11. 12. 13.
Alivisatos, A. (1996) Science 271, 933. Avnir, D., Levy, D., and Reisfeld, R. (1984) J. Phys. Chem. 88, 5956–5956. Avnir, D., Kaufman, V., and Reisfeld, R. (1985) J. Non-Cryst. Solids 74, 395–406. Sakka, S. (1996) Structure and Bonding, 85, Springer, Berlin. Brinker, C., and Scherer, G. (1990) Sol-Gel Science, Academic, London. Reisfeld, R. (2005) Polymery 51, 95–103. Rabinovich, E., and Wood, D. (1986) Fluorine in silica gels, in: Better Ceramics Through Chemistry II, Eds: Brinker, C., Clark, D., and Ulrich, D., Mat. Res. Soc., MRS, Pittsburgh, PA, 251. Pope, E., and Mackenzie, J. (1986) J. Non-Cryst. Solids 87, 185. Yariv, E., Reisfeld, R., Saraidarov, T., Axelrod, E., Rysiakiewicz-Pasek, E., and Wodnicka, K. (2002) J. Non-Cryst. Solids 305, 354. Zarzycki, J., Prassas, M., and Phalippou, J. (1982) J. Mater. Sci. 17, 3371–3379. Phil1lip, G., and Schmidt, H. (1984) J. Non-Cryst. Solids 63, 283. Lin, H., Bescher, E., Mackenzie, J., Dai, H., and Stafsudd, O. (1992) J. Mater. Sci. 27(20), 5523–5528. Zelner, M., Minti, H., Reisfeld, R., Cohen, H., and Tenne, R. (1997) Chem. Mater. 9, 2541–2543.
272
R. REISFELD
14. Gorer, S., Hodes, G., Sorek, Y., and Reisfeld, R. (1997) Mater. Lett. 31, 209–214. 15. Zelner, M., Minti, H. Reisfeld, R., Cohen, H., Feldman, Y., Cohen, S., and Tenne, R. (2001) J. Sol-Gel Sci. Techn. 20, 153–160. 16. Reisfeld, R., Weiss, A., Saraidarov, T., Yariv, E., and Ishchenko, A. (2004) Polym. Adv. Technol. 15, 291. 17. Reisfeld, R., Zelner, M., Saraidarov, T., and Minti, H. (2001) Advances in Energy Transfer Processes, in B. Di Bartolo and X. Chen, (eds.), World Scientific, New Jersey, London, Singapore, Hong Kong, pp. 341–358. 18. Minti, H., Eyal, M., Reisfeld, R., and Berkovic, G., (1991) Chem. Phys. Lett. 183, 277–282. 19. Efros, Al.L., and Efros, A.L., (1982) Fiz. Tekh Poluprovodn. (Sant-Peterburg) 16(7), 1209. 20. Brus, L.E. (1983) J. Chem. Phys., 79(11), 5566–5571. 21. Lipovskii, A.A., Kolobkova, E.V., Olkhovets, A., Petrikov, V.D., and Wise, F., (1999) Low-Dimensional Systems & Nanostructures (Amsterdam) Physica E: 5(3), 157. 22. Martucci, A., Innocenzi, P., Fick, J., and Mackenzie, J. D., (1999) J. Non-Crys. Solids, 244, 55. 23. Sashchiuk, A., Lifshitz, E., Reisfeld, R., Saraidarov, T., Zelner, M., and Willenz, A., (2002) J. Sol-Gel Tech. 24, 31. 24. Kang, I., and Wise, F., (1997) J. Opt. Soc. Am. B 14, 1632–1646. 25. Martucci, A., Fick, J., Schell, J., Battaglini, G., and Guglielmi, M., (1999), J. of Appl. Phys. 86(1), 79–87. 26. Saraidarov, T., Reisfeld, R., Sashchiuk, A., and Lifshitz, E., (2003) J. Sol-Gel Sci. Tech. 26, 533. 27. Saraidarov, T., Reisfeld, R., Sashchiuk, A., and Lifshitz, E., (2005) J. Sol-Gel Sci. Tech. 34, 137. 28. Saraidarov, T., Reisfeld, R., Sashchiuk A., and Lifshitz, E., (2007) Physica E 37,173–177. 29. Shweky, I., Aharoni, A., Mokari, T., Rothenberg, E., Nadler,, M., Popov, I., and Banin, U., (2006) Mater. Sci. and Eng. C, 26, 788–794.
ELECTROCHEMICAL SENSOR TECHNOLOGY BASED ON NANOMATERIALS FOR BIOMOLECULAR RECOGNITIONS
A. ERDEM* Ege University, Faculty of Pharmacy, Analytical Chemistry Department, Bornova, 35100 Izmir, TURKEY
Abstract – There have been many important technological advances for development of electrochemical approaches to monitor biomolecular interactions and recognition events in solution and at solid substrates. This investigation describes how novel surfaces modified with nanomaterials can produce selective and sensitive electrochemical DNA sensors and biological recognition surfaces.
Keywords: Electrochemical sensors, DNA, nanomaterials.
1. Introduction The detection of different biomolecules has a particular interest in genetics, pathology, criminology, pharmacogenetics, food safety and many other fields. Many important technological advances have been made for development of electrochemical approaches to monitor biomolecular interactions and recognition events in solution and at solid substrates.1–3 Various combinations of biological material associated with different types of transducers are an attracttive subject of research. A biosensor is a device that incorporates a biologically active layer at the surface as recognition element and converts the physical parameters of a specific biological interaction into a measurable analytical signal. It associates a bioactive sensing layer with any suitable transducer and
______ *
To whom correspondence should be addressed: A. Erdem, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
273
274
A. ERDEM
gives an output signal. Biomolecular sensing can be defined as the possibility of detecting analytes of biological interest. Employing an affinity layer, this can be a natural system or an artificial one, able to recognize a target molecule in a complex medium among thousands of others. The major processes involved in a biosensor system are analyte recognition, signal transduction and readout. Owing to their desirable characters, e.g., accuracy, speed, and easy automation, such analytical devices hold great promise for clinical and industrial food applications. DNA biosensor normally employs immobilized DNA probes as the recogniion element and measures specific binding processes such as the formation of DNA-DNA and DNA-RNA hybrids, and the interactions between proteins or ligand molecules with DNA at the sensor surface.4 The design of a genosensor involves the following steps5: (a) modification of the sensor surface to create an activated layer for the attachment of the DNA probe; (b) immobilization of the probe molecules onto the surface, preferably with controlled packing density and orientation; and (c) detection of target gene sequence by DNA hybridization at the sensor-liquid interface. The detection of specific DNA sequences provides the basis for detecting several of microbial and viral pathogens. Traditional methods for DNA sequencing based on the coupling of electrophoretic separations and radioisotopic (32P) detection are time consuming and labor intensive, and thus are not well suited for routine medical analysis, particularly for point-of-care tasks. Electrochemical sensors developed for biomolecular recognitions may greatly reduce the assay time and simplify its protocol. Such fast on-site monitoring schemes are required for quick preventive action and early diagnosis. The development of DNA hybridization biosensors holds great promise for obtaining sequence-specific information in connection with clinical, environmental or forensic investigations. Nucleic acid hybridization is a process in which inconsonant nucleic acid strands, with specific organization of nucleotide bases exhibiting complementary pairing with each other, form a stable duplex molecule6 under specific given reaction conditions. DNA hybridization biosensors can be employed for determining early diagnoses of infectious agents in various environments7,8 and for monitoring sequence-specific hybridization events directly9–14 based on the oxidation signal of guanine/adenine or using DNA intercalators which form complexes with the nitrogenous bases of DNA.15–23 DNA hybridization can also be detected by redox-active metal complexes.18,20 The novel surfaces modified with nanomaterials offer an excellent prospect for biological recognition surfaces in order to develop a more selective and sensitive electrochemical DNA sensor technology.
ELECTROCHEMICAL SENSORS 2.
275
The Applications of Electrochemical Sensor Technology Based on Nanomaterials
The development of advanced biosensors based on novel nanomaterials have been considered as important tools in the new approaches in the field of genomics, medical diagnosis, and drug-DNA interactions.34 Electrochemical nucleic acid sensors based on magnetic particles, nanoparticles labeled with metal tags, nanotubes and other nanomaterials have been recently investigated.11–14,24–36 The electrochemical detection of DNA based on magnetic particles11–14,29,33,35 as labeling with an enzyme,35 or using label free system,11–14,33 or combining with metal nanoparticles,29–31 brings the sequence specific detection of DNA hybridization observed in exceedingly low detection limits as resulting in efficient magnetic separation. An increasing interest has appeared in the development of simple, rapid and user-friendly electrochemical detection systems based on DNA sequence and mutant gene analysis, for instance early and precise diagnosis of infectious agents, for routine clinical tests.8–12,14,18,22 Wang et al. was reported a novel genomagnetic electrochemical assay related to BRCA1 breast-cancer gene based on label-free detection.11 An enzyme-linked sandwich hybridization was combined with electrochemical detection of DNA sequences related to BRCA1 gene by using magnetic particles labeled probe hybridizing to a biotinylated DNA target capturing a streptavidin-alkaline phosphatase (AP) enzyme, and consequently, 1-naphthol was measured as a product of enzymatic reaction in the presence of DNA hybridization.35 A genomagnetic assay using commercial magnetic particles developed as the first time for the electrochemical monitoring of detection of wild type hepatitis B virus (HBV) DNA in polymerase chain reaction (PCR) amplicons in length 437-bp has been described by Erdem et al.14 In contrast to other similar methodologies earlier reported in the literatures, Erdem et al.33 produced the streptavidin coated magnetic nanoparticles in the average diameter of 125 and 225 nm. It was exhibited that DNA hybridization can be realized onto magnetic nanoparticles carrying the probe oligonucleotides with the target sequences within the medium, and it can effectively followed by the guanine oxidation signal using an electrochemical nucleic acid sensor in order to detect DNA quite sensitively and selectively, with this less time-consuming, and cheaper label-free electrochemical technique as the first time by using these magnetic nanoparticles tailor-made by Erdem et al.33 in comparison to traditional techniques reported in the related literatures,18,20,37,38 in which several external indicators Co(phen)3]3+, di(2,2ƍ-bipyridine)osmium (III) complexes, methylene blue, etc. have been used. The application of magnetic assay in connection with a metal nanoparticle was presented for electrochemical DNA detection using screen printed electrodes (SPEs).20 The selectivity of this assay was
276
A. ERDEM
also checked co-existing of the larger excess of different point mutations in oligonucleotides and noncomplementary DNA sequences. The versatility of the Carbon-Carbon bond presents the opportunity for attaching different functional groups to the end of the carbon nanotube (CNT) that offers potential for CNTs to be used as a new material for sensors in (bio) chemical applications.39 Direct electrochemistry of DNA guanine and adenine at a multi-walled carbon nanotube (MWNT) modified glassy carbon electrode (GCE) provided significantly enhanced voltammetric signals in comparison to unmodified GCE. Wang detected 100 fmol of breast cancer BRCA1 gene by using enhanced guanine oxidation signal at a MWNT modified GCE.26 Pedano et al. have recently fabricated carbon nanotube paste electrode by using MWNT for adsorption and electrochemical oxidation of nucleic acids.27 Yim et al.28 designed DNAzyme-MWCNT conjugates by the help of streptavidin modification onto CNT conjugates using carbodimide chemistry in order to bind biotinylated DNAzyme to streptavidin coated surfaces properly and selectively. The detection of hybridisation of DNAzyme was measured fluorescently in the presence of fluorescein. A simple and sensitive electrochemical method based on CNT modified disposable graphite electrodes for the detection of DNA and label-free DNA hybridization was performed by using the signal enhancement of the guanine oxidation signal without any modifications in the native bases or any external labeling as the first time by Erdem et al.32 They showed that the label-free detection of DNA based on the signal enhancement of guanine oxidation was using CNT modified transducers; such as, GCE and graphite pencil graphite electrode (PGE). It has been shown that both CNT modified transducers displayed an attractive voltammetric performance over their bare ones, the modified PGE compared favorably to the commonly used CNT modified GCE electrode. The modification of disposable PGE with CNTs demonstrated here has been very simple, fast and consequently, a novel DNA biosensor tested in this study offers some important advantages such as being inexpensive, simple and direct electrochemical assay in more reproducible and sensitive results with a good degree of selectivity. In comparison to the studies performed using CNTs in the literatures,26 there was no time consuming surface chemistry applied for immobilization of DNA or any bead based amplification routes was followed here. 3. Conclusions Nanomaterials have unique chemical and physical properties that offer important possibilities for analytical chemistry. It is hoped that continued development in nanoscience through combined efforts in microelectronics, surface chemistry, molecular biology, and analytical chemistry will lead to the establishment of
ELECTROCHEMICAL SENSORS
277
sensor technology as a major component of analytical biochemistry.5 The integration of nanotechnology in combination with molecular biology and electrochemistry has been expected to create major advances in the area of electrochemical sensor technology. Urgent applications of these sensors will include directly quantifying DNA samples for use in sequencing or polymerase chain reactions (PCR), or pharmaceutical testing and quality control.8 They eventually could be applied to produce credit card-sized sensor arrays for clinical applications such as detection of some pathogenic bacteria, tumors, and genetic disease, or for forensics. ACKNOWLEDGEMENTS
A.E. acknowledges the financial support from TUBITAK; Project no.106S181, and she would like to express her gratitude to the Turkish Academy of Sciences (TUBA) as the associate member of TUBA for the support.
References 1. 2. 3. 4. 5. 6.
Wang J. (2000), Nucl. Acids Res., 28, 3011–3016. Palecek E., Fojta M. (2001), Anal. Chem., 73, 75A–83A. Erdem A., Ozsoz M. (2002), Electroanal., 14, 965–974. Pividori M.I., Merkoci A., Alegret S. (2000), Biosens. Bioelectron., 15, 291–303. Yang M., McGovern M.E., Thompson M. (1997), Anal. Chim. Acta, 346, 259–275. Bej A.K. (1996), Nucleic Acid Analysis: Principles and Bioapplications, Wiley-Liss, New York, Chapter 1, pp. 1–29. 7. Hodgson J. (1998), Nat. Biotech., 16, 725–727. 8. Millan K., Saraulo A., Mikkelsen S.R. (1994), Anal. Chem., 66, 2943–2948. 9. Wang J., Rivas G., Fernandes J. R., Lopez Paz J. L., Jiang M., Waymire R. (1998), Anal. Chim. Acta, 375, 197–203. 10. Erdem A., Pividori M. I., Del Valle M, Alegret S. (2004), J. Electroanal. Chem., 567, 29–37. 11. Wang J., Kawde A.-N., Erdem A., Salazar M. (2001), Analyst, 126, 2020–2024. 12. Erdem A., Pividori M. Isabel, Lermo A., Bonanni A., Del Valle M, Alegret S. (2006), Sensor Actuat. B-Chem. 114, 591–598. 13. Wang J., Flechsig G.-U., Erdem A., Korbut O., Gründler P. (2004), Electroanal., 16(11), 928–931. 14. Erdem A., Ozkan Ariksoysal D., Karadeniz H., Kara P., Sengonul A., Sayiner A.A., Ozsoz M. (2005), Electrochem. Commun., 7(8), 815–820. 15. Palecek E., Tomschik M., Stankova V., Havran L. (1997), Electroanal., 9, 990–997. 16. Karadeniz H., Erdem A., Gulmez B., Jelen F., Ozsoz M., Palecek E. (2006), Front. Biosci., 11, 1870–1877. 17. Fojta M., Doffkova R., Palecek E. (1996), Electroanal., 8, 420–426. 18. Erdem A., Kerman K., Meric B., Akarca U.S., Ozsoz M. (1999), Electroanal., 10, 586–588. 19. Erdem A., Ozsoz M. (2001), Anal. Chim. Acta, 437, 107–114.
278
A. ERDEM
20. Erdem A., Kerman K., Meric B., Akarca U.S., Ozsoz M. (2000), Anal. Chim. Acta, 422, 139–149. 21. Jelen F., Erdem A., Palecek E. (2002), Bioelectrochem., 55, 165–167. 22. Marrazza G., Chiti G., Mascini M., Anichini M. (2000), Clin. Chem., 46, 31–37. 23. Ozkan D., Erdem A., Kara P., Kerman K., Gooding J.J., Nielsen P.E., Ozsoz M. (2002), Electrochem. Commun., 4, 796–802. 24. Mc Farland A.D., Van Duyne R.P. (2003), Nano Lett., 3, 1057–1062. 25. Salem A.K., Chao J., Leong K.W., Searson P.C. (2004), Adv. Mater., 16, 268–271. 26. Wang J., Kawde A.N., Musameh M. (2003), Analyst, 128, 912–916. 27. Pedano M.L., Rivas G.A., Pedano M.L. (2004), Electrochem. Commun., 6, 10–16. 28. Yim T.J., Liu J., Lu Y., Kane R.S., Dordick J.S. (2005), JACS, 127, 12200–12201. 29. Wang J., Xu D., Polsky R. (2002), JACS, 124, 4208–4209. 30. Wang J., Guodong L., Polsky R., Merkoci A. (2002), Electrochem. Commun., 4, 722–726. 31. Wang J. Guodong L, Merkoci A. (2003), Anal. Chim. Acta, 482, 149–155. 32. Erdem A., Papakonstantinou P., Murphy H. (2006), Anal. Chem., 78, 6656–6659. 33. Erdem A., Sayar F., Karadeniz H., Guven G., Ozsoz M., Piskin E. (2007), Electroanal., 19, 798–804. 34. Wang J. (2005), Analyst, 130, 421–426. 35. Wang J., Xu D., Erdem A., Polsky R., Salazar M. (2002), Talanta, 56, 931–938. 36. Vaseashta A. (2003), J. Mater. Sci.-Mater. Electron., 14 (10/12) 653–656. 37. Ju H.X., Ye Y.K., Zhao J.H, Zhu Y.L. (2003), Anal. Biochem., 313, 255–261. 38. Ye Y.K., Zhao J.H., Yan F, Zhu Y.L., Ju X.H. (2003), Biosens. Bioelectron., 18,1501–1508. 39. Vaseashta, A. (2004), Carbon Nanotubes Based Sensors and Devices, in S. Bandopadhyay (ed.), Tata McGraw Hill, New Delhi/Calcutta/Sydney.
THE EFFECTS OF DOPING WITH ELEMENTS FROM THE IIA GROUP ON THE THERMAL AND ELECTRONIC PROPERTIES OF AMORPHOUS SELENIUM G. BELEV1, D. TONCHEV1*, S.O. KASAP1, AND H. MANI2 Department of Electrical and Computer Engineering, Electronic and Photonic Materials Research Laboratories, University of Saskatchewan, 57 Campus Drive, Saskatoon, S7N 5A9, CANADA 2 Anrad, 4950 Levy Street, St. Laurent, H4R 2P1, CANADA 1
Abstract – We have investigated the effects of dopants from Group IIA on the electrical and thermal properties of amorphous selenium (a-Se). We have found that additions in the parts per million range of Mg, Ca, and Ba can alter significantly both the electronic and the thermal properties of a-Se. Vacuum evaporated films from a-Se doped with a Group IIA element have demonstrated excellent n-like transport (electron ranges being much longer than the hole ranges). Differential scanning calorimetry (DSC) studies of the thermal stability of doped glasses and films have shown that these additives encourage crystallization.
Keywords: Doped a-Se, electrical properties, thermal properties, glass transition
1. Introduction Amorphous Se (a-Se) has lately received much attention due to its use as an x-ray photoconductor in recently commercialized x-ray image detectors.1 The actual photoconductor structure is an analog of a pin-structure.2 In such a structure, the p-layer transports holes but deeply traps electrons, the n-layer has good electron transport but deeply traps holes, and the i-layer transports both types of carriers. Such different transport properties are achieved by
______ *
To whom correspondence should be addressed. S. Kasap, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
279
280
G. BELEV ET AL.
suitably doping stabilized Se (alloyed with 0.2–2% wt As) material with the right dopant in the right amounts so as to ensure the desired type of electronic transport, while maintaining the stability of the photoconductor layer against crystallization. An n-like layer is used next to the positive contact to block the injection of holes or trap holes. The n-layer should have very long electron ranges and very short hole ranges. Such transport properties are typically achieved by doping a-Se with Na or other alkali elements from Group IA,2 which reduces the thermal stability of this layer, i.e. the alkali doped a-Se layer has a strong tendency to crystallize.3 It is useful to search for other possible n-type dopants for a-Se, and also examine their effect on the thermal stability of a-Se. In this work we present the effects of dopants from Group IIA of the periodic table (Mg, Ca and Ba) in parts per million amounts on the electrical and thermal properties of pure amorphous selenium (a-Se). 2. Materials and Methods Appropriate quantities of Mg, Ca and Ba were added to pure (99.999%) Se to obtain a-Se doped with Mg, Ca and Ba in amounts 300, 20 and 40 wt ppm respectively. Mg and Ca were added in their pure form, while Ba was added as BaSe. The details about the synthesis and quenching procedures are given in.4 The films were prepared by conventional vacuum thermal evaporation and their electronic transport properties were studied by time-of-flight conventional and interrupted field transient photoconductivity techniques.5 Differential scanning calorimeters DSC 2910 and DSC Q-100, TA Instruments, were used to evaluate the thermal properties of the bulk glasses and films produced from the latter. Both heating and cooling scans were used with a rate of 10 K/min. The average sample size was 20 mg. 3. Results and Discussion The effects of Mg, Ca and Ba on electronic transport in pure a-Se are summarized in Table 1. The data presented in the table clearly demonstrate that, as a result of Group IIA doping, the transport in the a-Se film is converted from p-like (PhWh >> PeWe) behavior in the starting pure Se to n-like (PhWh << PeWe) behavior in the doped films. Consequently Mg, Ca and Ba can be used as efficient n-type dopants for a-Se. The exact explanation of the observed effects on the charge transport is unknown but is probably related to the possible formation of nanocrystalline inclusions that act as deep hole traps or the modification of valence alternation pairs by the dopants.
THERMAL AND ELECTRONIC PROPERTIES OF GROUP IIA DOPED A-SE
281
TABLE 1. Charge transport changes in a-Se as a result of doping with elements from Group IIA. The added amounts of Mg, Ca and Ba are 300, 20 and 40 ppm by weight respectively.
Pure Se
Holes Mobility, Lifetime, Ph [cm2 V-1 s-1] Wh [Ps] 0.14 >150
Se:Mg Se:Ca Se:Ba
TL TL TL
Material
<0.5 ~4–5 <1
Electrons
Ph W h
Mobility,
[cm2 V-1] 2.1 × 10-5
Pe [cm2 V-1 s-1] We [Ps]
[cm2 V-1]
TL
10
TL
TL TL TL
0.0033 0.004 0.0035
>800 >350 >250
>2.6 × 10-6 >1.3 × 10-6 >0.8 × 10-6
Lifetime,
Pe W e
TL stands for “trap limited” transport meaning that the TOF waveform represents only a decay at all experimentally accessible electric fields and time scales.
For pure a-Se we do not observe a crystallization peak during the cooling scan (Figure 1). In addition, the thermograms for the first and the second heating scan of that material are very similar in the sense that we can clearly distinguish glass transition, crystallization and melting processes during both scans. These experiments demonstrate that cooling of the liquid pure Se at a rate of 10 K/min essentially results in a material that is glassy. The behavior of the doped samples is totally different. For all doped glasses and vacuum deposited films from the latter, the first heating scans have produced very similar thermograms to the one recorded for pure a-Se. Such thermograms show that both quenching of the molten material after the synthesis from 600°C to 20°C by immersing the ampoule in tap water and the condensation
Figure 1. DSC measurements of Group IIA doped Se glasses. The thermograms of pure Se, Se:Mg, and Se:Ca samples are artificially shifted in the vertical direction by 4.5, 3.0 and 1.5 W/g respectively. The data clearly demonstrate that all doped samples fully crystallize during the cooling scan, while the pure Se glass sample remains essentially as glass (amorphous).
282
G. BELEV ET AL.
of the vapour on the substrate (60°C) during the vacuum deposi-tion process of the films results in the amorphous state for all doped materials. However, during the cooling scan, all doped materials show clearly pronounced crystallization peaks. The only observable phase transition during the second heating scan is the melting endotherm, which shows that the doped glasses have fully crystallized during the cooling scan; the critical cooling rate for the doped materials is therefore well above 10 K/min. Thus, the addition of Group IIA dopants to a-Se results in a significant reduction in the thermal stability of a-Se against crystallization. 4. Conclusions We examined the addition of Group IIA elements, Mg, Ca and Ba to pure a-Se. Such dopants have shown to produce a-Se material with excellent n-layer properties for potential use in a-Se x-ray photoconductor applications as a blocking layer next to the positive bias electrode. We have evaluated the thermal stability of these doped glasses and of films deposited from the latter using DSC experiments. Our results show that doping with Group IIA elements have a tendency to encourage crystallization in a-Se even in high speed cooling scans (10 K/min). Thus, the production of practical n-like a-Se layers based on Group IIA doping will require further work towards improving their long term stability. ACKNOWLEDGMENTS
We thank NSERC for financial support.
References 1. 2. 3. 4.
Kasap, S., and Rowlands, J. (2002) Proc. IEEE 90, 591–604 and references therein. Polischuk, B., Shukri, Z., Legros, A., and Rougeot, H. (1998) Proc. SPIE 3336, 494–504. Abkowitz, M., Jansen, F., and Melnyk, A. (1985) Philos. Mag. B51, 405. Belev, G., Tonchev, D., Fogal, B., Allen, C., and Kasap, S. (2007) J. Phys. Chem. Solids 68, 972–977. 5. Kasap, S., Polischuk, B., and Dodds, D. (1990) Rev. Sci. Instrum. 61, 2080.
NANOSCALE MATERIALS FOR HYDROGEN AND FUEL CELL SYSTEMS M. SUHA YAZICI* UNIDO - International Centre for Hydrogen Energy Technologies, Sabri Ulker Sk. 38/4 34015 Istanbul, TURKEY
Abstract – The emphasis put on future “Hydrogen Economy” requires technologies further develop around “production”, “storage” and “utilization” for durability and cost reduction. Extended life time, performance and higher efficiencies are reported for fuel cell technologies utilizing nanostructures (tubes, fibers, and particles). New nanocomposites, better catalyst utilizations, nanoscale engineering of interfaces are some of the examples of success in energy applications. This wide spread popularity will result advancement against issues preventing commercialization of hydrogen technologies and fuel cell systems.
Keywords: Energy, hydrogen, fuel cells, nanomaterials.
1. Introduction Nanostructures have been the focus of interest for many electrochemical applications due to their small size and good electrochemical properties. Of particular interest, hydrogen storage and fuel cell applications shown to be early adapters for practical use of nanotechnology. In hydrogen economy, hydrogen, together with electricity, will be produced from clean renewable energy sources and will be used instead of fossil fuels to satisfy all the energy needs. Changing of the entire energy system is a big endeavor on global scale, which has already started. Today, hydrogen based energy systems cost significantly higher than conventional systems. However, socioeconomic savings are higher as well. Therefore, efforts are underway to
______ *
To whom correspondence should be addressed: S. Yazici, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
283
284
M.S. YAZICI
utilize nano-approaches for efficient hydrogen and fuel cell technology implementation. Figure 1 shows relative scale of different nanostructures used in hydrogen storage and fuel cell material and component development. Constructions of structural materials, hydrogen storage media, catalyst and catalyst supports are some applications which will be covered below.
Figure 1. Comparison of scale for different morphologies.4
2. Discussions 2.1. HYDROGEN STORAGE
Carbon-based nanoporous materials such as activated carbons (AC), singlewalled carbon nanotubes (SWNTs) and metal–organic frameworks (MOFs) have been widely investigated as hydrogen adsorption and storage materials.1 The reported values of the hydrogen storage capacity of nanofibers ranged from less than 1 wt % to several tens of weight percentages at moderate pressures and temperatures.2 Most of the results indicate no practical value of pursuing nanofibers for hydrogen storage applications. US Department of Energy (DOE) guidelines for 500 km driving of a fuel cell vehicle requires target storage of 6 wt % for 2010 and 9 wt % for 2015.3 The narrow pore size distribution of single-walled carbon nanotubes (SWNTs) makes them attractive candidates as adsorbents for hydrogen. Early experiments on SWNTs reported adsorbed amounts between 0 and 10 wt %. There is still controversy over percentage storage capacities reported in the literature due to contradicting results at different laboratories. Nevertheless, when value of 9 wt % is reached, application of nanotechnology for storage application will be commercially viable and become enabling technology.
NANOSCALE MATERIALS FOR HYDROGEN AND FUEL CELL SYSTEMS 285 2.2. FUEL CELLS
Greenhousegas Emissions / g CO2 / km
Fuel cells with their numerous possible applications may actually be the main driver for hydrogen economy offering cleaner, more-efficient alternatives to gasoline and other fossil fuels (Figure 2). Hydrogen powered fuel cell vehicles provide an emission-free mobility when the hydrogen production pathway is based on renewable sources.
200
Internal Combustion Engine
175 150
Hybrid Hybrid
125
Gasoline
Diesel
100 75
Fuel Cell Hybrid (Hydrogen from onsite natural gas steam reforming)
50 25 0
Fuel Cell Hybrid (Hydrogen from central Electrolysis powered by renewable energy sources)
140 150 160 170 180 190 200 210 220 230 240
Energy Consumption MJ/100km
Figure 2. Greenhouse gas emissions from different transportation engines.9
A fuel cell is an electrochemical system that converts chemical energy of fuel (typically hydrogen) directly into electricity. A fuel cell is like a battery but with constant fuel and oxidant supply. The fuel cells have been used in the space program, but the interest in terrestrial applications emerged in the 1990s. Fuel cells powered by pure hydrogen emit no harmful pollutants and significantly more energy efficient than combustion-based power generation technologies. A conventional combustion-based power plant typically generates electricity at efficiencies of 33–35%, while fuel cell plants can generate electricity at efficiencies of up to 60%. When fuel cells are used to generate electricity and heat (cogeneration), they can reach efficiencies of up to 85%. The fuel cells are still too expensive for most applications (several thousand U.S. dollars per kilowatt). This is due to the amount and kind of materials, manufacturing processes and manufacturing volumes. Cost must be reduced while increasing efficiency for transportation and stationary applications. For PEM systems designed for distributed generation this means that by 2010 units must cost no more than US $45/kW (to be reduced to US$30/kW by 2015) and exhibit a 40% electrical efficiency and 40,000 h durability.
286
M.S. YAZICI
Nanostructures are effectively used to reach above goals by targeting different functional materials used in fuel cell operation. Nanostructures are balancing requirement for cost effectiveness for fuel cell commercialization with durability and performance benefits. 2.3. RESEARCH PERSPECTIVE
Electrochemical activity is the driving force behind search for new ways to boost performance. There is also relationship between catalytic activity and surface area. Nanostructures of platinum have been used as the most active catalyst material for several decays in many electrochemical applications. Now, fuel cell systems are at the cross-road for commercialization if cost reduction is realized by effectively reducing amount of platinum used on electrodes without jeopardizing performance. Currently, Vulcan XC-72R is the most commonly used catalyst support. However, this carbon black structure may prevent threephase contact when catalyst particles sink into carbon pores smaller than 2 nm. Nanotubes and fibers seem to be the best option to increase reaction area and utilization through accessible surface and high aspect ratio. This fact is schematized in Figure 3.4
Figure 3. Schematics of carbon black and nanofiber with catalyst on.4
There are different methods to functionalize nanotubes with catalyst for fuel cell functionality with controlled dispersion, adhesion and conductivity parameters. In addition to electrochemical methods, electroless plating, microwave activation and colloid deposition are some to mention. Real value of each approach depends on fuel cell current densities obtained at high cell voltages. Electroless plating prove to be the most efficient method to uniformly distribute catalyst. Electrophoretic approach is useful when template synthesis or semiconductive structures are used for deposition. Deposition of phosphors, zeolites, semiconductors, carbon nanotube is some examples can be found in the literature.
NANOSCALE MATERIALS FOR HYDROGEN AND FUEL CELL SYSTEMS 287
There are also efforts to displace platinum with low cost options utilizing advances in nanosynthesis5 (Figure 4). These efforts focus around putting different metal nanostructures, as cost effective replacement to displace platinum, on nanocarbon support. These metals are selected from oxidation and corrosion resistant structures to prevent leaching and degradation during fuel cell operation.
Figure 4. Metals (B, Si, Ti, Ta, Mo, Zr, W, Hf) synthesized as carbides with nano-carbon structure for fuel cell catalyst applications.5
Schmoeckel et al.6 used a catalyst support system where a nanostructure film consisting of single layer of a dense oriented crystalline whiskers with number densities around 5 billion per square centimeter with aspect ratios on the order of 10 (Figure 5). They have shown orders of magnitude of more stability with this structure compare to Pt/C support. Considering current state-of-the-art lifetime being limited around 5,000 h, this effort is a significant step towards realizing commercial lifetime requirements.
Figure 5. Nanowhiskers with nanocatalysts for fuel cell electrode.6
288
M.S. YAZICI
Carbon nano-materials are also applied to the gas diffusion layers (GDL) of fuel cells. Main functions of gas diffusion layers are distribution of reactants to the active site of electrode, management of water supplied and/or generated and enhancement of electrical contact between the electrode and the bipolar plates. In some research,7,8 nano-fibers and nano-tubes are adopted to construct thinner layers for higher gas permeability and good electric conductivity. Figure 6 shows comparison of SWNT based GDL to conventional materials in performance.
Mass Specific Current Density (A / mg)
Methanol anode Fuel Cell Data,Normalized for Catalyst Content 0.4 Nanofiber Paper
0.35
SWNT on Toray paper Toray paper
0.3
Data taken in triplicate
0.25 0.2 0.15 0.1 0.05 0
0
0.2
0.4
0.6
0.8
1
Potential (V vs.DHE)
Figure 6. Fuel cell performance data with various GDL structures including SWNT on carbon paper.4
Perforated expanded graphite sheets with varying TPI (tips per square inch, defining the density and dimensions of perforations) are ideal substrates to produce a carbon nanofiber or nanotube reinforced composite for GDL functionality.10 Figure 7 shows a magnified view of a graphite sheet with nearly 400 openings per square centimeter. The result is a material with greater than 80% “open area” on one surface and only 3–30% open area on the other. The above structures are reinforced and functionalized by putting nanofibers for three-dimensional diffusion and mass transfer. This structures makes in-plane and through-plane diffusion possible for incoming reactants and outgoing product water.
NANOSCALE MATERIALS FOR HYDROGEN AND FUEL CELL SYSTEMS 289
Figure 7. Expanded graphite sheet with 400 perforations/cm2.
There are also engineering aspect of membrane development with nanostructures embedded, but those efforts are still in lab stage for future functionality improvements. 3. Conclusions The transition to a “Hydrogen Economy” has already begun and the world is already moving toward acceptance of hydrogen as a viable alternative carrier of energy. Nanostructures are expected to help realizing hydrogen economy through performance and cost improvements. For hydrogen to become a viable fuel, technologies to convert hydrogen into useful energy must be further improved to increase performance and reduce cost. Continued R&D is needed for the production and storage of hydrogen from renewable sources so that hydrogen from non-carbon emitting sources such as solar and wind energy, will become cost-competitive. Nanotechnology practically offers improvements on material utilization and reliability in fuel cell applications.
References 1. 2. 3. 4.
M. Becher et al., C. R. Physique 4 (2003). P. Benard and R. Chahine, Scripta Materialia, 56 (2007) 803 R. Strobel et al., Journal of Power Sources 159 (2006) 781–801 D. Firsich, 2nd MEA manufacturing symposium, Dayton, OH (2006)
290 5. 6. 7. 8. 9. 10.
M.S. YAZICI
S. McKenzie, 2nd MEA manufacturing symposium, Dayton, OH (August 24, 2006), A.K. Schmoeckel et al., Fuel Cell Seminar Proceeding, pp. 293–296 (2006). A.M. Kannan and L. Munukutla, GDL, Journal of Power Sources, 167 (2007) 330 S. Park et al., GDL, Journal of Power Sources, 163 (2006) 113 Tran, 2nd MEA manufacturing symposium, Dayton, OH (August 24, 2006) M.S. Yazici and D. Krassowski, Fuel Cell Seminar Proceeding: Fuel Cell: Progress, Challenges and Markets, pp. 117–120 (2005)
APPLICATION OF Fe-NANOSCALE MATERIALS USEFUL IN THE REMOVAL OF ARSENIC FROM WATERS M. VACLAVIKOVA*1, G. GALLIOS2, K. STEFUSOVA1, S. JAKABSKY1, AND S. HREDZAK1 1 Institute of Geotechnics, Slovak Academy of Sciences, Watsonova 45, SK-043 53 Kosice, SLOVAKIA 2 Department of Chemical Technology & Industrial Chemistry, School of Chemistry, Aristotle University of Thessaloniki (Box 116), GR-54124, Thessaloniki, GREECE
Abstract – Two iron minerals (akaganeite and magnetite) were synthesized according to standard procedures and nanosized particles were obtained. The minerals were tested for their ability to remove As from water streams at pH value 3.5, with and without electrolyte (0.1 M conc.). Their capacities were calculated to 50 and 30 mg As/g of solid for akaganeite and magnetite, respectively. The sorption isotherms were modeled with Freundlich and Langmuir equations and good agreement was observed. Zeta potential studies have shown that both minerals sorbed As specifically, probably due to chemical forces between As oxyanions and surface Fe.
Keywords: Arsenic, akaganeite, magnetite, nanoparticles, water treatment, sorption, modeling, thermodynamic diagrams
1. Introduction Nanotechnology is the science of the extremely small size materials. It is developing fast with a large impact on the ways we produce materials. The very small sizes and the respectively very large surfaces per unit mass of the nanoscale materials, affect significantly all processes that depend on surface
______ *
Correspondence should be addressed to: M. Vaclavikova, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
291
292
M. VACLAVIKOVA ET AL.
properties. Reaction kinetics are also greatly enhanced. The use of nanosize instead of larger particles in sorption processes multiplies their efficiency to selectively remove a wide range of compounds from contaminated water streams. Arsenic is a highly toxic material and is considered as a priority pollutant from US-EPA and World Health Organization. The hazardous health effects of As were known worldwide with the news coverage of the case of As poisoning in Bangladesh and India. For this reason, the maximum contaminant level (MCL) of As in drinking water, was decreased recently from 50 to 10 ȝg/L both in US and EU. However, there are still many community systems that contain amounts of As above the MCL. Several physico-chemical and biological methods have been proposed for the removal of As. Sorption processes have the ability to remove trace amounts of As from water streams and have receive considerable attraction worldwide. Many different sorbents including bacteria have been tested.1 The more efficient were the iron based ones.2 An overview of the available techniques for As removal from water streams has been recently given by Vaclavikova et al.3 This contribution discusses the preparation and characterization of nanosized akaganeite and magnetite and their application in the removal of As from water streams. 2. Experimental 2.1. MATERIAL AND METHODS
All chemicals used were on analytical grade. Deionized water was used in all experiments. Akaganeite has been prepared by hydrolysis of partially neutrallized Fe(III) chloride solution adding of 1M NaOH solution to 1M FeCl (OH/Fe = 0.75). Akaganeite cannot be prepared at pHs above 5, because the OH- ion is far more competitive than the chloride ion for structural sites. The chloride content of akaganeite varied between 1% to 7%, which was reduced by dialysis.4 Magnetite particles were synthesized by co-precipitation of Fe (III) and Fe (II) in the presence of NH4OH. A solution of mixture of Fe (II) and Fe (III) ions in molar ratio 1:2 was prepared from FeCl3 and FeSO4. An equal volume of 1M aqueous ammonia solution was then added to the iron mixture. The suspension was finally decantated using deionized water and centrifuged to remove the solid material.5 The clear samples were then collected for further analysis. The specific surface area was determined by the low temperature nitrogen adsorption method using a GEMINI 2360 apparatus. Powder X–ray diffraction were performed on Philips X’Pert Pro X–ray diffractometer (CuKD radiation). Zeta potential measurements were conducted on a microelectrophoretic zeta potential analyzer (Apparatus Mark II, Rank Brothers). Twenty-five milliliters of
Fe-NANOSCALE MATERIALS
293
solutions containing 100 mg L–1 of arsenic with 2 g L–1 of akaganeite or magnetite were shaken for 24 h, over a range of pH values 2–9 and then the zeta potential of the particles was measured. Finally, the sorption properties of akaganeite and magnetite nanoparticles towards arsenic have been investigated. Batch-type experiments were performed in solutions with initial concentrations of arsenic in the range of 20–200 mg L–1 and sorbent dosage of 2 g L–1. All sorption experiments were carried out at ambient temperature (22 + 1°C) in an orbital shaker at pH 3.5. The effect of ionic strength on As removal was examined using 0.1 M NaNO3 as electrolyte. The experimental data were modeled with Langmuir and Freundlich type sorption isotherms. The corresponding data and model isotherms are given in Figures 2 and 3. The sorption parameters are given in Tables 1 and 2. 2.2. RESULTS AND DISCUSSIONS
2.2.1. Material study Akaganeite produced by the above method consisted of narrow rod crystals with size 20–50 nm in length and surface area 146.9 m2 g–1. XRD analysis (not shown here) confirmed that the major phase was akaganeite. A minor phase of ferrihydride has also been recognized, as the result of precipitation of the remaining Fe (III) in the supernatant due to the rise of pH during the dialysis.4 The magnetite produced by the co-precipitation method, consisted of particles 10–40 nm in length and a BET surface area of 84.01 m2 g–1. XRD analysis confirmed the magnetite structure. SEM images of the material (not shown here) have shown the presence of cluster aggregates. The magnetite particles have been agglomerated during the air-drying to grains of size 20–900 nm.5 2.2.2. Thermodynamic study The solution chemistry of As (V) is quite complex and depends on the presence of other substances (i.e. cations, anions, solids) and the redox conditions that prevail. If a complete analysis of the solution is known then a thermodynamic equilibrium diagram can be constructed, which is very helpful for explaining sorption behavior. For our model, aqueous system that consisted only of As (V) in a system open to atmosphere, the thermodynamic equilibrium diagram (see Figure 1) has been constructed with the aid of the computer program Mineql+. It is observed that, up to pH 1, the neutral H3AsO4 species predominate. As the pH increases, deprotonation occurs and more negatively charged species are produced. At pH 3.5 to 5.5, the single negatively charged H 2 AsO 4
294
M. VACLAVIKOVA ET AL. Arsenic 100 ppm, pO2=0.68 (atmospheric) 100
Fraction [%]
80
H3AsO4
60
H2AsO4 HAsO4
40
AsO4
-
2-
3-
20
0 0
2
4
6
8
10
12
14
pH
Figure 1. Arsenic distribution in aqueous system.
predominate. As the pH increase continues, its concentration decreases and the double charged HAsO 42 predominates at pH values 7 to 11. Finally at pH above 12, the AsO 43 predominates. So, according to pH the relative ratio of charged species changes and affects sorption behaviour. 2.2.3. Sorption study The sorption isotherms of arsenic uptake on akaganeite (2 g of sorbent/liter of solution) at pH 3.5 with 0.1M NaNO3 and without electrolyte are given in Figure 2. The points represent the experimental sorption data and the lines represent the model fits of the Freundlich equation. We observed a good agreement to the model with coefficients of determination (R2) 0.99 and 0.97, respectively. The adsorption parameters are given in Table 1. The sorption capacity of akaganeite towards arsenic at pH 3.5 is around 50 mg/g, which is considered good and almost double the average value presented in recent literature.3 A small positive effect of ionic strength on sorption efficiency is also observed. TABLE 1. Sorption parameters – akaganeite. Freundlich parameters
Ionic strength Deionized water 0.1 M NaNO3
K 27.42
b 0.10
R2 0.99
28.63
0.12
0.97
Fe-NANOSCALE MATERIALS
295
The presence of electrolyte did not affect the final capacity but it slightly changed the slope of the curve. Fitting the experimental data to Freundlich and Langmuir equations has shown that the two systems behave differently. The results are presented in Table 2. The data without electrolyte were fitted better with the Langmuir model. The coefficient of determination (R2) was 0.97, showing a good agreement of the model with the experimental data. In the case of experiments with 0.1 M NaNO3 electrolyte, the data were fitted better by a Freundlich type equation (R2 = 0.96). Comparing the data in Figures 2 and 3 it is observed that the sorption capacities of the two iron minerals are different. Even though, akaganeite has a larger capacity, the sorption results of magnetite are considered more promising from the practical point of view. In the removal of soluble pollutants from water streams by sorption methods the solid/liquid (S/L) separation step is very important for an efficient operation. Nanosized materials are very small to be used in sorption columns and their small size makes S/L separation very difficult and expensive as the common methods can not be applied. Magnetite possesses good magnetic properties and the nanosized magnetite particles can be easily separated from water with a magnetic separator. TABLE 2. Sorption parameters – magnetite. Ionic strength
Langmuir parameters –1
Qmax [mg g ] 31.27
Deionized water
R2 0.97
K 0.15 Freundlich parameters
0.1 M NaNO3
K 15.91
R2 0.96
b 0.14
55 50
Akaganeite
45 40
Uptake (mg/g)
35 30 25 20
pH 3.5 deionized water ... Freundlich 0.1 M NaNO3 ....... Freundlich
15 10 5 0
20
40
60
80
100
120
Ceq(mg/L)
Figure 2. Sorption of As (V) on akaganeite.
296
M. VACLAVIKOVA ET AL.
30
Magnetite
25
Uptake (mg/g)
20
15
10
pH 3.5 deionized water ... Langmuir 0.1 M NaNO3 ....... Freundlich
5
0 0
20
40
60
80
100
120
Ceq(mg/L) Figure 3. As sorption by magnetite.
The surface charge of the mineral plays an important role in sorption processes, especially when charged species are considered. As discussed previously, As (V) has different negative surface charge according to pH. A study of the zeta potential would help to explain the sorption behavior of the respective material. For this reason, the zeta potential of the solids as a function of pH was measured in the presence and absence of arsenic. It was found that the point of zero charge (PZC) of akaganeite with water only was at pH 7.8. With 100 mg L–1 As in the system, the zeta potential curves changed dramatically and the PZC shifted to pH 3.5, showing clearly a specific sorption of As on akaganeite. For magnetite, the PZC was 6.5 with water only. With 100 mg L–1 As the whole curve was shifted to negative values and no PZC was observed in the pH range studies (2.0–12.0). So, the mechanism of sorption is different for the two iron based sorbents. However, both show specific sorption of As, as it is clearly shown by the zeta potential results. 3. Conclusions The two iron minerals synthesized and studied (akaganeite and magnetite) were proved to be efficient sorbents for As removal from water solutions. Maximum capacities observed were around 30 and 50 mg As/g of solid for magnetite and
Fe-NANOSCALE MATERIALS
297
akaganeite respectively, which are considered good and are over the average capacity reported in the recent literature. Magnetite nanoparticles with their magnetic properties seem to be a promising material for economic and effective removal of arsenic from water streams. The sorption of As on both iron minerals was of specific nature (probably chemical bonding of As on surface Fe), as was clearly shown by the zeta potential study. ACKNOWLEDGEMENTS
This work got financial support from project of Science and Technology Assistance Agency contract No. APVT-51-017104 and NATO Collaborative Linkage Grant EST.EAP.CLG 981103.
References 1. Katsoyiannis, I., Zouboulis, A.I. (2006), Comparable evaluation of conventional and alternative methods for the removal of arsenic from contaminated groundwaters, Rev. Environ. Health, 21: 25–43. 2. Deliyanni, E.A., Peleka, E.N., Gallios, G.P., Matis K.A. (2008), A critical review of the separation of arsenic oxyanions from dilute aqueous solution (the contribution of LGICT), Int. J. Environ. Pollut. 3. Vaclavikova, M., Gallios, G.P., Hredzak, S., Jakabsky, S. (2007), Removal of arsenic from water streams: an overview of available techniques, Clean Technol. Environ. Policy, DOI 10.1007/s10098-007-0098-3. 4. Schwertmann, U., Cornell, R.M. (2000), Iron Oxides in the Laboratory: Preparation and Characterization, 2nd edn., Willey-VCH, Weinheim, pp. 113–140. 5. Vaclavikova, M., Jakabsky, S., Hredzak, S. (2004), Magnetic nanoscale particles as the sorbents for removal of heavy metal ions, in: Nanoengineered Nanofibrous Materials, NATO Science Series II. Mathematics, Physics and Chemistry, Vol 169, Guceri, S., Gogotsi, Y.G, Kuznetsov, V. eds., Kluwer, Dordrecht, pp. 481–486.
NANOPATTERNING USING THE BIOFORCE NANOENABLER K. ARSHAK*, O. KOROSTYNSKA, AND C. CUNNIFFE Electronics and Computer Engineering Department, University of Limerick, Limerick, IRELAND
Abstract – This paper discusses the opportunities, offered by novel nanopatterning facilities, namely, BioForce NanoeNablerTM (NeN), in the area of sensors development, with the focus on microsensor arrays for biological, environmental and medical fields. The NeN can deliver attolitre to picolitre volumes of liquid, such as small molecules, biomolecules, including proteins and nucleic acids, nanoparticles, reactive solutions and so forth, with a high degree of spatial accuracy. It is envisaged that the reduction in the sensors size would result in their new advanced functionalities.
Keywords: BioForce NanoeNablerTM, sensors arrays, nanopatterning; microdevices, point-of-care diagnostics
1. Introduction Nanostructuring surfaces in order to improve the quality of determinations, in terms of detection limit and signal-to-noise ratio, had received a great attention in the last years.1–3 The possibility to routinely pattern surfaces on the micronand submicron spatial scales with biomaterials opens the door for a development of a vast spectrum of ultraminiaturized bioanalytical tests and devices, with special focus on point-of-care diagnostics. For example, the work is ongoing towards creation of a diagnostic biochip that uses just a few cells of blood for critical biomedical analysis.4 The pH sensor has many uses in chemistry, biology, environmental monitoring, especially water quality control and so forth. Advances in semiconductor sensor technology, medical diagnostics and health care needs boosted a rapid research into miniaturized pH sensors, which can be used as well for in-vivo
______ *
To whom correspondence should be addressed: K. Arshak, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
299
300
K. ARSHAK, O. KOROSTYNSKA, AND C. CUNNIFFE
diagnostics. This paper reports on a novel nanopatterning technology offered by BioForce NanoeNablerTM, which is tested for developing various sensors, including pH and bacteria, for applications in the biological, environmental and medical fields. The NanoeNablerTM can deliver attoliter to picoliter volumes of liquid with a high degree of spatial accuracy. The operation of the sensing elements is based on the properties of polymers, which exhibit a change in their electrical characteristics (such as conductivity, potential or capacitance), on exposure to solutions with different concentrations of pH value.5 A number of different polymers are being used to form the sensor arrays (fabricated using novel technology offered by BioForce NanoeNablerTM), with each array element having unique selectivity and sensitivity properties. The strategy of simultaneous measurement of a number of sensor arrays relies on the application of pattern recognition techniques, similar to the one in e-nose systems. 2. Applications Numerous applications benefit from the micro/nano patterning using BioForce NanoeNablerTM, some examples are listed below: x
The first category of the applications deals with the development of molecular detection devices, usually to detect protein biomarkers, nucleic acids or pathogens. They often have some kind of sensor, which needs to be functionalized by adding a small affinity capture domain, such as antibody or single stranded DNA. These sensors are sometimes couples with micro-fluidics for sample handling to create lab-on-chip devices.
x
The second category of applications is oriented more towards studying single cells and creating bioassays for extremely small volumes. These applications can provide insight for the researchers working in drug discovery and diagnostics. The most popular application from this group includes patterning molecules onto the surfaces that can subsequently be used to study cellular interactions with the patterned features.
x
The third category of NeN applications involves patterning non-biological materials onto surfaces to create physical structures. The authors of this paper are with the Micro Electronics and Semiconductor Research Centre in the University of Limerick, Ireland and are working on a number of projects, mainly oriented towards shrinking previously developed sensors system to the nanoscale level.
BIOFORCE NANOENABLER
301
For example, a nanotechnology-based sensors system is being developed to detect and monitor the quality of food. Attention is focused on one type of bacteria group, namely the Bacillus cereus group, which is commonly found in liquid, milk powder and mixed food products. An array of sensors is used to achieve more accurate reading results. Material properties of various formulations are the key parameter to allow for fast sampling rates, high discrimination and fast recovery times and so forth. Devices exhibit a dynamic resistive or capacitive response as the mechanism through which bacteria are identified. The prototype of this system is illustrated in Figure 1.
Figure 1. A prototype of the nanotechnology-based sensors system for the quality of food monitoring.
There are a number of parameters that have to be considered when using this novel nanopatterning technology. Following substrates were tested: BioForce SindexTM chip (Si coated with Au), alumina, glass and transparent polymer substrate. The materials were back loaded onto a Surface Patterning Tool (SPT) via hand pipetting. Material-substrate interaction is vital and patterning surface should be hydrophilic. This was achieved by UV-ozone surface treatment. Additionally, polymer/solvent composition should be adjusted to allow uniform patterning. Figure 2 presents optical image of 4 × 4 array consisting of 1–2.5 Pm spots distanced by 5 Pm. These are printed on a BioForce SindexTM chip; individual silicon pads measure 100 × 100 Pm. There is a laser-based force feedback system to ensure reproducible printing. Figure 3 depicts an image of “University of Limerick” logo, printed in an identical manner, with 1–2.5 Pm spots distanced at 5 Pm. Figure 4a and b illustrate the dots and line, respectively, patterned using pH sensitive polymer solution that forms a basis for miniaturized sensors array system. Printed dots measure ~2.32 Pm in diameter.
302
K. ARSHAK, O. KOROSTYNSKA, AND C. CUNNIFFE
The BioForce NanoeNablerTM system uses a liquid dispensing process via specially designed surface patterning tool (SPT), which is microfabricated cantilever with an integrated passive microfluidic system.6,7 Fluid loaded into the reservoir flows down the microchannel by capillary flow until it reaches the gap at the end of the SPT. During the deposition process, which typically takes less than 100 ms, SPT end touches the surface and a volume of fluid is instantly transferred.
Figure 2. Optical image of 4 × 4 array, consisting of 1–2.5 Pm spots distanced by 5 Pm. These are printed on a BioForce SindexTM chip; individual silicon pads measure 100 × 100 Pm.
Figure 3. “University of Limerick” logo, consisting of 1–2.5 Pm spots that are distanced at 5 Pm.
BIOFORCE NANOENABLER
a)
303
b)
Figure 4. (a) Dots of pH sensitive polymer based solution that forms a basis for miniaturized sensors array system. These dots measure ~2.32 Pm in diameter and are printed on transparent flexible substrate. (b) Demonstrated ~52 Pm long line of polymer material, drawn by the cantilever with 1 Pm opening. When printing transparent materials on transparent substrates, the visualization of the printing is impaired and dyes could be used to enhance process visibility.
We use an array of sensors to achieve more accurate reading results, where sensors would differ in dimensions and/or material composition. The optimal number of sensors and their parameters will be determined, based on material properties of various formulations. Suitable sensor materials will be formulated for printing as the active film layers, using nano-size precursors. The response of various materials will be optimised by designing and selecting the most effective topologies that allow for fast sampling rates, high discrimination and the best cost-effectiveness. ACKNOWLEDGEMENTS
This work was supported by the Irish Research Council for Science, Engineering and Technology (IRCSET): funded by the National Development Plan. The authors would like to thank BioForce Nanosciences Incorporated, Ames, IA 50011 USA, for their support and provided information.
References 1. Lieberzeita, P., Gazda-Miareckaa, S., Halikiasa, K., Schirka, C., Kaulingb, J., and Dickert, F. (2005), Sens. Act. B-Chem. 111–112, 259–263. 2. Lieberzeit, P., Schirk, C., Glanznig, G., Gazda-Miarecka, S., Bindeus, R., Nannen, H., Kauling, J., and Dickert, F. (2004), Superlattice Microstruct. 36(1–3), 133–142. 3. Kurowski, A., Schultze, J., Lüth, H., and Schöning, M. (2002), Sens. Act. B-Chem. 83(1–3), 123–128. 4. www.bioforcenano.com
304
K. ARSHAK, O. KOROSTYNSKA, AND C. CUNNIFFE
5. Arshak, A., Gill, E., Arshak, K., Korostynska, O., and Cunniffe, C. (2007), Proc. 30th IEEE ISSE Conference, Cluj-Napoca, Romania, 9–12 May 2007, 213–218. 6. Vengasandra, S., Lynch, M., Xu, J., and Henderson, E. (2005), Nanotechnology 16, 2052–2055. 7. Xu, J., Lynch, M., Huff, J., Mosher, C., Vengasandra, S., Ding, G., and Henderson, E. (2004), Biomedi. Microdevices 6(2), 117–123.
ELECTROCATALYSTS AND ELECTRODE DESIGN FOR BIFUNCTIONAL OXYGEN/AIR ELECTRODES V. NIKOLOVA1, P. ILIEV1, K. PETROV1*, T. VITANOV1, E. ZHECHEVA1, R. STOYANOVA1, I. VALOV2, AND D. STOYCHEV2 1 Institute of Electrochemistry and Energy Systems, Bulgarian Academy of Sciences, Acad. G. Bonchev str., bl, Sofia, BULGARIA 2 National Institute of Materials Physics, P.O. Box G7, ROMANIA
Abstract – Electrocatalysts and appropriate electrode designs have been studied with respect to the development of Bifunctional Air/Oxygen Electrode (BAE). Three groups of catalysts have been prepared: (i) CuxCo3-xO4; (ii) thin films of Co-Ni-Te-O and Co-Te-O and (iii) CoxOv/ZrO2 films. Different catalysts deposited on classical and originally designed GDE were compared by their electrochemical performances.
Keywords: Bifunctional air/oxygen electrode, gas diffusion, electrocatalysts.
1. Introduction Unitized Regenerative Fuel Cells are dual mode energy storage systems, combining water splitting electrolyzer for hydrogen and oxygen generation and for their subsequent conversion into electrical energy. A key for such systems is the development of Bifunctional Air/Oxygen Electrode (BAE). Oxygen electrode is known to be a strongly irreversible system with a high activation overvoltage in aqueous solutions.1 The choice of electrocatalysts for oxygen reactions is hence limited to mixed valence oxides of Co, Ni and Mn with a spinel and perovskite crystal structure. Different methods for electrocatalysts preparation, ______ *
To whom correspondence should be addressed: K. Petrov, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
305
306
V. NIKOLOVA ET AL.
viz: low temperature synthesis using inorganic and organic precursors2; vacuum evaporation,3 electrochemical deposition4 have been proposed. Since reaction proceeds at the phase boundary, it is important to expand the surface area in contact with the electrolyte and the O2 molecules, dissolved or gaseous. This study includes, preparation of GDE with different design, preparation and deposition of catalysts as nanopowders or thin films. 2. Experimental 2.1. PREPARATION OF CATALYSTS AND DIFFUSION ELECTRODES
The catalysts were synthesized by different methods. CuxCo3-xO4 powders were prepared by thermal decomposition of mixed nitrate and carbonate precursors. Thin films of Co-Ni-Te-O and Co-Te-O were deposited by vacuum co-evaporation of Co, Ni and TeO2 on pressed hydrophobic carbon black representing gas-supplying layer.5 The films with various thicknesses (100–450 nm) were prepared by co-evaporation of Co, Ni and TeO2 on stationery substrates under vacuum better than 10–4 Pa. TeO2 is the source of oxygen for CoO and NiO synthesis during the co-deposition. CoxOy and ZrO2 films were consequently obtained by electrochemical deposition from an absolute Ethyl alcohol containing 2.3 M LiCl as an electro-conducting additive to which 0.3 M ZrCl4 or CoCl2 were added.6 The films were deposited on stainless steel gauze (SS) and on nickel foam (Ni). The cathodic deposition was performed in a voltastatic regime at 22 V for ZrO2 and 8 V per CoxOy layers. The cell was kept at a constant temperature of 13oC. The deposition time was 30 min, for a thickness – ~2 am. The electrochemical behavior of the powder catalysts were studied on bilayered GDE, as shown in Figure 1a, with S = 10 cm–2, consisting of gassupplying layer of hydrophobic carbon black with a loading of 100 mg/cm2 and an active layer containing 50 mg/cm2 catalyst and 18 mg/cm2 of XC-30. Current collector
A
Catalytic thin film
B
Catalytic layer Gas-supplying layer Figure 1. Schematic view of GDE design: (a) – bilayered gas diffusion electrode; (b) – GDE catalyzed with vacuum and electrochemically deposited catalysts.
ELECTROCATALYSTS FOR BOE
307
The carbon black was hydrophobized with Teflon. The electrodes were prepared by mechanically pressing at 300 kg/cm2 and 3000C. The GDE catalyzed with vacuum deposited catalysts were prepared by direct deposition of the catalyst on gas-supplying layer, thus creating a ready-to-use gas diffusion electrode (Figure 1b). Third design of GDE (similar to Figure 1b) was used for catalysts prepared electrochemically. They were first deposited on Ni foam and than mechanically pressed on gas-supplying layer of hydrophobic carbon black (XC-30) with a loading of 100 mg/cm2. 2.2. CATALYSTS CHARACTERIZATION
The phase composition of the synthesized catalysts, their morphology and surface structure were studied by physical methods for bulk and surface analysis. Scanning electron microscopy (SEM), TEM, XRD and XPS spectroscopy were applied. The specific surface area was measured by low temperature nitrogen adsorption according to BET. The electrochemical characterization of the electrodes was performed by steady-state i-E curves and charge-discharge cycle life tests. The GDE’s were tested in a three-electrode cell at room temperature. The electrolyte was 3.5 M KOH, Ni plate was used as counter electrode. The potential was measured versus Hg/HgO reference electrode in 3.5 M KOH. 3. Result and Discussion 3.1. PHYSICAL CHARACTERIZATION OF THE CATALYTIC FILMS
In Table 1 are summarized the physicochemical characteristics of the catalysts studied. The spinal structure of CuxCo3-xO4 powders synthesized by the nitrate method (0 x 0.9) and by the carbonate one (x 0.5) was confirmed by XRD, with a higher surface area (BET) and a smaller particle size of the last samples. Data are presented only for the samples with the best electrochemical behavior (Table 1). The XPS analysis and TEM study showed that depending on atomic ratio R(Co+Ni)/Te and RCo/Ni the catalyst Co-Ni-Te-O treated at 3000C is amorphous or nanocrystalline with Co2+, Ni2+, Te4+ and Te0 on the surface. From Selected area electron diffraction (SAED) in TEM the as-deposited catalytic films Co-Te-O are generally amorphous with very small nanocrystals of cubic CoO, embedded in the amorphous matrix. The oxidation state of the species is Co2+, Te4+ and Te0. No changes in the chemical state of the films were found after thermal annealing up to 3000C.
V. NIKOLOVA ET AL.
308
TABLE 1. Physicochemical characteristics of chemically synthesized, vacuum- and electrochemically deposited catalysts. Composition
T (oC)
Structure
XPS
(a/Ao)
Cu0,2Co2,8O4 (nitrate) Cu0,3Co2,7O4 (carbonate) CoO-TeO2 (vacuum) Co3O4/ZrO2 electrochem.
S/BET
Particles
(m2 g–1)
size (μm)
350
8,080
–
5–10
0,05–0,1
400
8,080
–
70–80
<0,05
–
nm
–
–
25
Hexagonal CoO nanocrystals, amorphous matrix
25 400
Co2+, Te4+,Te0 Co2+/Co3+ = 2,4 Co2+/Co3+ = 0.34
–
3.2. ELECTROCHEMICAL STEADY-STATE MEASUREMENTS
The electrochemical activity of the different catalysts is shown in Figure 2 which compares Cu0.2Co2.8O4 powder, electrochemically deposited CoxOy/ZrO2 and vacuum deposited Co-Ni-Te-O, Co-Te-O thin films, tested in oxygen. Clearly GDE with Co-Te-O deteriorates when working in air. The electrochemical parameters of the curves for all catalysts are tested at low current density. The activity increases in the order: Co-Te-O d CoxOy/ZrO2 < Cu0.3Co2.7O4 (carbonate) < Cu0.2Co2.8O4 (nitrate), and in OE reaction: Co-Ni-Te-O CoxOy/ZrO2 | Co-Te-O Cu0.3Co2.7O4 (carbonate) Cu0.2Co2.8O4 (nitrate). At low current density we can explain the better performance of powder catalysts for both reactions as due to the high catalyst loading (50 mg/cm2) compared to the vacuum and electrochemically deposited samples (see Table 2). The activity of the vacuum evaporated films is ascribed to nanoparticle effect. 2000 1800 1600
E, mV (RHE)
1400
Cu0.2Co0.8O4 (air) CoxOy/ZrO2 (air) Co-Ni-Te-O (oxygen Co-Te-O (air) Co-Te-O (oxygen)
1200 1000 800 600 400 200 0 0
10
20
i, mA/cm
30
40
50
2
Figure 2. Steady-state current-voltage curves of Cu0.2Co2.8O4 powder, Co-Ni-Te-O, Co-Te-O and CoxOy/ZrO2 gas-diffusion electrodes in air or O2, 3.5 M KOH.
ELECTROCATALYSTS FOR BOE
309
TABLE 2. Electrochemical characteristics of CuxCo3-xO4, Co-Ni-Te-O, Co-Te-O, CoxOy/ZrO2 gas diffusion electrodes at low current density in air or in O2 (Tafel region), E vs. RHE. Composition
Cu0,2Co2,8O4 (air) Cu0,3Co2,7O4 (air) Co-Ni-Te-O (O2) Co-Te-O (air) CoxOy/ZrO2 (air)
Loading (mg cm-2) 50 50 0,07 0,06 0.06
i = 6 mA cm-2 (Ec/mV*) (Ea/mV*) 850 1,240 820 1,410 700 1,550 665 1,505 680 1,520
bc (mV dec-1) ba (mV dec-1)
60 100 66 63 220
150 150 128 47 85
TABLE 3. Electrode potentials of CuxCo3-xO4, Co-Ni-Te-O, Co-Te-O and CoxOy/ZrO2 gasdiffusion electrodes in air or in O2 ( i = 40 mA cm-2) *E vs. RHE. Composition Cu0.2Co2.8O4 (air) (nitrate) Cu0.3Co2.7O4 (air) (carbonate) Co-Ni-Te-O (O2) RCo/Ni = 4.5 CoO-TeO2 (air) RCo/Te = 1.5 CoxOy/ZrO2 (air)
Loading (mg cm-2) 50
Ec (mV)*
Ea (mV)*
770
1,280
50
630
1,430
0.07
500
1,950
0.06
100
1,600
0.06
515
1,680
At higher current density (i = 40 mA cm-2, Table 3) the order of electrode activity in OE reaction is the same as that at low current density, while in OR reaction the order is: Co-Te-O CoxOy/ZrO2 Cu0.3Co2.7O4 Cu0.2Co2.8O4. This order of electrochemical performances most probably is determined by transport limitations of the OR reaction. It seems that the diffusion conditions in the active layers from powder catalysts as well as from electrochemically deposited thin films ensures the same order of performances even though the catalyst loading on these electrodes differs by three orders of magnitude. This is probably due to a maximum use of catalytic material and is a substantial advantage from a practical view-point. At the same conditions the electrode performance of vacuum deposited Co-Te-O most probably deteriorates due to non-optimized three phase boundary. Gas-diffusion electrode, catalyzed with Cu0.3Co2.7O4 was tested in a real metal hydride – air battery: charge and discharge current density was ich = idis = 20 mA/cm2 for the bifunctional GDE.
V. NIKOLOVA ET AL.
310
Time of charge (oxygen evolution) was 16 h and time of discharge (oxygen reduction) was 8 h. The 200 charge-discharge cycles with stable characteristics were achieved (Figure 3). 1600
E, mV (RHE)
1400 1200 1000 800 600 0
20
40
60
80 100 120 140 160 180 200 Number of cycles
Figure 3. Charge- discharge cycles of Cu0.3Co2.7O4-air gas-diffusion electrode; ich = idis = 20 mA cm–2, 25oC, 3.5 M KOH.
4. Conclusion Gas-diffusion electrodes catalyzed with a loading of 50 mg/cm2 Cu0.2Co2.8O4 and Cu0.3Co2.7O4 powders exhibit a good apparent catalytic activity in oxygen reduction and evolution reactions. GDE’s with Cu0.3Co2.7O4 powder exhibits stable IV characteristics after 200 charge-discharge cycles in air gas-diffusion electrodes. The vacuum and electrochemically deposited Co-Ni-Te-O, Co-Te-O and CoxOy/ZrO2 catalytic films exhibit high mass activity and electrochemical stability in both oxygen reduction and evolution reactions, despite minimal catalyst loading of about 0.05–0.07 mg cm–2. These types of electrodes have less catalyst loading which is a substantial advantage from a practical point of view.
References 1. 2. 3. 4. 5. 6.
O’Sullivan, E., and Calvo, E. (1987) In “Electrode Kinetic Reaction”, R.G. Compton, Editor, Elsevier, Amsterdam, 274. Nikolov, I., Darkaoui, R., Zhecheva, E., Stoyanova, R., Dimitrov, N., and Vitanov, T. (1997) J. Electroanal. Chem., 429, 157. Nkeng, P., Koenig, J., Gautier, J., Chartier, P., and Poilerat, G. (1996) J. Electroanal. Chem., 402, 81. Ismail, J., Ahmed, M., and Vishnu Kamath, P. (1991) J. Power Sources, 36, 507. Rashkova, V., Kitova, S., Konstantinov, I., and Vitanov, T. (2002) Electrochim. Acta, 47, 1555. Valov, I., Stoychev, D., and Marinova, T. (2002) Electrochim. Acta, 47.
PREPARATION OF MAGNETIC CHITOSAN NANOPARTICLES FOR DIVERSE BIOMEDICAL APPLICATIONS D. KAVAZ, T. ÇIRAK, E. ÖZTÜRK, C. BAYRAM, AND E.B. DENKBAù* Hacettepe University, Department of Chemistry, Biochemistry Division, Beytepe, Ankara, TURKEY
Abstract – Polymeric nanoparticles with magnetic properties can be potentially used in many fields such as biotechnology, separation processes, optoelectronic, catalysts and/or sensors, medical diagnosis and therapy. In this respect, biopolymers give promising trends due to their excellent biocompatibility and biodegradability. Therefore in this study, magnetic chitosan/Fe3O4 nanoparticles were prepared according to the procedure based on the ionic gelation of chitosan with tripolyphosphate anions. The formation of the particles was a result of the interaction between the negatively charged groups of the tripolyphosphate and the positively charged amino groups of chitosan. The prepared samples were observed by atomic force microscopy to obtain information about the morphology. The mean particle size of the nanoparticles was determined by dynamic light scattering measurements. Nanoparticles were spherical in shape with a particle size range of about 250–400 nm according to obtained data. Magnetic properties of the nanoparticles were determined by using ESR and VSM.
Keywords: Chitosan, magnetic nanoparticles, biotechnology, tripolyphosphate, ESR, VSM
1. Introduction Paramagnetic carriers, ranging from micro sized particles to nano sized colloids have shown to be applied in increasing applications in various fields of biotechnology such as enzyme immobilization, separation for purification, drug targeting, optoelectronic, catalysts, medical diagnosis and therapy.1 Magnetic nanoparticles have several advantages such as particle size, large surface area
______ *
To whom correspondence should be addressed: E. Denkbas, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
313
314
D. KAVAZ ET AL.
which can be properly modified to attach with biological agents. They have magnetic response so that they can be manipulated by an external magnetic field gradient.2 Ensuring biocompatibility and non-toxicity so as to meet the requirements for these applications, iron oxide based particles, e.g. magnetite, are commonly used as the magnetically responsive component of commercially available magnetic microspheres. Magnetic carriers are most commonly manufactured from polymers, since they have a variety of surface functional groups which can be tailored to specific applications. Natural and synthetic polymers (i.e. calcium alginate, polystyrene, polyacrylamide, polyvinyl alcohol, nitrocellulose, polyvinyl butyral, chitosan) have been used in the preparation of magnetic carriers for years.3 Chitosan, deN-acetylated analog of chitin, is a heteropolysaccharide (polycationic, hydrogel). It is non-toxic, hydrophilic, biocompatible, biodegradable, anti-bacterial and remarkable affinity for many biomacromolecules.4 Due to the presence of both hydroxyl and amine groups in its structure, chitosan can be chemically modified to be used as novel separation media. Chitosan has been applied to many fields, such as metal adsorption, enzyme immobilization, protein adsorption and the controlled release of drugs.5–6 In this study, magnetic chitosan nanoparticles were prepared and characterized for application in biotechnology. Morphology, particle size and magnetic properties of the magnetic chitosan nanoparticles were evaluated using selected parameters; molecular weight of the chitosan and the Fe3O4 content, and cross-linker concentration. 2. Materials and Methods 2.1. REAGENTS
Chitosan polymers with different molecular weights (i.e. 150, 450 and 650 kDa; low molecular weight (LMW), medium molecular weight (MMW) and high molecular weight (HMW), respectively) were obtained from Fluka (Switzerland). Aqueous acetic acid (Carlo Erba, Italy) solutions were used as solvent for the chitosan polymers and sodium triphosphate pentabasic was used as the crosslinker. Iron (II, III) oxide, nanopowder, purity >98%, from Sigma Aldrich and all chemicals were of analytical grade and no further purification was required. 2.2. PREPARATøON OF MAGNETøC CHøTOSAN NANOPARTøCLES
Chitosan with the deacetylation degree of 75–85% and the viscosity of 20–200 cP was dissolved at 0.5% (w/v) with 1% (v/v) acetic acid. pH of the chitosan gel raised to 4.6 with 10 N NaOH. Fe3O4 was added in chitosan gel and
MAGNETIC CHITOSAN NANOPARTICLES
315
sonicated for 15 min. Chitosan nanoparticles formed spontaneously upon addition of aqueous tripolyphosphate solution to chitosan solution including Fe3O4 under mechenical stirrer at 2,000 rpm and mixed for further 1 h after addition of tripolyphosphate. (Chitosan to TPP weight ratio is 6:1 and chitosan to TPP volume ratio is 3:1). The particles were then incubated at room temperature for 20 min. Nanoparticles were purified by centrifugation at 9,000 g for 30 min at 5°C. Supernatants were discarded, and the chitosan nanoparticles were extensively rinsed with distilled water to remove any sodium hydroxide. Nanoparticles were resuspended in ultrapure water. 2.3. CHARACTERIZATION OF MAGNETIC CHITOSAN NANOPARTICLES
2.3.1. Morphology The morphological characterization of the magnetic chitosan nanoparticles were carried out with a scanning electron microscope (SEM, Jeol, Japan). A 1.5 ml aqueous suspension of chitosan nanoparticle was dropped onto a sample holder (a stap) and placed in a vacuum oven at room temperature for 24 h to dry. The samples were coated with gold, and then SEM micrographs were obtained. 2.3.2. Nanoparticle size Size of magnetic chitosan nanoparticles were measured with Zeta-Sizer (Malvern 3000). In brief, the dried nanoparticles were suspended in filtered deionized water and sonicated to prevent particle aggregation and to form uniform dispersion of nanoparticles. Effect of chitosan molecular weight, Fe3O4 content and chitosan:TPP weight ratio on particle size were investigated. 2.3.3. Magnetic properties Magnetic properties of the nanoparticles were evaluated using vibrating-sample magnetometer (LDJ 9600 VSM). A certain amount of magnetic chitosan nanoparticles was balanced and placed in magnetometer. The degree of magnetism of nanoparticles was then determined by applying an increasing magnetic field over the microspheres and the results were used to calculate the magnetic quality of the nanoparticles.
D. KAVAZ ET AL.
316
3. Results and Discussion 3.1. MORPHOLOGY OF THE MAGNETIC CHITOSAN NANOPARTICLES
The morphology of the magnetic chitosan nanoparticles was investigated by using scanning electron microscopy (SEM). SEM micrographs of magnetic chitosan nanoparticles were shown in Figure 1. Magnetic chitosan nanoparticles appear separated as illustrated in Figure 1 and have 250–400 nm of average size.
Figure 1. SEM micrograph of magnetic chitosan nanoparticles.
3.2. NANOPARTICLE SIZE
The impact of chitosan molecular weight, Fe3O4 content and chitosan-TPP mass ratio on nanoparticle size were investigated. According to Zeta-Sizer results nanoparticles were having an average size of 250–450 nm with a rather narrow size distribution. 3.2.1. Effect of chitosan molecular weight on nanoparticle size The molecular weight of the chitosan was varied at 150 kDa (LMW), 450 kDa (MMW) and 650 kDa (HMW) for the investigation of the effects of molecular weight on magnetic nanoparticle size. The stirring rate was 2,000 rpm and chitosan:TPP weight ratio was 6:1 in all of the experiments. The effect of chitosan molecular weight on particle size at three different molecular weights was shown in Figure 2. As seen in the graph the size of the nanoparticles was increased by the increase in molecular weight of chitosan.
MAGNETIC CHITOSAN NANOPARTICLES
317
60
Size (nm)
50 40 30 20 10 0
LMW
MMW
HMW
Chitosan Molecular Weight Figure 2. Effect of chitosan molecular weight on nanoparticle size.
3.2.2. Effect of chitosan:TPP weight ratio on nanoparticle size The effect of chitosan to TPP weight ratio on nanoparticle size was also very prominent showing a linear increase of size with increasing chitosan to TPP weight ratio within the tested chitosan to TPP ratio range as seen in Figure 3. This linear relationship provides a simple processing window for manipulating and optimising the nano size for intended applications. The stirring rate was 2,000 rpm and chitosan molecular weight was 450 kDa in all of the experiments. 600
size (nm)
500 400 300 200 100 0 1
2
3
TPP concentration
Figure 3. Effect of chitosan:TPP weight ratio on nanoparticle size.
D. KAVAZ ET AL.
318
3.2.3. Fe3O4/chitosan ratio on nanoparticle size The Fe3O4/chitosan ratio was varied in the range of 1:1–4:1 (w/w) for the investigation of the effects of the Fe3O4/chitosan ratio on size of the nanoparticles. The stirring rate was fixed at 2,000 rpm and MMW chitosan was used in all of the experiments. The obtained results were summarized in Figure 4. The size of the nanoparticles was increased with increasing Fe3O4 content.
440
50 25
s ize (nm )
420 400 380 360
15 10
340 320 1 Fe3O4 content (mg)
Figure 4. Effect of Fe3O4/chitosan ratio on nanoparticle size.
3.3. MAGNETIC PROPERTIES
The magnetic properties of the magnetic chitosan nanoparticles were evaluated using a vibrating-sample magnetometer (VSM) and electron spinning resonance (ESR). The Fe3O4/chitosan ratio was varied in the range of 1:4–1:1 (w/w) for the investigation of the effects of the Fe3O4/chitosan ratio on magnetic properties of the nanoparticles. The stirring rate was fixed at 2,000 rpm and MMW chitosan was used in all of the experiments. ESR results were shown in Table 1 and VSM results were shown in Figure 5. As can be seen the magnetic quality increases significantly with Fe3O4 content in Table 1 and Figure 5. The most important aspect for the magnetic carrier is a sufficient magnetic field intensity to excite all the dipole moment of each gram of magnetic carrier. This value defines the magnetic quality of the carriers and is in the range of 8–20 kG for various applications.2 The magnetic field intensity for magnetic chitosan nanoparticles was found to be in the range of 850–1,050 G in this study.
MAGNETIC CHITOSAN NANOPARTICLES
319
TABLE 1. ESR results of magnetic chitosan nanoparticles having different Fe3O4 content. Sample 1 2 3 4
Fe3O4 content (g) 0.050 0.025 0.015 0.010
Hr ( G) 1,050 1042.5 1,021 875
G 2.29 2.27 2.25 2.1647
Figure 5. VSM result of magnetic chitosan nanoparticles having different Fe3O4 content.
4. Conclusion In this presented study magnetic chitosan nanoparticles were prepared with ionic gelation method for biomedical applications. Morphology, size and magnetic properties of the magnetic chitosan nanoparticles were evaluated using selected parameters; molecular weight of the chitosan and the Fe3O4 content, and crosslinker concentration. Magnetic chitosan nanoparticles appeared separated and had 250–400 nm of average size. The size of the nanoparticles was increased by the increase in molecular weight of chitosan and Fe3O4 content; decreased by the increase in cross-linker concentration. The magnetic property of the nanoparticles was also increased by the increase in Fe3O4 content. Magnetic chitosan nanoparticles are promising as a potentially good material for diverse biomedical applications such as separation processes, optoelectronic, catalysts and/or sensors, medical diagnosis and therapy with good economical aspects and good magnetic quality.
320
D. KAVAZ ET AL.
References 1. Hamoudeh, M., and Fessi, H. (2006), Journal of Colloid and Interface Science, 300, 584–590. 2. Hütten, A., Sudfeld, D., Ennen, I., Reiss, G., Hachmann, W., Heinzmann, U., Wojczykowski, K., Jutzi, P., Saikaly, W., and Thomas, G. (2004), Journal of Biotechnology 112, 47–63. 3 Tartaj, P., Morales, M., Gonzalez-Carreno, T., Veintemillas-Verdaguer, S., and Serna, C. (2005), Journal of Magnetism and Magnetic Materials 290–291, 28–34. 4. Chang, Y., Chang, S., and Chen, D. (2006), Reactive & Functional Polymers 66, 335–341 5. Jia, Z., Yujun, W., Yangcheng, L., Jingyu, M., and Guangsheng, L. (2006), Reactive & Functional Polymers 66, 1552–1558. 6. Chang, Y., and Chen, D. (2005), Journal of Colloid and Interface Science 283, 446–451.
ANOMALOUS BEHAVIOR OF CARBON FILLED POLYMER COMPOSITES BASED CHEMICAL AND BIOLOGICAL SENSORS K. ARSHAK1*, C. CUNNIFFE1, E. MOORE1, AND A. VASEASHTA2 1 Department of Electronic and Computer Engineering, University of Limerick, Limerick, IRELAND 2 Department of Physics & Graduate program in Physical Sciences Marshall University, One John Marshall Drive Huntington, WV 25575-2570, USA
Abstract – This paper details results obtained from an array of drop coated room temperature carbon filled polymer composite sensors showing anomalies in the initial responses of the devices. The paper details the construction of the sensors, and the theory of their operation. The manufacture techniques are introduced along with the experimental configuration used to acquire the results.
Keywords: Conducting polymer composites, sensors, sensor systems
1. Introduction Conductivity sensors include both conducting polymer and metal oxide sensors both of which give a change of resistance upon exposure to Volatile Organic Compounds (VOCs).1 The mechanism that leads to the resistance change is different for both types of sensors. However, the physical structure of the sensor remains the same.2 In the case of Conducting Polymer Composite (CPC) sensors the sensing material is made from a polymer mixed with carbon black and suspended in a solution if it is deposited using drop coating. CPC sensors may use as conducting filler either carbon black or polypyrrole and an insulating matrix constitute a non-conductive organic polymer.3 The sensing material is
______ *
To whom correspondence should be addressed: K. Arshak, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
321
322
K. ARSHAK ET AL.
deposited over the metal electrodes to form an electrical connection through which the resistance is measured.2 Conducting polymer sensors do not require a heater as they can operate at room temperature, so they are easier to make. Also, as they are resistive elements the interface circuitry is less complex and makes them more suitable for portable instruments.1 2. Experimental Investigation The materials were deposited between Copper (Cu) electrodes onto an alumina substrate using a drop coating technique. A 500 nl syringe was used to deposit approximately 100 nl of the material over the electrode. A custom teÀon sampling head was used to encapsulate the sensor array. In order to characterize the responses of the sensors under varying conditions and using varying solvents a custom testing apparatus was used. Central to the system is a Bronkhorst Controlled Evaporator Mixer (CEM) which delivers vaporized liquid at its output for use in the chamber. A National Instruments Data Acquisition Card (DAQ) card model number PCI-MIO-16E-4 is also installed in the PC. This card was used to measure the resistance of the sensors in the chamber through the analog inputs. The sensor array encapsulated in the sampling head is placed in the 10 l volume insulated steel chamber. The environment inside the chamber can then be controlled via the PC. 3. Results The permeation of gases into the polymer bulk leads to an expansion in the insulating polymer matrix, which pushes the Carbon Black (CB) conducting particles apart thus decreasing the conductivity of the material. This expansion mechanism effectively reduces the fractional volume of conducting particles in the device. The conductivity of these materials in response to compression or expansion can be explained by the percolation theory. As the volume fraction of the conductive filler reaches the percolation threshold the first continuous conducting filament is completed through the polymer, which results in a large fall in the resistivity of the material. Large effects on the conductivity of these devices are evident when they are operating around the percolation threshold. Compression leads to increased conductivity and expansion leads to reduced conductivity. It was shown previously by charting the resistivity against percentage CB loading that the percolation threshold for a Poly(vinyl acetate) (PVAc)/CB composite sensor was found with a 6% CB loading. Further experiments with CB loadings of 6%, 8% and 10% showed the sensor with the 8% CB loading outperformed the others.4 The graph in Figure 1 shows the anomalous behavior of the sensor at approximately 40 s of exposure. These abnormal
POLYMER COMPOSITES BASED CHEMICAL & BIOLOGICAL SENSORS 323
responses start to happen when the gas concentration penetrating the polymer bulk is approximately 8,000 parts per million (ppm). The instability of the sensor usually happens when the sensor is operated around the percolation threshold. For this study single finger, non-parallel electrodes were used. The sensing material was drop coated onto the electrodes and comprises Ethyl Cellulose (EC) with 6.5% CB. The baseline resistances of the sensors ranged from 21.1 to 1,200 k, with sensor S1 having a baseline resistance of 1,200 k and outperforming the other sensors in the array. Using fractional baseline manipulation removes the multiplicative effect of different currents passing through the sensors ultimately giving the maximum percentage change in resistance. Figure 1 shows the data after baseline manipulation. Fractional baseline manipulation was used for the data analysis as it can help reduce the effects of additive and multiplicative drift. The result is a dimensionless normalized response. Figure 2 plots the data for a single sensor S1 on array A66 in response to increasing concentrations of propanol in the range 5,000–20,000 ppm in step increments of 3,000 ppm. The plot shows the sensor operational below the percolation threshold in response to 5,000 ppm. As the concentration increases from 5,000 through to 11,000 ppm the transition from below the percolation threshold to above is evident. An oscillation is present at 11,000 and 20,000 ppm as the sensor approached the percolation threshold. Across the entire concentration range there is finite attenuation in the response as the sensor transitions to above the percolation threshold. This operation is evident across all sensors at the in the same concentration region.
Figure 1. Fractional baseline manipulation applied to the raw data response of sensor array (EC Sensor with 6.5% CB) when exposed to propanol.
324
K. ARSHAK ET AL.
Figure 2. ǻV/V% responses of sensor exposed to propanol in the concentration range 5,000– 20,000 ppm.
4. Conclusion The results in this paper have highlighted anomalous behavior of carbon filled polymer composites based chemical and biological sensors. The results show the anomalous behavior of the sensor occurs after approximately 35 s, and when the concentration of the VOC is at approximately 8,000 ppm. References 1. Nagle, H., Gutierrez-Osuna, R., and Schiffman, S. (1998) The how and why of electronic noses, IEEE Spectrum, 35(9), 22–34. 2. Arshak, K., Moore, E., Lyons, G., Harris, J., and Clifford, S. (2004) A review of gas sensors employed in electronic nose applications, Sensor Review, 24(2), 181–198. 3. Albert, K., Lewis, N., Schauer, C., Sotzing, G., Stitzel, S., Vaid, T., and Walt, D. (2000) Cross-reactive chemical sensor arrays, Chemical Reviews, 100(7), 2595–2626. 4. Arshak, K., Cavanagh, L., Moore, E., Clifford, S., Harris, J., Cunniffe, C., and Lyons, G. (2004) Response of poly(vinyl acetate)\carbon black composites to ethanol vapour and temperature. In Proceedings of the International Conference on Microelectronics, 24 I, 181–184, Nis, Serbia and Montenegro.
POLY(N-ISOPROPYLACRYLAMIDE) (PNIPAM) BASED
NANOPARTICLES FOR IN VITRO PLASMID DNA DELIVERY N. OZDEMIR1*, A. TUNCEL2, M. DUMAN2, D. ENGIN3, AND E.B. DENKBAS4 1 Erciyes University, Chemistry Department, Kayseri, TURKEY 2 Hacettepe University, Chemical Engineering Department, Beytepe, Ankara, TURKEY 3 Gazi University, Faculty of Medicine, Ankara, TURKEY 4 Hacettepe University, Chemical Department, Beytepe, Ankara, TURKEY
Abstract – In this study; poly(n-isopropylacrylamide) P(NIPA) based nanoparticles were prepared by dispersion polymerization technique. Prepared nanoparticles were characterized by their morphology and chemical point of view using different techniques. Morphological evaluations of the nanoparticles were taken by using an atomic force microscope (AFM). Zeta potential and the particle size of NIPA based nanoparticles in aqueous solutions were determined with DLS (Dynamic Light Scattering) technique at different pHs and different temperatures. MTT studies were carried out to confirm the non-toxic character of the nanoparticles. In the transfection and expression studies; the plasmid DNA for Green Fluorescence Protein (GFP) expressing was used as a model plasmid DNA and the HeLa cells were used as the model cell line.
Keywords: Gene delivery, polymeric nanoparticles
1. Introduction The completion of a working draft of the human genome project paved the way for a greater understanding of diseases. It is now possible to treat diseases of genetic origin by-administering healthy copies of mutated (disease) genes1 or promote a protective immune response by administering genes encoding for
______ *
To whom correspondence should be addressed: N. Ozdemir, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
325
326
N. OZDEMIR ET AL.
specific antigens. A potential approach to the treatment of genetic disorders in man is called gene therapy. Gene therapy is the transfer of nucleic acids to the somatic cells of an individual with a resulting therapeutic effect.2 For this purpose DNA must be transported inside the cell, into the nucleus of the target cell. One of the most important factors for successful gene therapies is utilize a carrier that delivers the genes into cells for the production of therapeutic proteins.3 A great deal of research and development is being applied to obtain better viral and non-viral carriers. Viral carriers are effective but they have some disadvantages such as immunogenicity, a risk of mutation and relatively small capacity for therapeutic DNA etc.4,5 Non viral carriers especially polymeric carriers are the most attractive and promising carriers among the non-viral gene delivery systems. One class of the polymeric carriers consists of stimuli responsive polymers. Stimuli responsive polymers are polymers that respond with dramatic property changes to small changes in their environment. According to the stimuli they respond to, they can be classified as temperature-, pH-, ionic strength-, light-, electric-, and magnetic field sensitive. Such polymers may be used in numerous applications, including phase separations, affinity precipitations, bioactive surfaces, permeation switches and bioreactors.6–10 They have recently been the target of increasing interest especially in biomedical applications (i.e., controlled drug release, immunodiagnostics, gene therapy) and others (i.e., immobilization of enzyme and cell, affinity separation).11–20 In recent years, an increasing interest has been devoted to the preparation of N-isopropylacrylamide (NIPA) based thermosensitive particles for diagnosis applications and gene delivery. Because of the subcellular size, nanoparticles effectively endocytosed by cells which could result in higher cellular uptake of the entrapped or adsorped biomolecules. For these reasons, nanoparticles have become an important area of research in the field of biomolecule (protein, peptide, DNA etc.) delivery. Nanoparticles were introduced as a possible means to improve site specific drug delivery in the 1970s. Only in recent times, extensive research has been performed to investigate polymeric nanoparticles responsive to subtle changes in environmental conditions. Especially, polymeric nanoparticles have attracted much attention as a delivery vehicle in biomedical applications due to enhanced permeation, specific cell targeting, and long circulation in blood.10 The sub-micron size of nanoparticles offers a number of distinct advantages over microparticles or complexes in gene delivery. Due to this huge surface area, reliable and reproducible results were obtainable regarding adsorption of biomolecules onto them. Nanoparticles, in general, have also relatively higher intracellular uptake compared to microparticles and complexes. As a consequence of their subcellular size, they can be easily endocytosed/phagocytosed by cells, with a resulting cell internalization and higher cellular uptake of the adsorbed or encapsulated DNA.10
PNIPAM BASED NANOPARTICLES FOR IN VITRO PLASMID
327
2. Experimental Details 2.1. MATERIALS
The monomer, N-isopropylacrylamide (NIPA), the cross-linking agent, N, N-methylenebisacrylamide (MBA), the initiator, 2-2'Azobis(2-methylpropionamidine) dihydrochl-oride (APDH) and the comonomer, vinylprydine (VP) were purchased from Sigma-Aldrich Company, USA. HeLa cell and L929 cell lines used in cell culture studies were obtained from the tissue culture collection of Foot and Mouth Disease Institute of Ministry of Agriculture and Rural Affairs of Turkey. In the gene delivery studies, the plasmid GFP was used. This plasmid was reproduced in Esherichia coli and purified by column chromatography (Qiagen Plasmid Giga Kit). Plasmid DNA was stored at í20°C until transfection and gene expression experiments. 2.2. PREPARATION OF PARTICLES
Poly(NIPA) and poly(NIPA-co-VP) particles were synthesized by dispersion polymerization with two different recipes. The polymerization was carried out at 70°C for 4 h, with a shaking rate of 120 cpm in a temperature-controlled shaker. The particles were extensively washed with distilled water by following a centrifugation-decantation protocol. 2.3. CHARACTERIZATION OF PARTICLES
The morphologies of the nanoparticles were characterized by an atomic force microscope (AFM). For the determination of particle size in aqueous solution, Zeta-sizer NarioS (Malvern Instruments, England) was used. The size of particles was measured as a function of pH and temperature. The particle size measurements were made at the temperatures between +10°C and 50°C. The effect of NIPA based particles on the cell viability was investigated by the MTT assay. In these test, the nanoparticle concentration was changed between 0.001– 0.1% w/w at body temperature (37°C). In vitro cytotoxicity of PNIPA based particles was tested by using L929 mouse fibroblasts. The MTT assay was performed according to the method of Mosmann.21 The HeLa cells were used as the model cell line in gene transfection studies. In the transfection and expression studies; the plasmid DNA for Green Fluorescence Protein (GFP) expressing was used as a model plasmid DNA. Cellular uptake of the plasmid DNA adsorbed nanoparticles were confirmed by light and fluorescence microscope.
N. OZDEMIR ET AL.
328
3. Results and Discussion AFM micrographs are shown in Figure 1. As seen in the micrographs, the nanoparticles are fairly mono disperse. According to the results reported in Figure 2, particle size decreased with an increase in temperature (from +10ºC to 50ºC) at pH 7. To evaluate the suitability of nanoparticles as gene carriers for plasmid DNA, it is essential to obtain information on the surface positive charge density. The surface charge of the plasmid DNA delivery system is the most critical factor affecting transfection efficiency. The ȗ-potential of NIPA based nanoparticle was measured as a function of temperature (from +25ºC to 45ºC) and pH 7 in an aqueous solution. The result is given in Figure 3.
Zeta potential (mV)
Figure 1. The AFM images of polyNIPA and poly(NIPA-co-VP) nanoparticles.
600 500 400
pH:7 NPVP1
300
pH:7 NPVP2
200 100 0 0
20
40
60 o
Temperature ( C)
Figure 2. The effect of temperature on size of NPVP1 and NPVP2 at pH 7.
PNIPAM BASED NANOPARTICLES FOR IN VITRO PLASMID
329
Zeta potential (mV)
60 50 40
pH:7 NPVP1
30
pH:7 NPVP2
20 10 0 15
35 Temperature
Figure 3. The effect of temperature on ȗ-potential of NIPA based nanoparticles encoded NPVP1 and NPVP2 in an aqueous solution at pH 7.
105 100 95 90 85 80 75
1 0.1
24 h
48 h
NPVP2
NPVP1
NPVP2
NPVP1
NPVP2
0.01 NPVP1
Cell viability %
The variation of cell viability for poly(NIPA-co-VP) nanoparticles by time is given in Figure 4. The NIPA based nanoparticles was not caused any observable toxicity in the examined polymer concentration range, especially for 24 h. Figure 5 gives the representative micrographs of HeLa cells taken with the light and fluorescence microscopes, respectively. Two representative micrographs are shown in Figure 5a and b, which clearly demonstrates expression of GFP gene.
72 h
Polymers
Figure 4. The effect of polyNIPA and poly(NIPA-co-VP) particle concentration on the cell viability of L-929 mouse fibroblast cells (the legends show the concentrations).
N. OZDEMIR ET AL.
330
a
b
Figure 5. Representative micrographs showing expression of GFP gene. (a) Light microscopy image of HeLa cells; (b) fluorescence microscopy image of HeLa cells transfected with poly (NIPA-co-VP) nanoparticles.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Rubanyi, G. M. (2001) Mol. Aspect Med. 22,113–142. Panyam, J., and Labhasetwar, V. (2003) Adv. Drug Deliv. Rev. 55, 329–347. Niidome, T., and Huang, L. (2002) Gene Ther. 9, 1647. Liu, F., and Huang, L. (2002), J. Control Release 78, 266. Liu, W., and Yao, K. (2002), J. Control. Release 83, 1. Tuncel, A., and Ozdemir, A. (2000) J. Biomater. Sci. Polymer Edn. 11, 817. Elaissari, A., and Bourrel, V. (2001) J. Magn. Magn. Mater. 225, 151. Kikuchi, A., and Okana, T. (2002) Prog. Polym. Sci. 27, 1165. Taniguchi, T., Duracher, D., Delair, T., Elaissari, A., and Pichot, C. (2003) Colloid Surf. B 29, 53. Ozdemir, N, Tuncel, A., Kang, M., and Denkbas, E. (2006) J. Nanosci. Nanotechnol. 6, 2804. Hinrichs, W., Schuurmans-Nieuwenbroek, N., Van de Wetering, P., and Hennink, W. (1999) J. Control. Release 60, 249. Chilkoti, A., Dreher, M., Meyer, D., and Raucher, D. (2002) Adv. Drug Deliver Rev. 54, 613. Elaissari, A., Rodrigue, M., Meunier, F., and Herve, C. (2001) J. Magn. Magn Mater. 225, 127. Takeda, N., Nakamura, E., Yokoyama, M., and Okano, T. (2004) J. Control. Release 95, 343. Eeckman, F., Moes, A., and Amighi, K. (2003) J. Control. Release 88, 105. Kopecek, J. (2003) Eur. J. Pharm. Sci. 20, 1. Eeckman, F., Moes, A., and Amighi, K. (2002) Int. J. Pharm. 241, 113 Hinrichs, W., Schuumans-Nieuwenbroek, N., Van de Wetering, P., and Hennink, W. (1999) J. Control. Release 60, 249. Yoo, M., Sung, Y., Lee, Y., and Cho, C. (2000) Polymer, 41, 5713. Leroux, J., Allemann, E., De Jaeghere, F., Doelker, E., and Gurny, R. (1996) J. Control. Release 39, 339. Mossman, T. (1983) J. Immunol. Methods 65, 55.
RAPID, CONTACTLESS AND NON-DESTRUCTIVE TESTING OF CHEMICAL COMPOSITION OF SAMPLES O. IVANOV1*, A. VASEASHTA2 , AND L. STOYCHEV1 1 Geordi Nadjakov Institute of Solid -State Physics, Bulgarian Academy of Science, 72 Tzarigradsko Chaussee, Sofia, BULGARIA 2 Nanomaterials Processing & Characterization Laboratories, Marshall University, Huntington, WV 25575, USA
Abstract – Our results demonstrate that a new effect can be induced in each solid in a wide spectral range of electromagnetic irradiation. In the present manuscript we prove experimentally that one of the possible applications of this effect is for an express contactless control of the chemical composition of a series of samples, in this case, coins. The method has wide applicability ranging from defense and homeland security to several applications requiring rapid and nondestructive identification of chemical composition.
Keywords: Chemical composition test, non-destructive testing, field-matter interaction.
1. Introduction An effect called “Surface photo charge effect” (SPCE) is very attractive for various practical applications. It is based on the fact that, on irradiation with an electromagnetic field all investigated solids are charged with an alternating potential difference, which frequency is equal to the frequency of the incident electromagnetic wave. Every solid generates a specific signal.1 The voltage is measured contactless between the irradiated solid and a second solid whose potential is assumed to be zero.2,3 An important feature of the effect is that it is strongly depends on the state of the irradiated solid. Some of the applications experimentally tested by us so far include: retrieval of information of defects,
______ *
To whom correspondence should be addressed: O. Ivanov, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
331
332
O. IVANOV, A. VASEASHTA, AND L. STOYCHEV
irregularities and impurities on the surface4; investigation of surface electrical characteristics5; remote retrieval of information on a solid object and its displacement; studies on the surface electron state in semiconductors6; sensor for gases, liquids, and vapours. The method may be used to study fluids and various processes taking place in them.7 Furthermore, the presence of impurities in liquids can be monitored and a level meter for liquids can be designed8; deposition of solution components on a surface can be observed – deposition of limestone from water on a metal surface. Also gasoline octane numbers can be determined. The measurements are rapid and contactless.1 2. Experimental A schematic of the experimental set up is illustrated in Figure 1. The sample (2) was irradiated with a modulated electromagnetic field (1). The voltage, induced by the SPCE, was measured with the help of the electrode (3), which is in close proximity to the sample (2). The signal in such cases was measured using a lock-in nanovoltmeter (4). We provide a basic scheme of the device because of its practical application for developing apparatus and potential for patent infringement and intellectual property protection. It is possible to use radiation in different ranges from the electromagnetic spectrum. In this particular case, we applied an irradiating electromagnetic field with a frequency of 144 kHz and used a lock-in nanovoltmeter as a measuring device. One of the features of SPCE, is that every solid generates a specific signal, which amplitude signal depends on the chemical composition of the solid. This naturally renders an elucidation that we could quickly and without any physical contact monitor the composition of a variety of samples. As an interesting application, we selected a variety of coins as they are easily accessible, widely used, and the investigations associated with them are of significant interest. At the same time, however, there are a limited number of methods for their rapid
Figure 1. A sketch of an experimental set-up for SPCE – observation: 1 – electromagnetic field source, 2 – irradiated solid, 3 – signal-measuring electrode, 4 – measuring device.
CONTACTLESS TESTING OF CHEMICAL COMPOSITION OF SAMPLES
333
and contact-less testing. The cash directorate of the Bulgarian National Bank (BNB) provided the coins for this investigation. The object of our investigation was coins of 50 stotinki – subject to widespread counterfeiting, according to the information we received from the BNB. Initially, a large number of genuine samples, taken out of common circulation in order to determine the range of the amplitudes of signals generated by coins was investigated. Later, we measured various counterfeits, kindly presented to us by BNB. The comparison of the results showed unambiguously that identification of real and counterfeit coins is fully possible. Subsequently, we measured also a large number of series consisting of 10–15 coins each comprising of unknown number of counterfeits and genuine samples without any cleaning procedure, as were presented to us by the BNB. The amplitudes of the signals for the genuine samples were with some differences, due to variations in the chemical composition of the original coins. But they are set in a separate area despite of the differences. If there is a standardization of chemical composition, this relative wide area, in which are set the signals of the genuine coins, can be reduced strongly. The amplitudes of electrical signals caused from SPCE measured over counterfeits were significantly different from these for original coins. A comparison of the results showed that identification of real and counterfeit coins by the proposed method is clearly discernible, primarily due to the chemical composition. This is shown by Figure 2, which is one axis graphics. On the axis are set the amplitudes of electrical signals cased from SPCE measured over 100 investigated coins. Three of them are counterfeits and they are easy to be seen on the graphics because they are out of genuine coins area signals. It is even possible to identify several different types of counterfeits, so providing the source of origin. Since the measured signals are instantaneous, the method in principle finds application where an instantaneous identification is necessary. However, the effect is very sensitive and the whole device together with all the necessary electronic sections has to be placed in a dense metal shield. We believe that without any difficulties the detector could be designed relatively economically and small in size, so as to be installed in any coin-operated machine.
Figure 2. The amplitudes of electrical signals cased from SPCE, measured over 100 investigated coins: x – genuine coins; x – counterfeits coins.
334
O. IVANOV, A. VASEASHTA, AND L. STOYCHEV
The method however cannot be used to determine the exact amount of various chemical components present in any given unknown sample. It can only be applied to detect whether or not the sample meets a certain standard. It is possible to trace change in a certain impurity in a series of samples provided that the rest of the components are kept constant. This would require a preliminary calibration. 3. Future Applications The proposed method has several applications and could readily be employed in places such as banks, shops, coin-operated machines and so on. Another application of SPCE is for example, monitoring various types of absorbing filters for gases and liquids to determine when they need to be replaced. This is possible due to the fact that the absorption (of the filtered substances) changes the chemical composition of the filter, which can be detected with SPCE. When security is concerned, analysis of drinking water from the public water distribution system or that distributed to an army in field operations is technically feasible. Changes in the chemical composition due to contamination can be detected by relevant variation in the SPCE signal. Very often for the producing special details in industry it is important to know if there are changes in the chemical composition of the material. The offered method can be applied to control on every separate detail.
References 1. Ivanov, O. (2006), Sensor applications of field-matter interactions, Encyclopedia of SENSORS, American Scientific Publishers, California, vol 9, pp. 165–197. 2. Pustovoit, V., Borissov, M., and Ivanov, O. (1989), Phys. Lett. A 135, 59–61. 3. Ivanov, O., Mihailov, V., Pustovoit, V. Abbate, A., and Das, P. (1995), Opt. Commun. 113, 509–512. 4. Pustovoit, V., Borissov, M., and Ivanov, O. (1989), Solid State Commun. 72, 613–619. 5. Das, P., Mihailov, V., Ivanov, O., Guergiev, V., Andreev, S., and Pustovoit, V. (1992), IEEE Electron. Device Lett. 13, 291–293. 6. Ivanov, O., and Konstantinov, L. (1999), Appl. Surf. Sci. 143, 101–103. 7. Ivanov, O., and Konstantinov, L. (2000), Surf. Rev. Lett. 7, 211–212. 8. Ivanov, O. (2001), Sens. Actuators B 75, 210–212.
SYNTHESIS AND APPLICATION OF METAL-CONTAINING SILICAS K. KATOK*, V. TERTYKH, AND V. YANISHPOLSKII Institute of Surface Chemistry of National Academy of Sciences of Ukraine, 17 Naumov Str., 03164 Kyiv-164, UKRAINE
Abstract – Gold nanoparticles were obtained by in situ reduction with silicon hydride groups grafted to the silochrom C-120 and aerosil A-300 silica surface. Such gold-containing silicas have been applied for hydrogen oxidation.
Keywords: Silica, modified surface, silicon hydride groups, gold, in situ reduction, hydrogen oxidation
1. Introduction Gold nanoparticles have been widely studied because of their unusual optical, catalytic and electronic properties.1 Their properties are very different from the corresponding bulk material in particular spherical gold nanoparticles exhibit a single plasmon resonance in the visible region of the spectrum.2 However, unsupported nanoparticles are thermally unstable. Hence, the immobilization of these nanoparticles on the supports, such as polymer latex particles,3 zeolite,4 and porous silica5 to form composite particles is of great interest. Although gold is the most inert of all metallic elements, it has interesting properties as a heterogeneous catalyst. For preparation of supported metal nanoparticles the chemically modified silochrom C-120 and aerosil A-300 silica type were applied in this work. The surface of such matrices contained chemically bound groups possessing reducing properties towards metals which are allocated after hydrogen in the electromotive series. This approaches can allow to obtain metal nanoparticles in the close vicinity to the appropriate surface compounds and, thus, to fulfill definite
______ *
To whom correspondence should be addressed: K. Katok, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
335
336
K. KATOK, V. TERTYKH, AND V. YANISHPOLSKII
anchorage of the formed nanoclusters. It is expedient to use obtained goldcontaining mesoporous silicas as catalysts in hydrogen oxidation. 2. Experimental Reducing reagents immobilized on the external surface of aerosil A-300 and silochrom C-120 silicas (specific surface area 273 m2/g and 114 m2/g correspondingly) were used for in situ reduction of nanoparticles of metals at the interaction with chloroauric acid in the mild conditions. For this silica matrices were modified with triethoxysilane in the presence of acetic acid with the object of producing grafted silicon hydride groups ({SiH) on the silica surface. Then modified silica was impregnated with HAuCl4 at room temperature and was dried for 24 h in an oven at 150qC. Presence of silicon hydride groups grafted to the silica surface was confirmed by FTIR spectroscopy data (NEXUS FT-IR). Nanoparticles of gold were identified by the transmission electron microscopy (JEM-1OOCXII) and X-ray powder diffraction (DRON-4-07, CuKD-radiation). UV-Vis spectra of metal-containing composites were recorded on Carl Zeiss Jena spectrophotometer. 3. Results and Discussion Silicas with silicon hydride groups in the grafted modifying layer were used for synthesis of noble metal nanoparticles. In the IR-spectrum of the initial silica (Figure 1, curve 1), a sharp band is observed in the region of O-H vibration at 3,750 cm-1, which originates from isolated OH groups of the silica surface.
Figure 1. FTIR spectra of aerosil (a) and silochrom (b): 1 – initial silica, 2 – silica with grafted {SiH groups, 3 – silica with reduced nanoparticles of gold.
SYNTHESIS AND APPLICATION OF METAL-CONTAINING MATRICES
337
After modification of silicas with triethoxysilane in presence of acetic acid (samples A-300-H and C-120-H) in the spectrum of the silicon hydridecontaining silicas (Figure 1, curve 2) this band is absent and, a broad band is seen at 2,240 cm–1 indicating evidence for the fact that attachment of {SiH groups occurred. Interaction between modified silicas with chloroauric acid (samples C-120-Au and A-300-Au) is accompanied by decreasing of intensity of absorption band 2,240 cm–1 relating to {SiH groups and appearance of absorption band of silanol groups 3,750 cm–1 is observed (Figure 1, curve 3). In the high-angle region at 2ș = 30–90q of X-ray diffraction patterns typical peaks of the face-centered cubic structure of metallic gold are observed, which are ascribed to (111), (200), and (220) diffractions (Figure 2a). The crystallite sizes of nanoparticles were determined by the Scherrer equation.6
Figure 2. (a) – XRD patterns of aerosil A-300-Au (1), silochrom C-120-Au; (b) – TEM of aerosil A-300-Au.
According to TEM data average diameters of formed gold particles fall within the range of 2–27 nm for A-300-Au and 22–35 nm for C-120-Au composites. The inset in Figure 2b, which exhibits an electron diffraction pattern, gives evidence of metallic gold formation. Thirty milligrams of metal-containing composites were dissolved in 5 ml solution of 5 M sodium hydroxide for measuring of UV-Vis spectra of colloid gold. A strong absorption at approximately 540 nm is observed for colloid gold and gold-containing silica (Figure 3). It demonstrated that the AuCl4í ions have been totally reduced into gold clusters. The catalytic activity of samples A300-Au, C-120-Au in oxidation of H2 with molecular O2 was investigated under flowing conditions with chromatographic control of the reaction mixture composition under atmospheric pressure (LKhM-8MD chromatograph with thermal conductivity detector). Various gas
K. KATOK, V. TERTYKH, AND V. YANISHPOLSKII
338 1.0
1.0
a
0.8
b
0.8 A
A
0.6 0.4
0.4
0.2 0
0.6
300
400
500 600 L. nm
700
800
900
0.2
300
400
500 600 L. nm
700
800
Figure 3. UV-Vis spectra of aerosil A-300-Au (a), silochrom C-120-Au (b).
mixtures were used (20% O2, 1% H2 and 79% Ar). The activity of a given catalysts is usually expressed in terms of the temperature corresponding to 50% conversion of H2. The catalytic activity in H2 oxidation of these materials is increased with decrease of specific surface area of samples, hydrogen conversion degrees is gone down, i.e. the area of passing of reaction is moved to lower temperatures. A lower temperature indicates correspondingly greater catalytic activity. The highest catalytic activity in H2 oxidation was shown by sample of silochrom C-120-Au. 4. Summary A number of metallized composites, which contain highly dispersed metal, on their surface by reduction ions of gold or silver from its solutions over chemically modified silica matrices were obtained. Such an approach allow us to regulate size of the immobilized nanoclusters by varying structural characteristics of the initial matrices. Gold-containing silicas have shown catalytic activity in reaction of hydrogen oxidation.
References 1. 2. 3. 4. 5.
Cortie, M., and Van der Lingen, E. (2002) Mater. Forum 26, 1–14. Kreibig, U., and Vollmer, M. (1995) Optical Properties of Metal Clusters. (Berlin: Springer). Mayer, A., Grebrier, W., and Wannemacher, R. (2000) J. Phys. Chem. B 104, 7278–7285. Park, K., and Ihm, S. (2000) Appl. Catal. A-Gen 203, 201–209. Hornebecq, V., Antonietti, M., Cardinal, T., and Delapierre, M. (2003) Chem. Mater. 15, 1993–1999. 6. Genie, A. (1961) Radiographic Imaging of Crystals. (Moscow: State Publisher of PhysicoMathematical Literature).
SEMICONDUCTING GAS SENSORS, REMOTE SENSING TECHNIQUE AND INTERNET GIS FOR AIR POLLUTION MONITORING IN RESIDENTIAL AND INDUSTRIAL AREAS
O. PUMMAKARNCHANA1*, V. PHONEKEO2, AND A. VASEASHTA3 1 Faculty of Science, Silpakorn University, Nakornpathom, 73000, THAILAND 2 Geoinformatics Center, Asian Institute of Technology, Klongluang, Pathumthani, 12120, THAILAND 3 On detail from Nanomaterials Processing & Characterization Laboratories, Marshall University, Huntington, WV, USA
Abstract – This research aims at developing a cost-effective and real time air pollution monitoring system with high reliability and less maintenance that utilizes numerical and air dispersion modeling in conjunction with inexpensive, state-of-the-art gas sensors, remote sensing methodologies, and Internet GIS. Conventional pollution detectors, installed by the Bangkok Pollution Control Department and WO3 sensors are employed for in-situ pollution measurements. The data obtained from the satellites sensors and measurements conducted on ground are used for numerical modeling by “Multiple Regressions” and for air dispersion modeling to investigate air pollutants distribution. The correlation and distribution of the air pollutants values are transferred to wireless GIS network system using Web Map Service for rapid and simultaneous dissemination of information on pollution levels at multiple sites.
Keywords: Internet GIS, real time air pollution monitoring, remote sensing, sensors.
______ * To whom corresponding
[email protected]
author
should
be
addressed:
O.
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
Pummakarnchana,
email:
339
340
1.
O. PUMMAKARNCHANA, V. PHONEKEO, AND A. VASEASHTA
Introduction
Industrialization, urbanization and rapid traffic growth has resulted in serious air pollution problems in some Asian countries, especially in central Bangkok, Thailand. Moreover certain rural practices, especially biomass burning, are causing severe air quality problems in the northern parts of the Thailand. The air quality levels indicated by PM10 data have increased above the National Air Quality Standard (120 μg m–3).1 The PM10 increase is mainly from forest fire and biomass burning in Northern parts of Thailand and along Thai borders; Burma and Lao People’s Democratic Republic. The areas selected for this investigation include Central Bangkok, Rayong city (“Map-Ta-Put” with a high population density) and Chiangmai cities, for which the air pollution is considerably large. The pollutants in Map-Ta-Put contaminating urban environment are VOCs causing the most widespread and acute health risks, as indicated by medical records of people living around the area. Some of the primary pollutants in Bangkok include oxides of Nitrogen, ozone, and PM10, 2.5, as reported by the PCD. Global NO2 column (molecule per square centimeters), as detected by the SCIMACHY and GOEM, clearly shows that the columns are increasing at a rapid pace in and around Bangkok.2 Many pollutants, such as CO, SO2, suspended particles in VOCs, etc., also exist in Bangkok, however, the present investigation focuses on monitoring and analyzing NOx, O3 and PMs as the representative pollutants. This research employs data obtained by WO3 semiconductor-type gas sensors in conjunction with satellite imagery as a cost and time-effective and efficient alternative for monitoring air quality as compared to the conventional air pollution monitoring systems.3 Moreover, the air quality monitoring unit, can be deployed anywhere and transmit data that can eventually be routed via a wireless GIS network system as a rapid monitoring tool to the general public and policymakers. The data from the gas sensors monitoring air pollutants is uploaded in real time via Personal Digital Assistant (PDA) to the air quality monitoring server as concentration values. Users can simply browse and view the air quality data in graphical format in real-time. Thus, the objectives of this investigation are twofold, viz.; (a): to apply a network of PDA coupled semiconductor type gas sensor network to obtain spatial distribution of pollution data. The data is then compared with the satellite data and air dispersion modeling for air pollution over the area under investigation; and (b): to assist national policymakers in establishing pollution awareness policies and priorities.
REMOTE SENSING AND INTERNET GIS FOR AIR POLLUTION
341
2. Method 2.1. METAL-OXIDE GAS SENSOR FOR AIR QUALITY MONITORING
Low cost, metal-oxide (WO3) semiconductor junction with Schottky structure type sensors was used. The WO3 based sensors were purchased from the Synkera Technologies and packaged in a commercial electronic package (P/N 7000006).4 Platinum was used as an electrode to measure conductivity changes of the sensor operating at a working temperature 250ºC.4 The NO2 concentrations in 0–7.34 ppm range were measured using varying gas mixtures of O3 and NOx by employing a ‘Programmable Multi-gas Calibrator’ model 5008. The gas calibrator was linked simultaneously to NOx gas analyzer using a “Dual Chamber Chemiluminesce” gas monitor to accurately acquire NO2 concentrations produced by the multi-gas calibrator. The ratio used for mixing gases was NOx:O3 (1:1) by gas volume. Besides flow meter “DC-lite” – which uses patented DryCal near-frictionless piston technology and photo-optic sensors were used to obtain volumetric flow readings quickly and accurately. The air and gas flow rate were adjusted at 70 ml min–1. Power source was provided by a 9 V and 1.2 mAH adapter. The sensor was inserted into a plastic chamber of 10 ml volume under continuous flow of testing gas mixtures at a constant flow rate. Before measuring NO2 gas each time, the sensor was stabilized for a minimum of 16 h at room temperature.5 2.2. AIR QUALITY MODELING
Industrial Source Complex Short Term-ISCST model, integrated with GIS and remote sensing techniques are employed to determine the affected areas over the study areas as demonstrated in Figure 1. The model predicts several air gaseous (NO2, NO, CO, SO2, O3) including PMs at various times and receptor points. ISCST was designed to support the EPA’s regulatory modeling programs. This ISCST accepts hourly meteorological data to define the conditions for plume rise, transport, diffusion, and deposition. The model estimates the concentration or deposition value of each source and receptor combination for hourly meteorological data, and calculates user-selected short-term averages. Required input data, hence, are (1) sources data; location, emission rate, release height above ground, stack gas velocity, and stack inside diameter, (2) hourly meteorological data including wind speed and direction, ambient temperature, stability class, cloud cover, ceiling height, and mixing height, and (3) receptor data; x, y, coordinates and elevated receptor height. Finally predicted air pollutant concentration values; high and maximum values and concurrent values of each averaging period for each day of data processed are viewed as contour lines,
342
O. PUMMAKARNCHANA, V. PHONEKEO, AND A. VASEASHTA
overlaid with population and land use in shapefile and aerial photo formats so as to identify its relation to pollutant concentrations. 2.3. AIR POLLUTANTS DETECTION USING SATELLITE IMAGERY
Moderate Resolution Imaging Spectroradiometer (MODIS) is used to monitor forest fire and biomass burning occurrence in the Northern part of Thailand. The active fire occurrence has been monitored using the MODIS Fire Information
Figure 1. Conceptual framework for real time air quality monitoring based WO3 gas sensors and satellite imagery conjunction with air dispersion modeling through Internet GIS.
System, generated daily MODIS Fire product (MOD14) from Geoinformatics Center.7 In order to understand the air quality conditions, the MODIS aerosol product (MOD04) was used to identify the air quality status, monitoring the ambient Aerosol Optical Thickness (AOT) over the oceans globally and over a portion of the continents. Daily Level 2 data are produced at the spatial resolution of a 10 × 10 km2 (at nadir)-pixel array. The product was generated in and distributed by the Level 1 and Atmosphere Archive and Distribution System6 (LAADS). Aerosol product employed in this investigation contains many scientific data sets. One of them is the AOT at 0.55 μm for both Ocean and Land. Moreover, the key target of SCIMACHY is to determine the global troposphere NO2 column. Nevertheless, the broad spectral and fine spatial resolution allows the estimation of NO2 emissions. SCIMACHY was launched on ENVISAT observing nadir and limb viewing with spatial resolution 60 × 30 km2. Therefore, the data obtained from these satellites sensors and measurements conducted on ground using the gas sensors will be used for numerical modeling in order to identify the correlation of the air quality levels.
REMOTE SENSING AND INTERNET GIS FOR AIR POLLUTION
343
3. Results 3.1. TYPICAL RESPONSES OF WO3 SEMICONDUCTING SENSOR TO NO2 GAS
The gas sensors were tested for sensitivity, selectivity, stability, and reproducibility. A representative data indicate that the sensors typically exhibit a linear response. For a typical sensor, when observed on a curve of voltage output versus concentration plot, the characteristic of sensor response displays a linear correlation between NO2 gas concentration and voltage output measured from the sensor, with a regression coefficient of 0.94, indicating correlation within experimental margin and the data show reproducible responses across this range of exposed concentrations. The circuitry in the sensor refreshes itself periodically for reproducible results.5 3.2. REAL TIME AIR QUALITY MONITORING BASED GROUND AND SATELLITE MEASUREMENTS AND GIS MODELLING
The air quality dissemination and monitoring model developed is applied in central Bangkok, Rayong and Chiagmai environment so as to retrieve and monitor real time air quality levels of NOx, O3 and PMs and also to continuously update information through wireless GIS and map-server service. The air quality levels, based on the NAAQ Standards, were overlaid with city base maps and satellite imagery. Internet users, hence, can browse and access the current and daily air quality based maps, related to geographic information, including districts, urban settlement, roads, railways, hydrology. The server is programmed such that the AQ events monitored over past 48 h can be queried at a given time. As investigated in the Northern part of Thailand and also in Laos and Myanmar, shown in Figure 2a, the MODIS image shows the relationship between the smoke from forest fire and biomass burning with the distribution of the AOT in range. The AOT varies from 0 to 1 μm. The highest value is shown in white, while the non-fired areas are approximately in 0–0.1 range as seen in Figure 2b. Further studies are still ongoing in order to estimate absorption and scattering properties as well as backscattering probability of the suspended matter. Such events, in conjunction with spatial distribution of tempe-rature, air currents, and VOCs and other pollutants concentration will be able to validate the MODIS satellite sensor based observation. Furthermore, hyper-spectral measurements in conjunction with the proposed optical model will provide a theoretical basis and practical basis for space based measurements and its correlation with observation on the ground, viz. quantifying as to how much aerosol is in the air by investigating the transmission of the sunlight. As an extension to this
O. PUMMAKARNCHANA, V. PHONEKEO, AND A. VASEASHTA
344
a
b
Figure 2. (a) Active fire occurrence in the Northern parts of Thailand detected by Aqua MODIS on March 11, 2007 at 06:59 GMT, (b) aerosol optical thickness for the detection of biomass burning (AOT in range (μm): white = 1, red = 0.9, yellow = 0.8, orange = 0.7, blue = 0.2, violet = 0–0.1).
project, using SCIMACHY to calculate NOx levels is in progress which will be further investigated so as to retrieve its correlation and observations on the ground together with the air quality dispersion model. 4. Conclusion and Recommendations We have shown that inexpensive WO3 semiconducting gas sensors can provide good performance characteristics, such as high sensitivity (<0.5–7.50 ppm), reproducibility, and reliable operation at relatively high temperatures irrespective of the moisture contents. The sensors in conjunction with Internet GIS is a viable tool to monitor pollution in industrial and residential environment. Remote sensing techniques using satellite imagery obtained from SCIMACHY, and MODIS can be applied to investigate real time industrial and residential air quality levels in conjunction with ground based measurements and air dispersion model.
References 1. The Pollution Control Department. Air Quality and Noise Management Bureau. (2007) Emergency and Pollution Situation Report. (February 8, 2007); http://pcd.go.th. 2. Richer, A., Burrows, J., Nüß, H., Granier, C., and Neimeir, U. (2005) Nature Letters 437, 129–132.
REMOTE SENSING AND INTERNET GIS FOR AIR POLLUTION
345
3. The Pollution Control Department, Air and Noise Quality Management Division. (2003) Air Viro System (He’s Co. Ltd, Bangkok, 2003), pp. 1–8. 4. Deininger, J., Williams, S., and Kostelecky, J. (2004) The Instrumentation, System and Automation Society, ISA Expo 2003 Technical Conference. (January 19, 2004); http:// www.isa.org. 5. Vaseashta, A., Vaclavikovam, M., Vaseashta, S., Gallios, G., Roy, P., and Pummakarnchana, O. (2007) Science and Technology of Advanced Materials 8, 47–59. 6. MODIS Fire Information System, Geoinformatics Center, Asian Institute of Technology. (2007) (March 11, 2007); http://geoinfo.ait.ac.th/mod14.
SELF-ASSEMBLED SYSTEM OF SEMICONDUCTOR AND VIRUS LIKE NANOPARTICLES
YU. DEKHTYAR1*, A. KACHANOVSKA1, G. MEZINSKIS1, 3 A. PATMALNIEKS2, P. PUMPENS3, AND R. RENHOFA 1
Biomedical Engineering and Nano Technologies Institute, Riga Technical University, Riga, LATVIA 2 Institute of Microbiology and Biotechnologies, University of Latvia, LATVIA 3 Biomedical Research and Study Centre, University of Latvia, Riga, LATVIA
Abstract – Virus like nanoparticles (VLP) are in use to be absorbed by cells to cause biological effects. To increase a local concentration of VLP, nanoparticlescarriers bringing the latter to the target cell could be employed. N-type and p-type Si semiconductor nanoparticles, to control adhesion of VLP were applied. Optical absorbance spectra and electron microscopy evidenced that VLP became connected to Si nanoparticles. Moreover, a density of the adhered VLP depended on the type of both semiconductor and VLP.
Keywords: Semiconductor nanoparticles, adhesion.
1. Introduction Modern treatment of humans has come to a scale of the targeted cell due to advantages by nanotechnologies that became “the manufacturing technology of the 21st century”.1 Virus like particles (VLP) are in use for cell therapy because VLP recognize and easy penetrate via the cell’s membrane. A typical diameter of VLP is around 25 nm. To do treatment more effective, a local (in a vicinity of the cell) concentration of VLP should be increased. To reach this, nanoparticles that perform the VLP carrier function could be in use.
______ *
To whom correspondence should be addressed: A. Kachanovska, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
347
YU. DEKHTYAR ET AL.
348
Interaction between the particles is supplied by adhesive mechanisms, the latter having electrostatic nature. To influence this, an adjusted surface charge of the nanoparticles-carrier could be employed. From this point of view n- or p-type semiconductor, that surface is originally charged, could become as the carrier. This presentation is devoted to explore a possibility to use n-Si or P-Si to adhere VLP. 2. Experimental The crystalline p-Si and n-Si having specific electrical resistance 1 and 2 cm (for p-Si and n-Si, correspondingly) were in use to prepare nanoparticles. The latter were fabricated because of mechanical milling by the porcelain balls. As the result a typical size of the particles around 100 nm was achieved. The size and a shape of the nanoparticles were verified due to electron and atomic force microscopy. For this the scanning electron microscope (SEM) S4800 and atomic force machine (AFM) SOLVER-PRO47 were in use, correspondingly. Three types (VLP1, VLP2 and VLP3) of VLP were selected for the experiment. The VLP were placed in a buffer solution and mixed with the semiconductor nanoparticles. The buffer was characterized with pH 7, 8. Drops of the prepared solutions were dried and further analysed by the above electron microscope. Alongside optical attenuation (OA) of the solutions was measured. The Helios Gamma UV/VIS spectrophotometer with photometric accuracy ±0.005A @1A was in use. 3. Results and Discussion SEM tests evidenced VLP were adhered on the Si nanoparticles. However, VLP3 demonstrated the highest capability for adhesion to p-Si in contrast with n-Si. Figure 1 demonstrates the example in this favour. Optical attenuation OA results are presented in Figures 2 and 3. All VLP types demonstrated the maximums of absorbance (ab. unit) at 235 and 260 nm. A
B
Figure 1. (A) SEM image of VLP3 on p-Si particle, (B) SEM image of VLP3 beside n-Si particle.
SEMICONDUCTOR AND BIOMOLECULAR SELFASSEMBLED SYSTEM 349
The Si particles in the buffer were characterized with the constant absorbance within the recorded wave length range. Mixture of VLP3 with p-Si did not have the maximum at 235 nm. However, the solution of VLP1 and VLP2 mixed n-Si did not have the same maximum. This means that the VLP1 and VLP2 were bonded to n-Si, in spite of VLP3 coupled with p-Si.
Figure 2. The OA spectra of VLP.
A
B
Figure 3. (A) OA spectra of the VLP mixture with p-Si. (B) OA spectra of the VLP mixture with n-Si.
350
YU. DEKHTYAR ET AL.
4. Conclusions The following conclusions were drawn from the investigation above. (a) VLP could be bonded to Si physically/chemically, (b) VLP1 and VLP2 demonstrates coupling with n-Si, but VLP3 – with p-Si nanoparticles, and (c) both p- and nSi nanoparticles could be in use as the carries of eligible VLP.
References 1. LaVan, D., Lynn D., and Langer, R. (2002) Moving smaller in drug discovery and delivery, Nat Rev Drug Discov 1, 77–84.
THERMAL STABILITY AND OPTICAL ACTIVITY OF ERBIUM DOPED CHALCOGENIDE GLASSES FOR PHOTONICS D. TONCHEV1*, K. KOUGHIA1, S.O.KASAP1, K. MAEDA2, T. SAKAI2, J. IKUTA2, AND Z.G. IVANOVA3 1 Department of Electrical Engineering (EE), Electronic and Photonic Materials Center, University of Saskatchewan, 57 Campus Drive, Saskatoon, S7N 5A9, CANADA 2 Department of Electrical Engineering, University of Miyazaki, Miyazaki, JAPAN 3 Institute of Solid State Physics, Bulgarian Academy of Sciences, Sofia, BULGARIA
Abstract – The glass transition and crystallization temperatures (Tg, Tc), heat capacity, thermal stability and glass uniformity of GeSGa, GeSeGa, Ge(SeTe)Ga chalcogenide glasses doped with Er3+ by the addition of Er2S3 have been investigated by conventional differential scanning calorimetry (DSC) and Temperature-Modulated DSC (TMDSC). While some of the glasses have two crystallization peaks, these glasses were nonetheless optically actively and uniform. Essential optical properties have been evaluated, such as the photoluminescence (PL) intensity and lifetime as a function of the glass composition. We present typical results to emphasize some of the important characteristics of these systems and discuss trends within a glass system; and also highlight differences between glass systems.
Keywords: Rare earth doping, chalogenide glass, DSC, photoluminescence.
1. Introduction Chalcogenide glasses offer a number of potential applications in photonics as reviewed by Zakery and Elliott.1 Among chalcogenide glasses, Ge and Ga containing glasses have been the subject of close interest, because of their ability
______ *
To whom correspondence should be addressed: D. Tonchev, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
351
352
D. TONCHEV ET AL.
to dissolve Er3+ ions, their high refractive indices, and relatively high glass transition and crystallization temperature as discussed recently.2 The purpose of this work is to examine the glass transition and crystallization (hence the thermal stability,) and the photoluminescence (PL) properties of GeSGa, GeSeGa and Ge(Se,Te)Ga glasses, and identify useful host compositions. We limit ourselves to selected properties to emphasize trends within a given glass system, and also mention differences between these three different systems. 2. Experimental Bulk glasses were prepared from predetermined mixtures of pure Ge, Ga, Se, S and Te synthesized by conventional rapid quenching of the melts in ice-water. For certain stoichiometric GeSGa compositions, GeS2 and Ga2S3 were initially synthesized as starting materials. Commercial Er2S3 was used as the dopant. DSC and TMDSC experiments were performed as described previously.3 The PL measurements were the same as described in Kasap et al. (2007).2 The composition, homogeneity and structure of the samples were checked by EDX/SEM, WDX/SEM electron microscopy and X-ray diffraction. 3. Results Figure 1 shows two crystallization exothermic peaks in the DSC thermogram of an Er3+-doped (GeS2)75(Ga2S3)25 glass. Such double peaks normally indicate separation out from the matrix by crystallization. The crystallization peaks are closer to each other for Er-doped samples than for non doped samples. While the higher Tc decreases with Er2S3 concentration, the lower Tc increases. 2.0
510.13°C
Heat Flow (W/g)
1.5
1.0
0.5
487.23°C
392.73°C
0.0
403.19°C(I) 476.94°C 34.40J/g
415.41°C
-0.5 360 Exo Up
380
400
420
440
460
Temperature (°C)
480
507.00°C 40.53J/g
500
520
540
Universal V3.7A TA Instruments
Figure 1. DSC of 2.1 mol % Er2S3 doped (GeS2)75(Ga2S3)25 glasses.
ERBIUM DOPED CHALCOGENIDE GLASSES FOR PHOTONICS
353
Heat capacity vs. temperature TMDSC curve (not shown) evinces only one Tg for these glasses with two crystallization peaks. There is no indication of the presence of any crystalline inclusions from X-ray diffraction studies. Further, even though there are two crystallization peaks, the Er3+-doped (GeS2)75(Ga2S3)25 can be nonetheless doped with up to 1.5 at. % Er3+ and the glasses are optically active as discussed below. The latter system appears homogeneous and comparatively stable, and can dissolve and activate Er3+ ions. Some of the GeSeGa and most of the Ge(SeTe)Ga (with Te < 10 at. %) glasses show similar tendencies. We observe that both Tg and Tc in (Ge30(SeS)1-x(Te)x)94Ga6 glasses (with some S only from Er2S3) as Te is substituted for Se remain relatively unchanged and then rapidly decrease at Te > 5 at. % . In addition, at high Te concentrations (>10 at. %) we observe two crystallization peaks; a possible indication of phase separation. We have also observed crystalline inclusions on the X-ray diffraction pattern, and significantly reduced optical activity. Thus (Ge30(SeS)1-x(Te)x)94Ga6 glasses with 5 at. % Te or higher amounts of Te are less stable. The crystallization and glass transition vs. Ga content behavior in (GeSe2)1-y (Ga2Se3)y, y = 0 0.3, glasses exhibit less stable regions as the Ga content becomes large. The structure of GeSeGa glasses were studied by DSC, Raman scattering, and XPS.4 The Ga content beyond which the (GeSe2)1-y(Ga2Se3)y glasses become less stable, where Tc Tg shows a sharp fall, is about 12 at. % Ga. We have also observed a good correlation between thermal and structural properties and optical properties (PL intensity) as shown in Figure 2.
PL Intensity (a.u.)
0
2
4
6
8
10
12
14
16
18
G a C o n t e n t ( a t .% )
Figure 2. Photoluminescence intensity dependence of 1 at. % Er3+ doped (GeSe2)1x(Ga2Se3)x chalcogenide glasses at 1,538 nm on Ga (at. %) content. The dashed line is a guide to the eye.
354
D. TONCHEV ET AL.
PL intensity increases with the Ga content as more Er3+ become activated until the Ga content reaches about 10–12 at. %. Beyond the latter value there is a sharp fall in the PL intensity, and the structure is less stable. Er3+-doped GeSGa, GeSeGa and Ge(SeTe)Ga glasses exhibit the expected characteristic PL emission spectra, as shown, as an example, in Figure 3, for 1.8, 2.1 and 2.4 mol % Er2S3 doped (GeS2)75(Ga2S3)25 glasses.
PL intensity (arb. un.)
1 .0
E r2S
3
0 .8 0 .6
m ol % 2 .4 2 .1 1 .8
0 .4 0 .2 0 .0 1450
1500
1550
1600
1650
W a v e le n g th ( n m ) Figure 3. Examples of photoluminescence spectra from Er3+-doped (GeS2)75(Ga2S3)25 glasses.
The PL line-shape is virtually independent of the doping level, with minor variations which can be attributed to self-absorption effects, as discussed elsewhere.2 Most of the erbium doped GeSeGa and GeSGa-based glass compositions we have fabricated were optically active and have shown excellent photoluminescence properties, provided the Ga content in the structure was sufficient. In particular stoichiometric compositions have shown better thermal and optical properties. We have observed a good correlation between the maximum active Er3+ dopants that can be dissolved homogenously and the Ga content in GeSGa and GeSeGa glasses. In the case of GeSeGa glasses the Ga to Er3+ concentration ratio should be typically more than ~6; the glass containing 6 at. % Ga can dissolve up to about 1 at. % Er3+. We have also measured the PL lifetime, and found values in the milliseconds range for both Er2S3 doped stoichiometric GeSGa and GeSeGa compositions; such long lifetimes are obviously highly desirable for optical amplifier applications, along with high a concentration of dissolved and activated Er3+. Table 1 summarizes some of the important experimental observations.
ERBIUM DOPED CHALCOGENIDE GLASSES FOR PHOTONICS
355
TABLE 1. Summary of various properties of GeSGa, GeSeGa, and Ge(SeTe)Ga glasses doped with Er2S3. Glass system
Properties
(GeS2)1y(Ga2S3)y Stoichiometric y = 0.2, 0.25
Can be doped with up to 1.5 at. % Er3+. Tg |420 qC (y = 0.25). Two crystallization peaks when non-stoichiometric glasses but one Tg. Characteristic PL emission band, which increases with Er up to 1.5–2 at. % Er3+; and good PL lifetime 1.5–3.5 ms. Tg initially decreases from 410qC to 370qC from 0 to 2 at. % Er3+, and then constant between 2–12 at. % Ga. Beyond 12 at. % Ga, stability decreases rapidly with Ga. Er3+ dissolution and activation depends on the [Ga]/[Er3+] ratio, which should be more than 6. Characteristic PL emission band and good PL lifetime 1–2 ms, up to 2 at. % Er3+. Non-stoichiometric glasses with two Tc and one Tg. Tg and Tc remain constant up to 1 at. %Te, and then both decrease linearly with Te. Thermal stability decreases as Te content increases. Non-stoichiometric Te content samples have two crystallization peaks. Low optically activity when X-ray detected crystalline state.
(GeSe2)1y(Ga2Se3)y Stoichiometric y = 0–0.3
(Ge30 (Se1-xTex)70)94Ga6 doped with 1 at. % Er
4. Discussion Based on DSC, Raman scattering, and XPS measurements, Maeda and Lucovsky4 have recently discussed the structure of stoichiometric (GeSe2)1y(Ga2Se3)y glasses as the Ga content is increased. Most important compositional range is the stoichiometric range that is able to dissolve substantial amounts of Er3+. The reason for these host glasses being able to dissolve Er3+ readily can be easily understood by noting the stoichiometric composition behaves as a pseudobinary alloy (GeSe2)1-y (Ga2Se3)y, in which the two-fold coordinated Se atoms can bridge pairs of Ge-atoms, pairs of Ga-atoms, and Ga and Ge atoms as well. The structure has neutral GeSe4/2 groups, negatively charged GaSe4 (or equivalently Se3GaSe1) groups, and Ga3+ ions. In the absence of Er3+, three Se3Ga Se1 groups are required to neutralize each six-fold coordinated Ga3+ ion. The bonding coordination of Er in oxides, sulfides and selenides is 6 (six),5 the same as the coordination of Ga in the alloys. In addition, the formal chemical valence of Er is 3, the same as Ga. The case of Er incorporation into the (GeSe2)1-y (Ga2Se3)y alloys, and the difficulty of incorporating trivalent Er into GeSe2, derives from the differences in formal chemical valence between Er and Ge, 3 as opposed to 4, as well as the differences in bonding coordination, Er = 6 as opposed to Ge = 4, and the differences in the electronegativities of Er and Ge, about 1.2 for Er, and 1.8 for Ge. The incorporation of Er into the alloys
356
D. TONCHEV ET AL.
may be best understood by treating the Er doped system as a pseudoternary alloy: (GeSe2)1-y-G(Ga2Se3)y (Er2Se3)G, where G << y, for example 0.01 compared to 0.2–0.3. The relative values of the electronegativity will force Er into the system as an Er3+ ion, with the compensation negative charge residing on the negatively charged GaSe4 (Se3GaSe1) groups. As mentioned above, the glasses in which Te substitutes for Se are less stable. It is well known that the atomic size has a direct influence on the glass stability. While both (GeS2)1y(Ga2S3)y and (GeSe2)1-y(Ga2Se3)y glasses are relatively stable, the (Ge30 (Se1xTex)70)94Ga6 glasses are less stable and form crystals at high Te compositions because of the large atomic size of Te; As2Te3 is a poorer glass former than As2Se3. Most of Te containing glasses we have prepared contained some degree of micro and nanocrystalline inclusions which differed in size and concentration in the amorphous matrix. Nonetheless these glasses including those which exhibited two crystallization peaks (except >10 at. % Te) were optically active and so were (GeS2)1-y(Ga2S3)y glasses. 5. Conclusions Both stoichiometric GeSGa and GeSeGa glasses can be very good hosts for Er3+ dopants. We can dissolve up to 2 at. % Er3+ in GeSGa and in GeSeGa glasses. Most of GeSGa and Ge(SeTe)Ga based chalcogenide glasses we have studied have two crystallization peaks but only a single heat capacity step change type of glass transition. GeSGa glasses are nonetheless optically active when doped with Er2S3, and exhibit typical PL spectra and good lifetimes in the milliseconds range. Replacement of Se with Te in GeSeGa glasses leads to less stable glasses in which both Tg and Tc decrease with the Te content beyond 1 at. % Te addition. The Ga content in these glasses is critical to dissolving and activating the Er3+. At high Ga concentrations, beyond about 12 at. % Ga, the GeSeGa glasses become less stable.
References 1. Zakery, A., and Elliott, S. (2003) J. Non-Cryst. Solids, 330, 1 and references therein. 2. Kasap, S., Koughia, K., Munzar, M., Tonchev, D., Saitou, D., and Aoki, T. (2007) J. Non-Cryst. Solids, 353, 1364 and references therein. 3. Maeda, K., Sakai, T., Tonchev, D., Munzar, M., Ikari, T., and Kasap, S. (2005) Mater. Sci. Eng. B, 122, 20. 4. Maeda, K., Sakai, T., Sakai, K., Ikari, T., Munzar, M., Tonchev, D., Kasap, S., and Lucovsky, G. (2007) J. Mater. Sci. Mater. Electron, 18, 367–370. 5. Cotton, F., and Wilkenson, G. (1972) Advanced Inorganic Chemistry, third ed. Interscience, New York.
XRD STUDY OF PULSED LASER DEPOSITED AlN FILMS WITH NANOSIZED CRYSTALLITES S. BAKALOVA1*, A. SZEKERES1, A. CZIRAKI2, E. GYORGY3, S. GRIGORESCU3, G. SOCOL3, AND I.N. MIHAILESCU3 1 Georgi Nadjakov Institute of Solid State Physics, Bulgarian Academy of Sciences, Tzarigradsko Ch 72, Sofia, BULGARIA 2 Eotvos Lorand University, Faculty of Solid State Physics, 1 Pazmany Peter str., 1117 Budapest, HUNGARY 3 Laser-Surface-Plasma Interactions Laboratory, Lasers Department, National Institute for Lasers, Plasma and Radiation Physics, PO Box MG-54, 77125 Bucharest-Magurele, ROMANIA
Abstract – The structure of pulsed laser deposited AlN films was investigated by X-ray diffractometry. The AlN films were deposited on (111) single-crystalline Si wafers in ambient nitrogen at a pressure of 0.1 Pa via ablation of an AlN target using KrF* excimer laser radiation (248 nm wavelength, t >= 7 ns) with 3.7 J/cm2 incident fluence. The obtained films had a polycrystalline structure with cubic phase nanocrystallites. The size of the crystallites, as estimated from the Bragg peaks, was about 55 nm slightly depending on the post-deposition cooling rate.
Keywords: Aluminium nitride, pulsed laser deposition, XRD, crystalline structure, nanocrystallites, XPS analysis.
1. Introduction Aluminium nitride (AlN), a wide-bandgap semiconductor material, possesses great potential for versatile applications ranging from electronics, acoustic wave devices, and photonic devices to antiwear coatings. Obtaining AlN films with
______ *
To whom correspondence should be addressed: S. Bakalova, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
357
358
S. BAKALOVA ET AL.
definite structure and crystalline quality is still a difficult task for most deposition techniques. Some comprehensive research is thus necessary to find out how exactly the crystallographic structure and degree of crystallinity of AlN films depend on the film preparation method and conditions. AlN generally crystallizes in either the hexagonal wurtzite (h-AlN) structure or the cubic zinc-blende (c-AlN) polytype. The wutrzite structure shown in Figure 1 differs from the cubic one mainly by the relative position of the third neighbours and beyond. The cubic phase has been reported to be theoretically metastable.1 Theoretical investigations of the thermal conductivity of AlN have shown that the energy difference between two kinds of phase structure of AlN is so small that both hexagonal and cubic phases are able to coexist in a realistic 2 material. Still, relevant data on the formation of each of the crystalline phases of AlN have been scant so far.
AI AI
AI N 107.7˚ AI N
N N
110.5˚
Figure 1. Locations of the Al and N atoms in the wurtzite structure.
AlN films deposited by reactive radio frequency and coherent magnetron sputtering have been reported to exhibit c-axis oriented hexagonal structure.3,4 A highly oriented (0001) polycrystalline h-AlN structure has been obtained by metal organic chemical vapour deposition.5 Films of AlN hexagonal polytype have been deposited by molecular-beam epitaxy.6 Pulsed laser deposition (PLD) method has resulted in AlN material in either cubic or hexagonal phase, depending on the preparation conditions. By varying deposition temperature and ambient pressure, different crystalline structures have been obtained. Vispute et al. reported the synthesis of h-AlN with (0001) orientation,7 while Wen-Tai Lin et al. obtained cubic AlN films.8 Both groups used silicon wafers as substrate. The metastable c-AlN was found in films synthesized by nitrogen-ion-assisted PLD9 and by solid state reaction.10
PLD AlN FILMS WITH NANOSIZED CRYSTALLITES
359
Our recent investigations are focused on the preparation of AlN films by pulsed laser deposition and their properties. In this study, PLD AlN thin films were analyzed by X-ray diffractometry (XRD), and the influence of the postdeposition cooling rate on the crystallographic structure and crystallite size was discussed. 2. Materials and Methods 2.1. SAMPLE PREPARATION
AlN films were synthesized by laser ablation from bulk AlN target by a KrF* excimer laser operating at 248 nm with a pulse duration of 7 ns, energy fluence of 3.7 J/cm2 and repetition rate of 2 Hz. To avoid piercing, the target was both rotated (0.3 Hz) and translated (along two orthogonal directions). Depositions were performed in low-pressure nitrogen at a dynamic pressure of 0.1 Pa. Single crystalline substrates of (111)Si were heated in the vacuum chamber (5 × 10-4 Pa) up to 800°C prior to deposition in order to remove the native oxide11 and were kept at this temperature during deposition providing good conditions for the crystalline growth of AlN.7 After applying 8,000–15,000 laser pulses, the samples were cooled down to room temperature with an average cooling rate of either ~25°C/min (fast cooling) or ~5°C/min (slow cooling). Details on the processing, including pulse numbers, cooling rates, and also the film thickness are presented in Table 1. TABLE 1. Number of laser pulses applied, post-deposition cooling rates and the corresponding films thickness. Sample series A B C
Laser pulses
Cooling rate (°C/min)
Thickness (nm)
8,000 15,000 20,000
25 25 5
560 1,200 1,400
2.2. DIAGNOSTICS
The crystalline phases in the AlN films were determined with a large-angle (24 = 0–90°) X-ray Philips X’Pert diffractometer working with Cu radiation (Ȝ = 0.154056 nm). The XRD patterns were identified using the ASTM database.12 Assuming a spherical shape of the crystallites, their average grain size was deduced from the Bragg peaks with Sherrer’s formula.13
360
S. BAKALOVA ET AL.
Chemical analysis using X-ray Photoelectron Spectroscopy (XPS) technique was applied to get information about the chemical bonding. 3. Results and Discussion For crystalline AlN there are two types of lattice structures: hexagonal and cubic ones. The deposited films crystallize in these two phases but in different proportions. Typical X-ray diffraction patterns of the PLD films are shown in Figure 2. These spectra are representative of polycrystalline structure with predominant cubic phase of AlN crystallites, the corresponding lattice parameter being a = 0.4045 nm (ASTM 46-120012). The hexagonal phase is also observed in the case of fast cooling and greater thickness – series B (see Figure 2b), but it is quite possible, that hexagonal crystallites are also presented in the thinner films of series A. However, due to the smaller thickness, the Bragg peaks corresponding to these crystalline phases are very weak for the thinner samples and the signal-to-noise ratio is greater (Figure 2a). For deposition of the films followed by slow post-deposition cooling (series C), the XRD patterns show the presence of solely cubic phase (Figure 2c) and a (111) preferential orientation of the cubic crystallites. Bragg peaks related to the hexagonal crystalline phase are not detected in this XRD spectrum. The shape of the Bragg peaks reveals the presence of residual non-uniform strains and/or crystallites with sizes in the nanometres range. Concerning the internal strains we do not have enough number of peaks to estimate the stress contribution to the peak broadening. However, we observe asymmetry of the XRD peaks, which is an indication of an inhomogeneous film structure. This is clearly shown in Figure 2b, where the splitted C(002) peak is magnifies and inserted into the figure. The related inhomogeneities may have been caused by lattice parameter or composition variations in the films. We deconvoluted the C(002) peaks using a Philips Pro’Fit commercial program in order to determine the lattice parameter difference and the corresponding crystallite size. Precise calculations of the crystallite size could not be made due to the instrumental sources of error and the contribution of the residual stress to the peak broadening. However, rough estimation can be made from the most intensive peaks by applying the Sherrer formula and assuming spherical shape of the crystallite grains. The crystallite size values, as inferred from the FWHM of the strongest C(111) XRD peaks, are given in Table 2. For the different sample series we obtained crystallite size of 48.9–55.4 nm with an accuracy of ±5 nm.
PLD AlN FILMS WITH NANOSIZED CRYSTALLITES
pulse=8000 o v=25 C/min
C
(111)
35
C (002)
Count (arb. unit)
a)
361
40
45
50
24 (degree) pulse=15000 o v=25 C/min
C
(002) (111)
Count (arb. unit)
b)
H
35
(002)
C
40
45
50
24 (degree) Counts (arb. unit)
c)
C(111)
pulse=20000 o v=5 C/min
C (111) peak
(002)
C
35
40
45
50
24(degree) Figure 2. XRD patterns of AlN films deposited with different number of laser pulses and at different post-deposition cooling rates.
S. BAKALOVA ET AL.
362
TABLE 2. Size of crystallites grown along C(111) preferential orientations. Sample series
A (fast cooling) B (fast cooling) C (slow cooling)
Peak center (24) 38.36 38.36 38.39
Lattice parameter (nm r 0.0005 nm) 0.4060 0.4064 0.4057
FWHM (24) 0.11 0.17 0.15
Crystallite size (nm r 5 nm) 55.4 48.9 54.2
For series B samples, as we mentioned above, the C(002) peak was splitted and, therefore, the crystallite size was calculated after deconvolution of this peak into two ones. The estimated diameter of the crystallite grains were 58.7 and 82.6 nm, respectively. The corresponding lattice parameters varied from 0.4066 to 0.4053 nm, most probably due to non-uniform strains in the structure. As mentioned in the Introduction, during PLD AlN films crystallize in either cubic or hexagonal phase depending on experimental conditions.7,8 The XRD analysis of samples revealed that under given deposition conditions we obtain AlN films with dominantly cubic structure. This is similar to results of Wen-Tai Lin et al.8 performing experiments at a Si substrate temperature of 630°C and N2 pressure of 20 Pa. In contrast other pulsed laser depositions at 700°C and in NH3 ambient7 resulted in hexagonal AlN films structure. The obtained AlN films were characterized with XPS in order to determine the nature of the chemical bonds. We measured the core level spectra of Al 2p and N 1s (Figure 3). The corresponding binding energies (BE) are presented in Table 3. For the slow cooled films we registered a maximum at 73.89 eV in the XPS spectrum. This energy position is reported to correspond to Al (2p3/2) spin-orbital split component of the transition in Al chemically bonded to N in AlN.14 The same observation was made in the case of fast cooling, where the Al peak appearing at 74.3 eV can be attributed to Al in AlN.15 However, we cannot exclude the possibility that at the surface we have also some oxygen and OH traces, because at this position of 73.9 eV Wagner et al.16 have attributed this maximum to Al (2p3/2) bonded with OH radical in Al(OH)3 configuration and Taylor17 has related the 74.3 eV maximum to Al in Al(OH)3 bond configuration. The maxima in the N 1s core level spectra (see Figure3a and b) appeared at 396.61 and 396.66 eV for fast and slow cooling rates, respectively, corresponded to N bonds in either AlN14 or NO/Al bond configurations reported in the literature.18 Taking into account that by the XRD method we did not detect any peaks corresponding to oxygen compounds in the films, we can conclude that even if there is any oxygen related material it is limited to the surface region only.
PLD AlN FILMS WITH NANOSIZED CRYSTALLITES
N 1s
20000 pulses slow cooling
Intensity (arb. unit)
Intensity (arb. unit)
a)
Al 2p
404 400 396 392
15000 puses fast cooling
b)
Al 2p
N 1s
404 400 396 392
80 78 76 74 72 70 68 66
Binding energy (eV)
363
8078767472706866
Binding energy (eV)
Figure 3. The XPS spectra of N 1s and Al 2p core levels in the slow cooled AlN samples (a) and fast cooled ones (b). TABLE 3. The BE of the peaks of Al 2p3/2 and N 1s in the case of slow and fast cooling and the related chemical bonding. Core level
Cooling rate
Al 2p3/2
Slow – 5°C/min
73.89
Fast – 25°C/min
74.34
Slow – 5°C/min Fast – 25°C/min
396.61 396.66
N 1s
BE (eV)
Bonding and related BE (eV) AlN Al(OH)3 AlN Al(OH)3 NO/Al AlN
– – – – – –
73.9 73.9 74.4 74.3 396.8 397.3
4. Conclusions The analysis of the XRD patterns of AlN films produced by low nitrogen pressure pulsed laser deposition at a nitrogen pressure of 0.1 Pa has revealed a cubic film structure with preferential (111) orientation of the crystallites. The average crystallite size was about 50 nm. Post-deposition cooling rates only affected the structural ordering. More specifically, fast quenching resulted in more disordered film structure. The chemical binding energies of the film surface were determined by XPS analysis and they were assigned to Al and N core levels, related to AlN, Al(OH)3 and NO/Al compounds.
364
S. BAKALOVA ET AL.
ACKNOWLEDGEMENTS
The authors acknowledge with thanks the support of this work under the 2004– 2007 Collaboration Agreement between the Bulgarian Academy of Sciences and Romanian Academy of Sciences.
References 1. Paisley, M., and Davis, R. (1993), J. Cryst. Growth. 127, 136–142. 2. Kitagawa, H., Shibutani, Y., and Ogata, S. (1995), Model. Simul. Mater. Sci. Eng. 3, 521–531. 3. Ho, C., Shing, T., and Li, P. (2004), Tamkang J. Sci. Eng. 7, 1–4. 4. Chu, K., Chao, C., Lee, F., and Huang, H. (2001), J. Electron. Mater. 30(1), 1–5. 5. Boo, J., Lee, S., Kim, Y., Park, J., Yu, K., and Kim, Y. (1999), Phys. Stat. Sol. (a) 176(1), 711–717. 6. Kipshidze, D., Schenk, H., Fissel, A., Kaiser, U., Schulze, J., Richter, W., Weihnacht, M., Kunze, R., and Kräusslich, J. (1999), Semiconductors 33(11), 1241–1246. 7. Vispute, R., Narayan, J., Wu, H., and Jagannadham, K. (1995), J. Appl. Phys. 77(9), 4724–4728. 8. Lin, W., Meng, L., Chen, G., and Liu, H. (1995), Appl. Phys. Lett. 66(16), 2066–2068. 9. Ren, Z., Lu, Y., Ni, H., Liew, T., Cheong, B., Chow, S., Ng, M., and Wang, J. (2000), J. Appl. Phys. 88(12), 7346–7350. 10. Petrov, I., Mojab, E., Powell, R., Greene, J., Hultman, L., and Sundgren, J. (1992), Appl. Phys. Lett. 60(20), 2491–2493. 11. Miyata, N., Shigeno, M., Arimoto, Y., and Ito, T. (1993), J. Appl. Phys. 74(8), 5275–5276. 12. Index to the Powder Diffraction File (2000), Published by Joint Committee on Powder Diffractions Standards. 13. Klug, H., and Alexander, L. (1962), X-Ray Diffraction Procedures, Wiley, New York, p. 491. 14. Taylor, J., and Rabalais, J. (1981), J. Chem. Phys. 75(4), 1735–1745. 15. McGuire, G., Schweitzer, G., and Carlson, T. (1973), Inorg. Chem. 12(10), 2450–2453. 16. Wagner, C., Passoja, E., Hillery, H., Kinisky, T., Six, H., Jansen, W., and Taylor, J. (1982), J. Vac. Sci. Technol. 21, 933. 17. Taylor, J. (1982), J. Vac. Sci. Technol. 20, 751. 18. Pashutski, A., and Folman, M. (1989), Surf. Sci. 216(3), 395–408.
FUNCTIONALIZATION OF MULTI-WALLED CARBON NANOTUBES (MWCNTS) M. MOHL, Z. KÓNYA*, Á. KUKOVECZ, AND I. KIRICSI Department of Applied and Environmental Chemistry, University of Szeged, Rerrich B. tér 1., 6720 Szeged, HUNGARY
Abstract – The surface of the CNTs can be chemically modified to impart a specific desired property. In general functionalization has usually been done by oxidation using HNO3, KMnO4, etc. The defects and the ends of the CNTs are thus functionalized by carboxyl groups. The current focus of the research is to develop a convenient and fast method for the functionalization of carbon nanotubes. We explore a covalent chemical strategy for functionalization of MWCNTs with amine-containing groups. The chemical interaction of the amine-containing group and the oxidized CNTs surface is confirmed by FT-IR and XPS.
Keywords: MWCNT, functionalization, amine-containing groups.
1. Introduction Carbon nanotubes are extremely promising materials for applications in materials science and medicinal chemistry. CNT consist of graphitic sheets, which have been rolled up into a cylindrical shape. Carbon nanotubes can be considered as attractive candidates in nanotechnological applications such as sensors or molecular tanks. However, the lack of solubility and the difficult manipulation in any solvents have imposed great limitation to the use of CNTs.1 2. Experimental The MWCNTs were prepared by a catalytic vapor decomposition process at 650ºC using acetylene as carbon source and 2.5% Co,Fe/MgO as catalyst.
______ *
To whom correspondence should be addressed: Z. Kónya, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
365
366
M. MOHL ET AL.
MWCNTs were first refluxed first in concentrated HNO3 for 24 h and second in H2O2 for 12 h. Afterwards the MWCNTs were sonicated in deionized water and KMnO4 solution was added drop wise. After 10 min vigorous stirring, citric acid solution was added to quench the KMnO4. It was filtrated, washed with deionized water and dried (A).2 To convert the carboxylic groups into – NH2 functional groups CNTs were first reacted with SOCl2 and after that with 1,8-diaminooctane (B). 3-aminopropyl-triethoxysilane (APTES) functionalized CNTs were prepared as follows. The as grown CNTs were first oxidized as described above, and then they were reacted with APTES dissolved in acetone under ultrasonic treatment (C).3 Rh containing CNTs were prepared as follows. The as grown CNTs were first oxidized as described above and after that they were treated with aqueous solution of Rh(III) in an ultrasonic bath for 2 h. Afterwards aqueous hydrogen peroxide solution was added drop wise under vigorous stirring and refluxed for 4 h at 80°C. The precipitate was filtered and washed by deionized water to neutral (D). 3. Result and Discussion The resultant composites were characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), scanning electron microscopy (SEM) and electron dispersion X-ray (EDX) analysis. The Figure 1 shows the C 1s and O 1s core levels of the oxidized CNTs (A) and the amine-functionalized CNTs (B). The spectra corresponding to the oxidized sample show a large peak at 284.4 eV (C = C bonds), a smaller peak at 291.0 eV (O-COO or –COO bonds), another large peak at 531.6 eV (C=O bonds) and a smaller at 533.1 eV (O-C bonds). In the aminated samples (Figure 1B) the following bonds were assigned: 284.3 eV (C = C), 535.8 eV and 532.1 eV (C = O or/and O-C bonds) and 401.1 eV (-NH2 bonds).4 In Figure 2A and B, the cylindrical structure of CNTs are clearly revealed and show that the sample does not contain any bundles or aggregation of APTES. On TEM image (Figure 2B), RhO2 particles can not be seen on the surface of the CNTs, probably caused by their very small size. Figure 3 shows the FT-IR spectra of oxidized (A), aminated (B) and APTESfunctionalized (C) samples. Acidic treatment resulted in the appearance of a new peak at 1,715 cm-1 (Figure 3B) that corresponds to the ›C=O stretching indicating the introduction of carboxylic groups. The peak around 1,640 cm-1 (Figure 3) -NH bending vibration of amine group. The large peak at 3,436 cm-1 can be assigned to the vibrational modes of the –OH groups. The appearance of new peaks: 1,117 cm-1 and 1,043 cm-1 corroborated to Si-O-Si stretching and
FUNCTIONALIZATION OF CARBON NANOTUBES
367
Si-O-C stretching can be due to the chemical interactions between silane and CNT. The XRD analysis result shown in Figure 3 demonstrates that the crystal structure of CNTs remains unchanged after supporting the rhodium oxide nanoparticles.
Figure 1. XPS plots of the oxidized CNT (A) and amine-functionalized sample (B).
Figure 2. TEM images of APTES functionalized nanotubes (left and middle frames) (C) and RhO2/CNT sample (right frame) (D).
368
M. MOHL ET AL.
Figure 3. FT-IR spectra of (A) oxidized, (B) aminated, and (C) APTES functionalized CNT and XRD diffraction patterns of RhO2 /CNT.
Figure 4. EDX analysis of APTES functionalized sample (left frame) (C), RhO2 /CNT sample (right frame) (D).
The EDX investigation of APTES-functionalized CNTs shows a high concentration of Si (Figure 4A), while on Figure 4B the presence of Rh can be seen. 4. Conclusion The functionalization methods described in this article have proved to be successful for MWCNTs. All the functionalized CNTs can prove to be suitable for gas sensor applications.
References 1. Tasis, D., Tagmatarchis, N., Bianco, A., and Prato, M. (2006) Chem. Rev. 106, 1105–1136 2. Kordas, K., Mustonen, T., Toth, G., Jantunen, H., Lajunen, M., Soldano, C., Talapatra, S., Kar, S., Vajtai, R., and Ajayan, P. (2006) Small 2, 1021–1025 3. Shanmugharaj, A., Bae, J., Lee, K., Noh, W., Lee, S., and Ryu, S. (2006) Comp. Sci. Technol. 67, 1813–1822 4. Okpalugo, T., Papakonstantinou, P., Murphy, H., McLaughlin, J., and Brown, N. (2005) Carbon 43, 153–161.
SONOCHEMICAL SYNTHESIS OF INORGANIC NANOPARTICLES J. KIS-CSITÁRI1,2, Z. KÓNYA2* AND I. KIRICSI1,2 Bay Zoltán Foundation for Applied Research, Pf.:46, 3515 Miskolc-Egyetemváros, HUNGARY 2 Department of Applied and Environmental Chemistry, University of Szeged, Rerrich B. tér 1, 6720 Szeged, HUNGARY 1
Abstract – Sonochemistry has applications across almost the whole breadth of chemistry. A number of theories were developed in order to explain how a 20 kHz sonic radiation can break chemical bonds. They all agree that main event in sonochemistry is cavitation, in other words the creation, growth and collapse of bubbles (so called hot spots) formed in the liquid. These hot spots have temperatures of roughly 5,000°C, pressures of about 500 atmospheres, and heating and cooling rates greater than 109 K/s. The enormous local temperature, pressure and the extraordinary heating/cooling rates generated by cavitational collapse provides an unusual mechanism for generating high energy chemistry. A novel sonochemical method for the continuous preparation of metal- and metal-oxide nanocrystalline materials has been developed. The products were characterized by transmission electron microscopy. We observed that the choice of the source metal salt, the reactant and the optimal usage of high-power ultrasound are both important in the formation of the nanostructures.
Keywords: Sonochemistry, cavitation, nanoparticles.
1. Introduction The chemical effects of ultrasound do not derive from a direct coupling of the acoustic field with chemical species on a molecular level. Instead, sonochemistry and sonoluminescence derive principally from acoustic cavitation: the formation, growth and implosive collapse of bubbles in liquids irradiated with highintensity ultrasound. Bubble collapse during cavitation serves as an effective means of concentrating the diffuse energy of sound: compression of a gas
______ *
To whom correspondence should be addressed: Z. Konya, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
369
370
J. KIS-CSITÁRI, K. KÓNYA AND I. KIRICSI
generates heat. When the compression of bubbles occurs during cavitation, heating is more rapid than thermal transport, creating a short-lived localized hot spot. The chemical effects of ultrasound can be divided into three general types: neat liquids; heterogeneous liquid-liquid; and heterogeneous liquid-solid systems. Over the past few years, the synthesis of inorganic materials has developed as one of the most important applications of sonochemistry.1 2. Experimental 2.1. SYNTHESIS OF MAGNETITE NANOPARTICLES
The synthesis of magnetite nanoparticles was carried out as follows: distilled water solution of FeCl2 and FeCl3 was sonicated using a high-intensity ultrasounic horn (Hielscher Ultrasound Homogenizater UIP1000, 20 mm,Ti horn, 20 kHz). During the sonication oleic acid was added. At the end of the synthesis a black solution of iron oxide was obtained. It was filtrated, washed with distilled water several times and the particles separeted by centrifugation. 2.2. SYNTHESIS OF SILVER NANOPARTICLES
The synthesis of polyvinylpyrrolidon (PVP)-stabilized silver nanoparticles was carried out as follows: distilled water solution of PVP was sonicated using high intensity ultrasonic horn (Hielscher Ultrasound Homogenizater UIP1000, 20 mm,Ti horn, 20 kHz). Solution of AgNO3 and NaBH4 were added and sonicated the mixture for 10 min. At the end of the synthesis a grey solution of silver was obtained. It was filtrared and washed with distilled water several times. 2.3. SYNTHESIS OF COPPER NANOPARTICLES
The synthesis of polyvinylpyrrolidon (PVP)-stabilized copper nanoparticles was carried out as follows: distilled water solution of PVP was sonicated using high intensity ultrasonic horn (Hielscher Ultrasound Homogenizater UIP1000, 20 mm,Ti horn, 20 kHz). Solution of CuSO4 and NaH2PO2 were added and sonicated the mixture for 10 min. A brown solution of copper nanoparticles was obtained which filtrated and washed with distilled water.
SONOCHEMICAL SYNTHESIS OF NANOPARTICLES
371
2.4. SYNTHESIS OF COBALT-OXIDE
The cobalt-oxide was prepared as followed: distilled water solution of (NH4)2SO4 was sonicated using high intensity ultrasounic horn (Hielscher Ultrasound Homogenizater UIP1000, 20 mm,Ti horn, 20 kHz). CoSO4 and Zn were added and sonicated the mixture for 10 min. A green-grey solution of cobalt-oxide was obtained. It was filtrated and washed with distilled water. 2.5. CHARACTERIZATION
For the characterization of the nanoparticles a Philips CM 10 transmission electron microscopy was used. 3. Results and Discussion It is easier said than done to prepare nanoparticles using sonochemistry. Easy, because the sonotrode gives enough energy to force the reaction between the reactants, however, to control the reaction by itself is not an easy job. In order to synthesize nanoparticles with narrow size distribution, we have to control several reaction parameters, like reaction temperature, ultrasonic energy, ultraonic amplitude, reaction time, etc. Usually the formation of the nanoparticles can be divided into two steps; the first is the reaction between the reactants, while the second is the actual growng of the nanoparticles.2 We have found that the most important question is the rate of the reaction. If we can separate the two main steps, we can easily and precisely control the size and the shape of the particles.3 In the sonochemistry owing to its very high energy this approximation is a hard task since it is almost impossible to separate the two main steps. We have found that one of the most important questions is to properly choose the reactant; if the basic reaction is too fast, the sonochemical effect is hardly controllable. However, if we can find a reaction, which is running only if we use ultrasound, but then the rate of the reaction is very fast, and we can introduce a proper strucure directing agent, the reaction and the growing of the nanoparticles can be quasi separated. Figure 1 shows different nanoparticles synthesized by sonochemistry driven reactions. The size distribution of the nanoparticles is quite good. According to the TEM measurements, the average particles sizes are 8.4, 16.3, 28 and 137 nm for magnetite, silver, copper and cobalt-oxide, respectively. It can be conluded that for magnetite and silver, we can precisely control the reaction, while for copper and cobalt-oxide, we still have to work on the synthesis parameters.
J. KIS-CSITÁRI, K. KÓNYA AND I. KIRICSI
372
(a)
(b)
(c)
(d)
Figure 1. TEM images of magnetite (a), silver (b), copper (c) and cobalt-oxide (d) nanoparticles prepared by sonochemical method.
4. Conclusions We reported in this short communication the one step sonochemical synthesis of magnetite, silver, copper and cobalt-oxide nanoparticles. We have found that controlling the reaction precisely a quite narrow size distribution of the nanoarticles can be achieved. Although still much work remains to be done to find the best synthesis paraeters in order to optimize the reactions, it is a simple and efficient way to produce nanoparticles, which may find applications in many fields.
References 1. Suslick, K., Didenko, Y., Fang, M., Hyeon, T., Kolberg, K., McNamara, W. III, Mdleleni, M., and Wong, M. (1999), Philos. Trans. Roy. Soc. Lond A 357, 335–353. 2. Puntes, V., Krishnan, K., and Alivisatos, A. (2001), Science 291, 2115–2117. 3. Konya, Z., Puntes, V., Kiricsi, I., Zhu, J., Alivisatos, A., and Somorjai, G. (2002), Nano Lett., 2, 907–910.
NOVEL TRANSPARENT MOLECULAR CRYSTALS OF CARBON G. KHARLAMOVA1*, N. KIRILLOVA2, A. KHARLAMOV2, AND A. SKRIPNICHENKO2 1 Kiev National Taras Shevchenko University, 14 Glushkov str., room. 89, 03187 Kiev, UKRAINE 2 Frantsevich Institute for Problems of Materials Science, NAS of Ukraine, 3, Krjijanovskogo str., 03680 Kiev, UKRAINE
Abstract – Novel molecular transparent thread-like crystals of carbon at evaporating powdery carbon and transformation of molecules of aromatic hydrocarbons are obtained.
Keywords: Carbon threads, SiC nanostructures, exothermal nanosynthesis, sublimation.
1. Introduction The great variety of compounds of carbon, in particular with hydrogen, is caused, first of all, ability of its atom to form chains ɋ-ɋ – bonds. Directions of these bonds, as believed earlier, are strictly fixed. In diamond and molecule of methane atoms are coupled by means of only sp3 – bonds, and in graphite and in a molecule of benzene are coupled by means of only sp2 – bonds. However opening of new allotropic state of carbon, in particular fullerenes and carbon nanotubes, convincingly have shown, that the corners between ɋ-ɋ – bonds can be a little bit distinct from standard. In molecule of phenyl ethylene C8H8 and spheroidal molecule of hydrocarbon – dodecahedrane C20H20 the corners are not standard. Two of three allotropic state of carbon are diamond and fullerene are transparent and both have cubic crystal structure.
______ *
To whom correspondence should be addressed: G. Kharlamova, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
373
374
G. KHARLAMOVA ET AL.
2. Experimental Investigations As diamond, and fullerene, as against graphite, are non-conductors. Isotropic crystals of diamond are colorless whereas isotropic crystals of fullerene ɋ60 are reddish. Here we represent original experimental results concerning obtaining and certification of the new forms of transparent carbon. The basic attention will be given to thread-like crystals of carbon, which in transmission polarized light are, as a rule, bright colored (Figure 1). The similar painted threads of length more 40 mm and diameter up to 50 mkm are formed in conditions two radically different processes. However here one of them we shall consider in more details.
Figure 1. Optical microscopy in transparent polarized light images of painted trans-parent threads formed simultaneously at synthesis of silicon carbide.
3. Characterization 3.1. OPTICAL CHARACTERIZATION
The transparent painted threads are revealed by means of optical micro-scopy in a powdery product of synthesis of silicon carbide, where as initial reagents the powders of silicon and carbon are used. The study of these threads in transmission polarized light has shown that they are anisotropic crystals with a parameter of two-refraction approximately equal 1,575. (We shall note that the crystals of quartz are also anisotropic with a parameter of two-refraction close to 1,575).
MOLECULAR CRYSTALS OF CARBON
375
Figure 2. X-ray spectral analysis of transparent colored microthreads.
The element composition of the revealed transparent threads was estimated by means of X-ray spectrometer (Camebax), which typical spectra are presented on Figure 2. From spectra follows that the transparent threads basically consist of carbon, instead of silicon and oxygen. Some ends of microthreads of carbon have extremely unusual structure (Figure 1). The research of a transparent micro thread by means of scanning electronic microscopy also has shown that all carbon micro thread consists of great number of threads considerably smaller diameter and, hence, the micro thread is a rope. The chemical analysis of a painted micro thread by means of electron dispersive spectroscopy (EDS) has shown that this rope consists of more thin transparent threads from carbon. Painted transparent microthreads of carbon are formed parallel with growth of thread-like and tubular structures of silicon carbide. As we believe, the painted microthreads of carbon and nanothreads of silicon carbide grow with participating of atoms of evaporating carbon (Figure 3). It is possible, that new thread-like crystals of carbon grow immediately from intermediate thermostable molecules of carbon.
Figure 3. The scheme of growth process of carbon and SiC threads.
376
G. KHARLAMOVA ET AL.
The basic part of atoms of carbon goes on formation nanostructures of silicon carbide therefore microthreads of carbon is formed in very small amounts. The growth of anisotropic particles of carbon and silicon carbide, as well as many other substances,1–5 for example boron carbide, silicon and carbon nanostructure,6,7 synthesized by us, is fulfilled according to the mechanism of low temperature exothermic nanosynthesis. Exothermic nanosynthesis is self-accelerating growth of anisotropic particles of silicon carbide from atoms of evaporating substances, in particular of carbon and silicon. In vicinities of nanocenter the sublimation of initial powdery reagents essentially raises because of considerable growth of temperature at realizing of highly exothermic reaction of formation of silicon carbide. Let’s note that it was accepted earlier to consider that the interaction between solid reagents is accomplished according to diffusion the mechanism. The stage of diffusion of atoms of silicon or carbon through a layer of a formed product of reaction (silicon carbide) is considered limiting in this process. The painted transparent microthreads of carbon were obtained following a detailed study of process of thermal transformation of molecules aromatic hydrocarbons, in particular benzene and toluene. As it is visible (Figure 4), these microthreads, at least, on color and morphology are very similar to the previous microthreads, which grow at evaporation of carbon. Crystal-optical investigations have shown that these thread-like crystals are also anisotropic with a parameter of two-refraction close 1,575. The study of chemical structure of these microthreads by means of the X-ray analyzer has shown that they consist also basically of carbon. Pulverized threads of carbon are investigated by means of X-ray diffractional analysis (Figure 5) and infrared (IR) spectroscopy. .
Figure 4. Optical microscopy in transparent polarized light images of painted trans-parent threads formed at thermal transformation of aromatic hydrocarbons.
MOLECULAR CRYSTALS OF CARBON
377
Figure 5. X-ray diffraction patterns of coloured transparent threads of carbon.
4. Conclusions Threads have original hexagonal crystalline structure with the following parameters: a = 0.498 nm and c = 0.826 nm. In IR spectrum the group of lines from 1,157 up to 1,724 sm-1 is visible which are usually responsible for skeletal fluctuations of ɋ-ɋ – bonds in aromatic hydrocarbons. (In IR spectrum fullerene ɋ60 in this range there are three lines: 1,182; 1,428 and 1,539 sm-1). The lines in the range less 900 sm-1 can be referred to deformational fluctuations of aromatic ɋ-ɇ – bonds.
References 1. Kharlamov, A. I., Kirillova, N. V., Karachevtseva, L. A. and Kharlamova, A. A. (2003) Theoretical and Experimental Chemistry. 39 (6) pp.374–379. 2. Kharlamov, A. I. and Kirillova, N. V. (2002) Theoretical and Experimental Chemistry. 38 (1) pp. 59–63. 3. Kharlamov, A. I., Kirillova, N. V. and Kaverina, S. V. (2003) Theoretical and Experimental Chemistry. 39 (3), pp. 141–146. 4. Kholmanov, I., Kharlamov, A. I. and Milani, P., et al. (2002) Journal of Nanoscience and Nanotechnology. 2 (5), pp. 453–456. 5. Kharlamov, A. I., Kirillova, N. V. and Kaverina, S. V. (2002) Theoretical and Experimental Chemistry. 38 (4), pp. 232–236. 6. Kharlamov, A. I., Kirillova, N. V. and Ushkalov, L. N. (2006) Theoretical and Experimental Chemistry 42 (2), pp. 90–95. 7. Kharlamov, A. I., Loythenko, S. V., Ʉirillova, N. V., Kaverina, S. V. and Fomenko, V. V.
(2004) Report of Academia of Science of Ukraine (1), pp. 95–100 (Russian).
HYDROGEN MICROSENSOR BASED ON NIO THIN FILMS I. FASAKI1,2, M. ANTONIADOU1,2, A. GIANNOUDAKOS1,2, M. STAMATAKI1, M. KOMPITSAS1*, F. ROUBANIKALANTZOPOULOU2, I. HOTOVY3, AND V. REHACEK3 1 National Hellenic Research Foundation, Theoretical and Physical Chemistry Institute, Vasileos Konstantinou Ave. 48, 11635 Athens, GREECE 2 National Technical University of Athens, School of Chemical Eng. 9 Iroon Polytechniou St., 15780 Zografou, Athens, GREECE 3 Slovak University of Technology, Ilkovicova 3, 812 19 Bratislava, SLOVAK REPUBLIC
Abstract – A multitude of industries use H2 either as part of their process or as a fuel. All these applications motivate nowadays the development of hydrogen sensor devices which enable its safe and controlled use. Since H2 is explosive above the lower explosion limit at 40,000 ppm, devices which permit the detection of its presence and measure its concentration become indispensable. In this work, we present a microsensor based on NiO thin films produced with dc reactive magnetron sputtering on GaAs, with an incorporated Pt heater, all on a DO-8 package ready for use. The microsensor was tested to H2 concentrations 5,000 and 10,000 ppm at different working temperatures. The change of the electrical resistance of NiO thin films was the signal for hydrogen sensing. The response of the sensor was not proportional to concentration of the gas neither to the working temperature.
Keywords: NiO, H2, gas sensors 1. Introduction Hydrogen as an industrial gas is being used by a multitude of industries. Some of the major industries today are chemical industries (refining crude oil, plastics, reducing environment in float glass industry, etc.), food industry (hydrogenation
______ *
To whom correspondence should be addressed: M. Kompitsas, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
379
380
I. FASAKI ET AL.
of oils and fats), semiconductor industry (as processing gas in thin film de-position and in annealing atmosphere), transportation (as fuel in fuel cells, rockets for space vehicles) and use H2 either as part of their process or as a fuel.1 All these applications necessitate the development of hydrogen sensor devices which enable its safe and controlled use. Since hydrogen is explosive above the lower explosion limit (LEL – 40,000 ppm) devices which permit the detection of its presence and its concentration become indispensable.2 Nickel oxide (NiO) is frequently considered as a model for p-type semiconductors. It is a wide band-gap (Eg | 4 eV) transition metal oxide, with a cubic rock-salt structure and antiferromagnetic properties below its Néel temperature, 523 K.3 Due to their excellent chemical stability NiO films have a wide range of applications as catalysts,4 electrochromic display devices5 and fuel cells.6 Moreover, recent works have shown that thin NiO films are attractive sensing material in gas and humidity detection devices.7,8 NiO thin films have been deposited by different techniques, including chemical self-assembly,9 sol-gel,10 RF,11 DC sputtering12 and recently pulsed laser deposition (PLD).6,13,14 The preparation method is fundamental in determining the microstructure and consequently the functional properties of the synthesized materials. In this work we demonstrate the response of NiO thin films to hydrogen. 2. Experimental The different layers of the microsensor were deposited by dc reactive magnetron sputtering on a GaAs substrate (Figure 1). By using a suitable mask and photolithographic process, platinum integrated heater having a shape of meander was realized. A layer of polyimide was deposited on Pt heater for electrical isolation. At the top NiO thin films were deposited. The microsensor was placed on a DO-8 package ready for use. The sample was mounted inside a gas test chamber which was evacuated at 10–2 mbar. The chamber was filled with dry air and then heated at different temperatures. The microsensor was tested at 5,000 ppm (working temperatures 210ºC, 240ºC and 280ºC) and 10,000 ppm (working temperatures 185ºC and 205ºC) of H2. The concentration of hydrogen was calculated using the partial pressures of the sensing gas and air in the chamber. The change of the electrical resistance of NiO thin films was the signal for hydrogen sensing. 3. Results and Discussion The response of the microsensor at 5,000 ppm (working temperature 240ºC) and at 10,000 ppm (205ºC) of H2 is seen at Figure 2. The response of the sensor
HYDROGEN NIO GAS SENSOR
381
a b)
c
Figure 1. (a) The structure of the microsensor, (b) image of the surface of the microsensor, (c) photograph of the microsensor.
was not proportional to the concentration of the gas neither to the working temperature. 0,45
0,20
o
5000 ppmH2 at 240 C
0,40
air
0,14
0,30
0,12
ǻR/Ro
ǻR/Ro
0,25 0,20
H2
H2
0,10
H2
H2
0,08 0,06
0,10
0,04
0,05
0,02
0,00
0,00 0
a)
air
air
0,16
0,35
0,15
o
10000 ppmH2 at 205 C
0,18
air
20
40
60
Time (min)
80
100
0
b)
15
30
45
60
75
Time (min)
Figure 2. (a) The response of the sensor at 5,000 ppm, (b) the response of microsensor at 10,000 ppm.
The increase of the electrical resistance was expected, considering that NiO is a p-type semiconductor and hydrogen is a reducing gas. As it is known, the NiO p-type conductivity is due to the non-stoichiometry of the prepared samples, in which vacancies occur in cation sites, i.e. the NiO films showed a metal deficiency.15 Atmospheric oxygen is expected to be present on the surface of NiO as O2,adsí and Oadsí negative charged chemical species. The high coverage with adsorbed oxygen species causes an increase in the concentration of the holes of the NiO film and an increase in its conductivity. The presence of H2 causes a decrease of the electrical conductivity, because H2 reacts with adsorbed oxygen and forms water vapor, injecting electrons in the NiO p-type semiconducting film2:
382
I. FASAKI ET AL.
2H2,gas + O2,adsí ļ 2H2Ovap + eí H2,gas + Oadsí ļ H2Ovap + eí 4. Conclusions A NiO microsensor has been developed by DC Magnetron Sputtering. The sensor’s response to H2 down to 5,000 ppm has been recorded. The resistance increased as expected for a p-type semiconductor in a reducing gas atmosphere.
References 1. Fuel Cell Standards Committee (2001) “Basic Consideration for Safety of Hydrogen Systems”, Technical Report ISO TC 197 N166, International Standards Organization. 2. Hotovy, I., Huran, J., Siciliano, P., Capone, S., Spiess, L., and Rehacek, V. (2004) Sens Actuators B, 103, 300. 3. Seehra, M., and Giebultowicz, T. (1988) Phys. Rev. B 38, 11898. 4. Blaumer, M., and Freund, H. (1999) Prog. Surf. Sci. 61, 127. 5. Jiao, Z., Wu, M., Qin, Z., and Xu, H (2003) Nanotechnology 14, 458. 6. Chen, X., Wu, N., Smith, L., and Ignatiev, A. (2004) Appl. Phys. Lett. 84, 2700. 7. Shi, J., Zhu, Y., Zhang, X., and Baeyens, W. (2004) Trends Anal. Chem. 23, 1. 8. Ando, M., Sato, Y., Tamura, S., and Kobayashi, T. (1999) Solid State Ionics 121, 307. 9. Wang, Y., Ma, C., Sun, X., and Li, H. (2004) Micropor. Mesopor. Mater. 71, 99. 10. Jiao, Z., Wu, M., Qin, Z., and Xu, H. (2003) Nanotechnology 14, 458. 11. Souza Cruz, T., Hleinke, M., and Gorenstein, A. (2002) Appl. Phys. Lett. 81, 4922. 12. Lee, M., Seo, S., Seo, D., Jeong, E., and Yoo, I. (2004) Integr. Ferroelectr. 68, 19. 13. Zbroniec, L., Sasaki, T., and Koshizaki, N. (2005) J. Ceram. Process. Res. 6, 134. 14. Sasi, B., and Gopchandran, K. (2007) Nanotechnology 18, 115613. 15. Lee, M., Seo, S., Seo, D., Jeong, E., and Yoo, I. (2004) Integr. Ferroelectr. 68, 19–25.
DESIGN AND CHARACTERIZATION OF STYRENE-BASED PROTON EXCHANGE MEMBRANES D. EBRASU1*, I. PETREANU1, L. PATULARU1, I. STEFANESCU1, AND M. VALEANU2 1 National Research Institute of Cryogenics and Isotopic Technologies, Uzinei Street no. 4, 240050, Rm. Valcea, ROMANIA 2 National Institute of Materials Physics, P.O. Box MG-7, 077125, ROMANIA
Abstract – This paper deals with preparation of PEM, based on commercial block copolymer of the styrene-butadiene. The copolymer was structurally changed by sulfonation followed by cross linking, in order to design a Proton Exchange Membrane for Fuel Cells. The membranes were structural tested by FTIR Spectroscopy and Scanning Electron Microscopy. Ionic Exchange Capacity (IEC) and thermal behavior by Differential Scanning Calorimetry (DSC) were measured too.
Keywords: Membrane, polymer, fuel cells
1. Introduction The commercial breakthrough of fuel cells is hindered by the high price of fuel cell components. PEM is the most expensive part of the fuel cell. Lower prices will be achieved by developing new materials and improving performance. Perfluorinated copolymers are the current state-of-the-art proton exchange membranes (PEM).1 The typical operation temperature of PEMFCs is in the range of 60–90°C.2 This operating temperature range is currently limited by the perfluorinated proton exchange membrane. Tri-block copolymers produced ______ *
To whom correspondence should be addressed: D. Ebrasu, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
383
384
D. EBRASU ET AL.
by DAIS-Analytic Corporation which may be based on sulfonated styrene-coethyleneco-butylene have been described in the literature.3 These sulfonated Kraton®-type block copolymers are post-sulfonated. The stability of these aliphatic hydrocarbon copolymers is, in general, inferior to the current state-ofthe-art perfluorinated copolymers. For this reason, the DAIS membranes are being promoted for the low temperature (<60°C), portable power markets.4 In order to increase the working temperature till 80°C, we were preparing a new type of membrane as is shown below. 2. Experimental 2.1. MATERIALS/SYNTHESIS OF SULFONATED STYRENE-BASED POLYMER
The ion-exchange membrane was prepared using raw polystyrene–blockbutadiene 21% styrene (PBS) from Aldrich. The solvent was chloroform chromatographic grade from Merck. The sulfonation agent was made from sulfuric acid 97%, acetic anhydride and dichloroethane chromatographic grade from Merck. Methanol from Merck was used to stop the sulfonation reaction. Divinylbenzene (DVB) 80% mixture of isomers from Aldrich was used as reticulation agent and Dimyristylperoxidicarbonat – Degussa used as an initiator for reticulation. PBS (5.2 g) and chloroform (150 ml) were successfully added to a 500 ml, four-necked flask. The solubilization was done at 40ºC, using a mechanical stirrer. The mixture was stirred for 30 min then was left to cool down to room temperature. From this mixture 100 ml were taken and were deposited as film, called Non Sulfonated Membrane (NSM), using “Doctor Blade” technique and tested by FTIR spectroscopy. Sulfonation was done using a solution made from of sulfuric acid 97% (1.4 ml), acetic anhydride (3.2 ml) and dicloretan (20 ml). The sulfonation agent was added, at room temperature, slowly using a dropping funnel and continuously stirring (500 rpm) in order to prevent the secondary reactions. The mixture is warmed till 50ºC for 1 h in order to activate the sulfonation process. The reaction is stopped by adding methanol (15 ml) and the mixture is filtered. 2.2. SYNTHESIS OF THE CROSSLINKED POLYMERIC MEMBRANE
The synthesized sulfonated polystyrene–block-butadiene 21% styrene (300 ml) was placed in a 1,000 ml, four-necked flask to which divinyl-benzene was added (0.5 ml, based 5 wt %) that has the role of cross-linking. Dimyristylperoxidicarbonat (0.4 g) was disolved in methyl acetate (6 ml) and added to the mixture in two steps: 4 ml and after 1 h the rest of 2 ml. The reaction takes place under reflux and stirring. From this mixture 100 ml were taken and it was
STYRENE-BASED PROTON EXCHANGE MEMBRANES
385
deposited as a film, called Sulfonated Membrane (SM), using “Doctor Blade” technique and tested by FTIR. 2.3. CHARACTERIZATIONS
The structural characterization was achieved using Fourier-transform infrared (FT-IR) spectroscopy by Nicolet Magna 750 apparatus with 4 cm–1 resolution. The morphological characterizations of samples were performed using scanning electron microscopy on a Scanning Electron Microscope (SEM) by JEOL TEMSCAN 200 CX. The differential scanning calorimetry (DSC) spectra of sulfonated and polymer was obtained on SETARAM DSC Model 131. Measurements were performed over the temperature range of 25–200ºC at the heating rate of 5ºC/min in hermetically sealed aluminum pans. Membrane samples were allowed to attain steady state with the solvents and the sample pan conditioned in the instrument before running the experiment. Ionic Exchange Capacity (IEC) and degree of substitution (DS) has been measured by common techniques.5 3. Results and Discussion 3.1. FTIR
Figure 1 shows specific features for films denoted with NSM and SM respectively show specific features for a PBS. Indexed on Figure 1 can be observed from 4,000 to 400 cm–1 bands: (1) sulfonated membrane SM presents an absorption pick at 3,400 cm–1 corresponding to the hydroxyl group that is formed due to the oxidation during the synthesis; (2) 2,985, 2,914, 2,829 and 2,761 cm–1 for ȞCH and ȞCH2 stretching vibrations from substituted aromatic ring; (3) 1,630 cm–1, for C=C symmetric stretch; (4) 1,548, 1,494, 1,406 and 1,375 cm–1 corresponding to bending bands for CH2 and CH; (5) strong band from 970 cm–1 indicate the presence of 1,4-trans butadiene; (6) 916 and 881 cm–1 correspond to CH2 and CH out of plane bands from vinyl unit; (7) 820 cm–1 indicate the presence of poly-butadiene with vinyl configuration; (8) 677 and 595 cm–1 indicate the presence of cis-1,4 ale poly-butadiene. The most important feature is band at 1,100 cm-1 corresponding to R-SO3H sulphonic groups. 3.2. SEM
The morphology of the membranes investigated by SEM is shown in Figure 2. At higher resolution we may see that the polymer chains are cross linked, though at the lower resolution we may see some non homogeneity. Obviously, the structure should be improved.
D. EBRASU ET AL.
2761
3400
2985
1630
0.2
2914
1406 1548
1358
R-SO 3H
1494
0.3
677
Absorbance
0.4
595
0.5
1100
820
0.6
970
881
916
2829
386
SM
0.1
NSM 0.0 500
1000
1500
2000
2500
3000
3500
-1
Wavenumbers[cm ]
Figure 1. FT-IR spectra for two representative samples NMS and SM with indexed bands.
Figure 2. SEM pictures for Sulfonated Membrane (SM) at high (a) and low (b) resolutions.
3.3. DSC
The DSC spectrum of the sulfonated PBS is shown in Figure 3. The spectrum presents two Tg values are situated at 60ºC and 100ºC and a band corresponding to an exothermic effect. The first Tg is corresponding to the sulfonated polystyrene and the second one to the row polystyrene. The exothermic effect indicates a further reticulation of the sulfonated PBS. The non-neutralized free acid forms (–SO3H) of all the sulfonated polymers have the effect to lower Tg than the non-sulfonated ones. This may be ascribed to the structural changes introduced into the polymer on account of sulfonation. This effect is illustrated clearly by the DSC spectra for sulfonated PBS as shown in Figure 3.
STYRENE-BASED PROTON EXCHANGE MEMBRANES
387
Figure 3. DSC spectra for Sulfonated Membrane (SM).
3.4. ION EXCHANGE CAPACITY (IEC) AND DEGREE OF SUBSTITUTION (DS)
Ion exchange capacity provides an indication of the content of acid groups present in a polymer matrix which are responsible for the conduction of protons and thus is an indirect and reliable approximation of the proton conductivity. For sulfonated PBS, IEC value is 3.8 meq/g and DS is 30.7% that are com-parable or even higher that ones for Nafion. 4. Conclusions This paper presents the preliminary results obtained a commercial block copolymers of the styrene-butadiene (PBS). PBS subjected to the sulfonation process presents FTIR characteristic bands, glass transition temperature Tg and IEC corresponding to a sulfonated polymer. The membrane is mechanical stable but from the Tg values and from the presence of the exothermic effect it is clear that the crosslinking is not complete. The membrane is mechanical stable although extensive sulfonation may lead to proportional increase in water uptake. The membranes show good perspectives in polymer electrolyte fuel cell (PEMFC) application. The experiments will continue in this direction and there will be performed additional tests like conductivity measurements, life time, X-Ray.
388
D. EBRASU ET AL.
References 1. Zalbowitz, M., and Thomas, S. (1999) Fuel Cells: Green Power, Department of Energy, LAUR-99-3231. 2. Carrette, L., Friedrich, K., and Stimming, U. (2001) Fuel cells-fundamentals and applications, Fuel Cells 1(1), 5. 3. Ehrenberg, S., Serpico, J., Sheikh, A., Tangredi, T., Zador, E., and Wnek, G. (1997) In: Proceedings of the 2nd International Symposia on New Materials for Fuel Cell and Modern Battery Systems (Montreal, Canada). 4. Wnek, G., Rider, J., Serpico, J., and Einset, A. (1995) In: Proceedings of the 1st International Symposium on Proton Conducting Membrane Fuel Cells (Electrochem. Soc. Proceedings, 247). 5 Smitha, B., Sridhar, S., and Khan, A. (2003) Synthesis and characterization of proton conducting polymer membranes for fuel cells, Journal of Membrane Science 225, 63–76.
STRONTIUM-SUBSTITUTED HYDROXYAPATITE THIN FILMS GROWN BY PULSED LASER DEPOSITION C. CAPUCCINI1*, F. SIMA2, E. AXENTE2, E. BOANINI1, M. GAZZANO3, A. BIGI1, AND I.N. MIHAILESCU2 1 Biomimetics and Materials Chemistry Laboratory, Department of Chemistry, Bologna, ITALY 2 Laser-Surface-Plasma Interactions Laboratory, Lasers Department, National Institute for Lasers, Plasma and Radiation Physics, PO Box MG-54, Bucharest-Magurele, ROMANIA 3 ISOF-CNR, c/o Department of Chemistry “G. Ciamician”, Bologna, ITALY
Abstract – Strontium substitution for calcium in the hydroxyapatite structure has lately attracted growing interest due to its beneficial effects on both bone formation and prevention of bone resorption. Coating Ti implants with Sr2+substituted hydroxyapatite is expected to enhance the bioactivity of the surface and stimulate bone apposition. To this end, we deposited thin films of hydroxyapatite with different substitutions of Sr2+ for Ca2+ on Ti substrates by Pulsed Laser Deposition (PLD). Solid solutions of Sr-Ca hydroxyapatites >Ca10-xSrxHA (x = 0–1)@ were prepared by direct synthesis in aqueous medium at 90°C. Sr2+ insertion led to a decrease of crystallinity degree, which accounted for the simultaneous reduction of the crystal dimensions. For PLD experiments, we used an UV excimer (KrF*) laser source (248 nm, ~7.4 ns) operating at a repetition rate of 2 Hz. The fluence during target irradiation was set at 2.4 J/cm2, and substrate temperature kept at 400°C. The depositions were performed from HA at different degrees of Sr2+ substitution for Ca2+ (x = 0; 0.1; 0.5; 1). All structures were post-treated in a H2O enriched atmosphere for 6 h. The results of structural and morphological characterizations carried out on the obtained structures indicated that the coatings, which adhered well to the substrates, were made of crystalline HA and contained strontium with a (Ca + Sr)/P molar ratio close to the stoichiometric value of HA.
______ *
To whom correspondence should be addressed: C. Capuccini, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
389
390
C. CAPUCCINI ET AL.
Keywords: Isomorphous substitution, strontium, hydroxyapatite, PLD
1. Introduction Apatites are widely spread in nature; in particular, the inorganic phase of the hard tissues of vertebrates can be assimilated to the synthetic hydroxyapatite Ca10(PO4)6(OH)2, CaHA. The similarity with biological apatites accounts for the high biocompatibility of synthetic CaHA. This bioactive ceramic material does not possess acceptable mechanical properties, as it is brittle in bulk. However, it does demonstrate significant potential for use as a coating on metallic orthopaedic and dental prostheses.1 At present, titanium and its alloys due to their excellent mechanical and biomedical properties are the most widely used materials for the production of metal implants.2 The presence of a CaHA coating improves osteointegration and creates a barrier to the release of metallic elements from the implant. Pulsed laser deposition has proved to be a competetive technique for growing thin calcium phosphate structures on metallic substrates. PLD utilizes a short, generally UV pulsed laser beam that is focused onto a rotating target placed inside a reaction chamber, where a controlled atmosphere can be maintained. The species that are expulsed by each subse-quent laser pulse form the coating as they reach the substrate, which can also be heated to a fixed temperature. The stoichiometry and crystallinity of the deposited material can be selected by a proper choice of the ablation and deposition parameters.3–7 The high stability and flexibility of the hydroxyapatite structure justify the wide variety of possible ionic substitutions.6 Among the bivalent cations that can replace calcium in CaHA, strontium has attracted a remarkable interest for its potential biological role. Strontium is present in the mineral phase of the bone, especially in the regions of high metabolic turnover,8 and its beneficial effect in the treatment of osteoporosis is well known.9 In vitro, strontium promotes the proliferation of osteoblasts and decreases the number and activity of osteoclasts10,11; in addition, strontium administration reduces bone resorption and stimulates bone formation.12–14 Strontium can replace calcium in the HA structure over the whole range of composition. The solid solutions that have been obtai-ned by hydrothermal methods or by treatment at high temperatures, display a linear variation with composition in the lattice parameters, whereas different data have been reported on the preferential substitution site of Sr for Ca in CaHA.15–17 To better clarify the interaction with HA structure of Sr, we previously synthesized and characterized Sr-Ca-HA solid solutions across the whole range of concentrations and proved that Sr when in low concentration showed an unexpected preference for site (1) of the HA structure.6 In this paper, we investigate the possibility to obtain thin CaHA films at different degrees of Sr
STRONTIUM-SUBSTITUTED HYDROXYAPATITE THIN FILMS
391
substitution for calcium, up to 10%. The films were grown on pure Ti substrates by pulsed laser deposition, which provides several advantages over other techniques for preparing HA thin films. Specifically, PLD allows precise control over HA growth parameters at low deposition temperatures. An appropriate selection of the deposition and post-deposition parameters can result in HA thin film microstructures and nanostructures having unique biological properties. One of the most important and successful biomedical applications of calcium phosphates is in coatings of endosseous implants.18 The deposition of calcium phosphate coatings on titanium seems to enhance the bioactivity of the surface, which improves fixation between hard tissue and the metal implant and stimulates bone apposition.19,20 The use of PLD has lately been extended to coating Ti substrates with calcium phosphate.20–23 Compared with other physical techniques, the use of laser light in PLD has the additional advantage of causing much lower pollution of the deposited material.3,4,24 The presence of Sr2+ ions is all the more interesting as awareness of their biological role has recently increased following the development of strontium ranelate, a drug that has been shown to reduce the incidence of fractures in osteoporotic patients.25–27 2. Synthesis and General Characterization (Ca–Sr) hydroxyapatites (Ca–Sr–HA) with Sr/(Ca + Sr) molar ratios in the range from 0 to 0.1 were synthesized in N2 atmosphere using 50 mL solutions with different Sr/(Ca + Sr) ratios that were prepared by dissolving the appropriate amounts of Ca(NO3)2·4H2O and Sr(NO3)2 in CO2-free deionized water and adjusting the pH to 10 with NH4OH. The total concentration of [Ca2+] + [Sr2+] was 1.08 M. The solution was heated to 90°C, and 50 mL of 0.65 M (NH4)2HPO4 solution, pH 10 adjusted with NH4OH, was added dropwise under stirring. The precipitate was maintained in contact with the reaction solution for 5 h at 90 °C under stirring, then centrifuged at 10,000 rpm for 10 min and washed repeatedly with distilled water. The product was dried at 37°C overnight. X-ray diffraction analyses were carried out using a PANalytical X’Pert PRO powder diffractometer equipped with a monochromator in the diffracted beam. Cu KĮ radiation was used (40 mA, 40 kV). The 2ș range was 10–60 at a scanning speed of 0.75/min. To evaluate the coherence lengths of the crystals and perform the line profile analysis, further X-ray powder data were collected over the 2ș range in the step scanning mode with a fixed counting time of 10s per 0.030/step. Calcium and strontium contents were determined using a Perkin Elmer AAnalyst 400 atomic absorption spectrophotometer (Ȝ(Ca) = 422.7 nm; Ȝ(Sr) = 460.7 nm). The samples were diluted to an appropriate volume with 10% lanthanum in 50% HCl, in order to suppress interferences. Phosphorus
392
C. CAPUCCINI ET AL.
content was determined spectrophotometrically in molybdovanadophosphoric acid using a Varian Cary50Bio instrument (Ȝ = 400 nm).28 For TEM investigations, a Philips CM 100 transmission electron microscope operating at 80 kV was used. A small amount of powder was dispersed in ethanol and submitted to ultrasonication. Furthermore, a drop of the calcium phosphate suspension was transferred onto holey carbon foils supported on conventional copper microgrids. 3. Thin Films Generation and General Characterization Disk shaped targets (13 mm in diameter and 1 mm in thickness) were manufactured by pressing the Sr-doped CaHA powders at 3 MPa and sintering at 380°C for 6 h The films were pulsed laser deposited on etched Ti substrates. In our experiments, we used an UV KrF* excimer laser source (Ȝ = 248 nm, IJ ~ 7.4 ns). The reaction chamber was evacuated down to a residual pressure of 10í4 Pa prior to every deposition. Films were deposited in 50 Pa water vapor flux on substrates heated to 400°C. The substrates were placed parallel to the targets 4 cm away from them. Fluence was set at 2.4 J cmí2 and we applied 25,000 subsequent laser pulses for each deposition. The as-deposited samples were submitted to annealing treatments in water vapor and ambient pressure for 6 h at temperatures identical to those applied during deposition. The average thickness of the obtained structures, measured by profilometry, was ~1ȝm. Grazing incidence XRD measurements were performed on the coatings with an X’Pert Philips Diffractometer using CuKĮ radiation and a grazing angle of 0.3–1.0°. The 2ș angles ranged from 10° to 40° with a 0.005°/s scanning speed. Morphological investigations of the synthesized products were conducted using a Philips XL-20 Scanning Electron Microscope. The samples were sputter coated with gold before examination. EDX analyses were also performed on uncoated specimens. 4. Results and Discussion The X-ray diffraction patterns of the solid products synthesized with different Sr/(Ca + Sr) molar ratios are shown in Figure 1. As all of the patterns show, they are made of hydroxyapatite as a unique crystalline phase. The patterns of the samples corresponding to CaHA (Sr0%) display well-defined sharp peaks in agreement with a high degree of their crystallinity. On the other
STRONTIUM-SUBSTITUTED HYDROXYAPATITE THIN FILMS
393
Figure 1. Powder X-ray diffraction patterns of (a) Sr0; (b) Sr5 and (c) Sr10. The strontium content in the solid phase is given in Table 2.
hand, the patterns of the samples containing both Ca and Sr generally exhibit broader diffraction peaks. This is in agreement with a reduced degree of crystallinity of the mixed Ca–Sr–HA, which suggests an increasing difficulty for CaHA to host a larger strontium amount (Ca 2+ ionic radius = 0.100 nm; Sr2+ ionic radius = 0.118 nm). The line broadening of the 0 0 2 and 3 1 0 reflections was used to evaluate the length of the coherent domains (IJh k l) both along the c-axis and perpendicular to it. The values IJh k l were calculated from the widths at half maximum intensity (ȕ1/2) using the Scherrer equation:
IJ
hkl
KO , E 1/ 2 cos T
(1)
where Ȝ is the wavelength, ș the diffraction angle and K a constant depending on crystal habit (chosen as 0.9). The silicon standard peak 111 was used to evaluate the instrumental broadening. The values of IJ002 shift from 469(±5) to 385(±5) Å, those of IJ310 decrease from 224(±6) to 153(±6) Å on increasing strontium content up to 10% (Table 1).
C. CAPUCCINI ET AL.
394
TABLE 1. Lengths of the coherent domains along the 002 and 310 directions. Samples
IJ0 0 2 (Å)
IJ3 1 0 (Å)
Sr0 Sr5 Sr10
469 (5) 396 (1) 385 (5)
224 (6) 217 (4) 153 (6)
The relative amount of strontium in the solid products, ȤSr, as evaluated through atomic absorption spectrometry is presented in Table 2 as a function of the strontium content in solution. The values of ȤSr increase on increasing Sr/(Ca + Sr) in the starting solution, in agreement with the quantitative incorporation of strontium in the solid phase. The ratio between the two cations in the solid phase is slightly smaller than that in the synthesis solution. The isomorphous substitution does not significantly affect the stoichiometry of HA, as can be deduced from the (Ca + Sr)/P molar ratio, which supports a mean value of 1.68 ± 0.03, very close to the stoichiometric value of 1.67, independently of the Ca and Sr contents. The TEM investigation results indicate that the morphology of the apatite crystals is little affected by the chemical composition. Ca–HA is constituted of plate-shaped crystals, with mean dimensions up to about 200 × 40 nm2 (Figure 2a). By low Sr2+ contents (0 ȤSr2 +d 0.1), the Ca–Sr–HA nanocrystals display more perturbed shapes and ill-defined edges (Figure 2b), in agreement with the lower degree of crystallinity that is revealed by the broadening of the X-ray diffraction peaks. TABLE 2. Relative amount of strontium in the solid products, evaluated through atomic absorption spectrometry reported as a function of strontium content in solution. Sample
Sr0 Sr5 Sr10
Sr/(Ca + Sr) molar ratio in solution 0 0.05 0.10
Sr/(Ca + Sr) = ȤSr molar ratio in the solid product 0 0.03 0.07
The X-ray diffraction patterns of the PLD thin films are shown in Figure 3. As those recorded from the powders, all these patterns indicate that the films are made of hydroxyapatite as a unique crystalline phase and display welldefined sharp peaks in agreement with a high degree of crystallinity achieved by the Pulsed Laser Deposition technique.
STRONTIUM-SUBSTITUTED HYDROXYAPATITE THIN FILMS
(a)
395
(b)
Figure 2. TEM micrographs of (a) Sr0; (b) Sr10. Scale bars = 200 nm.
Figure 3. Grazing incidence X-ray diffraction patterns of thin films deposited from (a) Sr0; (b) Sr5; (c) Sr10 samples. Asterisks indicate the reflections due to Ti.
The SEM investigation results (Figure 4) show that the apatite thin films exhibit a granular surface, with mean grain dimensions smaller than 1 Pm. The presence of Sr2+ does not affect the morphology of the coating, which is close to that previously obtained from apatites of different composition.29 The results of EDX indicate [(Sr/Sr + Ca)100] values of 7% and 10% for the thin films deposited from Sr5% and Sr10% powders, respectively. The data are clearly in excess with respect to the results of the chemical analysis carried out on the original powders. Although the EDX method is less reliable than atomic absorption spectroscopy (AAS), it confirms that PLD succeeds in preserving the same composition in the thin films as in the initial powder.
396
C. CAPUCCINI ET AL.
(a)
(b)
Figure 4. SEM micrographs of thin films deposited from (a) Sr0; (b) Sr10 samples. Scale bars = 5 Pm.
5. Conclusions The main contributions reported in this paper deal with (i) the chemical synthesis of Ca-Sr-HA powders and (ii) their use in nanostructured depositions onto Ti substrates by PLD. We therefore summarized our main results accordingly. I. We synthesized in N2 atmosphere (Ca–Sr) hydroxyapatite (Ca–Sr–HA) powders with Sr/(Ca + Sr) molar ratios in the range from 0 to 0.1. (1) The XRD patterns of the samples corresponding to CaHA (Sr0%) display well-defined sharp peaks in agreement with a high degree of crystallinity. The samples containing both Ca and Sr generally exhibit broader diffraction peaks, indicating a reduced degree of crystallinity of the mixed Ca–Sr–HA. (2) The TEM investigations indicate that the morphology of the apatite crystals is little affected by the chemical composition. Ca–HA consists of plate-shaped crystals with mean dimensions up to about 200 × 40 nm2. The Ca–Sr–HA nanocrystals with low Sr2+ contents display more perturbed shapes and ill-defined edges, in line with the lower degree of crystallinity the broadened X-ray diffraction peaks reveal. II. Nanostructured Ca-Sr-HA coatings were successfully grown on Ti substrates by PLD. (1) The XRD recorded patterns show that the films are made of hydroxyapatite as a unique crystalline phase and display well-defined sharp peaks. (2) The Sr:HA films exhibit a granular surface, with mean grain dimensions smaller than 1 ȝm, as seen from the SEM micrographs. The presence of Sr2+ does not affect the morphology of the coating, which remains close to that previously obtained from apatites of different composition. The EDX analyses reveal [(Sr/Sr + Ca) %100] values of 7% and 10% for the thin films deposited from Sr5% and Sr10% powders, respectively.
STRONTIUM-SUBSTITUTED HYDROXYAPATITE THIN FILMS
397
We conclude that PLD is fully appropriate for obtaining stoichiometric, well crystallized Ca-Sr-HA nanostructures compatible with further use in biomimetic Ti implants applications. ACKNOWLEDGEMENTS
We acknowledge with thanks the support provided to this research under theme 36 “New biomimetic calcium phosphate coatings for metallic implants” in the framework of the 2006–2008 Agreement on Scientific and Technological Co-operation between Italy and Romania.
References 1. Calvert P., De Rossi D., Petty M.C., Editors (2007) Material Science and Engineering C 27, 484. 2. Lacefield W. (1999) Advance in Dental Research 12, 21. 3. Chrisey D. and Hubler G., Editors (1994) Pulsed laser deposition of thin films, Wiley, New York. 4. Mihailescu I. and Gyorgy E. (1999) Pulsed laser deposition: an overview. In: Asakura T., Editor, Trends in optics and photonics. Springer series in optical science, Springer, Berlin, 201. 5. Bauerle D., Editor (1996) Laser processing and chemistry, Springer, Berlin. 6. Bigi A., Boanini E., Capuccini C., and Gazzano M. (2007) Inorganica Chimica Acta, 360, 1009. 7. Nelea V., Mihailescu I., and Jelinek M. (2007) Pulsed laser deposition of thin films: ApplicationsLED growth of functional materials. In: Robert Eason J., Editor, J. Wiley & sons Inc., Hoboken, New Jersey, pp. 421–456. 8. Blake G., Zivanovic M., and McEwan A. (1986) European Journal of Nuclear Medicine 12, 447. 9. Shorr E. and Carter A. (1952) Bulletin of the Hospital for Joint Diseases Orthopaedic Institute 13, 59. 10. Canalis E., Hott M., Deloffre P., Tsouderos Y., and Marie P. (1996) Bone 18, 517. 11. Chang W., Tu C., Chen T., Komuwes L., Oda Y., Pratt S., Miller S., and Shoback D., Endocrinology 40, 5883. 12. Grynpas M., Hamilton E., Cheung R., Tsouderos Y., Deloffre P., Hott M., and Marie P., Bone 18, 253 13. Marie P., Ammann P., Boivin G., and Rey C. (2001) Calcified Tissue International 69, 121. 14. Dahl S., Allain P., Marie P., Mauras Y., Boivin G., Ammann P., Tsouderos Y., Delmas P., and Christiansen C. (2001) Bone 28, 446. 15. Zhu K., Yanagisawa K., Shimanouchi R., Onda A., and Kajiyoshi K. (2006) Journal of the European Ceramic Society 26, 509. 16. Kikuchi M., Yamazaki A., Otsuka R., Akao M., and Aoki H. (1994) Journal of Solid State Chemistry 113, 373. 17. Bigi A., Bracci B., Cuisinier F., Elkaim R., Fini M., Mayer I., Mihailescu I.N., Socol G., Sturba L., and Torricelli P. (2005) Biomaterials 26, 2381.
398
C. CAPUCCINI ET AL.
18. Sun L., Berndt C., Gross K., and Kucuk A. (2001) Journal of Biomedical Materials Research Applied Biomaterials 58, 570. 19. Ducheyne P. and Qiu Q. (1999) Biomaterials 20, 2287. 20. Torrisi L. and Setola R. (1993) Thin Solid Films 227, 32. 21. Fernández-Pradas J., Sardin G., Clèries L., Serra P., Ferrater C., and Morenza J. (1998) Thin Solid Films 317, 393. 22. Arias J., Garcia-Sanz F., Mayor M., Chiussi S., Pou J., Leon B., and Perez-Amor M. (1998) Biomaterials 19, 883. 23. Antonov, E., Bagratashvili V., Popov V., Sobol E., Davies M., Tendler S., Roberts C., and Howdle, S. (1997) Biomaterials 18, 1043. 24. Bauerle D. (1996) Laser processing and chemistry, Springer, Berlin. 25. Meunier P., Lorenc R., and Smith I., et al. (2002) Osteoporosis International 13 (Suppl. 3), 66. 26. Meunier P., Roux C., Seeman E., Ortolani S., Badurski J., et al. (2004) The New England Journal of Medicine 350, 459. 27. Reginster J., Sawicki A., Devogelaer J., Padrino J., et al. (2002) Osteoporosis International 13 (Suppl. 3). 28. Quinlan K. and De Sesa M. (1955) Analytical Chemistry 27, 1626. 29. Gyorgy E., Torricelli P., Socol G., Iliescu M., Mayer L., Mihailescu I., Bigi A., and Werckman J. (2004) Journal of Biomedical Materials Research 71A, 353.
GROWING THIN FILMS OF CHARGE DENSITY WAVE SYSTEM Rb0.3MoO3 BY PULSED LASER DEPOSITION D. DOMINKO1*, D. STAREŠINIû1, K. BILJAKOVIû1, K. SALAMON1, O. MILAT1, A. TOMELJAK2, D. MIHAILOVIû2, J. DEMŠAR2, G. SOCOL3, C. RISTOSCU3, I.N. MIHAILESCU3, AND J. MARCUS4 1 Institute of Physics, HR-10001, Zagreb, P.O. Box 304, CROATIA 2 J. Stefan" Institute, Jamova 39, SI-1000, Ljubljana, SLOVENIA 3 National Institute for Lasers, Plasma and Radiation Physics, PO Box MG-54, Bucharest-Magurele, ROMANIA 4 Institut Neel, CNRS, BP 166, F-38042, Grenoble, FRANCE
Abstract – We prepared high-quality epitaxial thin films of charge density wave system Rb0.3MoO3 on MgO substrate. In continuation to the femtosecond spectroscopy performed in,1 new studies of femtosecond time-resolved Terahertz conductivity dynamics, necessary to directly probe the relaxation processes of photo-excited carriers, need high-quality thin films. Film morphology was characterized by EDS, AFM, and STM, while optical and electrical properties were studied using FTIR and UV-Vis spectrometers, and terahertz conductivity measurements, respectively. Certain degree of crystallinity has been observed in some films by x-ray diffraction.
Keywords: Thin films, blue bronze
1. Introduction Blue bronzes K0.3MoO3 and Rb0.3MoO3, the two most known CDW compounds, undergo a Peierls transition into 3D CDW ordered state at 183 K. Amplitudons and phasons, excitations of the amplitude and phase of the CDW complex order parameter were widely investigated,1 as they are at the origin of many fascinating CDW properties. Collective charge transport by moving CDW is probably
______ *
To whom the correspondence should be addressed: D. Dominko, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
399
400
D. DOMINKO ET AL.
the most known one. CDW sliding (and pining) was studied in bulk and only recently in film compounds.2 CDW properties on mesoscopic scales, as well as possible applications of the CDW effects, especially in thin films, are potentially attractive. 2. Motivation Femtosecond spectroscopy investigation on blue bronze gave first information on the collective (amplitudon and phason) and single particle excitations. With optical pump – terahertz probe in transmission geometry we could measure the dynamics of the photoexcited phason directly (since it is IR active and lies in this frequency range).
Figure 1. Arrhenius plot of temperature dependence of resistivity for various sample thickness.3
Finite size effects were investigated on 1D2,3 and 3D CDW systems only. TP and width of transition increase for wires thinner than transverse correlation length (observed on TaS3 only), Figure 1 As similar effects were observed in doped samples (doping induced/enhanced CDW), interplay doping/dimensionality can be an additional information. 3. Deposition UV excimer laser system has been used, with Ȝ = 248 nm and FWHM p 7 ns. Deposition rate is estimated to be §0.5 nm/pulse. Conditions that influence film quality are: (1) preferable commensurability between substrate surface lattice and lattice of film deposit, (2) higher substrate temperature results larger grains and higher corrugation, (3) oxygen pressure reduces oxygen loss during deposition,4 (4) higher deposition rate result in smaller grains. 5 Thin films of various thicknesses were grown by pulsed laser deposition applying 2,000 to
THIN FILMS OF BLUE BRONZE
401
6,000 subsequent laser pulses with an excimer laser system COMPexPro 205 (Lambda Physik). The depositions were performed in oxygen pressure 0.1–0.3 Torr on substrates heated to 400°C. The incident laser fluence was 2.5 J/cm2 (Table 1). It has been shown5 that using SrTiO3 [510] substrate, chains are aligned in one direction. With our deposition conditions we failed so far in obtaining well defined chain direction. TABLE 1 Deposition conditions.
Sample PO2 (Torr) Pulse No.
1 2 3 4 5 6
0.3 0.3 0.1 0.1 0.2 0.2
2,000 3,000 2,000 6,000 2,000 6,000
Other information
Substrate MgO (100) T (°C) 400 3 FluenceO2 (J/cm ) 2.5 Dsubstrate-target (cm) 5
4. Film Characterization 4.1. CROSS POLARISED MICROSCOPY
Figure 2 clearly shows homogeneous structure for sample 1, while samples 4 and 5 exhibit nonhomogeneous structure. In addition, leaf-like texture has been found in samples 3 and 4. Similar structures have been observed for films deposited in the same oxygen pressure, indicating its crucial roll for obtaining films with higher degree of crystallinity.
200X
200x200x
400
Figure 2. SEM images of samples 1, 4 and 5.
4.2. AFM, SEM, EDXS AND RBS
Due to high corrugation of the rest of the films only sample 5 could have been scanned by AFM. Walls were observed in surface structure. SEM results from previous works5 have shown clear substrate dependence of film quality. So far
402
D. DOMINKO ET AL.
SEM on our samples did not show similar chainlike features. Only Rb and Mo concentration ratios could be deduced from EDXS and RBS, due to oxygen contained in MgO substrate. We reached desired stoichiometry for Rb0.3MoO3. 4.3. X-RAY AND THZ CONDUCTIVITY
X-ray has been performed on samples 1, 2, 3 and 5. Figure 3 shows that sample 1 (as is the case for sample 2 as well) has the right peak as expected for blue bronze, at 11Û. GISAX showed oscillations for samples 3 and 5, which revealed smoother surface and consequently amorphous structure. Thus samples 1 and 2 are most likely to have crystalline structure of blue bronze. With GISAX film thickness of sample 5 has been estimated to be 90 nm. Two orders of magnitude lower conductivity at 1 THz is obtained from film 4, compared to bulk 4.
Figure 3. X-ray for sample 1 (upper) compared with the theory (lower). Different colors represent different angle regions.
5. Conclusions We deposited blue bronze films on MgO substrate with right stoichiometry. This substrate has been used because it is transparent to IR wavelengths needed for femtosecond spectroscopy investigation. Certain degree of crystallinity has been obtained from film deposition with higher oxygen pressure.
References 1. 2. 3. 4. 5.
Demšar, J. et al. (1999) Physical Review Letters 83, 800. Mantel, O. et al. (1999) Journal of Applied Physics 86, 4440. Zaitsev-Zotov, S. (2003) Microelectronic Engineering 69, 549. Travaglini, P. (1984) Physical Review B 30, 1971. Van Der Zant H. et al. (1996) Applied Physics Letters 68, 3823.
SINGLE CELL DETECTION WITH DRIVEN MAGNETIC BEADS B.H. MCNAUGHTON1, R.R. AGAYAN1,2, V.A. STOICA1, R. CLARKE1, AND R. KOPELMAN1,2* 1 University of Michigan, Department of Physics, Ann Arbor, MI 48109 2 University of Michigan, Department of Chemistry, 930 N. University, Ann Arbor, MI 48109-1055
Abstract – Shifts in the nonlinear rotational frequency of magnetic beads (microspheres) offer a new and dynamic approach for the detection of single cells. We present the first demonstration of this capability by measuring the changes in the nonlinear rotational frequency of magnetic beads driven by an external magnetic field. The presence of an Escherichia coli bacterium on the surface of a 2.0 Pm magnetic bead affects the drag of the system, thus changing the nonlinear rotation rate. Measurement of this rotational frequency is straightforward utilizing standard microscopy techniques.
Keywords: Nonlinear rotation, nonlinear dynamics, magnetic microspheres, magnetic beads, single cell detection
1. Introduction The ability to detect and measure single biological agents is of fundamental importance for rapid and accurate medical diagnostics. Recent investigations have focused on the development of micro and nanoscale oscillating systems as novel detection schemes that are both ultra-sensitive and rapid. Detection methods utilizing this technology offer a powerful and diverse group of extremely sensitive tools that have already demonstrated single biological agent detection.1–4 Microand nanoscale oscillators can be classified into several general categories, some ______ *
To whom correspondence should be addressed: R. Kopelman, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
403
404
B.H. MCNAUGHTON ET AL.
of which include resonant nanomechanical (NEM) cantilevers,3,4 rotationalbased oscillators bound to a substrate via carbon nanotubes,5–7 and fluid-based magnetically actuated systems.8–10 A key distinction of fluid-based magnetically actuated systems is that they exhibit a nonlinear behavior that enables a new sensing scheme – see Figure 1a, a scheme where sensitivity is unaffected by viscous losses11 (unlike with cantilevers,12–14 which work best in air or vacuum). In 1990, the seminal experimental and theoretical work of Helgesen et al. detailed the rotational dynamics of a pair of magnetic holes (non magnetic microspheres in a ferrofluid), in which nonlinear behavior was observed at sufficiently high external magnetic field rotation rates.15 Many groups followed this work, with various nonlinear rotation studies of small scale systems.10,16,17 For example, Shelton and coworkers used angular momentum from polarized light to torque a glass nanorod.17 They theorized and experimentally verified that the average rotation rate of the glass rod had a nonlinear dependence on the rotation rate of the polarized light and showed that the nonlinear rotation rate is dependent on the optical torque and fluidic drag of the system. The nonlinear rotation rate of magnetic beads has the same dependence on drag as a torqued glass nanorod. Indeed, Biswal and Gast showed that chains of paramagnetic beads are governed by similar rotational dynamics.16 Recently, Cebers and Ozols performed a rigorous theoretical analysis on single particle systems, but did not focus on applications of such systems. Other types of rotational systems that could be used for nonlinear rotation experiments include magnetic nanorods (or “nanowires”)18,19 and substrate-based rotational actuators (if the system experiences drag).5 While many systems have been shown to exhibit nonlinear rotational dynamics, few studies considered applications, such as single cell detection. To fill this research gap, we have studied the nonlinear rotation of magnetic microparticles and explored a number of applications.11,20,21 Indeed, the rotational dynamics of magnetic particles offer potential use in the detection of biological agents. We report on such an application, demonstrating single cell sensitivity. 2. Methods The orientation and rotation rate of magnetic particles can be measured because of a physical or optical asymmetry that the particle has, i.e. a “nanocap” on one side of a particle – shown in Figure 1b. A nanocap can be made by depositing a thin layer of a light attenuating metal, such as aluminum. Nonlinear rotation occurs at high frequencies when the phase-lag between an external rotating magnetic field and the dipole vector of an aligning magnetic particle becomes larger than ʌ/2. After this point, the magnetic particle cannot overcome the viscous
SINGLE CELL DETECTION
405
drag (to remain phase-locked with the external field’s rotational frequency) and thus “slips,” rotating asynchronously (nonlinearly) with the driving field, e.g. the average rotational frequency of the magnetic particle has a lower value than that of the driving field.11,20–22 This type of asynchrony also appears in the flashing of fireflies and Josephson junction voltage dynamics.23
Ab2Ab1
Objective Lens
Xenon Light Field: 5-20 Oe Rot:0-10 Hz Visc: 122 cSt
2.0 micron
˜
2 Mm
0˚
90˚
ll
ic Ce
Fluid
180˚
Emission / Reflection
Ab2Ab1
m
External Magnetic Field
m Average Nonlinear Rotation Rate
(b)
ii) Post-detection Slow Rotation
i) Pre-detection Fast Rotation Single Bacterium
(a)
Filter Cube
CCD interfaced to Image Analysis Software 270˚
Figure 1. (a) Schematic of the nonlinear rotation rate changes that a magnetic bead undergoes when bound to a bacterium. The magnetic bead is functionalized with a secondary antibody (Ab2) and a primary antibody (Ab1). The bottom series shows fluorescence microscopy images of a rotating 2.0 Pm magnetic bead with a single Escherichia coli bacterium attached. The dotted circle indicates the location of the magnetic bead. (b) Schematic of the microscopy setup used to perform single cell sensitivity measurements (reflection microscopy is shown, but fluorescent microscopy was also utilized).
The rotational frequency was measured using microscopy and image analysis techniques reported elsewhere.20,21 Measurements were performed in two homemade ~100 Pm thick fluidic cells: one contained magnetic bead solution with no bacteria present and the other contained magnetic beads with bacteria bound to their surfaces. Samples were mixed with glycerol before being placed in the fluidic cells to a glycerol-water mass fraction of 0.5. The approximation of the average nonlinear rotational frequency was determined by performing a discrete Fourier transform of the beads’ intensity fluctuations – see Figure 2b. This measurement was performed for 20 single magnetic bead without bacteria in one fluidic cell and for 20 single magnetic beads with one E. Coli bound to each surface – see Figure 2c. To obtain nonlinear rotation, the particles were rotated in an external field with a magnitude of approximately 10 Oe at a driving frequency of 2.5 Hz. The beads were ferromagnetic and obtained from Spherotech, Inc.
B.H. MCNAUGHTON ET AL.
406
3. Results and Discussion One of the physical properties that the nonlinear rotation rate depends on is fluidic drag. When a bacterium attaches to a nonlinear rotating magnetic particle, the particle’s volume and shape are drastically changed. This produces more drag and, therefore, the nonlinear rotation rate slows considerably. This technique has been used to measure a change of drag caused by the attachment of a 1.0 ȝm particle to a 1.9 ȝm nonlinear rotating magnetic bead.21 Here, the binding of a single bacterium, to a 2.0 ȝm sphere, is shown to be detectable; in fact, the nonlinear rotation rate slowed down, on average, by a factor of ~3.8 – see Figure 2. The technique is also dynamic in the sense that a change in drag causes a direct change in the nonlinear rotation rate; thus the growth of an attached bacterium would cause further changes in drag. While an entire range of frequencies can be scanned to determine the point of criticality (when the motion changes from linear to nonlinear, which is given by : C ), for a magnetic particle with an attached bacterium, as was done in Figure 2a, it is much faster and more straightforward to simply measure the value of the nonlinear rotation frequency, T , at a given external driving frequency of : . From the rotation rate of the magnetic bead and the rate of the external driving field, the critical frequency can be calculated, as
:C
T
1 2
> 2: T @
1 2
mB / NKV ,
(1)
where m is the strength of the magnetic moment of the bead, B is the external magnetic field amplitude, ț is the shape factor, Ș is the dynamic viscosity of the surrounding fluid and V is the bead volume. Figure 2b shows the results from such an approach, where a Fourier transform of the modulated signal was performed for a typical magnetic bead with a single bacterium attached to its surface, and for one without (the antibody used was anti-E. coli IgG with specificity for all “O” and “K” serotypes of E. Coli obtained from Cortex Biochem). The peaks indicate the average rotation rate, which directly corresponds to the critical frequency and, therefore, to the fluidic drag, as described in Eq. (1). Figure 2c shows the curves for the rotation rates of 20 particles with single attached bacteria and for 20 particles without bacteria. Individual bacterium attachment was confirmed by imaging the bacteria with fluorescent microscopy, where DsRed fluorescent protein vectors were used for transformation of the E. coli. The presence of the bacteria on the surface of the magnetic beads caused a measurable change in the average rotation frequency, namely the average frequency of the magnetic beads at a 0.72 Hz to T2 0.19 Hz , driving frequency of 4.0 Hz changed from T1
SINGLE CELL DETECTION
407
Figure 2. Nonlinear rotation dynamics of magnetic beads with and without bacterium attached. (a) The rotational response of a single magnetic particle with attached bacterium at various external driving frequencies, where the squares are data and the line is a theoretical fit. (b) The fast Fourier transform of the intensity fluctuations of a typical particle, with a bacterium attached (solid curve) and for one without (dashed curve). (c) The average nonlinear rotation frequency of 20 particles in a fluidic cell with bacteria present (solid curve) and a fluidic cell without bacteria (dashed curve).
a factor of ~3.8. This change in rotation frequency is similar in value to our previous measurements on a 1.0 Pm particle that was attached to a single ~1.9 Pm ferromagnetic bead.21 Once a bacterium is attached to a magnetic bead, this technique could also be used to monitor a single bacterium’s growth. Monitoring changes in nonlinear rotation rate could lead to the study of single bacterium growth dynamics and thus to rapid antibiotic susceptibility measurements. The ability to use the change in nonlinear rotation of magnetic particles to detect bacteria has been demonstrated. In summary, the nonlinear rotation frequency of 2.0 Pm magnetic beads changed on average by a factor of 3.8. These data show that a dynamic micro-oscillator is sensitive enough to detect a single bacterium in a fluidic environment. 4. Conclusions The ability to use the change in nonlinear rotation of magnetic particles to detect bacteria has been demonstrated. The nonlinear rotation frequency of 2.0 ȝm magnetic beads changed on average by a factor of 3.8. For the first time, we have shown that a dynamic micro-oscillator is sensitive enough to detect a single bacterium in a fluidic environment. ACKNOWLEDGMENTS
The authors would like to thank Fierke, Hernick, and Hurst for help with the bacteria growth and transformations, and NSF for funding (DMR # 0455330).
408
B.H. MCNAUGHTON ET AL.
References 1. Craighead, H. (2000) Science 290(5496), 1532. 2. Ekinci, K., and Roukes, M. (2005) Review of Scientific Instruments 76(6), 61101–61101. 3. Ilic, B., Czaplewski, D., and Zalalutdinov, M., et al. (2001) Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures 19, 2825. 4. Ilic, B., Yang, Y., and Craighead, H. (2004) Virus detection using nanoelectromechanical devices, Applied Physics Letters 85(13), 2604–2606. 5. Fennimore, A., Yuzvinsky, T., and Han, W., et al. (2003) Nature 424, 408–410. 6. Papadakis, S., Hall, A., and Williams, P., et al. (2004) Physical Review Letters 93(14), 146101. 7. Williams, P., Papadakis, S., and Patel, A., et al. (2002) Physical Review Letters 89(25), 255502. 8. Anker, J., and Kopelman, R. (2003) Applied Physics Letters 82(7), 1102–1104. 9. Fan, D., Zhu, F., and Cammarata, R., et al. (2005) Physical Review Letters 94, 247208. 10. Korneva, G., Ye, H., Gogotsi, Y., et al. (2005) Nano Letters 5(5), 879–884. 11. McNaughton, B., Agayan, R., and Kopelman, R. (2006) Arxiv preprint cond-mat/0610144. 12. Bhiladvala, R., and Wang, Z. (2004) Physical Review E 69(3), 36307. 13. Paul, M., and Cross, M. (2004) Physical Review Letters 92(23), 235501. 14. Vignola, J., Judge, J., Jarzynski, J., et al. (2006) Applied Physics Letters 88(4), 41921–41921. 15. Helgesen, G., Pieranski, P., and Skjeltorp, A. (1990) Physical Review Letters 64(12), 1425–1428. 16. Biswal, S., and Gast, A. (2004) Analytical Chemistry 76(21), 6448–6455. 17. Shelton, W., Bonin, K., and Walker, T. (2005) Physical Review E 71(3), 36204. 18. Lapointe, C., Cappallo, N., Reich, D., et al. (2005) Journal of Applied Physics 97(10), 10. 19. Tok, J., Chuang, F., and Kao, M., et al. (2006) Angewandte Chemie International Editon England. 20. McNaughton, B., Kehbein, K., Anker, J., et al. (2006) Journal of Physical Chemistry B 110, 18958–18964. 21. McNaughton, B., Agayan, R., Wang, J., et al. (2007) Sensors and Actuators B 121, 330–340. 22. Cebers, A., and Ozols, M. (2006) Physical Review E 73(2), 21505. 23. Strogatz, S. (1994) Nonlinear dynamics and chaos. (Addison-Wesley, Reading, MA).
ANTIMICROBIAL PROPERTIES OF TITANIUM NANOPARTICLES B.K. ERDURAL, A. YURUM, U. BAKIR, AND G. KARAKAS* Department of Chemical Engineering, Middle East Technical University, 06431 Ankara, TURKEY
Abstract – In the present study, nanostructured titania particles were synthesized using hydrothermal processing and their photocatalytic antimicrobial activities were characterized. Sol-gel synthesized TiO2 samples were treated with a two step hydrothermal treatment. The first stage treatment was the alkaline treatment with 10 M of NaOH for 48 h at 130ºC, followed with the second step which applied with distilled water for 48 h at 200°C. Scanning Electron Microscope (SEM) images showed that alkaline treatment yields lamellar structure particles from the sol-gel synthesized anatase. Further treatment of nanoplates with distilled water results in crystal growth and the formation of nano structured thorn like particles. The photocatalytic antimicrobial activities of samples were determined against Escherichia coli under solar irradiation for 4 h. It was observed that the samples treated under alkaline conditions have higher antimicrobial activity than the untreated samples.
Keywords: Titania, nanostructured, hydrothermal, photocatalytic, antimicrobial
1. Introduction Microbial contamination and growth on the surfaces are potential risks for human health. Various applications are utilized for disinfections of surfaces such as, detergents, alcohols and chlorine components. However these agents are ineffective for long term disinfection and are not environmentally benign. In addition to these methods, UV irradiation is also used for disinfection. This method is an effective, but a temporary and hazardous one requiring lower
______ *
To whom correspondence should be addressed: G. Karakas, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
409
B.K. ERDURAL ET AL.
410
wavelength UV-C light source. However, photocatalysis is an alternative to direct UV disinfection. Large band semiconductor metal oxides, e.g. TiO2, SnO2 and ZnO are potential alternatives because of their higher wavelength UV absorption (UVA) which exists in natural sunlight and artificial illumination.1–6 TiO2 photocatalysts with anatase structure generate strong oxidizing power under UV irradiation which corresponds to wavelengths less than 385 nm as a result of charge separation. The hole (h+) and electron (e–) pairs formed react with H2O and O2 over the surface and hydroxyl radicals (OH-) and super oxide ions (ƔO2-) are generated. The hydroxyl radicals are highly toxic towards microorganisms.7 The photocatalytic efficiency of many synthetically produced TiO2 samples depends on many structural parameters such as band gap, surface area, particle size and crystallinity. During the last decade, much effort has been devoted to the increase of the photocatalytic efficiency of TiO2 materials.8 The most important reason of low photocatalytic efficiency of TiO2 is the competition of hole/electron charge recombination reaction with charge separation and free radical production reactions. In present study, the effects of alkaline and neutral hydrothermal treatments were examined on the sol-gel synthesized TiO2 particles. 2.
Experimental
2.1. SUBSTRATE PREPARATION
In the synthesis, titanium tetraisopropoxide (TTIP, Aldrich, extra pure grade), ethanol (C2H5OH, 99.5%) and 35% HClaq. were used. TTIP was first dissolved in ethanol and hydrolyzed with dropwise addition of ethanol–aques hydrochloric acid solution in thermostated bath at 0ºC, under continuous stirring. The resulting sol was dried in the oven at 60ºC overnight, and after mashing the aggregates in mortar, the samples were calcined in air at 600ºC for 2 h. Afterwards the alkaline post treatment step was carried out hydrothermally by using 1.5 g of sol-gel synthesized samples with 100 ml of 10 N NaOH in autoclave sealed at 130ºC for 48 h. After cooling the autoclave, the suspension was filtered, washed repeatedly with 0.1 N HCl solution followed by washing with distilled water until a neutral filtrate (pH = 7) was obtained. After drying at 60ºC for 2 h, sample stored in a desiccator for further analyses and second stage post treatment. The second stage post treatment was applied to 1 g of sample with 100 ml distilled water and the mixture was autoclaved at 200ºC for 48 h. The treated samples were filtered and dried in air at 80ºC for 4 h.
ANTIMICROBIAL PROPERTIES OF TITANIUM NANOPARTICLES
411
2.2. CHARACTERIZATION
The photochemical antimicrobial activity of samples was measured against Escherichia coli. For this purpose microbial cell concentrations were determined by viable count procedure on agar plates after serial dilutions of the culture in 0.1% peptone water. Total reaction mixture volume was 20 ml. The reaction mixture contained 3.5 g/l TiO2 samples and E. coli about 103 cells/ml in 0.1% peptone water. The mixture was agitated with a magnetic stirrer at 250 rpm at ambient temperature while irradiated with artificial irradiation source which was applied vertically. Osram Ultra-Vitalux (Product number: 03313) 300 W bulb with similar spectral distribution to solar spectrum between 280 and 780 nm was used as artificial irradiation source. The light intensity over the test bench was adjusted to achieve 10 mW/cm2. Microbial inactivation was followed for 4 h with the removal of the samples of reaction mixture at various time intervals. 200 Pl of samples was directly spread onto agar plates and incubated at 35ºC for 24 h to determine the survivors by counting the colonyforming units (CFUs). Two sets of control experiments were carried out. “Light control” experiments were tested without TiO2 samples were performed with artificial light and dark control experiments were performed with TiO2 without artificial light. 3. Results and Discussion The effects of alkaline and neutral hydrothermal treatments on the photocatalytic antimicrobial performance of sol-gel synthesized TiO2 (SGS) samples were evaluated. The antimicrobial performances of the SGS TiO2 and its derivatives which was hydrothermally treated under alkaline conditions for 48 h are presented in Figure 1. A limited antimicrobial activity was observed for SGS sample. At the end of 4 h, only 70% of the microorganisms could be inactivated by using SGS. However, when SGS sample was treated with NaOH for 48 h (S48) nearly complete inactivation were achieved within 3 h. Our detailed analysis on the effect of hydrothermal treatment9 under alkaline conditions revealed that sol-gel synthesized titania particles are re-structured to trititanate nanoplates which have poor crystallinity. Therefore, complete inactivation of alkali treated samples could be explained by the formation of trititanate structure. Second stage of hydrothermal post treatment was applied to the S48 sample for 48 h (S48h48) with distilled water at 200ºC to test the effect of hydrothermal treatment under neutral conditions on the crystallinity and the photocatalytic activity. In our previous study we have reported that the increase of
412
B.K. ERDURAL ET AL.
Figure 1. Effect of alkaline hydrothermal treatment on antimicrobial efficiency of sol-gel synthesized samples under irradiation (10 mW/cm2) (x, SGS; o, S48; , dark; , under irradiation). (C/C0 = (number of survival microorganisms)/(number of initial microorganisms)).
crystallinity of anatase phase increase with hydrothermal post treatment period.9 Unexpectedly, the photocatalytic antimicrobial performance of S48h48 sample resulted in poorer activity than S48 sample in spite of S48h48 has better anatase crystallinity.9 When the activity of S48 samples were compared with S48h48 sample, it is clear that the initial inactivation rates are almost the same. However the initial activity of S48h48 sample has not been able to sustain after 20 min of irradiation. Surprisingly the activity was resumed after about 3 h and similar inactivation rate was observed in this period (Figure 2). Ineffective period in microbial inactivation might be attributed to the density of active surface sites and active species.9 The concentration of •OHads species depends on the consumption rate by microbial inactivation and production rate by charge separation (e- and h+). Over the ineffective period, the microbial inactivation rate might be decreased with the lower concentration of active species. With increasing neutral hydrothermal treatment time, more water may start to be adsorbed over the surface. On the other hand, when anatase is irradiated with UV, adsorbed molecular water species dissociate8 and the concentration of oxidative species (such as •OHads and H2O2) directly influences the photocatalytic microbial inactivation. To support this suggestion, S48h48 was agitated in peptone water for 4 h under irradiation and in dark separately. After 4 h period, E. coli was inoculated to the aqueous mixtures and antimicrobial activity of S48h48 was tested for 100 min.
ANTIMICROBIAL PROPERTIES OF TITANIUM NANOPARTICLES
413
0
ln(C/C0)
-1
-2
-3
-4
-5 0
50
100
150
200
250
Time (min)
Figure 2. Comparison of effects of alkaline hydrothermal treatment for 48 h (S48) (o), neutral hydrothermal treatment for 48 h (S48h48) (x), treatment with water and UV for 4 h (ź) on antimicrobial activity, dark; (), under irradiation ().
As shown in Figure 2, 80% microbial inactivation was achieved within 100 min over the sample which was treated with UV for 4 h. However, keeping the substrate in water in dark have no antimicrobial activity which indicates the surface sites are saturated with water species and UV radiation is essential for •OHads formation.9 4. Conclusion The photocatalytic antimicrobial activity of sol-gel synthesized TiO2 anatase particles can be enhanced with the hydrothermal treatment under alkaline conditions. Further hydrothermal post treatment under neutral conditions improves the crystallinity but not photocatalytic activity. However treatment under UV irradiation facilitates the photocatalytic activity by converting surface water species to active •OHads.
References 1. Matsunaqa, T., Tomoda, R., Nakajima, T., and Wake, H. (1985), FEMS Microbiol. Lett., 29(1–2), 211–214. 2. Saito, T., Iwas, T., Horis, J., and Morioka, T. (1992), J. Photochem. Photobiol. B, 14, 369–379. 3. Benedix, R., Dehn, F., Quaas, J., and Orgass, M. (2000), Lacer, 5, 157. 4. Hong, J., and Otaki, M. (2003), Biosci. J. Bioeng., 96(3), 298–303.
414
B.K. ERDURAL ET AL.
5. Erkan, A., Bakir, U., and Karakas, G. (2006), J. Photochem. Photobiol., 184(3), 313–321. 6. Huang, Z., Maness, P., Blake, D., and Wolfrum, E. (2000), J. Photochem. Photobiol. A: Chem., 130, 163–172. 7. Seven, O., Dindar, B., Aydemir, S., Metin, D., Ozinel, M., and Icli, S. (2004), J. Photochem. Photobiol. A., 165, 103–107. 8. Diebold, U. (2003), Surf. Sci. Rep., 48, 53–229. 9. Erdural, B., Yurum, A., Bakir, U., and Karakas, G. (2008), J. Nanosci. Nanotechnol. 8, 878–886.
CsHSO4/NANOOXIDE POLYMER MEMBRANES FOR FUEL CELL A. ANDRONIE1, A. MOROZAN1, C. NASTASE1, F. NASTASE1, A. DUMITRU1, S. VULPE1, A. VASEASHTA2, AND I. STAMATIN1* 1 3Nano-SAE Research Centre, University of Bucharest, PO Box MG-38, Bucharest-Magurele, ROMANIA 2 On detail from Nanomaterials Processing & Characterization Laboratories, Marshall University, Huntington, WV, USA
Abstract – Composite solid acid/nanooxide polymer membranes with good electrical behaviour can be obtained by thermocentrifugal field processing. The supporting polymer matrix is designed to embed solid acids or strong solid acids/ nanooxides making it appropriate for proton conducting composite membranes. Nanocomposites CsHSO4-YSZ/PAN membranes were investigated by Raman, FT-IR and SEM techniques. The dependence of the electrical conductivity on temperature is evaluated.
Keywords: Nanocomposite membrane, solid acid, superprotonic transition
1. Introduction Due to their high efficiency and low emissions, fuel cells have merged as attractive alternatives to combustion engine. With good electroactivity and waste heat treatment solid acid electrolytes have potential for implementing in novel fuel cells.1,2 SAFCs utilize an anhydrous, nonpolymeric proton-conducting electrolyte that can operate at slightly elevated temperatures.3 Their brittleness, the narrow temperature range of the superprotonic phase and chemical instability in hydrogen atmospheres (are reduced to form H2S) are the limiting factors for electrochemical application of these protonic electrolytes.4 For fuel cell
______ *
To whom correspondence should be addressed: I. Stamatin, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
415
A. ANDRONIE ET AL.
416
applications, CsHSO4 offers the advantages of anhydrous proton transport and high temperature stability.1 In the present study, we developed composite CsHSO4/YSZ embedded in a stable polymer matrix of polyacrylonitrile. The effect of doping the ionic salt with nanooxide particle is the ionic conductivity enhancement due to the interface interaction between the two species.3,5,6 2. Experimental Crystals of CsHSO4 samples were synthesized at room temperature from an aqueous solution of Cs2SO4 (Sigma–Aldrich) and H2SO4, Cs:SO4 = 1:2. The precipitated product with ethanol was then dried at 60°C for 3 h. The same procedure was followed for CsHSO4/YSZ synthesis, adding from the beginning a certain amount of YSZ nanopowder (ZrO2 contains 3% Y2O3, 100–120 m2/g, d < 100 nm, Sigma–Aldrich) (2.5, 5, and 7.5% (w/w) of YSZ to CsHSO4). The samples were indexed as follows: CsHSO4, CsHSO4/YSZ_2.5, CsHSO4/YSZ_5 CsHSO4/YSZ_7.5. Dimethylformamide (DMF)–10 wt % polyacrylonitrile (PAN) solutions (DMF-PAN) were prepared and then mixed with CsHSO4/YSZ by ultrasonic processing until homogeneous mixtures were formed. Thin membranes (about 100 Pm) as continuous and pore-free films were formed from each mixture (CsHSO4 or CsHSO4/YSZ: PAN = 1:4) in thermocentrifugal field up to 2000C. The corresponding sample indexing is: CsHSO4-PAN, CsHSO4/YSZ_2.5-PAN, CsHSO4/YSZ_5-PAN, CsHSO4/YSZ_7.5-PAN. 3. Results and Discusions 3.1. SEM IMAGING
SEM (FEI-Quanta 400) reveals a sponge-like morphology of CsHSO4/YSZ composite at different scale as shown in Figure 1.
Figure 1. SEM imaging of CsHSO4/YSZ_2.5 at different scales.
CsHSO4/YSZ POLYMER MEMBRANE
417
3.2. FT-IR AND RAMAN MEASUREMENTS
The Raman (Raman NRS–3100 JASCO) and FT-IR (Jasco FT/IR – 6200 Spectrometer) spectra of CsHSO4/YSZ compounds are shown in Figure 2. The Raman peaks around 1,000 cm–1 observed in all the spectra are basically attributable to stretching vibration of sulfate ions. Stretching modes between 800 and 1,300 cm–1 and bending modes between 400 and 600 cm–1 were observed in the cases of Cs2SO4, but also of CsHSO4 and its YSZ composites. The FT-IR spectra of Cs2SO4, CsHSO4 and its YSZ composites show stretching modes due to sulfate ions between 800 and 1,300 cm–1.
Figure 2. (Left) Raman spectra of: (a) Cs2SO4, (b) CsHSO4, (c) CsHSO4/YSZ_2.5, (d) CsHSO4/ YSZ_5, (e) CsHSO4/YSZ_7.5; (right): FT-IR spectra of: (a) Cs2SO4, (b) CsHSO4, (c) CsHSO4/ YSZ_2.5, (d) CsHSO4 /YSZ_5, (e) CsHSO4/YSZ_7.5.
3.3. ELECTRICAL MEASUREMENTS
The electrical conductivity measurements were performed using Keithley 2400, and a Faraday box with a temperature controller.7 The conductivity of CsHSO4 and CsHSO4/YSZ composite polymer membranes was measured between 353 K and 463 K, as shown in Figure 3. A reduction of the jump in conductivity at the phase-transition temperature has been obtained. Increasing the amount of YSZ leads to a reduction of the conductivity at low temperature.
418
A. ANDRONIE ET AL.
Figure 3. Graphical representation of the conductivity for membranes: CsHSO4-PAN, CsHSO4/ YSZ_2.5-PAN, CsHSO4/YSZ_5-PAN, CsHSO4/YSZ_7.5-PAN.
4. Conclusions Composite solid acid/nano-oxide polymer membranes can be easily obtained by thermo-centrifugal field processing. The introduction of YSZ increases the low-temperature proton conductivity of the material and reduces the jump in conductivity at the phase-transition temperature. Further studies are required for obtaining a proton conducting composite membrane appropriate for commercial fuel cells. ACKNOWLEDGEMENTS
The authors acknowledge research support Grant CEEX-MENER # 704/2006.
References 1. Haile, S., Boysen, D., Chisholm, C., and Merle, R. (2001), Nature, 410, 910–913. 2. Uda, T., and Haile, S. (2005), Electrochemical and Solid-State Letters, 8(5), A245–A246. 3. Ponomareva, V., and Shutova, E. (2005), Solid State Ionics, 176, 2905–2908. 4. Baranov, A., Grebenev, V., Khodan, A., Dolbinina, V., and Efremova, E. (2005), Solid State Ionics, 176, 2871–2874. 5. Ponomareva, V., and Lavrova, G. (1998), Solid State Ionics, 106, 137–141. 6. Colomban, P. (1992), Protonic Conductors (Cambridge University Press, Cambridge). 7. Nastase, C., Mihaiescu, D., Nastase, F., Moldovan, A., and Stamatin, I. (2004), Synthetic Metals, 147, 133–138.
IV AND CV CHARACTERISTICS OF MULTIFUNCTIONAL ILMENITE-HEMATITE 0.67FeTiO3-0.33Fe2O3 C. LOHN1, W.J. GEERTS1*, C.B. O’BRIEN1, J. DOU2, P. PADMINI2, R.K. PANDEY2,3, AND R. SCHAD2 1 Texas State University-San Marcos, Department of Physics, 601 University Drive, San Marcos, TX 78666, USA 2 The University of Alabama, Department of Physics, Tuscaloosa, AL 35487, USA 3 The University of Alabama, Department of Electrical and Computer Engineering, Tuscaloosa, AL 35487, USA
Abstract – We investigated the IV and CV properties of an [(FeTiO3)0.67 (Fe2O3) 0.33] epitaxial thin film. The four point probe (4pp) measurements revealed that the material has a linear IV relation and has a resistivity of approximately 0.56 : cm. In contrast, the two point (2pp) measurements are highly non-linear suggesting the existence of Schottky barriers. The CV data suggest that the material under the contacts is depleted. From the corrected CV data, the carrier concentration is found to be of the order of 1023/cm3.
Keywords: Ilmenite-hematite, multi-functional materials, CV, IV
1. Introduction In conventional electronics, devices manipulate charge in order to store and transfer data. The emerging technology of spintronics uses the phenomenon of electron spin to encode data. Spintronic devices, such as the spin MOSFET first proposed by Datta and Das,1 combine charge transport with spin-dependent effects that arise from the interaction of the charge and properties of the magnetic materials. A major obstacle of spintronics development has been in finding materials with both ferromagnetic and semiconductor properties above room
______ *
To whom correspondence should be addressed: W.J. Geerts, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
419
420
C. LOHN ET AL.
temperature. The ideal ferromagnetic semiconductor should have the Curie point well above the device operating temperature, have high mobility, exhibit a spinsplit band structure and be controllably doped to produce p-type or n-type material. The prediction of Curie temperatures much above room temperature in ferromagnetic semiconductors2 has led to the approach of doping nonmagnetic semiconductor material with magnetic ions. Alternatively, one can investigate materials that are ferro- or ferrimagnetic above room temperature and have semiconductor properties. Individually, ilmenite (FeTiO3) and hematite (Į–Fe2O3) are antiferromagnetic insulators, but compositions of ilmenite-hematite (IH) [(FeTiO3)(1-x)-(Fe2O3)x] are ferrimagnetic over a wide composition range. 3,4 IH systems with a Curie point above room temperature have been demonstrated.5 Furthermore, adjusting the composition can produce p-or n-type material.6 Although the magnetic properties of IH systems have been well documented, data on its dielectric constant, its mobility, and the type of contact it makes to metals are scarce in literature. IH is a multifunctional material; its unique magnetic and dielectric properties make the interpretation of results obtained with standard semiconductor characterization techniques far from trivial. In this paper we attempt to characterize some of the electric properties of the IH system that will be important for a potential device. The focus is on the characterization of metal semiconductor contacts by IV and CV analysis. 2. Experimental Details 2.1. SAMPLE PREPARATION
The IH film was deposited on a single crystal sapphire substrate (10 × 10 mm) by pulsed laser deposition (PLD). The deposition was carried out in an argon atmosphere of 10–3 Torr and at a substrate temperature of 750°C.7 The epitaxial [(FeTiO3)0.67(Fe2O3)0.33] (IH-33) thin film has a thickness of 92 nm as determined by XRR. An Al2O3 layer (142 nm) with a width of approximately 4 mm was deposited at the center of the sample by PLD, leaving on both sides a strip of approximately 3 mm of exposed IH. Six silver contacts were made on the perimeter of the exposed IH thin film using silver epoxy (three on each surface area). A 7th contact was made on top of the oxide in the center of the sample by sputtering Pt through a mask of approximately 2 × 2 mm. Since the oxide contained pinholes or cracks, this center contact should be considered to be a metal-semiconductor contact of which we do not know the effective surface area. The Ag contacts were annealed at 120°C for 20 min to reduce their resistance. The area, the perimeter, and the position of each contact were determined by using a digital top view image of the sample loaded into the Canvas software
IV AND CV CHARACTERISTICS OF ILMENITE-HEMATITE
421
program (error <0.1 mm). The data presented in this paper was measured from Ag contacts 2, 5, 6 and Pt contact 7. 2.2. ELECTRICAL MEASUREMENT SETUPS
2.2.1. IV setup The 2pp and 4pp measurements were taken using an HP 4145B Semiconductor Parameter Analyzer controlled by Metrics ICS software. The sample was mounted in a HP 16058A test fixture. The fixture was closed during the experiments, reducing noise and preventing light from interfering with the experiment. For the 4pp measurements, the voltage sensing contacts were each connected to a 0 ampere current source SMU whose large input impedance, 1012 , guaranteed that the current into the voltage contacts was negligible. 2.2.2. CV setup The CV measurements were performed using a HP-4192A LF Impedance Analyzer controlled by Metrics ICS software with a HP-16048A Test Lead connected to a parallel port connector on the door of a Delta Design 6400 Temperature Chamber. The sample was placed in a test socket mounted on PC board which was connected to the test leads inside the chamber door. Prior to measuring, the zero offsets were performed. CV measurements (parallel and series mode) were taken with a voltage bias sweep of -5–5 V, an oscillation amplitude of 0.1 V, and at a frequency of 100 kHz. 3. Measurement Data and Corrections 3.1. ELECTRICAL TRANSPORT DATA
We determined the resistivity (U) of our sample from 4pp measurements:
U
RCF
Vmeasured t I measured
(1)
Where t is the film thickness and RCF is the calculated correction factor. Using the technique of images and assuming a two dimensional current distribution in our thin film similar to Schroder,8 we calculated the RCFs for all possible 4pp configurations. No attempt was made to correct for contact size effects. The IV relations for all possible 4pp configurations are shown in Figure 1a and appear to be linear. The non-linearity was smaller then 1% for all measurement configurations except for configurations I27V56 and I56V27 that showed larger
C. LOHN ET AL.
422
non-linearity and large RCFs, i.e. 53. Ignoring those results we determined the sample to have a resistivity of 0.56 :-cm which is consistent with literature.5,6 We also measured the 2pp IV relation. The measured 2pp voltage consists of the voltage drops over the contacts and the film. The latter one can be calculated from the sheet resistance and the 2pp correction factor, i.e.:
Vmeasured 2pp
Vc1 Vc 2 Vthin film
Vc1 Vc 2
Us RCF2 pp
I measured
(2)
We estimated the RCF2pp from the 4pp correction factor assuming the electrodes were positioned at the center (current electrodes) and the perimeter (voltage probes) of the two contacts. The IV curve corrected for the thin film resistance, shown in Figure 1b, is clearly non-linear. This can mean two things: (1) intermixing of metal and IH under the contacts creates a non-ohmic alloy; (2) the difference in workfunctions between metal and IH creates a Schottky barrier. In the rest of the paper we will assume that the Ag forms a Schottky contact to the IH-film. Vcontact is the sum of the contact potential over both a forward and a reverse biased Schottky contact. We can model those contacts by a perfect Schottky contact parallel with a conductance G. The reverse biased junction is expected to be ohmic and will add an extra series resistance 1/G to the circuit. The series resistance as determined from the I/(dI/dV) versus I graph8 is approximately 1.4 times larger than the series resistance calculated from the measured sheet resistance.
Figure 1. (a) 4pp IV relation; (b) 2pp IV relation corrected for rs.
3.2. CV DATA
The raw 100 kHz CV data measured in series mode is presented in Figure 2a together with the AC resistance data. The presented data is strongly skewed by the limitations of using a two component measurement model to characterize a
IV AND CV CHARACTERISTICS OF ILMENITE-HEMATITE
423
three component device. Since our Schottky contacts are leaky and the IH has a large resistance both the series resistance (rs) as well as the parallel conductance have to be taken into account. We corrected the raw CV data for the rs calculated from the measured sheet resistance using the following equations:
Gcor jZCcor Vcorrected
ZCs2 Rs rs Z C ( Rs rs ) 1 2
2 s
2
jZ
Cs
Z 2Cs2 Rs rs 1 2
Vmeasured rs * I measured
(3)
(4)
where Rs and Cs are the raw measurement data measured in series mode. The corrected data is shown in Figure 2b. The double peak has disappeared and the curve changed from concave up to concave down. A similar type of correction was done on the data measured in parallel mode resulting in the same corrected capacitance data. The corrected capacitance appears to decrease as a function of the bias voltage. There might be two different explanations for this behavior: (1) the dielectric constant Hr depends on the applied bias voltage (C ~ Hr); (2) a depletion layer is formed under the contacts whose thickness increases with the applied bias voltage resulting in a smaller capacitance (C ~ 1/d). In the rest of this paper we assume that a depletion layer is formed under the metal contacts. The shape of the CV curve can be understood by realizing that the measured capacitance consists of the capacitance of the forward biased junction in series with the capacitor of a reverse biased junction. Since Cforward >> Creverse the measurement data will mainly show the behavior of the smaller of the two, i.e. Creverse. Also the conductance appears to depend on the bias voltage and is maximum for zero bias voltage. This is in qualitative agreement with the non-linear IV characteristics of the contacts. We used the minimum and maximum capacitance in Figure 2b to calculate the maximum and minimum depletion depth.
Figure 2. (a) Impedance analyzer raw data; (b) CV data corrected for series resistance.
424
C. LOHN ET AL.
Assuming the dielectric constant to be 30 and the surface area to be equal to the contact area, we found dmax = 20 Pm and dmin = 5 Pm. This suggests that the whole area under the contacts might be depleted and that the effective contact area is proportional to the contact’s perimeter times the film thickness. We plotted the square of the reciprocal capacitance per area as a function of the bias voltage and determined the carrier concentration in the IH film from the slope8 to be of the order of 1023/cm3. This number is in agreement with former results of Dou7 and suggests that the crude assumptions we made in this paper bare some truth. Calculations using a device simulator to get a better idea on the size and shape of the depletion area as well as measurements on specially designed test structures that will give a more accurate value of the capacitance are planned for the near future. ACKNOWLEDGEMENTS
We would like to thank the US ONR (# N00014-03-1-0358) and the US DoE (#DE-FG02-03ER46039) for their sponsorships. One of the authors, Lohn acknowledges NATO, NSF and Texas State University for travel grants. References 1. Datta, S., and Das, B. (1990) Electric analog of the electro-optic modulator, Appl. Phys. Lett. 56(7), 665–667. 2. Dietl, T., Ohno, H., Matsukura, F., Cibert, J., and Ferrand, D. (2000) Zener model description of ferromagnetism in zinc-blende magnetic semiconductors, Science 287(5455), 1019–1022. 3. Ishikawa,Y., and Akimoto, S. (1957) Magnetic properties of the FeTiO3-Fe2O3 solid solution series, J. Phys. Soc. Jpn. 12(10), 1083–1098. 4. Ishikawa, Y. (1958) Electrical properties of FeTiO3-Fe2O3 solid solution series, J. Phys. Soc. Jpn. 13(1), 37–42. 5. Hojo, H., Fujita, K., Tanaka, K., and Hirao, K. (2006) Epitaxial growth of room-temperature ferromagnetic semiconductor thin films based on the ilmenite-hematite solution, Appl. Phys. Lett. 89, 082509. 6. Zhou, F., Kotru, S., and Pandey, R. (2002) Pulsed laser-deposited ilmenite–hematite films for application in high temperature electronics, Thin Solid Films 408, 33–36. 7. Dou, J., Navarrete, L., Kale, P., Padmini, P., Pandey, R., Guo, H., Gupta, A., and Schad, R. (2007) Preparation and characterization of epitaxial ilmenite-hematite films, J. Appl. Phys. 101, 053908. 8. D. K. Schroder (1998) Semiconductor Material and Device Characterization (Wiley, New York).
ELECTRODEPOSITION OF BI1-XSBX NANOWIRES AS AN ADVANCED MATERIAL FOR THERMOELECTRIC APPLICATIONS J.E. WEBER1,2, W.G. YELTON3, AND A. KUMAR1,2* Department of Mechanical Engineering, Nanomaterials and Nanomanufacturing Research Center, University of South Florida, Tampa, FL USA 2 Department of Mechanical Engineering, Nanomaterials and Nanomanufacturing Research Center, University of South Florida, Tampa, FL USA 3 Photonics Microsystems Technologies, Sandia National Laboratories, Albuquerque, NM USA 1
Abstract – This paper focuses on the electrodeposition of high density nanowire arrays in porous anodic aluminum oxide (AAO) templates. A two step anodization technique was used to develop a template for Bi1-xSbx nanowire growth directly on Si substrates. Uniform pores with virtually no grain boundaries were achieved. Fundamental electrochemistry experiments on Bi3+, Sb3+, and both cations in dimethyl sulfoxide (DMSO) were carried out to characterize the ideal chronopotential pulse and to determine the thermodynamic, diffusion and mass transport issues when plating Bi1-xSbx into nano dimension pore structures. Such chronopotentiometry resulted in uniform nanowire growth within porous channels of the AAO template.
Keywords: Nanowire, electrodeposition, thermoelectric application
______ *
To whom correspondence should be addressed: A. Kumar, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
425
426
J.E. WEBER, W.G. YELTON, AND A. KUMAR
1. Introduction The measure of thermoelectric efficiency is based on a dimensionless ZT figure of merit:
ZT
S 2V
N
T
(1)
where S = Seeback coefficient, ı = electrical conductivity, and ț = thermal conductivity. Currently, bulk Bi1-xSbx with 12% Sb will yield a ZT = 0.88 at 80 K. However, Dresselhaus and coworkers1 have predicted a ZT § 1.25–1.5 at a wire diameter ~35–45 nm and Sb concentration ~ 10–15%. Once synthesized, the Bi1-xSbx nanowire arrays would serve as the P-type semiconducting leg in a thermoelectric device. The nanowire composition is controlled by controlling the plating bath composition and by manipulating the deposition potential. To get a large ZT, it is necessary to have large Seeback coefficients, small electrical conductivity, and small thermal conductivity. The fabrication of well aligned nanowire arrays is a vital project for the realization of novel microelectronic and nano electrical mechanical systems (NEMS). Anodic aluminum oxide (AAO) porous membranes produced on silicon substrates are suitable for a diverse field of applications, including template scaffolds for nanowire growth.2 In this work, aluminum films were anodized to create a porous template in which Bi1-xSbx nanowires were fabricated via electrodeposition. 2. Methods Smooth Nd-doped Al films were grown on a Si substrate on top of a stress-free Pt layer, which bonds to the Si substrate via a W adhesion layer. A two-step anodization process was used to obtain the porous templates. The aluminum films were anodized in 3% oxalic acid chilled at 2°C. The initial anodization was at 50 V for less than 1 min. This quick anodization step provided the rough template. Immediately following the first anodization, the films were etched in a 5% H3PO4, CrO3 bath at room temperature. The time of this etch varied from 10 min to 30 s in order to achieve an optimal etch of the oxide layer. The second anodization step began at 50 V, 2°C with a systematic potential reduction in order to more efficiently etch the barrier oxide layer. Electrochemical experiments were performed using a Radiometer Analytical Voltalab 40 electrochemical analyzer. All electrochemical measurements were executed in a standard three electrode system at room temperature in a solution of dimethyl sulfoxide (DMSO) due to the high solubility of Sb salt in DMSO.
SYNTHESIS OF BI1-XSBX NANOWIRE ARRAYS
427
3. Results Cyclic voltammetry was used to characterize the kinetic limitations of plating Bi3+ Sb3+ into nanoscale pores. It is necessary to know where reduction takes place and what the rate of reduction is (current density, or speed). Figure 1 shows that Bi has a large reduction potential (3 mA/cm2 @ –500 mV) when compared to Sb (–0.5 mA/cm2 @ –500 mV). Electrochemical impedance spectroscopy (EIS) was used to determine the ideal pulse parameters for nanowire growth. This technique was utilized to help explore the electron transfer rate from the electrode to the ion. The Nyquist plot (Figure 2) was used to determine the diffusion coefficient (2.4 × 107 cm/s2) of Bi3+ Sb3+. The time constant was also extrapolated to be 159 ms and the
Figure 1. CV’s of Bi3+, Sb3+, and both cations in DMSO at room temp. Potential vs. Ref: Hg:HgCl (sat KCl).
Figures 2 and 3. Nyquist plot of 50 mM Bi3+ and 50 mM Sb3+ in DMSO at room temperature. Chronopotentiometric pulse used for electrodeposition of Bi1-xSbx in a porous alumina template.
428
J.E. WEBER, W.G. YELTON, AND A. KUMAR
activation current was 419.2 ȝA. To summarize, a pulse of at least 419.2 ȝA for the duration of 159 ms was needed in order to begin the nanowire deposition process. The Bi1-xSbx nanowires were deposited via chronopotentiometry. The application pulse was at –1.5 mA for 250 ms. A small etching pulse of 0.25 mA for 2 s was also used as shown in Figure 3. 4. Discussion Chronopotentiometry was utilized to pulse the deposition potential in order to avoid a large Nernst diffusion layer and create more crystalline nanowires. The short on-pulse helped emulate nanoscale features during the growth process. It was found that by including a small etch pulse a better array of nanowires were produced. The reason for this was twofold. First, the 2 s etch allowed the concentration of ions at the interphase to return to the concentration of the bulk solution. This step helped maintain the correct Sb concentration in the nanowires. The 0.25 mA on-pulse also aided in the etch of any nanowires that have grown up and out of the porous channels. This additional step kept the surface smooth and the current uniformly distributed in the pores during deposition. Figure 4 illustrates that a smaller deposition pulse will create a more uniform growth rate and smoother template surface. To conclude, fundamental electrochemical studies were carried out on Bi3+, Sb3+, and both cations in order to determine any mass transport issues that may arise during nanowire plating. EIS was used to determine the ideal electrodeposition parameters and the Bi1-xSbx were successfully fabricated in an anodic aluminum oxide (AAO) template.
Figure 4. (a) Top-view SEM of Bi1-xSbx nanowires grown out of AAO template at low resolution and at right (b) same sample at high resolution.
SYNTHESIS OF BI1-XSBX NANOWIRE ARRAYS
References 1. 2.
Rabin, O., Lin, Y., and Dresselhaus, M. (2001) Appl. Phys. Lett. 79, 81–83. Liu, C. et al. (2003) Adv. Mater. 15, 838–841.
429
A SOLID STATE NANO-GENERATOR: CONCEPT, DESIGN AND THEORETICAL ESTIMATIONS M. VOPSAROIU1*, M.G. CAIN1, V. KUNCSER2, AND J. BLACKBURN1 1 National Physical Laboratory, Hampton Road, Module G9/A05Teddington TW11 0LW, UK 2 National Institute for Materials Physics, 77125, Bucharest – Magurele, ROMANIA
Abstract – Nano-technology is a very attractive area of research and innovation because it allows the current trends in miniaturization to continue. The transition from micro scale to nano scale devices has already taken place in many applications such as electronics, magnetic recording and nano-biophysics. However, as we scale down the size of the structures and devices, it becomes obvious that the classical behavior will break down at the nano-scale and an interesting superposition of classical and quantum effects will emerge. Therefore, the validity of classical physics is questioned and many aspects of physics are now being revisited from the point of view of nano-technologies. In line with the new developments in miniaturization and nano-technologies, we propose in this letter a simple mechanism that applies the Faraday effect at the nano-scale in order to create a possible solid-state energy nano-generator device. The proposed nanogenerator functionality is based on what we shall call the Super-Paramagnetic Electromotive Force (SPEF) effect. This has the potential to produce a very small voltage on short time scales by converting directly thermal energy at room temperature to electromotive energy without the need for external work or mechanical motion.
Keywords: Thermal energy, superparamagnetism, nano-generator, Faraday effect
______ *
To whom correspondence should be addressed: M. Vopsaroiu, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
431
M. VOPSAROIU ET AL.
432
1. Introduction and Concept Description The Faraday effect states that an electromotive voltage can be induced in an N turn coil, by a time fluctuating magnetic flux. Mathematically the electromotive voltage generated (e) is proportional to the rate of change of the magnetic flux through the circuit:
e
N
dI dt
N
d BA cos T dt
(1)
where I is the magnetic flux, B is the magnetic field, A is the area of the coil and T the angle between the magnetic field vector and the normal to the coil plane. According to Eq. (1) the only requirement to induce a voltage in the coil is the time variation of either B, A, T or any combination between them. This is typically achieved by mechanical periodic motion through the rotation of the coil in a constant magnetic field or the magnetic field around a fixed coil. Hence, mechanical work is converted into electrical energy via the Faraday effect (i.e. electromagnetic induction). The main idea of this letter is to reduce the size of the system down to the nano-meter range and then to apply the same principle. However, when we scale the system down to the nano-meter range, the main advantage is that the requirement for mechanical motion is no longer necessary and a time variation of the magnetic flux can be engineered by other means. Let us imagine that the magnet is scaled down to the size of a single domain ferromagnetic nano-particle having an ellipsoidal shape. A magnetic single-domain state can be achieved in a nanoparticle by reducing its volume below a critical size where the magnetic exchange interactions dominate the magnetostatic interactions. The combined shape and magneto-crystalline anisotropies will generate an easy axis parallel to the long axis of the ellipsoid forcing the net magnetic moment (m) of the particle parallel to this easy axis. It is assumed here that the magneto-crystalline is dominant over the shape anisotropy. The nano-particle system is in effect a small magnetic dipole, which can generate at a distance r a magnetic field B given in absolute value by Eq. (2).
B
P0 m 4Sr 3
1 3 cos D 2
1 2
(2)
where P0 is the magnetic permeability of vacuum and m is the magnetic dipole moment and D is angle between the easy axis and the r direction. We now assume a single nano-coil surrounding the magnetic nano-particle (Figure 1). Just as in the classical Faraday effect, a time dependent magnetic flux through the coil is
SUPER-PARAMAGNETIC ELECTROMOTIVE FORCE (SPEF)
433
Figure 1. Schematic diagram of the system nano-coil/superparamagnetic nano-particle. The long axis of the particle can be either in the plane of the coil or perpendicular to it. Thermal jumps between the two equilibrium states +m and –m along the easy axis will generate a time dependent magnetic flux through the nano-coil.
required to produce an electromotive force. In order to generate the time dependent magnetic flux, we can make use of the super-paramagnetic state of the nanoparticle. In the super-paramagnetic state the nano-particle possesses an intrinsic magnetic moment that randomly fluctuates between the two 0° and 180° equilibrium states due to the thermal activation (see Figure 2). This is achieved in nanoparticles for which the magneto-crystalline energy term (KuV) is dominated by the thermal Boltzman energy (KBT) at a temperature T z 0. Because the anisotropy term depends on the volume of the particle, by reducing its size, a super-paramagnetic state is achieved at room temperature. Also the magnetocrystalline energy term (KuV) can be further reduced by choosing ferromagnetic materials with anisotropy energy constant Ku as small as possible. Therefore, from the magnetic point of view, we deal with a magnetic dipole, which fluctuates along the easy axis direction. Recalling relation (2), a thermally induced time fluctuating magnetic dipole m(t) would generate at a distance r a time dependent magnetic field B(t), which in turn leads to a time dependent magnetic flux, I (t ) , through the nano-coil. The Faraday effect is induced, except that in this case there is no mechanical work or motion involved and the time dependent magnetic flux is achieved by converting the thermal energy of the environment to thermal fluctuations of the magnetization via the super-paramagnetic effect. Hence, the surrounding room temperature thermal energy is converted directly into electromotive energy and we propose to name this effect: The Super-Paramagnetic Electromotive Force (SPEF) effect.
M. VOPSAROIU ET AL.
434
Energy
KBT
m
+m
0
60
F
120
180
Figure 2. Energy (E = KuVsin2I) of a magnetic nano-particle in zero applied field as a function of the angle I between its magnetization and the easy axis. There are two energy minima (0°, 180°) with the energy barrier KuV between them. When the thermal Boltzman energy KBT is large enough to overcome the barrier, thermal jumps between the two energy minima occur randomly.
2. Theoretical Estimations and Discussions Let us assume the following relation describing the stochastic time fluctuating magnetic moment: m(t) = m cos(Zt) + f(t)
(3)
where m is the absolute value of the magnetic moment and f(t) is a stochastique noise parameter superimposed on to a harmonic oscillation, Z = 2S/W and W is the super-paramagnetic relaxation time given by the relation:
W
1
# 2 Q 0 e
K uV K BT
(4)
where Q0 # 109 s–1 and represents the attempt frequency to overcome the energy barrier. Using the relation m = MV, where M is the magnetization of the material and V is the volume of the particle, combined with relations (2) and (3), the magnitude of the magnetic field can now be expressed as:
B(t )
1 P 0 MV cos(Zt ) 2 2 1 3 cos D 4Sr 3
(5)
where for simplicity we have ignored the stochastic noise parameter. The approximate induced electromotive voltage through a nano-coil of radius L when the particle is located at the centre of the coil with the long axis normal to the plane of the coil (i.e. D = 90°) is then:
SUPER-PARAMAGNETIC ELECTROMOTIVE FORCE (SPEF)
e
dI dt
P 0 MV 2L
Z sin(Zt )
emax sin(Zt )
435
(6)
In order to estimate the maximum induced voltage (emax) we now consider a soft NiFe nano-particle. The following parameters are used: M = 1 T = 107/4S A/m, Ku = 102 J/m3, nano-coil radius L = 1.5 × 10–7 m. The calculations have been made for a relaxation time of W = 10-8 s and the thermal energy at room temperature is KBT = 4.14 × 10–21 J. Using relation (4), the volume of the NiFe particle has been estimated as: V # 1.24 × 10–22 m3. Using these values and relation (6), the estimated induced maximum voltage by a single nano-particle is and emax = 259 nV. Assuming a single Cu nano-coil of radius L = 150 nm, cross section area of A = 10–15 m2 and resistivity U = 2.24 × 10–8 : m, we can estimate the electrical resistance of the nano-coil as: R = U2SL/A = 21 :. Hence, the maximum current flowing through the Cu nano-coil will be Imax = 12.3 nA. However, in order to minimize the power loss through electron collisions in the coil, an interesting and challenging exercise would be to use carbon nano-tubes instead of Cu. These have been intensively studied and they are known as ballistic 1D conductors1 with mean free path extending to micrometers. By using a high conductivity carbon nano-tube coil the Joule heat loss in the coil could be minimized and the efficiency of the nano-generator could be maximized. The system nano-coil/super-paramagnetic nano-particle will behave in effect as a nano AC voltage generator by converting the surrounding room temperature thermal energy into electrical energy. The average frequency of the induced voltage from a particle is given by its super-paramagnetic relaxation time. Although the induced voltage/current appear to be very small, it is important to mention that our estimations are for a single nano-particle inside a single Cu nano-coil. Besides using carbon nano-tubes instead of Cu, other options to increase the power output are to optimize the particle material, the shape/size of the particle and coil, the use of an N turns nano-coil instead of a single nanocoil and a large number of nano-particles. Of course, one has to consider the technical challenges in fabricating such a device (for example we have ignored the auto-inductance effects, coil impedance, thermal Johnson noise, stochastic nature of the fluctuations, etc.) but the purpose of the paper is to communicate this idea rather than to give specific solutions to the practical realization. Most important fact is that the SPEF effect used with the nano-generator has the potential to produce a very small amount of electromotive work from the thermal energy of the environment. Since the electromotive work is produced directly from the heat of the environment, this appears to violate the second law of thermodynamics. However, new theoretical and experimental scientific developments have shown that the second law is applicable to large macroscopic systems while its violation is possible in small systems on short time scales. This has been
436
M. VOPSAROIU ET AL.
widely reported in the literature: Second law of thermodynamics “broken” [NewScientist, 19 July 2002; Second law broken, Nature July 23, 2002] and the justification is based on the Fluctuation Theorem developed by Evans et al.,2,3 which has been proven both theoretically4 and experimentally.5,6 In support of the idea presented in this letter we would like to quote a paragraph from reference3 that implies exactly the possibility of converting heat into work as in the SPEF effect: “If the work performed during the duty cycle of any machine is comparable to the thermal energy per degree of freedom in the system (i.e. KBT), the Fluctuation Theorems say that there is a significant probability that the machine will actually run backwards. In violation of the Second Law, heat energy from the surroundings can be converted into useful work to provide sufficient energy for the machine to run in reverse” p. 1580. 3. Conclusion The Super-paramagnetic Electromotive Force (SPEF) effect has been proposed in this paper. This works by converting the thermal energy of the environment into a time variation magnetic flux via the super-paramagnetic effect, which in turn is used to create electromotive work in a nano-coil. Although the proposed energy nano-generator is very promising, the practical realization and optimization of such a device remains at this stage a challenge for the nano-fabrication research groups around the world. If successfully fabricated, the nano-generator has possible applications as in-situ power source for nano/molecular devices and for fundamental studies of relaxation phenomena in nano-magnetic systems.
References 1. Poncharal, P., Berger, C., Yi, Y., Wang, Z., and de Heer, W. (2002), J. Phys. Chem. B 106, 12104. 2. Evans, D., Cohen, E., and Morris, G. (1993), Phys. Rev. Lett. 71, 2401. 3. Evans, D., and Searles, D. (2002), Adv. Phys. 51, 1529–1585. 4. Evans, D., and Searles, D. (1994), Phys. Rev. E 50, 1645–1648. 5. Wang, G., Sevick, E., Mittag, E., Searles, D., and Evans D. (2002), Phys. Rev. Lett. 89, 050601. 6. Carberry, D., Reid, J., Waang, G., Sevick, E., Searles, D., and Evans, D. (2004), Phys. Rev. Lett. 92, 140601.
APPLICATIONS OF STATISTICAL PHYSICS TO MIXING IN MICROCHANNELS: ENTROPY AND MULTIFRACTALS M. KAUFMAN1*, M. CAMESASCA2, I. MANAS-ZLOCZOWER2, L.A. DUDIK3, AND C. LIU3 1 Department of Physics, Cleveland State University, Cleveland, OH.44115, USA 2 Department of Macromolecular Science, Case Western Reserve University, Cleveland OH.44106, USA 3 Electronics Design Center, Case Western Reserve University, Cleveland OH.44106, USA
Abstract – We apply rigorous measures of mixing based on entropy in conjunction with fractals to the field of microfluidics. First we determine the entropy and multifractal dimensions of images of mixing a fluorescent and a non-fluorescent fluid in a microchannel. We find the microstructures to be self-similar (fractals). Second we propose a new approach for patterning the walls of microchannels using the Weierstrass function. We have evidence from numerical simulations that by properly adjusting the dimension of the Weierstrass function one can get microfluidic devices that exhibit better mixing than the current ones.
Keywords: Microfluidics, entropy, fractals, mixing 1. Introduction Microfluidic systems operate in a pressure driven flow regime with no moving parts to drag the fluids. Since the flows are laminar and diffusion is in general small, mixing can be achieved in such devices by patterning the channel walls.1 To design microchannels that are efficient mixers it is important to develop rigorous assessment and quantification tools of mixing. To this end we proposed2 to use Shannon and Renyi entropies. To further our understanding of mixing, we also characterize the geometric structures generated by the flow by using
______ *
To whom correspondence should be addressed: M. Kaufman, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
437
M. KAUFMAN ET AL.
438
multifractal dimensions. We apply the entropic and multifractal tools on the published1 images of mixing of a fluorescent and a non-fluorescent fluid in a microchannel. Stroock et al.1 have achieved mixing in microchannels by patterning the bottom wall of the duct in a device called the Staggered Herringbone Mixer (SHM). We investigate3 a non-periodic patterning design based on the Weierstrass fractal function.4 We present numerical simulations of three channels designed accordingly and a fourth were the patterning is periodic. We quantify the degree of mixing by using the mixing entropy. We also report here on our work to develop an experimental prototype of a fractal microchannel mixer. 2. Renyi-Shannon Entropy and Multifractal Dimensions The Shannon5 entropy S measures mixing as it is uniquely determined from the Khinchin axioms6: it depends on the probability distribution p only; the lowest entropy corresponds to one of the p’s being 1, i.e. no mixing; the largest value for the entropy is achieved when all p’s are equal to each other, i.e. perfect mixing; S is additive over partitions of the outcomes. For an experiment with N outcomes the Shannon entropy is: N
S
¦ p j ln p j
(1)
j 1
If the last Khinchin axiom is relaxed to consider only statistically independent partitions, Rényi7 found that the information entropy is replaced by a one-variable function:
S (E )
§ N · 1 ln ¨ ¦ p j E ¸ 1 E © j 1 ¹
(2)
For ȕ = 1 the Renyi entropy equals the Shannon entropy. Multifractals8,9 are self-similar complex structures characterized by a spectrum of dimensions d(E). The multifractal dimensions are obtained4 from the linear dependence of the Renyi entropy S(E) on ln(N). The slope equals d(E)/D, where D = 2 is the embedding dimension. In Section 3 we apply this formalism to images of binary fluids mixing in a microchannel.1 In view of the additivity axiom, the Shannon entropy for a multicomponent system can be expressed as:
S
S (locations) Slocations ( species)
(3)
MIXING IN MICROCHANNELS
439
where: N
Slocations ( species)
¦ ª¬ p S ( species) º¼ j
j
(4)
j 1
C
S j ( species)
¦ ª¬ pc / j ln pc / j º¼
(5)
c 1
N
S (locations )
¦ ª¬ p j ln p j º¼
(6)
j 1
where: pj is probability of a particle to be in bin #j and pc/j is probability of a particle to be of species c conditional on being in bin #j. In section 4 we analyze the mixing of two incompressible fluids. Since the density at any location does not change in time, the entropy S(location) is a time constant. Thus we concentrate on Slocation(species) which measures the local quality of mixing of the two fluids averaged over space. Since 0 Slocation(species) ln(C), we normalize the species entropy conditional on locations Slocation(species)/ln(C) to get an index of mixing varying between 0 for perfect segregation and 1 for perfect mixing. 3. Fractal Analysis of Images We analyze the six structures generated in the staggered herringbone mixer and reported by Stroock et al.1 in terms of their multifractal dimensions. The six images are converted to grayscale. Each pixel j has a gray scale value xj varying between 0, black, and 255, white. x measures the local concentration of the fluorescent (white) fluid. Renyi entropies are calculated for each image by substituting in Eq. (2) the probabilities computed using the reading from each pixel:
pj
xj
¦x
(7)
i
The multifractal dimensions are obtained from a multi-scale analysis of the pictures. For each picture we considered 20 scales of observation. The slope of the linear dependence of entropy on the logarithm of the number of bins yields the dimension. The high values of the Pearson correlation coefficient, in all cases considered, point to the fact that the structures are self-similar (multifractals).
440
M. KAUFMAN ET AL.
In Figure 1 we show the Renyi entropies for ȕ = 0, 1, and 2 for the six sections presented in Stroock et al.,1 measuring the spatial distribution of the fluorescent (white) fluid. The qualitative conclusion that mixing progresses from Sections 1 to 6 is confirmed quantitatively by the monotonic increase in entropy. In Figure 2 we show two sections from Stroock et al.1 and the corresponding multifractal dimensions function d(ȕ). The dimensions of the later section are higher, again consistent with the higher degree of mixing in the later section.
Figure 1. Renyi entropies versus section; symbol × for ȕ = 0; symbol + for ȕ = 1 (Shannon entropy); symbol square for ȕ = 2.
Figure 2. Sections #4 and #6 of Stroock et al. and the corresponding multifractal dimensions functions d(ȕ).
MIXING IN MICROCHANNELS
441
4. Microchannels with Fractal Grooves 4.1. WEIERSTRASS FUNCTION AND SURFACE PATTERNING
We recently proposed3 a new geometric design to pattern the microchannel wall. On one hand we preserve Stroock et al.1 asymmetric V shaped grooves, as they generate two stagnation points in the cross sections, and thus fluid rotation in clockwise and counter-clockwise direction at the same time. On the other hand, unlike Stroock et al., we locate the grooves in a non-periodic pattern along the channel axis. To this end we use the Weierstrass function to generate the coordinates of the “V” grooves tips. The Weierstrass function4 is:
W ( x)
f
¦
cos 2n x
n 0
2
n (2 D )
(8)
Its fractal dimension is D is easily controlled parametrically. We evaluate the function at M points x j j 2MS for j = 1, 2 …M. The coordinates of the grooves vertices are (xj, W(xj)). We also apply a linear transformation on each coordinate separately to insure that the M vertices occupy the length, 5,000 ȝm, and the width, 200 ȝm, of the channel. To generate asymmetric V-shaped ridges these vertices are connected to the lateral walls. As shown in Figure 3, to better approximate a more complex function, of higher dimension D, one needs a larger sampling rate M. Both the dimension D and the sampling rate M are important parameters. In our exploratory work we used a fixed M = 50 for three dimensions D = 1.25, 1.50, 1.75, labeled respectively: “W1.25”, “W1.50”, “W1.75”. In Figure 4 we show the W1.25 channel. All the channels have a length of 5,000 ȝm, width of 200 ȝm and height of 150 ȝm. Each ridge has a height of 50 ȝm. We have also studied numerically the SHM channel of Stroock et al.1 The flow field for the four different geometries is obtained by solving numerically the continuity equation for incompressible fluids and the NavierStokes equation of motion with no-slip boundary conditions:
G Gº G ª wv G U « v v » p P2 v ¬ wt ¼ G v 0
(9) (10)
In Eqs. (9) and (10) v is the velocity vector, ȡ the density, t the time, ȝ the viscosity and p the pressure. The fluid is Newtonian with density of 103 kg/m3
442
M. KAUFMAN ET AL.
Figure 3. Weierstrass function plots for the cases of D = 1.25, D = 1.50, D = 1.75. In each case we show M = 50 vertices (circles).
Figure 4. The “W1.25” channel.
and viscosity of 10–3 Pa s. We integrate numerically the velocity field, to get the trajectories of 26,000 massless non-interacting particles of two species (13,000 blue and 13,000 red), and compute the mixing entropy defined in Eq. (4). In Figure 5 the entropy is calculated using 1,200 bins. The mixing efficiency as measured by the mixing entropy follows the following order: W1.25 > W1.50 > mSHM > W1.75.
MIXING IN MICROCHANNELS
443
Figure 5. Normalized mixing entropy versus distance along channel. The W1.25 and W1.50 exhibit higher mixing entropies than the SHM.
4.2. MANUFACTURE THE MICROCHANNEL WITH WEIERSTRASS PATTERNING
A physical model of the channels is needed in order to validate the computer model. A method to produce a three dimensional representation of the grooves in a microchannel that also had fluid inlet and outlets was developed. Since the dimensions of the microchannel are small (width of the channel is approximately 1 mm with the raised grooves having height and width of 50 μm), it was necessary to use a photolithographic technique along with other fabrication methods in order to create the device. The final device created is shown below in Figure 6. It consists of a Pyrex substrate with the 50 micrometer high ridges, a 200 μm high patterned plastic sheet with adhesive on both sides, a piece of acrylic with holes and tubing connectors. The ridges need to be at the bottom of a channel in
Figure 6. Fabricated microchannel.
444
M. KAUFMAN ET AL.
which fluid will flow. Creating a straight walled channel can be difficult using conventional microfabrication techniques. In order to create a testing device that would allow for different thicknesses of channels we adapted a technique used in the production of commercial multi-layered polymer sensors such as the glucose sensors used by diabetics. The technique involves patterning different layers of plastic and then laminating them together to form a monolithic device. 5. Conclusions Mixing can generate complex structures. A fractal analysis of images complements the entropic mixing characterization. Future calibration work is needed to study the influence of the type of image file (.bmp, .jpeg, .tif) on the estimated fractal and entropic measures. Microchannels with patterned surfaces using a fractal function have been designed and characterized for their mixing efficiency. We found numerically that two of the mixers designed using the fractal approach in patterning are better mixers than the SHM of Stroock et al.1 One would expect the higher fractal dimension (more complex) structure to yield higher mixing. The results of Figure 5 show the opposite trend. We believe that the sampling rate M needs to be higher for the higher D structure. This problem needs further study.
References 1. 2. 3. 4. 5. 6. 7. 8. 9.
Stroock, A. et al. G. (2002) Science 295, 647–651. Camesasca, M. et al. (2005) J. Micromech. Microeng. 15, 2038–2044. Camesasca, M. et al. (2006) J. Micromech. Microeng. 16, 2298–2311. Schroeder, M. (1991) Fractals, Chaos, Power Laws (New York, W. H. Freeman). Weaver, W., and Shannon, C. (1963) The Mathematical Theory of Communications (Urbana, IL, Illinois University Press). Khinchin, A. (1957) Mathematical Foundations of Information Theory (New York, Dover). Rényi, A. (1960) Theory of Probability (Amsterdam, North-Holland). Mandelbrot, A. (1982) The Fractal Geometry of Nature (New York, W. H. Freeman). Grassberger, P., and Procaccia, I. (1984) Physica D 13, 34–54.
SYNTHESIS AND CHARACTERIZATION OF CARBON SUPPORTED Pd AND PtPd CATALYSTS FOR DMFCs A. MOROZAN1*, A. DUMITRU1, C. MIREA1, I. STAMATIN1, F. NASTASE1, A. ANDRONIE1, S. VULPE1, C. NASTASE1, AND A. VASEASHTA2 1 3Nano-SAE Research Centre, University of Bucharest, P.O. Box MG-38, 077125, Bucharest-Magurele, ROMANIA 2 On details in Washington DC from Nanomaterials Processing and Characterization Labs. Marshall University West Virginia, USA
Abstract – Carbon supported Pd and PtPd catalysts were prepared by chemical reduction of metal precursors with sodium borohydride. Multi-wall carbon nanotubes (MWNTs) were used as carbon support. The composition of prepared catalysts was given by EDX technique. Cyclic voltammetry (CV) measurements were used to evaluate the electrocatalytic activities for methanol electrooxidation after their immobilization onto carbon paper.
Keywords: Catalyst, methanol oxidation, MWNTs.
1. Introduction Direct methanol fuel cell (DMFC) is one of the most advanced low-temperature fuel cells owing to the choice for many portable/micro power applications. However, there are some limitations such as the operational temperature, sluggish methanol oxidation kinetics, methanol crossover through the membrane electrolyte and the requirement for a greater loading (density) of catalyst to electrodes.1,2 Pt catalyst for hydrogen oxidation is not suitable for methanol due to CO formed as a reaction intermediate product, irreversibly absorbed to the Pt surface, rapidly lowering its activity.3 The electrocatalytic system Pt-Ru for methanol oxidation reduces the CO-poisoning by Ru, which removes COabs as CO2 (gas).4
______ *
To whom correspondence should be addressed: A. Morozan, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
445
446
A. MOROZAN ET AL.
It is well established that bimetallic systems, with Pt as one of the components, give substantial tolerance compared to Pt alone, to the presence of CO in the fuel stream. Another noble metal with catalytic activity is Pd, which has the remarkable ability to store and release substantial amounts of hydrogen.5 For practical electrodes of a DMFC, carbon nanotubes (CNTs) are considered to be the most appropriate potential supports, especially for heterogeneous catalysts to be employed in liquid-phase reactions. Comparing with the most widely used Vulcan XC-72R carbon support, CNTs have significantly higher electronic conductivities and extremely high specific surface area.6 Moreover, higher activity of the oxygen reduction reaction and better performance of the DMFC compared to the catalysts supported on commercial carbons, was achieved.7,8 2. Experimental MWNTs from Shenzhen Nanotech PortoCo., Ltd., China (d<10 nm, length: 5–15 Pm, P: t95%, specific surface area: 40–300 m2/g) were used. To improve the catalyst dispersion on CNTs, the activation step is essential. MWNTs were refluxed with HNO3 (3M) (Merck reagent) for 24 h, followed by washing with de-ionized water until neutral pH and then drying in air for 30 min at 100°C.9 The Pd/MWNTs and PtPd/MWNTs catalysts were prepared by a conventional borohydride reduction method using NaBH4 (15 mM) and metal precursors, PdCl2 (0.5M) and H2PtCl6 (10%) – (Sigma-Aldrich products). The composition of the prepared catalysts was determined by energy-dispersive X-ray analysis (EDX) in a FEI-Quanta 400 scanning electron microscope provided with a microanalyser. The electrochemical measurements experiments were performed with a VoltaLab 40 (PGZ301, RADIOMETER) using a three-electrode cell. Pt mesh and Hg/Hg2SO4 (saturated by K2SO4 solution) were used as the counter and reference electrode, respectively. The working electrode was carbon paper painted with catalysts ink prepared by directly mixing catalyst powder with acetone in an ultrasonic bath. The cyclic voltammetry was carried out in a solution containing 0.5M H2SO4, 0.5M H2SO4 with 2M methanol, and 0.5M H2SO4 with 4M methanol, at room temperature. 3. Results and Discusions The results by EDX analyses have indicated that the final compositions were 71:29 for C:Pd and 57:28:15 for C:Pt:Pd (wt %). EDX show small discrepancies when compared with the nominal compositions expected from the relative amounts of precursors used in the preparation of the metal catalysts.
Pd AND PtPd CATALYSTS
447
Figure 1 illustrates the cyclic voltammograms of the PtPd/MWNT and Pd/ MWNT in 0.5 M H2SO4 solution saturated with nitrogen (a) and oxygen (b). The solution is first degassed, by bubbling nitrogen gas for about 15 min to record the background current–voltage curves. Both electrocatalysts, PtPd/MWNT and Pd/MWNT, have a well-defined hydrogen adsorption-desorption region in acid media. PtPd/MWNT has a good electroactivity towards oxygen reduction reaction (ORR), oxygen reduction occurs at ca. 0.3 V. The limiting current of PtPd/MWNT in 0.5 M H2SO4 solution saturated with oxygen is 1.3 higher than in 0.5 M H2SO4 solution saturated with nitrogen. The electroactivity of Pd/MWNT under the same conditions is not obvious towards ORR.10,11
Figure 1. Cyclic voltammograms of PtPd/MWNT and Pd/MWNT in 0.5 M H2SO4, saturated with nitrogen (a) and oxygen (b); 5 mV/s.
Figure 2. Cyclic voltammograms of PtPd/MWNT and Pd/MWNT in: 0.5 M H2SO4 + 2 M methanol, 0.5 M H2SO4 + 4 M methanol.
448
A. MOROZAN ET AL.
Figure 2 illustrates the cyclic voltammograms of the PtPd/MWNT and Pd/MWNT in 0.5 M H2SO4 with 2 M methanol and 0.5 M H2SO4 with 4 M methanol. Pd/MWNT do not show any peak current attributed to the methanol oxidation, whereas the PtPd-based catalyst exhibit a high peak current related to the oxidation of methanol. 4. Conclusions Single as well as bimetallic nanoparticles on carbon support have attracted a lot of attention for their potential as catalysts for DMFCs. Carbon nanotubes supported with Pd and PtPd were synthesized by a rapid approach. The electrocatalytic activity of PtPd/MWNT for ORR in an acid medium was demonstrated through CV. To reduce the amount of expensive platinum loading in the electrodes, PtPd/MWNT is a good choice proven by the high peak current related to the oxidation of methanol. ACKNOWLEDGEMENTS
This work was financially supported by contract CBP.EAP.RIG 982391 and by Grant CEEX-MENER No. 760/2006.
References 1. Ilic, D., Holl, K., Birke, P., Wohrle, T., Birke-Salam, F., Perner, A., and Haug, P. (2006) J. Power Sources, 155, 72–76. 2. Wasmus, S., and Küver, A. (1999) J. Electroanal. Chem., 461(1–2), 14–31. 3. Mehta, V., and Cooper, J.S. (2003) J. Power Sources, 114, 32–53. 4. Kawaguchi, T., Sugimoto, W., Murakami, Y., and Takasu, Y. (2005) J. Catal., 229, 176–184. 5. Lewis, F.A. (1967) The Palladium Hydrogen System, Academic, London. 6. Leei, K., Zhang, J., Wang, H., and Wilkinson, D.P. (2006) J. Appl. Electrochem., 36, 507–522. 7. Ebbesen, T.W., and Ajayan, P.M. (1992) Nature, 358, 220–222. 8. Li, W.Z., Xie, S.S., Qian, L.X., Chang, B.H., Zou, B.S., Zhou, W.Y., Zhao, R.A., and Wang, G. (1996) Science, 274, 1701–1704. 9. Colomer, J.F., Piedigrosso, P., Fonseca, A., and Nagy, J.B. (1999) Synth. Met., 103, 2482–2483. 10. Paun, T.V.P., Paun, M.A., Toma, A., and Ciucu, C. (2008) Mater. Plastic, 45(1), 57. 11. Berlic, C., Barna, E., and Ciucu, C. (2007) J. Optoelectronic Adv. Mater, 9(12), 3854.
THEORETICAL STUDY OF THE ADSORBED SMALL MOLECULE ON TWISTED NANOTUBES BY ATOMIC SCALE SIMULATIONS V. CHIHAIA,1* A. GHITA,1 B.-S. SEONG2, AND S.-H. SUH3 1 Institute of Physical Chemistry “Ilie Murgulescu”, Splaiul Independentei, 202, Bucharest, 060021, ROMANIA 2 HANARO Center, Korea Atomic Energy Research Institute, P.O. Box 150, Yusong, Taejon, 305-600, KOREA 3 Department of Chemical Engineering, Keimyung University, Taegu, 704-701, KOREA
Abstract – The adsorption of the hydrogen and methane molecules on the untwisted and twisted carbon nanotubes of type (5,5) are investigated by atomic scale simulation, using the empirical bond order (REBO) force field.1 It is found that the adsorptions of both molecules on the untwisted nanotube are endothermic processes, while the adsorptions on the twisted nanotube become exothermic and moreover the molecules are dissociating on some of the adsorption sites.
Keywords: H2, CH4, Adsorption, carbon nanotube.
1. Introduction Having exceptional mechanical and electrical properties, the carbon nanotubes are one of the most challenging materials, offering great applications including nanoelectronic devices, energy and molecule storage, chemical probes and biosensors. Sumio Ijima identified for the first time the carbon nanotubes as byproducts of arch discharge experiments.2 A nanotube can G be seen as a graphene sheet rolled into a cylinder along a periodical vector C , characterized by two integer parameters m and n. The cylinder radius R depends also on m and n. The pair (n,m) characterizes the type of nanotube as it determines the helical arrangements of the carbon hexagonal rings. The properties of the nanotubes
______ *
To whom correspondence should be addressed: A. Ghita, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
449
450
V. CHIHAIA ET AL.
depend on their type (radius and chirality) and may be modified by creation of local defects and by mechanical deformation. The adsorption of atoms and molecules on nanotubes is another way to modify their properties.3 However, the molecule-nanotube interactions are too weak to fix the molecules on the nanotubes.4 Stronger interactions are observed on the ends of the nanotubes,5 but more enhanced interactions are obtained by the nanotube distortions6 or by radical bombardments.7 It appears that the lateral pressing of the nanotube adsorption is not too efficient in the increasing of the nanotube reactivity.8 Srivastva et al.6 found by REBO calculations that the locally controlled deformations produced by the bending and the twisting of the nanotube significantly enhance the interaction of the atomic hydrogen with the nanotubes (10,10). Herein we report the results of the atomic scale simulations with the REBO potential, designed to investigate the enhancement of interactions of molecular hydrogen and methane with the single wall carbon nanotube (SWNT) of type (5,5) by the nanotube twisting. 2. Computational Methods The total energy and the interatomic forces are calculated by the analytical potential REBO in the second parameterization6 for short-range interactions ( r d 5 Å) and by 6–12 Lennard–Jones potential9 for the long-range interactions, using the atomic scale simulation code Brenner.10 The REBO potential contains a repulsive and an attractive component, which are essentially potentials of two-body Morse type. The attractive part is weighted by a factor that depends on the neighborhood of each interacting atom and that considers the local geometry of the atoms. The present study implies two atomic scale simulation tools, the geometry optimization and the molecular dynamics (MD). The full and partial geometry optimizations are done by the conjugate gradient method, using the cartesian coordinates of the atoms as variables. The optimization procedure is considered satisfactory when the maximum gradient of the total energy has a value bellow the threshold of 10–5 kcal/mol/Å. The MD calculations are carried out in a canonical ensemble (the parameters N-number of atoms, V-volume and T-temperature are constants). The temperature is controlled by the Berendsen technique.11 The Newton equation of motion of the atoms is integrated by the predictor-corrector method based on the Nordsieck-Gear algorithm,12 with a time step of 0.5 fs. An armchair carbon nanotube (5,5) oriented with its axis along the z-axis and of A = 60 Å long, consisting in 500 carbon atoms, is generated by the code Dommelen.13 The structure of the nanotube is fully optimized. In further procedures, five percents of the nanotube at the both ends are treated as rigid
SMALL MOLECULE ADSORPTION ON DEFORMED NANOTUBES
451
(rigid areas): the atoms in this area are considered fixed in the geometry optimization and MD calculations, but considering their interaction with the other atoms. The atoms in the rest of the nanotube (in the area between zmin = –45·A% and zmax = 45·A%) are fully optimized or considered active in the MD calculations. This area is called active area. The twisted nanotube used in this study is prepared in several stages. Starting from the optimized geometry, the nanotube is twisted in steps of 'T = 10°: the rigid areas are rigidly rotated with the angle 'T in opposite directions and the atoms from the active areas are rotated around z-axis by the angle:
T ( zi )
zi z mean 'T zmax zmin
(1)
where zi-is the coordinate of the atom i, zmean = (zmin + zmax)/2 indicates the middle of the nanotube. After each twisting step, the geometry of the nanotube is optimized excepting the carbon atoms in the rigid areas. After 20 steps a first jump in the nanotube total energy is observed when the nanotube become buckled as in a helicoidal ribbon. The different ribbon regions (ridge, intermediate, middle) are characterized by different curvatures. After further twisting steps the structure becomes more compact. Some 5-7-7-5 dislocations formed by the Stone-Wales transformation14 are observed. A new jump in the total energy is observed at the step 30, when the nanotube collapse and some carbon atoms become four coordinated. We have selected this structure for the adsorption study. To reduce the tensions from the twisted nanotube, the nanotube is heated to 1,000 K in a molecular dynamics stage for 500 ns. The total energy is decreasing from –6.02 eV, reaching the convergence around the value of–6.51 eV after 300 ns. The StoneWales defects are healed. The geometry of the nanotube is optimized again. The resulted structure corresponds to a very regular buckled nanotube, where all the carbon atoms are 3-coordinated in hexagonal rings. We investigate the adsorption of a hydrogen molecule and methane above the different adsorption sites on the untwisted nanotube and twisted carbon nanotube (5,5) (see Figure 1), starting from a height of 3.0 Å. On an armchair nanotube four adsorption sites can be identified in a hexagonal ring: (i) atop site – over a carbon atom, (ii–iii) two bridge sites – above the middle of a carbon-carbon bond which is perpendicular and oblique, respectively to the nanotube axis, and (iv) hole site – in the center of the hexagonal ring. The hexagonal rings are equivalent far from the ends of an untwisted nanotube, but on the twisted one they have different curvatures. The binding energies of the molecule M on the untwisted and twisted nanotube is given by the equation,
V. CHIHAIA ET AL.
452
Figure 1. The untwisted (a) and 300o-twisted (b) SWNT of the type (5,5) considered in the study. The red atoms at the end of the nanotube represent the rigid areas and the arrows indicate the twisting directions. The light colored atoms indicate the area of the adsorption sites, which are presented magnified in figures (c) and (d).
Eb
E (M+NT) E (M) E (NT)
(2)
where, E (M+NT) is the total energy of the optimized system formed by the molecule M adsorbed in a specific adsorption site on the nanotube excepting the atoms from the rigid area, E (M) and E (NT) are the total energy of the isolated molecule M and of the nanotube, respectively. 3. Results The analyze of the distribution of the distances between neighbor carbon atoms shows a large dispersion of the distances in the case of twisted nanotube (0.32 Å) compared with the dispersion of the untwisted nanotube (0.09 Å). The average values of the carbon-carbon distances are 1.49 and 1.43 Å, for the twisted and untwisted nanotubes, respectively. The larger values of the inter-atomic distances for the twisted nanotubes correspond to the atoms at the ridges of the ribbon and the smaller values to the carbon atoms in the middle areas of the ribbon. The hydrogen molecule adsorbs physically on the untwisted nanotube on the atop sites, as the binding energies have very small values (few tenths of kilocalories per mole). The adsorbed molecule is oriented perpendicularly or parallely to the nanotube surface contained in a perpendicular plane to the nanotube axis (horizontal plane). It is located at a distance H2 mass-center – adsorption site of about 2.10 or 1.79 Å, corresponding to the two orientations. The H-H distance dHH is 0.74 Å as in case of the experimental value for the free H2 molecule while the distance between the closest hydrogen to the adsorption site and the adsorption site is about 1.80 Å in both situations. For the adsorption over the bridge sites, the hydrogen molecule prefers two orientations: (i) perpendicular to the nanotube surface; and (ii) parallel to the nanotube surface and perpendicular to the middle of the C-C bond, with a minimum distance H-Cnanotube
SMALL MOLECULE ADSORPTION ON DEFORMED NANOTUBES
453
of 1.34 Å and 1.80 Å, respectively. The bonding energy for the case (i) is about 2.18 kcal/mol while in the case (ii) is about few tenths of kcal/mol. The H-H distance is about 0.91 and 0.74 Å, corres-ponding to the two adsorptions. In the case of the adsorption on the hole site, the hydrogen molecule adsorbs in the horizontal plane, parallel to the ring at a height of 1.31 Å, with a H-Cnanotube distance of 1.80 Å. The distance H-H is 0.74 Å. The methane molecule adsorbs physically on the untwisted nanotube on the atop sites as the binding energies have very small values. The molecule adsorbs oriented like in the staggered or in the eclipsed arrangement. Three C-H bonds of methane are oriented towards the center of the three hexagonal rings that contain the adsorption site, in the first case or they are superposing over the bonds formed by the carbon from the adsorption site with its neighbours, in the second case. In both situations the fourth C-H is oriented upwards. The closest hydrogens to the nanotube surface are located to the distance of about 1.80 Å from the neighbor carbon atoms. The distance Cmethane – adsorption site Cnanotube is about 2.00 and 2.16 Å, corresponding to the two adsorptions types. In the adsorptions above the bridge sites the methane molecule is approaching to the surface with one or two hydrogen atoms oriented towards the carbon atoms neighbors to the adsorption sites. The corresponding distance H-Cnanotube is about 1.80 Å when H is near one carbon atom or about 1.33 Å when is near two carbon atoms of the nanotube. In the latter case the distance H-Cmethane increases to 1.31 Å. There are several possibilities of methane adsorption over hole site but the most energetic adsorption (EB = 3.6 kcal/mol) identified by us is with two hydrogen atoms contained in the vertical plane that are oriented towards the middle of the C-C bonds of the hexagonal ring contained in the vertical plane at a distance of 1.80 Å from the Cnanotube atoms that form these bonds. The two distances H-Cmethane become 1.31 Å. In the case of the adsorptions on the twisted nanotube, the adsorption mechanisms of the two molecular species are more varied. Very often the disruption of some C-C bonds in the neighboring of the adsorption sites is observed to occur simultaneously with the adsorption. The energies associated to the two processes, together with the tension energy associated to the nanotube twisting are competing each other. The histograms of the adsorption energies corresponding to various adsorption sites on the untwisted and twisted nanotube (5,5) are presented in Figure 2 for the hydrogen and methane molecules, respectively. It can be observed that the adsorptions of both molecules are endothermic (Eb > 0, the system receives energy from the surroundings). By the nanotube twisting a large fraction of the investigated adsorptions become exoenergetic (Eb < 0, the system releases energy to the surroundings) for both
454
V. CHIHAIA ET AL.
Figure 2. The histogram of the binding energies of the molecular hydrogen (left) and of the methane (right) with untwisted (circles) and twisted nanotube (5,5). In the case of the twisted nanotube, the distribution is decomposed on the location of the adsorption site on the formed ribbon (ridge, intermediate, and middle regions).
molecular species (41.2% for H2 and 40.3% for CH4), while endothermic adsorptions are less frequent (27.5% for H2 and 32.2% for CH4). Because of the relatively small number of investigated adsorption sites comparing with the total number of the adsorption sites our results may be biased, but it is clear that the nanotube twisting produces an important number of exothermic adsorptions. The hydrogen molecule may adsorb non-dissociate, especially on the ridge regions of the ribbon-like nanotube. For hydrogen molecule the H-H distances are of 0.74 Å, when one of hydrogen atom is close to one Cnanotube at dHCtube = 1.80 Å, but may be larger, about 0.97 Å, when the hydrogen atom is equally closed to two Cnanotube at dHCtube = 1.32 Å. The hydrogen molecule dissociates (dHH > 1.5 Å) in very many of the investigated adsorptions. The hydrogen atoms diffuse on nanotube being finally captured by a Cnanotube where forms a strong covalent bond (dHCtube = 1.09 Å). The adsorption with the highest binding energy (–6.4 kcal/mol) is identified in the case of the adsorption over a bridge site located in the ridge region of the ribbon (see the bond 202–211 in Figure 1d). The hydrogen molecule dissociates over the bond C211-C202 that is breaking also. The free hydrogen atoms form strong covalent bonds with the two carbon atoms (dHCtube = 1.09 Å). The methane molecule is strongly affected by the presence of the twisted nanotube, at least one H-Cmethane bond being weakened (dHCmethane = 1.31 Å). Depending on the neighboring and the local tension inside nanotube one or more hydrogen atoms are disrupted from methane and diffuse on the nanotube surface before to form strong H-C bonds (dHCtube = 1.09 Å). We have identified a particular case of the bridge adsorption (see the bond 233–244 in Figure 1d), with exceptional high binding energy –20.7 kcal/mol, where one H-Cmethane bond is weakened (dHCmethane = 1.31 Å) and another is broken (dHCmethane = 8.45 Å). Some of the C-C bonds are broken and a local rearrangement takes place. The free hydrogen is diffusing far from the adsorption site,
SMALL MOLECULE ADSORPTION ON DEFORMED NANOTUBES
455
where it finds a proper Cnanotube in the other ridge of the ribbon, forming a new H-C bond (dHCtube = 1.09 Å). 4. Conclusions Atomic scale simulations have been performed in order to study the effects of the twisting of the nanotubes in the case of the H2 and CH4 molecules on the carbon nanotubes (5,5). They show a strong enrichment of the interaction of the molecules with the nanotube by the twisting of the nanotube. The interaction of the molecules of both species with the twisted nanotube become stronger, the adsorption being associated with the molecule dissociation and with the breaking of the carbon-carbon bonds neighbors to the adsorption sites. Some variations of the adsorption strength depending on the region where the adsorption site is located on the twisted nanotube may be observed for both adsorptions. The nanotube twisting may be a more efficient way to functionalize the nanotubes than the lateral or longitudinal deformation or than the bending because the deformation produced by twisting is regularly along the entire nanotube. Like in case of any empiric potential, the quality of REBO potential may be questionable, even if it was applied with success for different carbonic systems where the covalent bonds are broken or formed. Therefore, we plan to check the present conclusions using calculation methods with higher accuracy. Also, we plan to extend our study to different types of nanotubes, with various nanotube lengths and twisting degrees.
References 1. Brenner, D.W. et al. (2002) J. Phys.: Condens. Matter. 14, 783–802. 2. Iijima, S. (1991) Nature 354, 56–58. 3. Kong, J., Franklin, N.R., Zhou, C., Chapline, M.G., Peng, S., Cho, K., and Dai, H. (2000) Science 287, 622–625; Sumanasekera, G.U., Adu, C.K.W., Fang, S., and Eklund, P.C. (2000) Phys. Rev. Lett. 85, 1096–1099; Collins, P.G., Bradley, K., Ishigami, M., and Zettl, A. (2000) Science 287, 1801–1804. 4. Stan, G., and Cole, M.W. (1998) J. Low Temp. Phys. 110, 539–544; Ahn, C.C., Ye, Y., Ratnakumar, B.V., Witham, C., Bowman, R.C., Jr., and Fultz, B. (1998) Appl. Phys. Lett. 73, 3378–3380; Ajayan, P.M., Ebbesen, T.W., Ichihashi, T., Iijima, S., Tanigaki, K., and Hiura, H. (1993) Nature 362, 522–525.
456
V. CHIHAIA ET AL.
5. Tsang, S.C, Harris, P.J.F., and Green, M.L.H. (1993) Nature, 362, 520–522; Tsang, S.C., Chen, Y.K., Harris, P.J.F., and Green, M.L.H. (2002) Nature 372, 159–162; Sloan, J., Hammer, J., Zwiefka-Sibley, M., and Green, M.L.H. (1998) J. Chem. Soc., Chem. Commun. 3, 347–348; Satishkumar, B.C., Govindaraj, A., Mofokeng, J., Subbanna, G.N., and Rao, C.N.R. (1996) J. Phys. B: At. Mol. Opt. Phys. 29, 4925–4934. 6. Srivastava, D., Brenner, D.W., Schall, J.D., Ausman, K.D., Yu, M.F., and Ruoff, R.S. (1999) J. Phys. Chem. B 103, 4330–4337. 7. Ni, B., and Sinnott, S.B. (2000) Phys. Rev. B 61, R16343–R16346. 8. Dang, S., Ozturk, Y., Chiraci, S., and Ildirim, T. (2005) Phys. Rev. B 72, 155404. 9. Lennard-Jones, L.E. (1924) Proc. R. Soc. London, Ser. A 106, 441–462. 10. http://www.eng.fsu.edu/~dommelen/research/nano/brenner 11. Berendsen, H.J.C., Van Gunsteren Postma, W.F.A., Nola, D., and Haak, J.R. (1984) J. Chem. Phys. 81, 3684–3690. 12. Gear, C.W. (1971) Numerical Initial Value Problems in Ordinary Differential Equations (Prentice-Hall, NJ), p. 148. 13. http://www.eng.fsu.edu/~dommelen/research/nano/brenner/tools.html 14. Stone, A.J. and Wales, D.J. (1986) Chem. Phys. Lett. 128, 501–503.
THE DEFECT STRUCTURE OF COPPER INDIUM DISULFIDE D. PERNIU1*, A. DUTA1, AND J. SCHOONMAN2 1 Transilvania University of Brasov, ROMANIA 2 Delft University of Technology, THE NETHERLANDS
Abstract – The paper presents the defect chemistry of CuInS2 (CIS) in relation to deviations from stoichiometry and molecularity. The Kröger–Vink notation is used for the point defect description. A first approach for the construction of a qualitative three-dimensional Brouwer diagram is presented. In the case of ternary materials, like CIS, the deviation from molecularity is expressed in the 3D-Brouwer diagram by the variation of the activity of the metal sulfide and the deviation from stoichiometry by the variation of the sulfur partial pressure. This model can be used as an important tool for tailoring the material synthesis in order to obtain desired opto-electrical properties.
Keywords: Copper Indium Disulfide, intrinsic defect, 3D-Brouwer diagram.
1. Introduction The I-III-VI2 chalcopyrite-structured compounds and related solid solutions are currently of interest for use as visible-light absorbing layers in thin-film and 3D nano-structured solar cells. The fundamental optical and electrical properties of CuInS2 (CIS) thin films have been investigated, in view of practical applications e.g. in solar cells. Despite the fact that numerous papers have reported the results of research on solar cell applications using CIS as absorber material, insight into the defect chemistry of the material is rather limited. The purpose of the work presented in this paper is the presentation of the defect chemistry of CIS in relation to deviations from stoichiometry and from molecularity. This model can be used as an important tool for tailoring the material synthesis in order to obtain desired opto-electrical properties. The Kröger–Vink notation is used for the point defect description.1 In this paper we present a first approach for the construction of a qualitative three-dimensional Brouwer diagram,
______ *
To whom corresponding author should be addressed: D. Perniu, email:
[email protected]
A. Vaseashta and I.N. Mihailescu (eds.), Functionalized Nanoscale Materials, Devices and Systems. © Springer Science + Business Media B.V. 2008
457
458
D. PERNIU, A. DUTA, AND J. SCHOONMAN
which is considered an important tool in understanding the properties in relation to the defect chemistry of this material. 2. Defect Formation Reactions in CIS The literature reports the existence of defects playing an important role in the optical and electrical properties of the I-III-VI2 materials, and CIS in particular.2–6 The point defects presented for the CIS material are of several types and summa' rized as follows: cation and anion vacancies, i.e., VCu ;VIn''' ;VSxx , interstitials, i.e., x xxx '' '' Cu i ; Ini ; S i and cation-anion anti-site disorder, i.e., Cu Inxx ; InCu , while the ' x electronic defects are also considered, i.e., e ; h . The intrinsic defect formation mechanisms are described using the Schottky, Frenkel, or anti-Frenkel mechanisms. The CIS defect formation mechanisms have been discussed previously.7 In order to develop the CIS defect chemistry model, several approximations are being used. Firstly, it is assumed that the thermal generation of Frenkel disorder is the preferred intrinsic defect formation mechanism. The lattice reaction equations and the expressions of the mass action laws for the defect formation in the In and Cu – ion sublattice are presented as follows, i.e., the groups of Eqs. (1) and (2). The anti-Frenkel disorder is represented by the group of Eq. (3). In each case, the notation (a) is attributed to the intrinsic lattice reaction, (b) to the equilibrium constant, and (c) to the electroneutrality condition for the specific reaction:
In Inx Vix l In ixxx VIn'''
(1a)
>In @ >V @
(1b)
x xx i
K F,In
''' In
>In @ >V @ xxx i
''' In
(1c)
x ' Cu Cu Vix l Cu ix VCu
K F,Cu
>Cu @ >V @ x i
' Cu
(2a) (2b)
>Cu @ >V @
(2c)
SSx Vix l Si'' VSxx
(3a)
x i
' Cu
DEFECT STRUCTURE OF COPPER INDIUM DISULFIDE
>S @ >V @ >S @ >V @ xx S
'' i
K aF , S
xx S
'' i
459
(3b) (3c)
Here, it is further assumed that the thermal generation of Frenkel disorder in the copper sublattice is energetically favorable. CIS is a ternary material and can be formed from the binary metal sulfides Cu2S and In2S3 and its formation is given by reaction (4a) with reaction equilibrium constant (4b). Cu2S + In2S3 ĺ 2CuInS2
K CIS
1/ 2 a Cu a 1In/22S3 2S
(4a) (4b)
CuInS2 is the ideal composition, which is just a special case. In general, it is expected that CIS will have a range of compositions by changing the molecularity and/or stoichiometry,3 which results in extrinsic defect formation mechanisms, presented as follows. The deviation from stoichiometry, that is the sulfur incorporation in or loss from the CIS lattice, can be expressed by lattice reaction (5a) and Eq. (5b),
1 S 2 ( g ) VSxx l SSx 2h x 2
>h @ >V @ p
(5a)
x 2
KS
xx S
(5b)
1/ 2 S2
For explaining the deviations from molecularity, the equations for Cu2S and In2S3 incorporation, i.e., the Cu-rich/In-rich material (equations in groups (6) and (7)) can be written as follows: CuInS 2 x Cu 2 S 2 o 2Cu Cu 2VIn''' S Sx 3VSxx
K Cu2 S
>V @ >V @ ''' 2 In
xx 3 S
(6b)
a Cu2 S
CuInS 2 In 2 S 3 2 o 2VCu' 2 In Inx 3S Sx V Sxx
K In 2S3
(6a)
>V @ >V @ 2 ' Cu
a In 2S3
xx S
(7a) (7b)
D. PERNIU, A. DUTA, AND J. SCHOONMAN
460
Ternary compounds, such as CIS, may exhibit both a deviation from stoichiometry and from molecularity and extrinsic ionic/electronic point defects are generated in the case of such deviations. The groups of equations labeled (8) and (9) express the lattice reactions, the laws of mass action, and the electro-neutrality equation for the Cu-rich material and In-rich material,
Cu 2 S
(8a)
3 CuInS 2 x S 2 ( g ) 2 o 2Cu Cu 2V In''' 4 S Sx 6h x 2
>V @ >h @ ''' 2 In
K Cu2 S , S
(8b)
x 6
aCu2 S p S32/ 2
> @ >h @
(8c)
x
3 VIn'''
1 ' In 2S 3 S 2 ( g ) CuInS 2 o 2VCu 2In Inx 4SSx 2h x 2
(9a)
>V @ >h @ 2 ' Cu
K In 2S3 , S
a In 2S3 p
>V @ >h @
x 2
(9b)
1/ 2 S2
(9c)
x
' Cu
The electronic equilibrium and equilibrium constant are as follows (10a and b):
null l e ' h x
(10a)
>e @ >h @
(10b)
Ki
x
'
3. The Brouwer Diagram The concept of the 3D Brouwer diagram is presented in Figure 1. The diagram represents the variations of a particular defect concentration as function of an external parameter. In the case of a ternary material, like CIS, the deviation from molecularity is expressed by the variation in activity of the metal sulfide and the deviation from stoichiometry by the variation in the sulfur partial pressure. In order to construct the 3D-Brouwer diagram the formal electroneutrality condition is the starting point. For the specific case of CIS, the electroneutrality condition is given by the following Eq. (11):
>V @ 3>V @ >e @ >Cu @ 3>In @ >h @ 2>V @ ' Cu
''' In
'
x i
xxx i
x
xx S
(11)
DEFECT STRUCTURE OF COPPER INDIUM DISULFIDE
Ig [defect]
Deviation from stoichiom etry
Deviation from molecularity
Ig acu2s
461
Ig ps2
Ig aIn2S3
Figure 1. The 3D Brouwer diagram concept for the CIS material.
The diagram comprises different regimes, which are discriminated by a Brouwer approximation per specific regime. The choice of the Brouwer approximation, which is the reduced electroneutrality condition for the specific regime, is made considering the most probable defect formation reaction. In the following, the Brouwer regimes are characterized by the reduced electroneutrality conditions and the concomitant defect concentrations. Based on the Brouwer approximations and considering the equations expressing the laws of mass action for the lattice reactions, the defect concentrations were calculated for each regime in the diagram. 3.1. THE INTRINSIC REGIME
The preferred reaction is the Frenkel disorder in the copper cation sublattice, with the electroneutrality condition (2c). Considering this assumption, the concentrations of the point defects can be calculated (Eqs. (12a), (13a), and (14a)):
>V @ >Cu @ K >V @ >In @ K >e @ >h @ K ' Cu
x i
1/ 2 F ,Cu
(12a)
''' In
x xx i
1/ 2 F , In
(13a)
'
x
1/ 2 i
(14a)
2 with K F1 /,Cu ²² K F1 /, In2 ²² K i1 / 2 .
3.2. THE EXTRINSIC REGIME
If an excess of one of the binary metal sulfides is present in CIS, the material exhibits a deviation from molecularity. A deviation always leads to ionic point defects. Lattice reactions (8a) and (9a) describe the extrinsic disorder taking
D. PERNIU, A. DUTA, AND J. SCHOONMAN
462
into account both a deviation from molecularity and stoichiometry. In the 3D Brouwer diagram, the deviation from molecularity is presented in the front part of the diagram. Here, the defect concentrations are plotted versus the metal sulfide activity, taking the partial sulfide pressure constant. The deviation from stoichiometry is represented in the laterals and the defect concentrations are plotted versus the sulfur partial pressure, the activity of the metal sulfide being constant. The diagram contains two parts: one describing the Cu-rich materials, and the other the In-rich materials. In the Cu-rich region, reaction (8a) is predominant, and Eq. (8c) stands for the Brouwer approximation. In the In-rich region, the predominant reaction is (9a) and the Brouwer approximation is presented in Eq. (9c). The defect concentrations are calculated from the equations expressing the laws of mass action as a function of the sulfur partial pressure. The expressions for the defect concentrations are given in Eqs. (12b–18b) for the Cu-rich material and in Eqs. (12c–18c) for the In-rich region. For the Cu-rich material,
>Cu @ 3 K K K >V @ 3 K K >In @ 3 K K >V @ 3 K >V @ 3 K K >h @ 3 K >e @ 3 K K x i
1 CIS
1/ 4
F , Cu
1 / 4
' Cu
1/ 2 In 2 S 3 , S
CIS
xxx i
1 / 2 In 2 S 3 , S
3/ 4
F , In
3 / 4
''' In
xx S
x
1/ 4
(14b)
2
(15b)
2
1/ 4 1 / 8 aCu S pS 2
(16b)
2
1/ 8 3 / 16 aCu S pS 2
(17b)
2
1 / 8 Cu 2 S , S
i
1 / 8 3 / 16 aCu S pS 2
1/ 8 3 / 16 aCu S pS 2
1/ 4 Cu 2 S , S
1 / 4
'
(13b)
2
1 / 8 Cu 2 S , S
1/ 8 Cu 2 S , S
(12b)
2
1 / 8 5 / 8 1 / 16 K Cu aCu S pS 2 2S,S
1/ 8 Cu 2 S , S
1 S
1/ 2
1/ 8 5/8 1 / 16 K Cu aCu S pS 2 2S ,S
1 / 8 3 / 16 aCu S pS 2
(18b)
2
For the In-rich material,
>Cu @ K K K >V @ K K a >In @ K K K K >V @ K K K x i
' Cu
xx x i
1/ 4 In2 S 3 , S
F , In
''' In
1 / 4 In2 S 3 , S
F ,Cu
3 / 2 CIS
3/ 2 CIS
1 / 2 CIS
1/ 2 CIS
1 / 2 Cu 2 S , S
1/ 2 Cu 2 S , S
1/ 4 1 / 8 aCu S pS 2
1 / 4 Cu 2 S
p1S/28
3/ 4 In2 S 3 , S
3 / 4 In2 S 3 , S
(12c)
2
5 / 4 aCu pS23 / 8 2S
5/ 4 aCu pS32/ 8 2S
(13c) (14c) (15c)
DEFECT STRUCTURE OF COPPER INDIUM DISULFIDE
>V @ K K >h @ K >e @ K K xx S
1 S
x
1/ 2 In2 S 3 , S
1/ 4 In2 S 3 , S
i
(16c)
1 / 2 1 / 4 K CIS aCu S pS 2 2
(17c)
1/ 2 1/ 4 1/ 8 K CIS aCu S pS 2
1 / 4 In2 S 3 , S
'
463
2
(18c)
1 / 2 1/ 4 1 / 8 K CIS aCu S pS 2 2
The defect concentrations are plotted versus the variable parameter considered for each regime. The slopes are the result of the derivations of the defect concentrations versus the metal sulfide activity or sulfur partial pressure. Figure 2a and b presents the two views of the 3D-Brouwer diagram. It is to be mentioned that on the diagram faces the extreme conditions are represented, while in the interior of the diagram both the deviation from molecularity and the deviation from stoichiometry can be viewed.
lg [defect]
lg [defect]
Cui
lg pS2
a
Vcu
VIn Vcu Ini Vs h e lg acu2s
Cui
Ini Vs
VIn h
lg pS2 e
lg aIn2S3
lg acu2s
lg aIn2S3
b
Figure 2. (a) The 3D-Brouwer diagram showing the Cu-rich regime, (b) the 3DBrouwer diagram showing the In-rich regime.
4. Conclusions The defect chemistry of the chalcopyrite-structured material CuInS2 is governed mainly by extrinsic defects as a result of deviations from molecularity and stoichiometry. Keeping the partial pressure of sulfur constant, expressions have been derived for the concentrations of the point defects. In a similar approach, expressions for the defect concentrations have been derived under the condition that the activity of one of the binary sulfides is kept constant. Here, it is assumed that the intrinsic disorder is described by a Frenkel mechanism in the copper-ion sublattice. The expressions for the defect concentrations have been used to obtain the slopes in a three-dimensional 3D-Brouwer diagram. This study
464
D. PERNIU, A. DUTA, AND J. SCHOONMAN
describes for the first time a 3D-Brouwer diagram for this chalcopyrite material. It is anticipated, that using the present approach, the 3D-Brouwer diagrams for other members of the visible-light absorbing chalcopyrites can be constructed. The predominant defect structures will be used to explain electrical conductivity and optical experiments on CuInS2 with a deviation from molecularity and for a deviation from stoichiometry.
References 1. Kröger, F. (1964) The Chemistry of Imperfect Crystals, North-Holland, Amsterdam. 2. Maier, J. (2004) Physical Chemistry of Ionic Materials. Ions and Electrons in Solids, Wiley, Chichester. 3. Binsma, J. (1983) J. Phys. Chem. Solids 44, 237–244. 4. Ueng, H., and Hwang, H. (1990) J. Phys. Chem. Solids 51, 11–18. 5. Márquez, R., and Ricón, C. (1999) Mater. Lett. 40, 66–70. 6. Nanu, M., Schoonman, J., and Goossens, A. (2004) Thin Solid Films, 451–452, 193–197. 7. Perniu, D., Vouwzee, S., Duta, A., and Schoonman, J. (2007) J. Optoelectron. Adv. Mater. 9, 1568–1571.
ASI GROUP PHOTOGRAPH
465
ASI GROUP PHOTOGRAPH LEGEND
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 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
60
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. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60.
Adrian Ghita Cyril Popov Jan Boyd Emilia Pecheva Colm Cunniffe Olga Korostynska Dana Perniu Carmen Ristoscu Nicoleta Preda Damir Dominko Daniela Ebrasu Dan Tonchev Wolfgang Maser Brandon McNaughton Judit Kis-Csitari Marian Vopsaroiu Konstantin Petrov Adina Morozan Arezki Benfdila Melinda Mohl Jorgen Schou Ganna Kharlamova Ornprapa Pummakarnchana Vesselin Paunov Anna Kachanovska George Gallios Sigitas Tamulevicius Mr. Reisfeld Renata Reisfeld Uwe Kriebig Rodica Cristescu Felix Sima Ion N. Mihailescu Marina Turcan Ashok Vaseashta Doga Kavaz Alex G. Petrov Miroslava Vaclavikova Kseniia Katok Christopher S. Lohn Elena Lupan Andreea Purice Alexander Kukhta Tamer Cirak Pavel D’yachkov Oleksiy Kharlamov Yuriy Gnatyuk Adriana Andronie Jessica Weber Marie I. Baraton Marijus Brikas Temenuga Hineva Perica Paunovic Joop Schoonman Petia Petrova Aysegul Gokalp Silvia Bakalova Lyubomir Stoychev Nalan Ozdemir Ioanna Fasaki
467
SELECTED PHOTOGRAPHS TAKEN DURING THE ASI
469
470
SELECTED PHOTOGRAPHS
SELECTED PHOTOGRAPHS
471
472
SELECTED PHOTOGRAPHS
SELECTED PHOTOGRAPHS
473
LIST OF PARTICIPANTS Dr. ALVELO, Jose I. Technical Program Support Northrop Gruman Information Technology Defense Group 8211 Terminal Road, Suite 1000 Lorton, VA 22079, USA
[email protected] ANDRONIE, Adriana-Elena 3Nano-SAE Research Centre University of Bucharest PO Box MG-38, 077125 Bucharest-Magurele, ROMANIA
[email protected] AXENTE, Emanuel Lasers Plasmas and Photonic Processing Laboratory - LP3 CNRS – Université Aix-Marseille Campus de Luminy – Case 917 13288 Marseille Cedex 9, FRANCE
[email protected] Dr. AYVAZYAN, Gagik Semiconductor R&D Center Engineering Academy of Armenia 41 Arshakunyats, Technopark 0026 Yerevan, ARMENIA
[email protected] BAKALOVA, Silvia Institute of Solid State Physics Bulgarian Academy of Sciences Tzarigradsko Chaussee Blvd. 72 1784 Sofia, BULGARIA
[email protected]
Prof. BARATON, Marie-Isabelle SPCTS – UMR CNRS 6638 University of Limoges 87060 Limoges, FRANCE
[email protected] Prof. BATANI, Dimitri Dipartimento di Fisica“G.Occhialini” Università di Milano Bicocca Piazza della Scienza 3 20126 Milano, ITALY
[email protected] Dr. BENFDILA, Arezki LMPDS, Faculty of Electrical Engineering and Computer Science The University M. Mammeri BP 17 RP, 15000 Tizi-Ouzou, ALGERIA
[email protected] BRIKAS, Marijus Institute of Physics Savanoriu Ave. 231 LT-02300 Vilnius, LITHUANIA
[email protected] CIRAK, Tamer Hacettepe University Department of Chemistry Biopolymeric Systems Research Gr Beytepe, Ankara TURKEY
[email protected] CAPUCCINI, Chiara Biomimetics and Materials Department of Chemistry Bologna, ITALY
[email protected] 475
476
LIST OF PARTICIPANTS
Dr. CRISTESCU, Rodica National Institute for Lasers, Plasma and Radiation Physics Lasers Department, Laser-SurfacePlasma Interaction Laboratory Atomistilor 409, P.O. Box MG-36 Bucharest Magurele RO-077125, ROMANIA
[email protected] D’YACHKOV, N. Pavel Institute of General and Inorganic Chemistry Russian Academy of Sciences Leninskii pr. 31 119991 Moscow, RUSSIA
[email protected] DAMIR, Dominko Institute of Physics, HR-10001, P.O. Box 304 Zagreb, CROATIA
[email protected] Prof. DEKHTYAR, Yuri Riga Technical University Biomedical Engineering and Nanotechnologies Institute 1 Kalku str, Riga, LV1658, LATVIA
[email protected] ERDURAL, Beril Korkmaz Department of Chemical Engineering Middle East Technical University 06431 Ankara, TURKEY
[email protected] FASAKI, Ioanna National Hellenic Research Foundation
Theoretical and Physical Chemistry Institute Vasileos Konstantinou Ave. 48 11635 Athens, GREECE
[email protected] Prof. FONTCUBERTA, Josep Institut de Ciència de Materials de Barcelona-CSIC Campus UAB, Bellaterra 08193 Catalunya, SPAIN
[email protected] Dr. FORSEA, Anna Maria Dermatology Department Elias University Hospital Carol Davila University of Medicine & Pharmacy Bucharest, ROMANIA
[email protected] Dr. GALLIOS, George Aristotle University of Thessaloniki School of Chemistry Laboratory of General & Inorganic Chemical Technology (116) GR-540 06 Thessaloniki, GREECE
[email protected] GHITA, Adrian Institute of Physical Chemistry “Ilie Murgulescu” 202 Splaiul Indepentei Bucharest, ROMANIA
[email protected] GNATYUK, Yuriy Surface Photonics Laboratory O. Chuiko Institute of Surface Chemistry of NAS of Ukraine 03164 Kyiv, UKRAINE
[email protected]
LIST OF PARTICIPANTS
GOKALP, Aysegul Nanomaterials Processing & Characterization Laboratories Marshall University, One John Marshall Drive Huntington, WV 25575-2570, USA
[email protected] Dr. GURSAN ERDEM, Arzum Ege University Faculty of Pharmacy Analytical Chemistry Department 35100 Bornova, Izmir, TURKEY
[email protected] HINEVA, Temenuga Laboratory of Thin Films Technology University of Chemical Technology and Metallurgy Department of Physics 8 Kl. Ohridsky blvd 1756 Sofia, BULGARIA
[email protected] Dr. HORVÁTH, J. Zsolt Hungarian Academy of Sciences Research Institute for Technical Physics and Materials Science Budapest 114, P.O. Box 49, H1525, HUNGARY
[email protected] KACHANOVSKA, Anna Biomedical Engineering and Nano Technologies Institute Riga Technical University Riga, LATVIA
[email protected] Dr. KAUFFMAN, Miron Physics Department
477
Cleveland State University Cleveland, OH 44115, USA
[email protected] KATOK, Kseniia Institute of Surface Chemistry National Academy of Sciences of Ukraine 17 General Naumov Str., 03164 Kyiv, Ukraine
[email protected] KAVAZ, Do÷a Hacettepe University Chemistry Department Biochemistry Division Beytepe, Ankara, TURKEY
[email protected] KHARLAMOV, Oleksiy I.N. Frantsevich Institute for Problems of Materials Science Kiev, UKRAINE
[email protected] KHARLAMOVA, Ganna Kiev National Taras Shevchenko University Kiev, UKRAINE
[email protected] KIS-CSITÁRI, Judit Bay Zoltán Foundation for Applied Research Iglói út 2, 3519 Miskolc-Tapolca HUNGARY
[email protected] Prof. KREIBIG, Uwe I. Physical Institute IA RWTH Aachen GERMANY
[email protected]
478
LIST OF PARTICIPANTS
Dr. KUKHTA, V. Alexander Institute of Molecular and Atomic Physics National Academy of Sciences Nezalezhnastsi Ave. 70 220072 Minsk, BELARUS
[email protected] LOHN, S. Christopher Department of Physics Texas State University at San Marcos TX 78666, USA
[email protected] LUPAN, Elena Center of Optoelectronics Institute of Applied Physics Academy of Sciences of Moldova Chisinau MD-2028, REPUBLIC of MOLDOVA
[email protected] Prof. MASER, Wolfgang Department of Nanotechnology Instituto de Carboquímica (CSIC) C/Miguel Luesma Castán 4 E-50018 Zaragoza, SPAIN
[email protected] McNAUGHTON, H. Brandon Applied Physics Program University of Michigan Ann Arbor, MI, USA
[email protected] Prof. MIHAILESCU, Ion N. Director, National Institute for Lasers, Plasma and Radiation Physics Lasers Department, “Laser-SurfacePlasma Interactions” Laboratory
PO Box MG-54, RO-77125, Bucharest-Magurele, ROMANIA
[email protected] MOHL, Melinda Department of Applied and Environmental Chemistry University of Szeged, Rerrich B. tér 16720 Szeged, HUNGARY
[email protected] Dr. MOROZAN, Adina 3Nano-SAE Research Centre University of Bucharest PO Box MG-38, 077125 Bucharest-Magurele, ROMANIA
[email protected] NAIDENOVA, Tsvetelina Institute of Electronics Bulgarian Academy of Sciences 72 Tzarigradsko Chaussee blvd 1784 Sofia, BULGARIA
[email protected] ÖZDEMIR, Nalan Erciyes University Chemistry Department Kayseri, TURKEY
[email protected] Dr. PAPAKONSTANTINOU, Pagona Nanotechnology Research Institute (NRI) School of Electrical and Mechanical Engineering University of Ulster at Jordanstown Newtownabbey, County Antrim BT37 OQB Northern Ireland, UK
[email protected]
LIST OF PARTICIPANTS
Dr. PAUNOV, Vesselin N. Surfactant & Colloid Group Department of Chemistry University of Hull, UK
[email protected] PAUNOVIû, Perica Faculty of Technology and Metallurgy University Sts. Cyril and Methodius Skopje, R., MACEDONIA
[email protected] Dr. PECHEVA, Emilia Institute of Solid State Physics Bulgarian Academy of Sciences blvd. Tzarigradsko Chaussee. 72 1784 Sofia, BULGARIA
[email protected] Dr. PERNIU, Dana Transilvania University of Brasov Center: Product Design for Sustainable Development Brasov, ROMANIA
[email protected] Acad. Prof. PETROV, Alexander G. Institute of Solid State Physics Bulgarian Academy of Sciences 72 Tzarigradsko chaussee Sofia 1784, BULGARIA
[email protected] Dr. PETROV, Konstantin Institute for Electrochemistry and Energy Systems Bulgarian Academy of Sciences 1113 Sofia, BULGARIA
[email protected]
479
PETROVA, Petia Central Laboratory of Photoprocesses “Acad. J. Malinowski” Bulgarian Academy of Sciences “Acad. G. Bonchev”str., bl. 109, 1113 Sofia, BULGARIA petiakl@mail,bg Dr. PIVIN, Jean-Claude CSNSM, 91405 Orsay Campus, FRANCE
[email protected] Dr. POPOV, Cyril Institute for Nanostructure Technologies and Analytics (INA) University of Kassel, GERMANY
[email protected] Dr. PREDA, Nicoleta Roxana Optics and Spectroscopy Lab. National Institute of Materials Phys PO Box MG-7, BucharestMagurele, ROMANIA
[email protected] Dr. PUMMAKARNCHANA, Ornprapa Department of Environmental Science School of Science Silpakorn University Nakornoathom, 73000 THAILAND.
[email protected]
480
LIST OF PARTICIPANTS
Prof. REISFELD, Renata Department of Inorganic Chemistry The Hebrew University Jerusalem, ISRAEL
[email protected]
Lasers Department “Laser-Surface-Plasma Interactions” Laboratory PO Box MG-36, RO-77125, Bucharest-Magurele, ROMANIA
[email protected]
Dr. RISTOSCU, Carmen National Institute for Lasers, Plasma & Radiation Physics Lasers Department, “Laser-SurfacePlasma Interactions” Laboratory PO Box MG-36, RO-77125, Bucharest-Magurele, ROMANIA
[email protected]
Dr. STAMATIN, Ioan 3Nano-SAE Research Centre University of Bucharest P.O. Box MG-38 077125 Bucharest-Magurele, ROMANIA
[email protected]
Prof. SCHOONMAN, Joop Department of DelftChemTech Delft University of Technology Delft, THE NETHERLANDS
[email protected]
STOENESCU, Daniela National Institute for Cryogenics and Isotopic Technologies ICIT Rm. Valcea, ROMANIA
[email protected]
Prof. SCHOU, Jørgen Department of Optics and Plasma Research Risø National Laboratory Technical University of Denmark DK-4000 Roskilde, DENMARK
[email protected]
STOYCHEV, Lyubomir Geordi Nadjakov Institute of Solid State Physics, Bulgarian Academy of Science 72 Tzarigradsko Chaussee Blvd., 1784 Sofia, BULGARIA
[email protected]
Prof. SENTIS, Marc Lasers Plasmas and Photonic Processing Laboratory CNRS – Universities of AixMarseille Campus de Luminy – Case 917 13288 Marseille Cedex 9, FRANCE
[email protected]
Prof. SZÖRÉNYI, Tamas Research Group on Laser Physics University of Szeged PO Box 406, H-6701 Szeged, HUNGARY
[email protected]
SIMA, Felix National Institute for Lasers, Plasma and Radiation Physics
Prof. TAMULEVICIUS, Sigitas Institute of Physical Electronics Kaunas University of Technology Savanoriu 271 Kaunas LT50131, LITHUANIA
[email protected]
LIST OF PARTICIPANTS
Prof. TENNE, Reshef Department of Materials and Interfaces Weizmann Institute Rehovot 76100, ISRAEL
[email protected] Dr. TONCHEV, Dan Department of Electrical Engineering University of Saskatchewan Saskatoon, CANADA
[email protected] TOURLEIGH, Yegor V. Bioengineerings Department Lomonosov Moscow State University Moscow, RUSSIA
[email protected] ğURCAN, Marina Center of Optoelectronics Institute of Applied Physics Academy of Sciences of Moldova 1 Academiei Street Chisinau MD-2028, REPUBLIC Of MOLDOVA
[email protected] Dr. VACLAVIKOVA, Miroslava Institute of Geotechnics Slovak Academy of Sciences Watsonova 45
481
SK-043 53 Kosice, SLOVAKIA
[email protected] Prof. Dr. Ir. VASEASHTA, Ashok Director, Nanomaterials Processing & Characterization Labs. Department of Physics & Grad. Program in Physical Sciences Marshall University One John Marshall Drive Huntington, WV 25575-2570, USA
[email protected] Dr. VOPSAROIU, Marian National Physical Laboratory Teddington TW11 0LW, UK
[email protected] WEBER, Jessica E. Department of Mechanical Eng. Nanomaterials & Nanomanufacturing Research Ctr. University of South Florida, Tampa, FL, USA
[email protected] Dr. YAZICI, M. Suha Unido International Centre for Hydrogen Energy Technologies Sabri Ulker SK 38/4 CevizlibagZeytinburnu 34015 Istanbul, TURKEY
[email protected]
AUTHORS INDEX Agayan 403 Andronie 415, 445 Antoniadou 379 Arshak 299, 321 Axente 389 Bakalova 357 Bakir 409 Baraton 77 Batani 145 Bayram 313 Belev 279 Benito 101 Bigi 389 Biljakoviü 399 Blackburn 431 Boanini 389 Boyd 61 Cain 431 Caiteanu 27 Camesasca 437 Capuccini 389 Chihaia 449 ÇÕrak 313 Clarke 403 Cohen 51 Cunniffe 101, 321 Cziraki 357 Dekhtyar 169, 347 Delaporte 185 Demšar 399 Denkbas 313, 325 Dominko 399 Dou 419 Du 27 Dudik 437 Duman 325 Dumitru 415, 445 Duta 457 Ebrasu 383
Engin 325 Erdem 273 Erdural 409 Escoubas 27 Fasaki 379 Feldman 51 Gallios 291 Gazzano 389 Geerts 419 Geretovszky 121 Ghita 449 Giannoudakos 379 Grigorescu 357 Grojo 185 Gyorgy 357 Hotovy 379 Hredzak 291 Ikuta 351 Iliev 305 Ivanov 331 Ivanova 351 Jägermann 51 Jakabsky 291 Kachanovska 347 Karakas 409 Kasap 279, 351 Katok 335 Kaufman 437 Kavaz 313 Kharlamov 373 Kharlamova 373 Kiricsi 365, 369 Kirillova 373 Kis-Csitári 369 Kompitsas 379 Kónya 365, 369 Kopelman 403 Kopnov 51 Korostynska 299
483
484
AUTHORS INDEX
Koughia 351 Kukovecz 365 Kulisch 215 Kumar 425 Kuncser 431 Liaw 61 Liu 437 Lohn 419 Maeda 351 Manas-Zloczower 437 Mani 279 Marcus 399 Martínez 101 Maser 101 Mazingue 27 McNaughton 403 Meškinis 225 Mezinskis 347 Mihailescu, I. 27, 357, 389, 399 Mihailescu, C. 27 Mihailoviü 399 Milat 399 Mirea 445 Mohl 365 Moore 321 Morozan 415, 445 Moshkovich 51 Muñoz 101 Nastase, C. 415, 445 Nastase, F. 415, 445 Nikolova 305 O’Brien 419 Ozdemir 325 Öztürk 313 Padmini 419 Pandey 419 Patmalnieks 347 Patularu 383 Pereira 185 Perniu 457 Perrone 27 Petreanu 383
Petrov, A. 87 Petrov, K. 305 Phonekeo 339 Popov 215 Pummakarnchana 339 Pumpens 347 Rapoport 51 Rehacek 379 Reisfeld 257 Renhofa 345 Ristoscu 27, 399 Roubani-Kalantzopoulou 379 Sakai 351 Salamon 399 Schad 419 Schoonman 199, 457 Schou 241 Sentis 185 Seong 449 Sima 389 Skripnichenko 373 Socol 399 Späth 51, Stamataki 379 Stamatin 415, 445 Starešiniü 399 Stefanescu 383 Stefusova 291 Stoica 403 Stoitsas 199 Stoyanova 305 Stoychev, D. 305 Stoychev, L. 331 Suh 449 Szekeres 357 Szörényi 121 Tamuleviþius 225 Tenne 51 Tertykh 335 Tomeljak 399 Tonchev 279, 349 Tran 199
AUTHORS INDEX
Tuncel 325 Vaclavikova 291 Valeanu 383 Valov 305 Vaseashta 3, 321, 331, 339, 415, 445 Vitanov 305 Vopsaroiu 431
Vulpe 415, 445 Weber 425 Yanishpolskii 335 Yazici 283 Yelton 425 Yurum 409 Zak 51 Zhecheva 305
485
KEYWORD INDEX 3D-Brouwer diagram 457 A Ablation 121 ablation threshold flux 145 adhesion 347 adsorption 447 akaganeite 291 aluminium nitride 357 amine-containing groups 365 annealing 51 antimicrobial 409 applications 101 arsenic 291 B bifunctional air/oxygen electrode 305 BioForce NanoeNablerTM 299 biomedical applications 215 biomembrane 87 biosensors 87, 215 biotechnology 313 blue bronze 399 C carbon membrane reactors 199 carbon nanotube 445 carbon nanotubes 101 carbon threads 373 catalyst 445 cavitation 369 ceramic membranes 199 chalogenide glass 351 characterization 101 chemical composition test 331 chemical gas sensors 77 chitosan 313 clusters 121
CH4 449 collisional absorption 145 composites 101 conducting polymer composites 321 contact angle with water 225 copper indium disulfide 457 crater formation 145 crystalline structure 357 crystallization temperature 351 CV 419 D diamond like carbon films 225 dielectric barrier discharges 61 dispersions 101 DNA 273 doped a-Se 279 DSC 351 dual-electron spectroscopy 169 E electrical properties 279 electrocatalysts 305 electrochemical sensors 273 electrodeposition 425 energy 283 entropy 433 ESR 313 excimer lamps 61 exo-electron spectroscopy 169 exothermal nanosynthesis 373 F Faraday effect 431 field-matter interaction 331 flexoelectricity 87 fractals 437 FTIR spectroscopy 77
487
488
KEYWORD INDEX
fuel cells 283, 383 functionalization 101, 365 G gas diffusion electrodes 308 gas sensors 379 gas-surface interactions 77 gene delivery 325 glass transition 279 gold 335 greenhouse effect 199 H hydrodynamics 145 hydrogen 283 hydrogen oxidation 335 hydrogen production 199 hydrothermal 409 hydroxyapatite 389 H2 379, 449 I IF-WS2 51 ilmenite-hematite 419 inorganic-organic hybrid matrices 257 internet GIS 339 intrinsic defect 457 in situ reduction 335 ion beam synthesis 225 isomorphous substitution 390 IV 419 L laser 145 laser ablation 145 laser nano-particle interaction 185 laser processing 121 low temperature 61 M magnetic beads 403 magnetic microspheres 403
magnetic nanoparticles 313 magnetite 291 matrix-assisted pulsed laser evaporation (MAPLE) 241 mechanosensitivity 87 membrane 383 metal oxide films 27 methanol oxidation 445 microdevices 299 microfluidics 437 mixing 437 m-line technique 27 modeling 291 modified surface 335 multi-functional materials 419 MWCNT 365 MWNTs 445 N nanocomposite membrane 415 nanocrystalline diamond films 215 nanocrystallites 357 nanodot arrays 185 nano-generator 431 nanomaterials 273, 283 nanoparticles 51,121, 291, 369 nanopatterning 299 nanostructured coatings 27 nanostructured 409 nanostructures 169 nanowire 425 near-field enhancement 185 NiO 379 non-destructive testing 331 nonlinear dynamics 403 nonlinear rotation 403 O optical and electrical properties 257 optical gas sensor 27 optical properties 225
KEYWORD INDEX
P photocatalytic 409 photo-chemistry 61 photoluminescence 351 Plasma frequency 145 PLD 241, 389 point-of-care diagnostics 299 polymer 383 polymeric nanoparticles 325 photo-electron spectroscopy 169 production 101 pulsed laser deposition (PLD) 241, 357 purification 101 R raman scattering spectroscopy 225 rare earth doping 351 real time air pollution monitoring 339 remote sensing 339 S semiconductor metal oxides 77 semiconductor nanoparticles 347 semiconductor quantum dots and rods 257 sensor systems 321 sensors 321, 339 sensors arrays 299 SiC nanostructures 373 silica 335 silicon hydride groups 335 single cell detection 403 SiOx containing DLC 225 sol gel glasses 257 solid acid 415 sonochemistry 369
489
sorption 291 spatial coherence 145 Spitzer’s law 145 strontium 389 sublimation 373 superparamagnetism 431 superprotonic transition 415 surface chemical species 77 surface chemistry 77 T thermal energy 431 thermal properties 279 thermodynamic diagrams 291 thermoelectric application 425 thin films 121, 399 titania 409 tribology 51 tripolyphosphate 313 U ultraviolet 61 V VSM 313 W water treatment 291 X XPS (X-ray photoelectron spectroscopy) 51 XPS analysis 357 XPS study 225 XRD 357 Z zirconia- silica- polyurethane 257