NanoScience and Technology
NanoScience and Technology Series Editors: P. Avouris B. Bhushan D. Bimberg K. von Klitzing H. Sakaki R. Wiesendanger The series NanoScience and Technology is focused on the fascinating nano-world, mesoscopic physics, analysis with atomic resolution, nano and quantum-effect devices, nanomechanics and atomic-scale processes. All the basic aspects and technologyoriented developments in this emerging discipline are covered by comprehensive and timely books. The series constitutes a survey of the relevant special topics, which are presented by leading experts in the field. These books will appeal to researchers, engineers, and advanced students. Magnetic Microscopy of Nanostructures Editors: H. Hopster and H.P. Oepen Applied Scanning Probe Methods I Editors: B. Bhushan, H. Fuchs, S. Hosaka The Physics of Nanotubes Fundamentals of Theory, Optics and Transport Devices Editors: S.V. Rotkin and S. Subramoney Single Molecule Chemistry and Physics An Introduction By C. Wang, C. Bai Atomic Force Microscopy, Scanning Nearfield Optical Microscopy and Nanoscratching Application to Rough and Natural Surfaces By G. Kaupp Applied Scanning Probe Methods II Scanning Probe Microscopy Techniques Editors: B. Bhushan, H. Fuchs Applied Scanning Probe Methods III Characterization Editors: B. Bhushan, H. Fuchs Applied Scanning Probe Methods IV Industrial Application Editors: B. Bhushan, H. Fuchs
Nanocatalysis Editors: U. Heiz, U. Landman Roadmap of Scanning Probe Microscopy Editors: S. Morita Nanostructures – Fabrication and Analysis Editor: H. Nejo Applied Scanning Probe Methods V Scanning Probe Microscopy Techniques Editors: B. Bhushan, H. Fuchs, S. Kawata Applied Scanning Probe Methods VI Characterization Editors: B. Bhushan, S. Kawata Applied Scanning Probe Methods VII Biomimetics and Industrial Applications Editors: B. Bhushan, H. Fuchs Applied Scanning Probe Methods VIII Scanning Probe Microscopy Techniques Editors: B. Bhushan, H. Fuchs, M. Tomitori Applied Scanning Probe Methods IX Characterization Editors: B. Bhushan, H. Fuchs, M. Tomitori Applied Scanning Probe Methods X Biomimetics and Industrial Applications Editors: B. Bhushan, H. Fuchs, M. Tomitori
Bharat Bhushan Harald Fuchs Masahiko Tomitori
Applied Scanning Probe Methods IX Characterization
With 216 Figures and 25 Tables Including 27 Color Figures
123
Editors: Professor Bharat Bhushan Nanotribology Laboratory for Information Storage and MEMS/NEMS (NLIM) W 390 Scott Laboratory, 201 W. 19th Avenue The Ohio State University, Columbus Ohio 43210-1142, USA e-mail:
[email protected]
Professor Dr. Harald Fuchs Institute of Physics, FB 16 University of Münster Wilhelm-Klemm-Str. 10 48149 Münster, Germany e-mail:
[email protected] Professor Dr. Masahiko Tomitori Advanced Institute of Science & Technology School of Materials Science Asahidai, Nomi 1-1 923-1292 Ishikawa, Japan e-mail:
[email protected]
Series Editors: Professor Dr. Phaedon Avouris IBM Research Division Nanometer Scale Science & Technology Thomas J. Watson Research Center, P.O. Box 218 Yorktown Heights, NY 10598, USA Professor Bharat Bhushan Nanotribology Laboratory for Information Storage and MEMS/NEMS (NLIM) W 390 Scott Laboratory, 201 W. 19th Avenue The Ohio State University, Columbus Ohio 43210-1142, USA
Professor Dr., Dres. h. c. Klaus von Klitzing Max-Planck-Institut für Festkörperforschung Heisenbergstrasse 1, 70569 Stuttgart, Germany Professor Hiroyuki Sakaki University of Tokyo Institute of Industrial Science, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan Professor Dr. Roland Wiesendanger Institut für Angewandte Physik Universität Hamburg Jungiusstrasse 11, 20355 Hamburg, Germany
Professor Dr. Dieter Bimberg TU Berlin, Fakutät Mathematik, Naturwissenschaften, Institut für Festkörperphysik Hardenbergstr. 36, 10623 Berlin, Germany
ISBN 978-3-540-74082-7 DOI 10.1007/978-3-540-74083-4
e-ISBN 978-3-540-74083-4
NanoScience and Technology ISSN 1434-4904 Library of Congress Control Number: 2007937293 © 2008 Springer-Verlag Berlin Heidelberg This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typesetting: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig Production: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig Cover design: WMXDesign GmbH, Heidelberg Printed on acid-free paper 987654321 springer.com
Preface
The success of the Springer Series Applied Scanning Probe Methods I–VII and the rapidly expanding activities in scanning probe development and applications worldwide made it a natural step to collect further specific results in the fields of development of scanning probe microscopy techniques (Vol. VIII), characterization (Vol. IX), and biomimetics and industrial applications (Vol. X). These three volumes complement the previous set of volumes under the subject topics and give insight into the recent work of leading specialists in their respective fields. Following the tradition of the series, the chapters are arranged around techniques, characterization and biomimetics and industrial applications. Volume VIII focuses on novel scanning probe techniques and the understanding of tip/sample interactions. Topics include near field imaging, advanced AFM, specialized scanning probe methods in life sciences including new self sensing cantilever systems, combinations of AFM sensors and scanning electron and ion microscopes, calibration methods, frequency modulation AFM for application in liquids, Kelvin probe force microscopy, scanning capacitance microscopy, and the measurement of electrical transport properties at the nanometer scale. Vol. IX focuses on characterization of material surfaces including structural as well as local mechanical characterization, and molecular systems. The volume covers a broad spectrum of STM/AFM investigations including fullerene layers, force spectroscopy for probing material properties in general, biological films .and cells, epithelial and endothelial layers, medical related systems such as amyloidal aggregates, phospholipid monolayers, inorganic films on aluminium and copper oxides, tribological characterization, mechanical properties of polymer nanostructures, technical polymers, and nearfield optics. Volume X focuses on biomimetics and industrial applications such as investigation of structure of gecko feet, semiconductors and their transport phenomena, charge distribution in memory technology, the investigation of surfaces treated by chemicalmechanical planarization, polymeric solar cells, nanoscale contacts, cell adhesion to substrates, nanopatterning, indentation application, new printing techniques, the application of scanning probes in biology, and automatic AFM for manufacturing. As a result, Volumes VIII to X of Applied Scanning Probes microscopies cover a broad and impressive spectrum of recent SPM development and application in many fields of technology, biology and medicine, and introduce many technical concepts and improvements of existing scanning probe techniques. We are very grateful to all our colleagues who took the efforts to prepare manuscripts and provided them in timely manner. Their activity will help both
VI
Preface
students and established scientists in research and development fields to be informed about the latest achievements in scanning probe methods. We would like to cordially thank Dr. Marion Hertel, Senior Editor Chemistry and Mrs. Beate Siek of Springer for their continuous professional support and advice which made it possible to get this volume to the market on time.
Bharat Bhushan Harald Fuchs Masahiko Tomitori
Contents – Volume IX
13
Ultrathin Fullerene-Based Films via STM and STS Luca Gavioli, Cinzia Cepek . . . . . . . . . . . . . . . . . . . . . .
1
13.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
13.2
Basic Principles of STM and STS . . . . . . . . . . . . . . . . . .
2
13.3 13.3.1 13.3.2 13.3.3 13.3.4
Survey of Fullerene-Based Systems Bulk Properties . . . . . . . . . . . Electronic Structure . . . . . . . . . Alkali-Metal-Doped C60 . . . . . . Interaction of C60 with Surfaces . .
. . . . .
4 4 6 8 10
13.4
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18
14
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
Quantitative Measurement of Materials Properties with the (Digital) Pulsed Force Mode Alexander M. Gigler, Othmar Marti . . . . . . . . . . . . . . . . .
23
14.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
23
14.2 14.2.1 14.2.2 14.2.3 14.2.4 14.2.5
Modes of Intermittent Operation . . . . . . . . . . Destructive Versus Nondestructive Measurements The Pulsed Force Mode . . . . . . . . . . . . . . . Operating Principle of the Pulsed Force Mode . . Analog Pulsed Force Mode . . . . . . . . . . . . . Digital Pulsed Force Mode . . . . . . . . . . . . .
. . . . . .
24 25 26 26 32 32
14.3 14.3.1 14.3.2
Contact Mechanics Relevant for Pulsed Force Mode Investigations Hertz Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sneddon’s Extensions to the Hertz Model . . . . . . . . . . . . . .
33 33 36
14.4 14.4.1 14.4.2 14.4.3 14.4.4
Models Incorporating Adhesion . . . . . . . . . Data Processing . . . . . . . . . . . . . . . . . . Polymers . . . . . . . . . . . . . . . . . . . . . . Other Applications . . . . . . . . . . . . . . . . Pulsed Force Mode and Friction Measurements .
37 39 42 47 47
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . . .
. . . . .
. . . . .
VIII
Contents – Volume IX
14.4.5
Cell Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
14.5
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
51
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
52
15
Advances in SPMs for Investigation and Modification of Solid-Supported Monolayers Bruno Pignataro . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
15.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
15.2 15.2.1 15.2.2
SSMs and Their Preparation . . . . . . . . . . . . . . . . . . . . . Self-Assembled Monolayers . . . . . . . . . . . . . . . . . . . . . LB Monolayers . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57 57 60
15.3
Fundamental and Technological Applications of SSMs . . . . . . .
61
15.4 15.4.1 15.4.2
Characterization and Modification of SSMs . . . . . . . . . . . . . Characterization of SSMs . . . . . . . . . . . . . . . . . . . . . . . Modification of SSMs . . . . . . . . . . . . . . . . . . . . . . . . .
64 64 67
15.5 15.5.1 15.5.2
Latest Advances in SPMs and Applications for Imaging of SSMs . Dynamic SFM in the Attractive Regime . . . . . . . . . . . . . . . Dynamic SFM at Different Level of Interaction Forces . . . . . .
67 69 71
15.6 15.6.1 15.6.2 15.6.3 15.6.4
Nanopatterning by SPMs and SSMs . . . . . . Addition Nanolithography . . . . . . . . . . . Elimination and Substitution Nanolithography Nanoelectrochemical Lithography . . . . . . . 3D Nanolithography . . . . . . . . . . . . . . .
. . . . .
76 76 78 79 82
15.7
Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . .
84
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
16
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells Félix Rico, Ewa P. Wojcikiewicz, Vincent T. Moy . . . . . . . . . .
89
16.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
16.2 16.2.1 16.2.2
Principle of Operation . . . . . . . . . . . . . . . . . . . . . . . . . AFM Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . .
90 92 92
16.3 16.3.1 16.3.2 16.3.3 16.3.4
Cell Viscoelasticity . . . . . . . . . . . . AFM Tip Geometries . . . . . . . . . . . Elasticity: Young’s Modulus . . . . . . . Viscoelasticity: Complex Shear Modulus Cell Adhesion . . . . . . . . . . . . . . .
93 94 94 96 98
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
Contents – Volume IX
16.4
IX
Concluding Remarks and Future Directions . . . . . . . . . . . . .
104
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
105
17
Towards a Nanoscale View of Microbial Surfaces Using the Atomic Force Microscope Claire Verbelen, Guillaume Andre, Xavier Haulot, Yann Gilbert, David Alsteens, Etienne Dague and Yves F. Dufrêne . . . . . . . . . 111
17.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
111
17.2 17.2.1 17.2.2 17.2.3
Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . Visualizing Membrane Proteins at Subnanometer Resolution Live-Cell Imaging . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
112 112 112 113
17.3 17.3.1 17.3.2 17.3.3 17.3.4
Force Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . Customized Tips . . . . . . . . . . . . . . . . . . . . . . . . . Probing Nanoscale Elasticity and Surface Properties . . . . . Stretching Cell Surface Polysaccharides and Proteins . . . . . Nanoscale Mapping and Functional Analysis of Molecular Recognition Sites . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
116 116 117 119
. . .
120
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
123
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
124
17.4
18
Cellular Physiology of Epithelium and Endothelium Christoph Riethmüller, Hans Oberleithner . . . . . . . . . . . . . .
127
18.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
127
18.2 18.2.1 18.2.2
Epithelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport Through a Septum . . . . . . . . . . . . . . . . . . . . . In the Kidney . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
128 128 130
18.3 18.3.1 18.3.2 18.3.3 18.3.4
Endothelium . . . . . . . . . . Paracellular Gaps . . . . . . . Cellular Drinking . . . . . . . Wound Healing . . . . . . . . Transmigration of Leukocytes
. . . . .
136 137 139 142 143
18.4
Technical Remarks . . . . . . . . . . . . . . . . . . . . . . . . . .
144
18.5
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
145
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
145
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
X
Contents – Volume IX
19
Application of Atomic Force Microscopy to the Study of Expressed Molecules in or on a Single Living Cell Hyonchol Kim, Hironori Uehara, Rehana Afrin, Hiroshi Sekiguchi, Hideo Arakawa, Toshiya Osada, Atsushi Ikai . . . . . . . . . . . . . 149
19.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19.2
Methods of Manipulation To Study Molecules in or on a Living Cell Using an AFM . . . . . . . . . . . . . . . . AFM Tip Preparation To Manipulate Receptors on a Cell Surface . Analysis of Molecular Interactions Where Multiple Bonds Formed Measurement of Single-Molecule Interaction Strength on Soft Materials . . . . . . . . . . . . . . . . . . . . . . . . . . .
19.2.1 19.2.2 19.2.3 19.3 19.3.1 19.3.2 19.3.3 19.4 19.4.1 19.4.2
Observation of the Distribution of Specific Receptors on a Living Cell Surface . . . . . . . . . . . . . . . . . . . . . . Distribution of Fibronectin Receptors on a Living Fibroblast Cell Distribution of Vitronectin Receptors on a Living Osteoblast Cell Quantification of the Number of Prostaglandin Receptors on a Chinese Hamster Ovary Cell Surface . . . . . . . . . . . . .
150 151 151 153 155
. . .
156 156 159
.
161
Further Application of the AFM to the Study of Single-Cell Biology . . . . . . . . . . . . . . . . . . . . . . . . 164 Manipulation of Expressed mRNAs in a Living Cell Using an AFM 164 Manipulation of Membrane Receptors on a Living Cell Surface Using an AFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
173
20
What Can Atomic Force Microscopy Say About Amyloid Aggregates? Amalisa Relini, Ornella Cavalleri, Claudio Canale, Tiziana Svaldo-Lanero, Ranieri Rolandi, Alessandra Gliozzi . . . . 177
20.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
178
20.2 20.2.1 20.2.2 20.2.3 20.2.4 20.2.5
Techniques and Methods Used To Study Amyloid Aggregates . Optical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . X-ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . Nuclear Magnetic Resonance . . . . . . . . . . . . . . . . . . . Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . .
. . . . . .
181 182 183 185 185 186
20.3
Monitoring the Aggregation Process by AFM . . . . . . . . . . . .
188
20.4
Effect of Surfaces on the Aggregation Process . . . . . . . . . . .
190
20.5
Interaction with Model Membranes . . . . . . . . . . . . . . . . .
193
20.6
Physical Properties of Fibrils Obtained by AFM . . . . . . . . . .
197
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
201
. . . . . .
Contents – Volume IX
21
XI
Atomic Force Microscopy: Interaction Forces Measured in Phospholipid Monolayers, Bilayers and Cell Membranes Zoya Leonenko, David Cramb, Matthias Amrein, Eric Finot . . . .
207
21.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
207
21.2 21.2.1 21.2.2
Phase Transitions of Lipid Bilayers in Water . . . . . . . . . . . . Morphology Change During Lamellar Phase Transition . . . . . . Change in Forces During Phase Transition . . . . . . . . . . . . .
209 210 212
21.3 21.3.1 21.3.2
Force Measurements on Pulmonary Surfactant Monolayers in Air . Adhesion Measurements: Monolayer Stiffness and Function . . . . Repulsive Forces: The Interaction of Charged Airborne Particles with Surfactant . . . . . . . . . . . . . . . . . . . . . . . . . . . .
219 221 222
21.4 21.4.1 21.4.2
Interaction Forces Measured on Lung Epithelial Cells in Buffer . . Cell Culture/Force Measurement Setup . . . . . . . . . . . . . . . Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . .
224 225 227
21.5
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
230
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
231
22
Self-Assembled Monolayers on Aluminum and Copper Oxide Surfaces: Surface and Interface Characteristics, Nanotribological Properties, and Chemical Stability E. Hoque, J. A. DeRose, B. Bhushan, H. J. Mathieu . . . . . . . . .
235
22.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
236
22.2 22.2.1 22.2.2
Substrate Preparation . . . . . . . . . . . . . . . . . . . . . . . . . Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
238 238 239
22.3 22.3.1 22.3.2 22.3.3 22.3.4
Phosphonic Acid and Silane Based SAMs on Al and Cu SAM Preparation . . . . . . . . . . . . . . . . . . . . . Surface and Interface Characterization . . . . . . . . . . Nanotribological Properties . . . . . . . . . . . . . . . . Chemical Stability . . . . . . . . . . . . . . . . . . . . .
. . . . .
239 239 240 261 268
22.4
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
277
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
279
23
23.1
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
High Sliding Velocity Nanotribological Investigations of Materials for Nanotechnology Applications Nikhil S. Tambe, Bharat Bhushan . . . . . . . . . . . . . . . . . . .
283
Bridging Science and Engineering for Nanotribological Investigations . . . . . . . . . . . . . . . . . .
283
XII
Contents – Volume IX
23.1.1 23.1.2
Microtribology/Nanotribology . . . . . . . . . . . . . . . . . . . . Historical Perspective for Velocity Dependence of Friction . . . . .
23.2
Need for Speed: Extending the AFM Capabilities for High Sliding Velocity Studies . . . . . . . . . . . . . . . . . . . Modifications to the Commercial AFM Setup . . . . . . . . . . . . Friction Investigations on the Microscale/Nanoscale at High Sliding Velocities . . . . . . . . . . . . . . . . . . . . . . .
23.2.1 23.2.2 23.3 23.3.1 23.3.2 23.4
Microscale/Nanoscale Friction and Wear Studies at High Sliding Velocities . . . . . . . . . . . . . . . . . . . . . . . Nanoscale Friction Mapping: Understanding Normal Load and Velocity Dependence of Friction Force . . . . . . . . . . . . . Nanoscale Wear Mapping: Wear Studies at High Sliding Velocities . . . . . . . . . . . . . . . . . . . . . . .
284 284 285 287 298 300 301 303
Closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
308
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
309
24
Measurement of the Mechanical Properties of One-Dimensional Polymer Nanostructures by AFM Sung-Kyoung Kim, Haiwon Lee . . . . . . . . . . . . . . . . . . .
311
24.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
311
24.2
AFM-Based Techniques for Measuring the Mechanical Properties of 1D Polymer Nanostructures . . . . . . . . . . . . . . . . . . . .
312
24.3
Mechanical Properties of Electrospun Polymer Nanofibers . . . . .
318
24.4
Test of Reliability of AFM-Based Measurements . . . . . . . . . .
323
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
327
25
Evaluating Tribological Properties of Materials for Total Joint Replacements Using Scanning Probe Microscopy Sriram Sundararajan, Kanaga Karuppiah Kanaga Subramanian .
329
25.1 25.1.1 25.1.2
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Total Joint Replacements . . . . . . . . . . . . . . . . . . . . . . . Social and Economic Significance . . . . . . . . . . . . . . . . . .
329 329 330
25.2 25.2.1 25.2.2 25.2.3
Problems Associated with Total Joint Replacements Tribology . . . . . . . . . . . . . . . . . . . . . . . Materials . . . . . . . . . . . . . . . . . . . . . . . . Lubrication in Joints—the Synovial Fluid . . . . . .
. . . .
330 332 332 333
25.3
Conventional Tribological Testing of Material Pairs for Total Joint Replacements . . . . . . . . . . . . . . . . . . . . .
334
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Contents – Volume IX
XIII
25.3.1 25.3.2
Wear Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Friction Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
334 334
25.4 25.4.1 25.4.2
Scanning Probe Microscopy as a Tool to Study Tribology of Total Joint Replacements . . . . . . . . . . . . . . . . . . . . . . Nanotribology of Ultrahigh Molecular Weight Polyethylene . . . . Fretting Wear of Cobalt–Chromium Alloy . . . . . . . . . . . . . .
334 335 343
25.5
Summary and Future Outlook . . . . . . . . . . . . . . . . . . . .
347
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
348
26
Near-Field Optical Spectroscopy of Single Quantum Constituents Toshiharu Saiki . . . . . . . . . . . . . . . . . . . . . . . . . . . .
351
26.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
351
26.2
General Description of NSOM . . . . . . . . . . . . . . . . . . . .
353
26.3 26.3.1 26.3.2 26.3.3 26.3.4 26.3.5
NSOM Aperture Probe . . . . . . . . . . . . . . . . . . . . . . . Basic Process of Aperture Probe Fabrication . . . . . . . . . . . Tapered Structure and Optical Throughput . . . . . . . . . . . . Fabrication of a Double-Tapered Aperture Probe . . . . . . . . . Evaluation of Transmission Efficiency and Collection Efficiency Evaluation of Spatial Resolution with Single QDs . . . . . . . .
354 354 355 356 357 358
26.4 26.4.1 26.4.2 26.4.3
Single-Quantum-Constituent Spectroscopy . . . . . . . . . . . . . Light–Matter Interaction at the Nanoscale . . . . . . . . . . . . . . Real-Space Mapping of an Exciton Wavefunction Confined in a QD Carrier Localization in Cluster States in GaNAs . . . . . . . . . .
360 361 363 367
26.5
Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
370
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
371
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
373
. . . . . .
Contents – Volume VIII
1
Background-Free Apertureless Near-Field Optical Imaging Pietro Giuseppe Gucciardi, Guillaume Bachelier, Stephan J. Stranick, Maria Allegrini . . . . . . . . . . . . . . . . .
1
1.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1.2 1.2.1 1.2.2
Principles of Apertureless SNOM . . . . . . . . . . . . . . . . . . The Homodyne Apertureless SNOM Concept . . . . . . . . . . . . The Heterodyne and Pseudo-Heterodyne Apertureless SNOM Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 5
1.3
8
1.3.1 1.3.2 1.3.3 1.3.4
Interpretation of the Measured Near-Field Signal in the Presence of a Background . . . . . . . . . . . . . . . Noninterferometric Detection . . . . . . . . . . . . . . . . . Interferometric Detection . . . . . . . . . . . . . . . . . . . Artifacts in Apertureless SNOM and Identification Criteria New Techniques for Background Removal . . . . . . . . .
1.4 1.4.1 1.4.2 1.4.3 1.4.4 1.4.5
Applications of Elastic-Scattering Apertureless SNOM . . . . . . . Material-Specific Imaging . . . . . . . . . . . . . . . . . . . . . . Phase Mapping in Metallic Nanostructures and Optical Waveguides Tip-Induced Resonances in Polaritonic Samples . . . . . . . . . . Applications to Identification of Biosamples . . . . . . . . . . . . Subsurface Imaging and Superlensing . . . . . . . . . . . . . . . .
17 18 19 22 24 25
1.5
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2
. . . . .
. . . . .
. . . . .
. . . . .
9 9 12 14 17
Critical Dimension Atomic Force Microscopy for Sub-50-nm Microelectronics Technology Nodes Hao-Chih Liu, Gregory A. Dahlen, Jason R. Osborne . . . . . . .
31
2.1 2.1.1 2.1.2 2.1.3
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . AFM for Semiconductor and Data Storage Industries . . . . . Scanning Modes: Tapping Versus Deep Trench and CD Mode Specialty Probes . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
32 32 32 34
2.2
Reference Metrology System and Semiconductor Production . . .
34
. . . .
. . . .
XVI
Contents – Volume VIII
2.2.1 2.2.2 2.2.3
Requirements for Metrology Tools . . . . . . . . . . . . . . . . . . AFM as a Reference Metrology System . . . . . . . . . . . . . . . AFM as an In-Line Metrology System . . . . . . . . . . . . . . . .
37 37 39
2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5
Image Analysis for Accurate Metrology . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . Conventional Tip Characterization and Image Reconstruction CD Tip Shape Parameters . . . . . . . . . . . . . . . . . . . . CD Tip Shape Characterization Techniques . . . . . . . . . . CD Image (Reentrant) Reconstruction Algorithms . . . . . .
. . . . . .
40 40 41 44 44 47
2.4 2.4.1 2.4.2
Metrology Applications . . . . . . . . . . . . . . . . . . . . . . . . Examples within Process Control . . . . . . . . . . . . . . . . . . “Fingerprinting” of Sample Features . . . . . . . . . . . . . . . . .
52 52 54
2.5 2.5.1 2.5.2 2.5.3
Developments in Probe Fabrication . . . . . . . . . . . . . . Tip–Sample Interactions: Tip Shape, Stiffness, and Tip Wear Tip Wear and Surface Modification . . . . . . . . . . . . . . Application-Oriented Probe Designs . . . . . . . . . . . . . .
. . . .
60 61 64 67
2.6 2.6.1 2.6.2
Outlook: CD AFM Technologies for 45-/32-/22-nm Nodes . . . . Measuring Sub-50-nm Devices: System Requirements . . . . . . . Probe Technology for 45-/32-nm Structures . . . . . . . . . . . . .
70 70 71
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
73
3
. . . . . .
. . . .
. . . . . .
. . . .
Near Field Probes: From Optical Fibers to Optical Nanoantennas Eugenio Cefal`i, Salvatore Patan`e, Salvatore Spadaro, Renato Gardelli, Matteo Albani, Maria Allegrini . . . . . . . . . . .
77
3.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
77
3.2
Conventional Microscopy and Near-Field Optical Techniques . . .
78
3.3 3.3.1 3.3.2
The Probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aperture SNOM Probes . . . . . . . . . . . . . . . . . . . . . . . . The Apertureless Probe: Optical Nanoantennas . . . . . . . . . . .
84 85 118
3.4
Applications and Perspectives . . . . . . . . . . . . . . . . . . . .
127
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
129
4
Carbon Nanotubes as SPM Tips: Mechanical Properties of Nanotube Tips and Imaging Sophie Marsaudon, Charlotte Bernard, Dirk Dietzel, Cattien V. Nguyen, Anne-Marie Bonnot, Jean-Pierre Aimé, Rodolphe Boisgard . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
138
Contents – Volume VIII
XVII
4.2 4.2.1 4.2.2 4.2.3
CNT Tip Fabrication . . . . . . . . . . . . . . . . . . MWCNTs and Fusing . . . . . . . . . . . . . . . . . . SWCNTs and Direct Growth . . . . . . . . . . . . . . Controlling and Tailoring the Properties of CNT Tips
. . . .
140 141 144 148
4.3
4.3.4 4.3.5 4.3.6 4.3.7 4.3.8
Understanding the Mechanical Properties of CNT Tips: A Dynamical SPM Frequency Modulation Study . . . . . . . . . Mechanical Properties of CNTs . . . . . . . . . . . . . . . . . . Mechanical Properties of CNT Tips . . . . . . . . . . . . . . . . Mechanical Properties of CNT Tips in Dynamical Experiments: Competition Between Elasticity and Adhesion . . . . . . . . . . Experimental Signals . . . . . . . . . . . . . . . . . . . . . . . . Mechanical Properties of MWCNTs . . . . . . . . . . . . . . . . Mechanical Properties of SWCNTs: Main Adhesive Contribution Comparison of Mechanical Properties of CNTs . . . . . . . . . . Special cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . .
149 149 150
. . . . .
151 155 158 163 166 170
4.4 4.4.1 4.4.2
Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature Tour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using the Mechanical Properties of CNTs for Imaging . . . . . . .
174 174 174
4.5
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
176
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
178
4.3.1 4.3.2 4.3.3
5
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Scanning Probes for the Life Sciences Andrea M. Ho, Horacio D. Espinosa . . . . . . . . . . . . . . . . .
183
5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
183
5.2 5.2.1 5.2.2 5.2.3
Microarray Technology Microcontact Printing . Optical Lithography . . Protein Arrays . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
184 185 186 188
5.3 5.3.1 5.3.2 5.3.3
Nanoarray Technology . . . . . . The Push for Nanoscale Detection Probe-Based Patterning . . . . . . Alternative Patterning Methods . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
189 189 191 202
5.4
Nanoscale Deposition Mechanisms . . . . . . . . . . . . . . . . .
204
5.5 5.5.1 5.5.2
AFM Parallelization . . . . . . . . . . . . . . . . . . . . . . . . . . One-Dimensional Arrays . . . . . . . . . . . . . . . . . . . . . . . Two-Dimensional Arrays . . . . . . . . . . . . . . . . . . . . . . .
207 208 209
5.6
Future Prospects for Nanoprobes . . . . . . . . . . . . . . . . . . .
212
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
214
. . . .
. . . .
. . . .
. . . .
. . . .
XVIII
6
Contents – Volume VIII
Self-Sensing Cantilever Sensor for Bioscience Hayato Sone, Sumio Hosaka . . . . . . . . . . . . . . . . . . . . .
219
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
219
6.2
Basics of the Cantilever Mass Sensor . . . . . . . . . . . . . . . .
220
6.3
Finite Element Method Simulation of the Cantilever Vibration . . .
223
6.4 6.4.1 6.4.2
Detection of Cantilever Deflection . . . . . . . . . . . . . . . . . . Using a Position Sensor . . . . . . . . . . . . . . . . . . . . . . . . Using a Piezoresistive Sensor . . . . . . . . . . . . . . . . . . . . .
226 226 227
6.5 6.5.1 6.5.2
Self-Sensing Systems . . . . . . . . . . . . . . . . . . . . . . . . . Vibration Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . Vibration-Frequency Detection Systems . . . . . . . . . . . . . . .
232 232 232
6.6 6.6.1 6.6.2
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Molecule Detection in Air . . . . . . . . . . . . . . . . . . . Antigen and Antibody Detection in Water . . . . . . . . . . . . . .
233 233 238
6.7
Prospective Applications . . . . . . . . . . . . . . . . . . . . . . .
244
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
244
7
AFM Sensors in Scanning Electron and Ion Microscopes: Tools for Nanomechanics, Nanoanalytics, and Nanofabrication Vinzenz Friedli, Samuel Hoffmann, Johann Michler, Ivo Utke . . .
247
7.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
248
7.2 7.2.1 7.2.2 7.2.3 7.2.4
Description of Standalone Techniques . FEB/FIB Nanofabrication . . . . . . . . Cantilever as a Static Force Sensor . . . Cantilever as a Resonating Mass Sensor Nanomanipulation . . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
250 250 255 255 256
7.3 7.3.1 7.3.2 7.3.3 7.3.4 7.3.5 7.3.6
Fundamentals of Cantilever-Based Sensors Static Operation—Force Sensors . . . . . . Dynamic Operation—Mass Sensors . . . . Sensor Scaling . . . . . . . . . . . . . . . . Cantilever Calibration . . . . . . . . . . . . Temperature Stability . . . . . . . . . . . . Piezoresistive Detection . . . . . . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
257 257 258 263 264 265 267
7.4 7.4.1 7.4.2 7.4.3
Analytics at the Nanoscale . . . . . Nanomechanics . . . . . . . . . . . Cantilever-Based Gravimetry . . . . Atomic Force Microscopy in a SEM
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
268 268 276 282
7.5
Perspectives and Outlook . . . . . . . . . . . . . . . . . . . . . . .
283
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
284
. . . .
. . . .
. . . . .
. . . .
. . . .
Contents – Volume VIII
XIX
8
Cantilever Spring-Constant Calibration in Atomic Force Microscopy Peter J. Cumpson, Charles A. Clifford, Jose F. Portoles, James E. Johnstone, Martin Munz . . . . . . . . . . . . . . . . . . . 289
8.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
289
8.2
Applications of AFM . . . . . . . . . . . . . . . . . . . . . . . . .
291
8.3 8.3.1 8.3.2 8.3.3 8.3.4
Force Measurements and Spring-Constant Calibration . . . . Theoretical Methods . . . . . . . . . . . . . . . . . . . . . . V-Shaped Cantilevers and the Parallel-Beam Approximation . Dynamic Experimental Methods . . . . . . . . . . . . . . . . Thermal Methods . . . . . . . . . . . . . . . . . . . . . . . .
. . . . .
292 292 293 295 297
8.4 8.4.1 8.4.2
Repeatability in AFM Force Measurements . . . . . . . . . . . . . z-Axis Displacement Repeatability . . . . . . . . . . . . . . . . . . Cantilever Deflection Repeatability . . . . . . . . . . . . . . . . .
299 300 300
8.5
Microfabricated Devices for AFM Force Calibration . . . . . . . .
302
8.6
Lateral Force Calibration . . . . . . . . . . . . . . . . . . . . . . .
308
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
312
9
. . . . .
. . . . .
Frequency Modulation Atomic Force Microscopy in Liquids Suzanne P. Jarvis, John E. Sader, Takeshi Fukuma . . . . . . . . .
315
9.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
315
9.2 9.2.1 9.2.2 9.2.3
Instrumentation . . . . . . . . . Basic Setup for FM-AFM . . . Cantilever Excitation in a Liquid Cantilever-Deflection Detection
. . . .
318 318 319 320
9.3 9.3.1 9.3.2
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonbiological Systems . . . . . . . . . . . . . . . . . . . . . . . . Biological Systems . . . . . . . . . . . . . . . . . . . . . . . . . .
326 326 327
9.4 9.4.1 9.4.2 9.4.3 9.4.4 9.4.5
Theoretical Framework for Quantitative FM-AFM Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . . Decomposition of Interaction Force . . . . . . . . . . . . . Governing Equations . . . . . . . . . . . . . . . . . . . . . Fundamental Conditions on Interaction Force . . . . . . . . Determination of Forces . . . . . . . . . . . . . . . . . . . Validation of Formulas . . . . . . . . . . . . . . . . . . . .
. . . . . .
332 332 335 336 337 339
9.5 9.5.1 9.5.2
Operation in a Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . Governing Equations and Resonance Frequency in a Fluid . . . . . Validation of FM-AFM Force Measurements in a Liquid . . . . . .
340 340 342
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . . . .
. . . .
. . . . . .
. . . .
. . . . . .
XX
Contents – Volume VIII
9.6 9.6.1 9.6.2 9.6.3 9.6.4
Phase Detuning in FM-AFM . . . . . . . . . . . . . . . . . . . Governing Equations for Arbitrary Phase Shift . . . . . . . . . Coupling of Conservative and Dissipative Forces . . . . . . . . Operation of FM-AFM Away From the Resonance Frequency . Calibration of 90° Phase Shift . . . . . . . . . . . . . . . . . .
. . . . .
. . . . .
343 344 345 346 346
9.7
Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . .
348
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
349
10
Kelvin Probe Force Microscopy: Recent Advances and Applications Yossi Rosenwaks, Oren Tal, Shimon Saraf, Alex Schwarzman, Eli Lepkifker, Amir Boag . . . . . . . . . . . . . . . . . . . . . . . . 351
10.1
Kelvin Probe Force Microscopy . . . . . . . . . . . . . . . . . . .
351
10.2 10.2.1 10.2.2 10.2.3 10.2.4
Sensitivity and Spatial Resolution in KPFM . . . . . . . Tip–Sample Electrostatic Interaction . . . . . . . . . . . A Fast Algorithm for Calculating the Electrostatic Force Noise in KPFM Images . . . . . . . . . . . . . . . . . . . Deconvolution of KPFM Images . . . . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
354 354 356 359 360
10.3 10.3.1 10.3.2 10.3.3
Measurement of Semiconductor Surface States . . . . . Surface Charge and Band Bending Measurements . . . Measuring the Energy Distribution of the Surface States Organic Semiconductors: Bulk Density of States . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
362 362 365 368
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
374
11
. . . .
Application of Scanning Capacitance Microscopy to Analysis at the Nanoscale Štefan Lányi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
377
11.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
377
11.2 11.2.1 11.2.2 11.2.3 11.2.4
Capacitance Microscopes . . . . . . . . Resolution of Capacitance Transducers Stray Capacitance of the Probe . . . . . Sensitivity of the Probe . . . . . . . . . SCM Operation Modes . . . . . . . . .
. . . . .
379 382 387 392 399
11.3
Looking at the Invisible—Capacitance Contrast . . . . . . . . . . .
400
11.4
Semiconductor Analysis . . . . . . . . . . . . . . . . . . . . . . .
401
11.5
Other Semiconductor Structures . . . . . . . . . . . . . . . . . . .
405
11.6 11.6.1 11.6.2
Looking Deeper . . . . . . . . . . . . . . . . . . . . . . . . . . . . Impedance Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . Deep Level Transient Spectroscopy . . . . . . . . . . . . . . . . .
406 407 408
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
Contents – Volume VIII
XXI
11.7
Optimising the Experimental Conditions . . . . . . . . . . . . . .
414
11.8
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
416
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
417
12
Probing Electrical Transport Properties at the Nanoscale by Current-Sensing Atomic Force Microscopy Laura Fumagalli, Ignacio Casuso, Giorgio Ferrari, Gabriel Gomila . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
12.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
421
12.2
Fundamentals of Electrical Transport Properties: Resistance, Impedance and Noise . . . . . . . . . . . . . . . . . .
424
12.3 12.3.1 12.3.2 12.3.3
Experimental Setups for CS-AFM Conductive Probes . . . . . . . . . Operating Modes of AFM . . . . Current Detection Instrumentation
. . . .
429 429 431 432
12.4
Conductive Atomic Force Microscopy . . . . . . . . . . . . . . . .
434
12.5
Nanoscale Impedance Microscopy . . . . . . . . . . . . . . . . . .
437
12.6
Electrical Noise Microscopy . . . . . . . . . . . . . . . . . . . . .
443
12.7
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
445
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
446
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
451
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Contents – Volume X
27
Gecko Feet: Natural Attachment Systems for Smart Adhesion— Mechanism, Modeling, and Development of Bio-Inspired Materials Bharat Bhushan, Robert A. Sayer . . . . . . . . . . . . . . . . . .
1
27.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
27.2 27.2.1 27.2.2 27.2.3 27.2.4 27.2.5
Tokay Gecko . . . . . . . . . . . . Construction of Tokay Gecko . . . Other Attachment Systems . . . . Adaptation to Surface Roughness Peeling . . . . . . . . . . . . . . . Self-Cleaning . . . . . . . . . . .
. . . . . .
2 2 5 7 8 10
27.3 27.3.1 27.3.2
Attachment Mechanisms . . . . . . . . . . . . . . . . . . . . . . . Van der Waals Forces . . . . . . . . . . . . . . . . . . . . . . . . . Capillary Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12 12 13
27.4 27.4.1 27.4.2 27.4.3 27.4.4
Experimental Adhesion Test Techniques and Data Adhesion Under Ambient Conditions . . . . . . . Effects of Temperature . . . . . . . . . . . . . . . Effects of Humidity . . . . . . . . . . . . . . . . . Effects of Hydrophobicity . . . . . . . . . . . . .
. . . . .
. . . . .
14 15 17 18 18
27.5 27.5.1 27.5.2 27.5.3 27.5.4
. . . .
19 21 21 23
27.5.5 27.5.6
Adhesion Modeling . . . . . . . . . . . . . . . . . . . . . . . . . Spring Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . Single Spring Contact Analysis . . . . . . . . . . . . . . . . . . The Multilevel Hierarchical Spring Analysis . . . . . . . . . . . Adhesion Results for the Gecko Attachment System Contacting a Rough Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . Capillarity Effects . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesion Results that Account for Capillarity Effects . . . . . .
. . .
26 30 31
27.6 27.6.1 27.6.2 27.6.3 27.6.4 27.6.5
Modeling of Biomimetic Fibrillar Structures Fiber Model . . . . . . . . . . . . . . . . . . Single Fiber Contact Analysis . . . . . . . . Constraints . . . . . . . . . . . . . . . . . . Numerical Simulation . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . .
. . . . . .
34 34 34 35 39 41
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . .
. . . . . .
. . . . . .
. . . . . .
XXIV
Contents – Volume X
27.7 27.7.1 27.7.2
Fabrication of Biomimetric Gecko Skin . . . . . . . . . . . . . . . Single-Level Hierarchical Structures . . . . . . . . . . . . . . . . . Multilevel Hierarchical Structures . . . . . . . . . . . . . . . . . .
48 49 53
27.8
Closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
28
Carrier Transport in Advanced Semiconductor Materials Filippo Giannazzo, Patrick Fiorenza, Vito Raineri . . . . . . . . .
63
Majority Carrier Distribution in Semiconductors: Imaging and Quantification . . . . . . . . . . . . Basic Principles of SCM . . . . . . . . . . . . . Carrier Imaging Capability by SCM . . . . . . . Quantification of SCM Raw Data . . . . . . . . Basic Principles of SSRM . . . . . . . . . . . . Carrier Imaging Capability by SSRM . . . . . . Quantification of SSRM Raw Data . . . . . . . . Drift Mobility by SCM and SSRM . . . . . . . .
. . . . . . . .
64 64 67 70 78 81 81 85
Carrier Transport Through Metal–Semiconductor Barriers by C-AFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
28.3 28.3.1 28.3.2
Charge Transport in Dielectrics by C-AFM . . . . . . . . . . . . . Direct Determination of Breakdown . . . . . . . . . . . . . . . . . Weibull Statistics by C-AFM . . . . . . . . . . . . . . . . . . . . .
93 97 99
28.4
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
101
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
101
28.1 28.1.1 28.1.2 28.1.3 28.1.4 28.1.5 28.1.6 28.1.7 28.2
29
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
Visualization of Fixed Charges Stored in Condensed Matter and Its Application to Memory Technology Yasuo Cho . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
105
29.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
105
29.2
Principle and Theory for SNDM . . . . . . . . . . . . . . . . . . .
106
29.3
Microscopic Observation of Area Distribution of the Ferroelectric Domain Using SNDM . . . . . . . . . . . . . .
107
Visualization of Stored Charge in Semiconductor Flash Memories Using SNDM . . . . . . . . . .
109
29.5
Higher-Order SNDM . . . . . . . . . . . . . . . . . . . . . . . . .
110
29.6
Noncontact SNDM . . . . . . . . . . . . . . . . . . . . . . . . . .
111
29.7
SNDM for 3D Observation of Nanoscale Ferroelectric Domains . .
112
29.4
Contents – Volume X
29.8
XXV
29.8.5 29.8.6
Next-Generation Ultra-High-Density Ferroelectric Data Storage Based on SNDM . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of Ferroelectric Data Storage . . . . . . . . . . . . . . SNDM Nanodomain Engineering System and Ferroelectric Recording Medium . . . . . . . . . . . . . . . Nanodomain Formation in a LiTaO3 Single Crystal . . . . . . . . High-Speed Switching of Nanoscale Ferroelectric Domains in Congruent Single-Crystal LiTaO3 . . . . . . . . . . . . . . . . Prototype of a High-Density Ferroelectric Data Storage System . Realization of 10 Tbit/in.2 Memory Density . . . . . . . . . . . .
. . .
120 122 126
29.9
Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
128
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
129
29.8.1 29.8.2 29.8.3 29.8.4
30
. .
114 114
. .
116 117
Applications of Scanning Probe Methods in Chemical Mechanical Planarization Toshi Kasai, Bharat Bhushan . . . . . . . . . . . . . . . . . . . . .
131
30.1 30.1.1 30.1.2
Overview of CMP Technology and the Need for SPM . . . . . . . CMP Technology and Its Key Elements . . . . . . . . . . . . . . . Various CMP Processes and the Need for SPM . . . . . . . . . . .
131 131 134
30.2
AFP for the Evaluation of Dishing and Erosion . . . . . . . . . . .
137
30.3
Surface Planarization and Roughness Characterization in CMP Using AFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
141
Use of Modified Atomic Force Microscope Tips for Fundamental Studies of CMP Mechanisms . . . . . . . . . . .
144
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
149
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
149
30.4 30.5
31
Scanning Probe Microscope Application for Single Molecules in a π -Conjugated Polymer Toward Molecular Devices Based on Polymer Chemistry Ken-ichi Shinohara . . . . . . . . . . . . . . . . . . . . . . . . . .
153
31.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
153
31.2 31.2.1
Chiral Helical π-Conjugated Polymer . . . . . . . . . . . . . . . Helical Chirality of a π-Conjugated Main Chain Induced by Polymerization of Phenylacetylene with Chiral Bulky Groups . . . . . . . . . . . . . . . . . . . . . . Direct Measurement of the Chiral Quaternary Structure in a π-Conjugated Polymer . . . . . . . . . . . . . . . . . . . . . Direct Measurement of Structural Diversity in Single Molecules of a Chiral Helical π-Conjugated Polymer . . . . . . . . . . . . .
.
154
.
156
.
158
.
163
31.2.2 31.2.3
XXVI
31.2.4 31.3 31.3.1
Contents – Volume X
Dynamic Structure of Single Molecules in a Chiral Helical π-Conjugated Polymer by a High-Speed AFM . . . . . . . . . . .
166
Supramolecular Chiral π-Conjugated Polymer . . . . . . . . . . . Simultaneous Imaging of Structure and Fluorescence of a Supramolecular Chiral π-Conjugated Polymer . . . . . . . . . Dynamic Structure of a Supramolecular Chiral π-Conjugated Polymer by a High-Speed AFM . . . . . . . . . . . . . . . . . . .
169
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
181
31.3.2
32
169 177
Scanning Probe Microscopy on Polymer Solar Cells Joachim Loos, Alexander Alexeev . . . . . . . . . . . . . . . . . .
183
32.1
Brief Introduction to Polymer Solar Cells . . . . . . . . . . . . . .
184
32.2
Sample Preparation and Characterization Techniques . . . . . . . .
188
32.3 32.3.1
Morphology Features of the Photoactive Layer . . . . . . . Influence of Composition and Solvents on the Morphology of the Active Layer . . . . . . . . . . . . . . . . . . . . . . Influence of Annealing . . . . . . . . . . . . . . . . . . . . All-Polymer Solar Cells . . . . . . . . . . . . . . . . . . . .
. . . .
190
. . . . . . . . . . . .
190 193 199 201
32.4.2
Nanoscale Characterization of Properties of the Active Layer . . . Local Optical Properties As Measured by Scanning Near-Field Optical Microscopy . . . . . . . . . . . . Characterization of Nanoscale Electrical Properties . . . . . . . . .
201 203
32.5
Summary and Outlook . . . . . . . . . . . . . . . . . . . . . . . .
212
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
213
32.3.2 32.3.3 32.4 32.4.1
33
Scanning Probe Anodization for Nanopatterning Hiroyuki Sugimura . . . . . . . . . . . . . . . . . . . . . . . . . .
217
33.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
217
33.2
Electrochemical Origin of SPM-Based Local Oxidation . . . . . .
218
33.3 33.3.1 33.3.2 33.3.3
Variation in Scanning Probe Anodization . . . . . . . Patternable Materials in Scanning Probe Anodization . Environment Control in Scanning Probe Anodization . Electrochemical Scanning Surface Modification Using Cathodic Reactions . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . .
223 223 226
. . . . . . .
229
33.4 33.4.1 33.4.2 33.4.3
Progress in Scanning Probe Anodization . . . . . . . . . . . From STM-Based Anodization to AFM-Based Anodization Versatility of AFM-Based Scanning Probe Anodization . . In Situ Characterization of Anodized Structures by AFM-Based Methods . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .
232 232 233
. . . .
233
Contents – Volume X
XXVII
33.4.4
Technical Development of Scanning Probe Anodization . . . . . .
237
33.5 33.5.1 33.5.2 33.5.3
Lithographic Applications of Scanning Probe Anodization . . Device Prototyping . . . . . . . . . . . . . . . . . . . . . . . Pattern Transfer from Anodic Oxide to Other Materials . . . Integration of Scanning Probe Lithography with Other High-Throughput Lithographies . . . . . . . . . . Chemical Manipulation of Nano-objects by the Use of a Nanotemplate Prepared by Scanning Probe Anodization
. . . . . . . . .
239 239 240
. . .
247
. . .
248
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
251
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
251
33.5.4 33.6
34
Tissue Engineering: Nanoscale Contacts in Cell Adhesion to Substrates Mario D’Acunto, Paolo Giusti, Franco Maria Montevecchi, Gianluca Ciardelli . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
34.1
Tissue Engineering: A Brief Introduction . . . . . . . . . . . . . .
34.2
Fundamental Features of Cell Motility and Cell–Substrates Adhesion . . . . . . . . . . . . . . . . . . . . Biomimetic Scaffolds, Roughness, and Contact Guidance for Cell Adhesion and Motility . . . . . . . . . . . . . . . . . . . .
34.2.1 34.3
257 261 268
Experimental Strategies for Cell–ECM Adhesion Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
271
34.4
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
279
34.5
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
279
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
280
35
Scanning Probe Microscopy in Biological Research Tatsuo Ushiki, Kazushige Kawabata . . . . . . . . . . . . . . . . .
285
35.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
285
35.2 35.2.1 35.2.2 35.2.3 35.2.4 35.2.5 35.2.6
SPM for Visualization of the Surface of Biomaterials . . . Advantages of AFM in Biological Studies . . . . . . . . . AFM of Biomolecules . . . . . . . . . . . . . . . . . . . . AFM of Isolated Intracellular and Extracellular Structures AFM of Tissue Sections . . . . . . . . . . . . . . . . . . . AFM of Living Cells and Their Movement . . . . . . . . Combination of AFM with Scanning Near-Field Optical Microscopy for Imaging Biomaterials . . . . . . . . . . .
. . . . . .
286 286 287 289 292 292
. . . . .
294
SPM for Measuring Physical Properties of Biomaterials . . . . . . Evaluation Methods of Viscoelasticity . . . . . . . . . . . . . . . . Examples for Viscoelasticity Mapping Measurements . . . . . . .
296 296 299
35.3 35.3.1 35.3.2
. . . . . .
. . . . . .
. . . . . .
. . . . . .
XXVIII
35.3.3
Contents – Volume X
Combination of Viscoelasticity Measurement with Other Techniques . . . . . . . . . . . . . . . . . . . . . . . .
302
35.4
SPM as a Manipulation Tool in Biology . . . . . . . . . . . . . . .
304
35.5
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
306
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
306
36
Novel Nanoindentation Techniques and Their Applications Jiping Ye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
309
36.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
309
36.2 36.2.1 36.2.2
Basic Principles of Contact . . . . . . . . . . . . . . . . . . . . . . Meyer’s Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elastic Contact Solution . . . . . . . . . . . . . . . . . . . . . . . .
311 311 312
36.3
Tip Rigidity and Geometry . . . . . . . . . . . . . . . . . . . . . .
313
36.4 36.4.1 36.4.2 36.4.3
Hardness and Modulus Measurements Analysis Method . . . . . . . . . . . Practical Application Aspects . . . . Recent Applications . . . . . . . . .
. . . .
314 314 316 320
36.5 36.5.1 36.5.2
Yield Stress and Modulus Measurements . . . . . . . . . . . . . . Analysis Method . . . . . . . . . . . . . . . . . . . . . . . . . . . Recent Applications . . . . . . . . . . . . . . . . . . . . . . . . . .
324 324 326
36.6 36.6.1 36.6.2 36.6.3
Work-Hardening Rate and Exponent Measurements Analysis Method . . . . . . . . . . . . . . . . . . . Practical Application Aspects . . . . . . . . . . . . . Recent Applications . . . . . . . . . . . . . . . . . .
. . . .
329 329 333 335
36.7 36.7.1 36.7.2
Viscoelastic Compliance and Modulus . . . . . . . . . . . . . . . Analysis Method . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical Application Aspects . . . . . . . . . . . . . . . . . . . . .
336 336 339
36.8
Other Mechanical Characteristics . . . . . . . . . . . . . . . . . .
342
36.9
Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
343
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
343
37
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Applications to Nano-Dispersion Macromolecule Material Evaluation in an Electrophotographic Printer Yasushi Kadota . . . . . . . . . . . . . . . . . . . . . . . . . . . .
347
37.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
347
37.2 37.2.1
Electrophotographic Processes . . . . . . . . . . . . . . . . . . . . Principle and Characteristics of an Electrophotographic System . .
348 348
Contents – Volume X
37.2.2 37.3 37.3.1 37.3.2
XXIX
Microcharacteristic and Analysis Technology for Functional Components . . . . . . . . . . . . . . . . . . . . . . . . . . . .
352 352
. . .
353
. . .
355
Current Technology Subjects . . . . . . . . . . . . . . . . . . . . .
357
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
357
37.3.3 37.4
38
SPM Applications to Electrophotographic Systems . . . . . . Measurement of Electrostatic Charge of Toner . . . . . . . . Measurement of the Adhesive Force Between a Particle and a Substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . Observation of a Nanodispersion Macromolecule Interface —Toner Adhesion to a Fusing Roller . . . . . . . . . . . . .
349
Automated AFM as an Industrial Process Metrology Tool for Nanoelectronic Manufacturing Tianming Bao, David Fong, Sean Hand . . . . . . . . . . . . . . .
359
38.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
359
38.2 38.2.1 38.2.2 38.2.3 38.2.4
Dimensional Metrology with AFM . Dimensional Metrology . . . . . . . AFM Scanning Technology . . . . . AFM Probe Technology . . . . . . . AFM Metrology Capability . . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
361 361 362 367 367
38.3
Applications in Semiconductors— Logic and Memory Integrated Circuits . Shallow Trench Isolation Resist Pattern STI Etch . . . . . . . . . . . . . . . . . STI CMP . . . . . . . . . . . . . . . . . Gate Resist Pattern . . . . . . . . . . . Gate Etch . . . . . . . . . . . . . . . . FinFET Gate Formation . . . . . . . . . Gate Sidewall Spacer . . . . . . . . . . Strained SiGe Source/Drain Recess . . Pre-metal Dielectric CMP . . . . . . . . Contact and Via Photo Pattern . . . . . Contact Etch . . . . . . . . . . . . . . . Contact CMP . . . . . . . . . . . . . . Metal Trench Photo Pattern . . . . . . . Metal Trench Etch . . . . . . . . . . . . Via Etch . . . . . . . . . . . . . . . . . Via Etch . . . . . . . . . . . . . . . . . Roughness . . . . . . . . . . . . . . . . LWR, LER, and SWR . . . . . . . . . . DRAM DT Capacitor . . . . . . . . . . Ferroelectric RAM Capacitor . . . . . . Optical Proximity Correction . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
370 370 372 375 378 379 383 385 385 386 387 387 389 390 390 392 394 396 397 397 398 398
38.3.1 38.3.2 38.3.3 38.3.4 38.3.5 38.3.6 38.3.7 38.3.8 38.3.9 38.3.10 38.3.11 38.3.12 38.3.13 38.3.14 38.3.15 38.3.16 38.3.17 38.3.18 38.3.19 38.3.20 38.3.21
. . . . .
XXX
Contents – Volume X
38.4 38.4.1 38.4.2
Applications in Photomask . . . . . . . . . . . . . . . . . . . . . . Photomask Pattern and Etch . . . . . . . . . . . . . . . . . . . . . Photomask Defect Review and Repair . . . . . . . . . . . . . . . .
399 399 400
38.5 38.5.1 38.5.2
Applications in Hard Disk Manufacturing . . . . . . . . . . . . . . Magnetic Thin-Film Recording Head . . . . . . . . . . . . . . . . Slider for Hard Drive . . . . . . . . . . . . . . . . . . . . . . . . .
401 401 405
38.6 38.6.1 38.6.2
Applications in Microelectromechanical System Devices . . . . . Contact Image Sensor . . . . . . . . . . . . . . . . . . . . . . . . . Digital Light Processor Mirror Device . . . . . . . . . . . . . . . .
406 406 408
38.7
Challenge and Potential Improvement . . . . . . . . . . . . . . . .
408
38.8
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
409
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
411
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
413
Contents – Volume I
Part I
Scanning Probe Microscopy
1
Dynamic Force Microscopy André Schirmeisen, Boris Anczykowski, Harald Fuchs . . . . . . .
3
Interfacial Force Microscopy: Selected Applications Jack E. Houston . . . . . . . . . . . . . . . . . . . . . . . . . . . .
41
Atomic Force Microscopy with Lateral Modulation Volker Scherer, Michael Reinstädtler, Walter Arnold . . . . . . . .
75
Sensor Technology for Scanning Probe Microscopy Egbert Oesterschulze, Rainer Kassing . . . . . . . . . . . . . . . .
117
Tip Characterization for Dimensional Nanometrology John S. Villarrubia . . . . . . . . . . . . . . . . . . . . . . . . . .
147
2 3 4 5
Part II
Characterization
6
Micro/Nanotribology Studies Using Scanning Probe Microscopy Bharat Bhushan . . . . . . . . . . . . . . . . . . . . . . . . . . . .
171
Visualization of Polymer Structures with Atomic Force Microscopy Sergei Magonov . . . . . . . . . . . . . . . . . . . . . . . . . . . .
207
Displacement and Strain Field Measurements from SPM Images Jürgen Keller, Dietmar Vogel, Andreas Schubert, Bernd Michel . .
253
7 8 9
AFM Characterization of Semiconductor Line Edge Roughness Ndubuisi G. Orji, Martha I. Sanchez, Jay Raja, Theodore V. Vorburger . . . . . . . . . . . . . . . . . . . . . . . . . 277
10
Mechanical Properties of Self-Assembled Organic Monolayers: Experimental Techniques and Modeling Approaches Redhouane Henda . . . . . . . . . . . . . . . . . . . . . . . . . . .
303
XXXII
11 12
Contents – Volume I
Micro-Nano Scale Thermal Imaging Using Scanning Probe Microscopy Li Shi, Arun Majumdar . . . . . . . . . . . . . . . . . . . . . . . .
327
The Science of Beauty on a Small Scale. Nanotechnologies Applied to Cosmetic Science Gustavo Luengo, Frédéric Leroy . . . . . . . . . . . . . . . . . . .
363
Part III Industrial Applications 13
SPM Manipulation and Modifications and Their Storage Applications Sumio Hosaka . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
389
Super Density Optical Data Storage by Near-Field Optics Jun Tominaga . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
429
Capacitance Storage Using a Ferroelectric Medium and a Scanning Capacitance Microscope (SCM) Ryoichi Yamamoto . . . . . . . . . . . . . . . . . . . . . . . . . . .
439
Room-Temperature Single-Electron Devices formed by AFM Nano-Oxidation Process Kazuhiko Matsumoto . . . . . . . . . . . . . . . . . . . . . . . . .
459
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
469
14 15
16
Contents – Volume II
1 2 3 4 5
Higher Harmonics in Dynamic Atomic Force Microscopy Robert W. Stark, Martin Stark . . . . . . . . . . . . . . . . . . . .
1
Atomic Force Acoustic Microscopy Ute Rabe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
Scanning Ion Conductance Microscopy Tilman E. Schäffer, Boris Anczykowski, Harald Fuchs . . . . . . .
91
Spin-Polarized Scanning Tunneling Microscopy Wulf Wulfhekel, Uta Schlickum, Jürgen Kirschner . . . . . . . . . .
121
Dynamic Force Microscopy and Spectroscopy Ferry Kienberger, Hermann Gruber, Peter Hinterdorfer . . . . . .
143
6
Sensor Technology for Scanning Probe Microscopy and New Applications Egbert Oesterschulze, Leon Abelmann, Arnout van den Bos, Rainer Kassing, Nicole Lawrence, Gunther Wittstock, Christiane Ziegler . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
7
Quantitative Nanomechanical Measurements in Biology Małgorzata Lekka, Andrzej J. Kulik . . . . . . . . . . . . . . . . .
205
Scanning Microdeformation Microscopy: Subsurface Imaging and Measurement of Elastic Constants at Mesoscopic Scale Pascal Vairac, Bernard Cretin . . . . . . . . . . . . . . . . . . . .
241
Electrostatic Force and Force Gradient Microscopy: Principles, Points of Interest and Application to Characterisation of Semiconductor Materials and Devices Paul Girard, Alexander Nikolaevitch Titkov . . . . . . . . . . . . .
283
8
9
10
Polarization-Modulation Techniques in Near-Field Optical Microscopy for Imaging of Polarization Anisotropy in Photonic Nanostructures Pietro Giuseppe Gucciardi, Ruggero Micheletto, Yoichi Kawakami, Maria Allegrini . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
XXXIV
11
Contents – Volume II
Focused Ion Beam as a Scanning Probe: Methods and Applications Vittoria Raffa, Piero Castrataro, Arianna Menciassi, Paolo Dario .
361
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
413
Contents – Volume III
12
13
14
15
16
17
18 19
20 21
Atomic Force Microscopy in Nanomedicine Dessy Nikova, Tobias Lange, Hans Oberleithner, Hermann Schillers, Andreas Ebner, Peter Hinterdorfer . . . . . . .
1
Scanning Probe Microscopy: From Living Cells to the Subatomic Range Ille C. Gebeshuber, Manfred Drack, Friedrich Aumayr, Hannspeter Winter, Friedrich Franek . . . . . . . . . . . . . . . . .
27
Surface Characterization and Adhesion and Friction Properties of Hydrophobic Leaf Surfaces and Nanopatterned Polymers for Superhydrophobic Surfaces Zachary Burton, Bharat Bhushan . . . . . . . . . . . . . . . . . .
55
Probing Macromolecular Dynamics and the Influence of Finite Size Effects Scott Sills, René M. Overney . . . . . . . . . . . . . . . . . . . . .
83
Investigation of Organic Supramolecules by Scanning Probe Microscopy in Ultra-High Vacuum Laurent Nony, Enrico Gnecco, Ernst Meyer . . . . . . . . . . . . .
131
One- and Two-Dimensional Systems: Scanning Tunneling Microscopy and Spectroscopy of Organic and Inorganic Structures Luca Gavioli, Massimo Sancrotti . . . . . . . . . . . . . . . . . . .
183
Scanning Probe Microscopy Applied to Ferroelectric Materials Oleg Tikhomirov, Massimiliano Labardi, Maria Allegrini . . . . . .
217
Morphological and Tribological Characterization of Rough Surfaces by Atomic Force Microscopy Renato Buzio, Ugo Valbusa . . . . . . . . . . . . . . . . . . . . . .
261
AFM Applications for Contact and Wear Simulation Nikolai K. Myshkin, Mark I. Petrokovets, Alexander V. Kovalev . .
299
AFM Applications for Analysis of Fullerene-Like Nanoparticles Lev Rapoport, Armen Verdyan . . . . . . . . . . . . . . . . . . . .
327
XXXVI
22
Contents – Volume III
Scanning Probe Methods in the Magnetic Tape Industry James K. Knudsen . . . . . . . . . . . . . . . . . . . . . . . . . . .
343
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
371
Contents – Volume IV
23
Scanning Probe Lithography for Chemical, Biological and Engineering Applications Joseph M. Kinsella, Albena Ivanisevic . . . . . . . . . . . . . . . .
1
Nanotribological Characterization of Human Hair and Skin Using Atomic Force Microscopy (AFM) Bharat Bhushan, Carmen LaTorre . . . . . . . . . . . . . . . . . .
35
Nanofabrication with Self-Assembled Monolayers by Scanning Probe Lithography Jayne C. Garno, James D. Batteas . . . . . . . . . . . . . . . . . .
105
Fabrication of Nanometer-Scale Structures by Local Oxidation Nanolithography Marta Tello, Fernando García, Ricardo García . . . . . . . . . . .
137
Template Effects of Molecular Assemblies Studied by Scanning Tunneling Microscopy (STM) Chen Wang, Chunli Bai . . . . . . . . . . . . . . . . . . . . . . . .
159
Microfabricated Cantilever Array Sensors for (Bio-)Chemical Detection Hans Peter Lang, Martin Hegner, Christoph Gerber . . . . . . . .
183
Nano-Thermomechanics: Fundamentals and Application in Data Storage Devices B. Gotsmann, U. Dürig . . . . . . . . . . . . . . . . . . . . . . . .
215
Applications of Heated Atomic Force Microscope Cantilevers Brent A. Nelson, William P. King . . . . . . . . . . . . . . . . . . .
251
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
277
24
25
26
27
28 29
30
Contents – Volume V
1 2 3 4 5 6 7 8
9
10
Integrated Cantilevers and Atomic Force Microscopes Sadik Hafizovic, Kay-Uwe Kirstein, Andreas Hierlemann . . . . .
1
Electrostatic Microscanner Yasuhisa Ando . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
23
Low-Noise Methods for Optical Measurements of Cantilever Deflections Tilman E. Schäffer . . . . . . . . . . . . . . . . . . . . . . . . . . .
51
Q-controlled Dynamic Force Microscopy in Air and Liquids Hendrik H¨olscher, Daniel Ebeling, Udo D. Schwarz . . . . . . . .
75
High-Frequency Dynamic Force Microscopy Hideki Kawakatsu . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
Torsional Resonance Microscopy and Its Applications Chanmin Su, Lin Huang, Craig B. Prater, Bharat Bhushan . . . . .
113
Modeling of Tip-Cantilever Dynamics in Atomic Force Microscopy Yaxin Song, Bharat Bhushan . . . . . . . . . . . . . . . . . . . . .
149
Combined Scanning Probe Techniques for In-Situ Electrochemical Imaging at a Nanoscale Justyna Wiedemair, Boris Mizaikoff, Christine Kranz . . . . . . . .
225
New AFM Developments to Study Elasticity and Adhesion at the Nanoscale Robert Szoszkiewicz, Elisa Riedo . . . . . . . . . . . . . . . . . . .
269
Near-Field Raman Spectroscopy and Imaging Pietro Giuseppe Gucciardi, Sebastiano Trusso, Cirino Vasi, Salvatore Patanè, Maria Allegrini . . . . . . . . . . . . . . . . . . . 287
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
331
Contents – Volume VI
11
12
13
Scanning Tunneling Microscopy of Physisorbed Monolayers: From Self-Assembly to Molecular Devices Thomas Müller . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
Tunneling Electron Spectroscopy Towards Chemical Analysis of Single Molecules Tadahiro Komeda . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
STM Studies on Molecular Assembly at Solid/Liquid Interfaces Ryo Yamada, Kohei Uosaki . . . . . . . . . . . . . . . . . . . . . .
65
14
Single-Molecule Studies on Cells and Membranes Using the Atomic Force Microscope Ferry Kienberger, Lilia A. Chtcheglova, Andreas Ebner, Theeraporn Puntheeranurak, Hermann J. Gruber, Peter Hinterdorfer . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
15
Atomic Force Microscopy of DNA Structure and Interactions Neil H. Thomson . . . . . . . . . . . . . . . . . . . . . . . . . . . .
127
Direct Detection of Ligand–Protein Interaction Using AFM Małgorzata Lekka, Piotr Laidler, Andrzej J. Kulik . . . . . . . . .
165
Dynamic Force Microscopy for Molecular-Scale Investigations of Organic Materials in Various Environments Hirofumi Yamada, Kei Kobayashi . . . . . . . . . . . . . . . . . .
205
Noncontact Atomic Force Microscopy Yasuhiro Sugawara . . . . . . . . . . . . . . . . . . . . . . . . . .
247
Tip-Enhanced Spectroscopy for Nano Investigation of Molecular Vibrations Norihiko Hayazawa, Yuika Saito . . . . . . . . . . . . . . . . . . .
257
16 17
18 19
XLII
20
Contents – Volume VI
Investigating Individual Carbon Nanotube/Polymer Interfaces with Scanning Probe Microscopy Asa H. Barber, H. Daniel Wagner, Sidney R. Cohen . . . . . . . . .
287
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
325
Contents – Volume VII
21 22 23 24 25
Lotus Effect: Roughness-Induced Superhydrophobicity Michael Nosonovsky, Bharat Bhushan . . . . . . . . . . . . . . . .
1
Gecko Feet: Natural Attachment Systems for Smart Adhesion Bharat Bhushan, Robert A. Sayer . . . . . . . . . . . . . . . . . .
41
Novel AFM Nanoprobes Horacio D. Espinosa, Nicolaie Moldovan, K.-H. Kim . . . . . . . .
77
Nanoelectromechanical Systems – Experiments and Modeling Horacio D. Espinosa, Changhong Ke . . . . . . . . . . . . . . . .
135
Application of Atom-resolved Scanning Tunneling Microscopy in Catalysis Research Jeppe Vang Lauritsen, Ronny T. Vang, Flemming Besenbacher . . .
197
26
Nanostructuration and Nanoimaging of Biomolecules for Biosensors Claude Martelet, Nicole Jaffrezic-Renault, Yanxia Hou, Abdelhamid Errachid, François Bessueille . . . . . . . . . . . . . . 225
27
Applications of Scanning Electrochemical Microscopy (SECM) Gunther Wittstock, Malte Burchardt, Sascha E. Pust . . . . . . . .
259
Nanomechanical Characterization of Structural and Pressure-Sensitive Adhesives Martin Munz, Heinz Sturm . . . . . . . . . . . . . . . . . . . . . .
301
Development of MOEMS Devices and Their Reliability Issues Bharat Bhushan, Huiwen Liu . . . . . . . . . . . . . . . . . . . . .
349
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
367
28
29
List of Contributors – Volume IX
Rehana Afrin Department of Life Science, Tokyo Institute of Technology Nagatsuta, Midori-ku, Yokohama 226-8501, Japan e-mail:
[email protected] David Alsteens Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected] Matthias Amrein Department of Cell Biology and Anatomy, Faculty of Medicine University of Calgary, Canada e-mail:
[email protected] Guillaume Andre Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected] Hideo Arakawa Bio Nanomaterials Group, National Institute for Materials Science 1-1, Namiki, Tsukuba 305-0044, Japan e-mail :
[email protected] Bharat Bhushan Nanotribology Lab for Information Storage and MEMS/NEMS (NLIM) The Ohio State University 201 W. 19th Avenue, W390 Scott Laboratory, Columbus, Ohio 43210-1142, USA e-mail:
[email protected] Claudio Canale Physics Department, University of Genoa Via Dodecaneso 33, I-16146 Genoa, Italy e-mail:
[email protected]
XLVI
List of Contributors – Volume IX
Ornella Cavalleri Physics Department, University of Genoa Via Dodecaneso 33, I-16146 Genoa, Italy e-mail:
[email protected] Cinzia Cepek Laboratorio Nazionale TASC-INFM Strada Statale 14, km 163.5 Basovizza, I-34012 Trieste, Italy e-mail:
[email protected] David T. Cramb Department of Chemistry, University of Calgary, Canada e-mail:
[email protected] Etienne Dague Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected] James A. DeRose EMPA Ueberlandstrasse 129, CH-8600 Duebendorf, Switzerland e-mail:
[email protected] Y. F. Dufrene Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected] Eric Finot Institut Carnot de Bourgogne, UMR 5209 CNRS Université de Bourgogne, France e-mail:
[email protected] Luca Gavioli Dipartimento di Matematica e Fisica, Università Cattolica del Sacro Cuore Via dei Musei 41, I-25121 Brescia, Italy e-mail:
[email protected] Alexander Gigler Sektion Kristallographie, Ludwig-Maximilians-Universität München Theresienstraße 41/II, 80333 München, Germany e-mail:
[email protected] Yann Gilbert Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected]
List of Contributors – Volume IX
XLVII
Alessandra Gliozzi Physics Department, University of Genoa Via Dodecaneso 33, I-16146 Genoa, Italy e-mail:
[email protected] Xavier Haulot Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected] Enamul Hoque Department of Chemistry and Materials Science, Washington State University Pullman, WA 99164-4630, USA e-mail:
[email protected] Atsushi Ikai Department of Life Science, Tokyo Institute of Technology Nagatsuta, Midori-ku, Yokohama 226-8501, Japan e-mail:
[email protected] K. S. Kanaga Karuppiah Department of Mechanical Engineering, 2025 H. M. Black Engineering Building Iowa State University, Ames, IA 50011-2161, USA e-mail:
[email protected] Hyonchol Kim Division of Biosystems, Institute of Biomaterials and Bioengineering Tokyo Medical and Dental University 2-3-10, Kanda-surugadai, Chiyoda-ku, Tokyo, 101-0062, Japan e-mail:
[email protected] Sung-Kyoung Kim Department of Chemistry, Hanyang University 17 Haengdangdong, Seoul 133-791, Korea e-mail:
[email protected] Tiziana Svaldo-Lanero Physics Department, University of Genoa Via Dodecaneso 33, I-16146 Genoa, Italy e-mail:
[email protected] Haiwon Lee Department of Chemistry, Hanyang University 17 Haengdangdong, Seoul 133-791, Korea e-mail:
[email protected]
XLVIII
List of Contributors – Volume IX
Zoya Leonenko Department of Cell Biology and Anatomy, Faculty of Medicine University of Calgary, Canada e-mail:
[email protected] Othmar Marti Experimental Physics, Ulm University 89069 Ulm, Germany e-mail:
[email protected] Hans Jörg Mathieu EPFL – IMX, Station 12, 1015 Lausanne, Switzerland e-mail:
[email protected] Vincent T. Moy Department of Physiology & Biophysics, University of Miami 1600 NW 12th Ave., RMSB 5077, Miami, FL 33136, USA e-mail:
[email protected] H. Oberleithner University of Münster, Institute of Physiology II Robert-Koch-Straße 27b, 48149 Münster, Germany e-mail:
[email protected] Toshiya Osada Department of Life Science, Tokyo Institute of Technology Nagatsuta, Midori-ku, Yokohama 226-8501, Japan e-mail:
[email protected] Bruno Pignataro Dipartimento di Chimica Fisica “F. Accascina”, Università di Palermo V.le delle Scienze – Parco d’Orleans II, 90128 Palermo, Italy e-mail:
[email protected] Annalisa Relini Physics Department, University of Genoa Via Dodecaneso 33, I-16146 Genoa, Italy e-mail:
[email protected] Félix Rico Department of Physiology & Biophysics, University of Miami 1600 NW 12th Ave, RMSB 5077, Miami, FL 33136, USA e-mail:
[email protected] C. Riethmüller Institute of Physiology II, University of Münster Robert-Koch-Straße 27b, 48149 Münster, Germany e-mail:
[email protected]
List of Contributors – Volume IX
XLIX
Ranieri Rolandi Physics Department, University of Genoa Via Dodecaneso 33, I-16146 Genoa, Italy e-mail:
[email protected] Toshiharu Saiki Department Electronics and Electrical Engineering, Keio University 3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan e-mail:
[email protected] Hiroshi Sekiguchi Department of Life Science, Tokyo Institute of Technology Nagatsuta, Midori-ku, Yokohama 226-8501, Japan e-mail:
[email protected] Sriram Sundararajan Department of Mechanical Engineering, 2025 H. M. Black Engineering Building Iowa State University, Ames IA 50011-2161, USA e-mail:
[email protected] Nikhil S. Tambe Material Systems Technologies, GE Global Research, #122, EPIP, Phase 2 Hoodi Village, Whitefield Road, Bangalore 560066, India e-mail:
[email protected] Hironori Uehara Department of Life Science, Tokyo Institute of Technology Nagatsuta, Midori-ku, Yokohama 226-8501, Japan e-mail:
[email protected] Claire Verbelen Unité de Chimie des Interfaces, Université catholique de Louvain Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium e-mail:
[email protected] Ewa P. Wojocikiewicz Department of Physiology & Biophysics, University of Miami 1600 NW 12th Ave, RMSB 5077, Miami, FL 33136, USA e-mail:
[email protected]
List of Contributors – Volume VIII
Jean-Pierre Aimé CPMOH, Université Bordeaux 1 351 Cours de la Libération, 33405 Talence cedex, France e-mail:
[email protected] Matteo Albani Dipartimento di Ingegneria dell’Informazione, Università di Siena Via Roma 56, 53100 Siena, Italy e-mail:
[email protected] Maria Allegrini Dipartimento di Fisica “Enrico Fermi”, Università di Pisa Largo Bruno Pontecorvo 3, 56127 Pisa, Italy e-mail:
[email protected] Guillaume Bachelier Laboratoire de Spectrométrie Ionique et Moléculaire (LASIM) Université Claude Bernard – Lyon 1, Bât. A. Kastler 43 Boulevard du 11 Novembre 1918, 69622 Villeurbanne Cedex, France e-mail:
[email protected] Charlotte Bernard CPMOH, Université Bordeaux 1 351 Cours de la Libération, 33405 Talence cedex, France e-mail:
[email protected] Bharat Bhushan Nanotribology Lab for Information Storage and MEMS/NEMS (NLIM) The Ohio State University 201 W. 19th Avenue, W390 Scott Laboratory, Columbus, Ohio 43210-1142 USA e-mail:
[email protected] Amir Boag School of Electrical Engineering, Faculty of Engineering, Tel Aviv University Ramat-Aviv, 69978, Israel e-mail:
[email protected]
LII
List of Contributors – Volume VIII
Rodolphe Boisgard CPMOH, Université Bordeaux 1 351 Cours de la Libération, 33405 Talence cedex, France e-mail:
[email protected] Anne Marie Bonnot LEPES 25 av. des Martyrs, 38042 Grenoble cedex, France e-mail:
[email protected] P. Carl Institute of Physiology II, University of Münster Robert-Koch-Straße 27b, 48149 Münster, Germany e-mail:
[email protected] Ignacio Casuso Laboratory of Nanobioengineering, Barcelona Science Park and Department of Electronics, University of Barcelona C/ Marti i Franque 1, 08028-Barcelona, Spain e-mail:
[email protected] Eugenio Cefalì Dipartimento di Fisica della Materia e Tecnologie Fisiche Avanzate Università di Messina Salita Sperone n. 31, 98166 Messina, Italy e-mail:
[email protected] Charles Clifford Analytical Science Team, Quality of Life Division, National Physical Laboratory Teddington, Middlesex TW11 0LW, UK e-mail:
[email protected] Peter J. Cumpson Analytical Science Team, Quality of Life Division, National Physical Laboratory Teddington, Middlesex TW11 0LW, UK e-mail:
[email protected] Gregory A. Dahlen Research Scientist 112 Robin Hill Road, Santa Barbara, CA 93117, USA e-mail:
[email protected] Dirk Dietzel Physikalisches Institut, Universität Münster Wilhelm-Klemm-Straße 10, 48149 Münster, Germany e-mail:
[email protected]
List of Contributors – Volume VIII
Robert Eibl Plainburgstraße 8, 83457 Bayerisch Gmain, Germany e-mail:
[email protected] Horacio D. Espinosa Department of Mechanical Engineering, Northwestern University 2145 Sheridan Rd., Evanston, IL 60208-3111, USA e-mail:
[email protected] Giorgio Ferrari Dipartimento di Elettronica e Informazione, Politecnico di Milano P.za L. da Vinci 32, 20133 Milano, Italy e-mail:
[email protected] Vinzenz Friedli EMPA – Materials Science and Technology Feuerwerker Strasse 39, CH-3602 Thun, Switzerland e-mail:
[email protected] Takeshi Fukuma Department of Physics, Kanazawa University Kakuma-machi, Kanazawa, 920-1192, Japan email:
[email protected] Laura Fumagalli Laboratory of Nanobioengineering Barcelona Science Park and Department of Electronics, University of Barcelona C/ Marti i Franque 1, 08028-Barcelona, Spain e-mail:
[email protected] Renato Gardelli Dipartimento di Fisica della Materia e Tecnologie Fisiche Avanzate Università di Messina Salita Sperone n. 31, 98166 Messina, Italy e-mail:
[email protected] G. Gomila Laboratory of Nanobioengineering, Barcelona Science Park and Department of Electronics, University of Barcelona C/ Marti i Franque 1, 08028-Barcelona, Spain e-mail:
[email protected] Pietro Giuseppe Gucciardi CNR – Istituto per i Processi Chimico-Fisici, Sezione Messina Via La Farina 237, 98123 Messina, Italy e-mail:
[email protected]
LIII
LIV
List of Contributors – Volume VIII
Andrea Ho Department of Mechanical Engineering, Northwestern University 2145 Sheridan Road, Evanston, IL 60208-3111, USA Samuel Hoffmann EMPA – Materials Science and Technology Feuerwerker Strasse 39, CH-3602 Thun, Switzerland e-mail:
[email protected] Sumio Hosaka Department of Production Science and Technology, Gunma University 1-5-1 Tenjin-cho, Kiryu, Gunma, 376-8515, Japan Suzanne P. Jarvis Conway Institute of Biomolecular and Biomedical Research University College Dublin, Belfield, Dublin 4, Ireland e-mail:
[email protected] James Johnstone DEPC, National Physical Laboratory Hampton Road, Teddington TW11 0LW, UK e-mail:
[email protected] P.M. Koenraad Semiconductor Physics Group Department of Applied Physics, Eindhoven University of Technology P.O. Box 513, 5600 MB Eindhoven, The Netherlands e-mail:
[email protected] Stefan Lanyi Institute of Physics, Slovakian Academy o Sciences Dubravska cesta 9, SK-845 11 Bratislava, Slovakia e-mail:
[email protected] Eli Lepkifker School of Electrical Engineering, Faculty of Engineering, Tel Aviv University Ramat-Aviv, 69978, Israel e-mail:
[email protected] Hao-Chih Liu SPM Probe Scientist 112 Robin Hill Road, Santa Barbara, CA 93117, USA e-mail:
[email protected] Sophie Marsaudon CPMOH, Université Bordeaux 1 351 Cours de la Libération, 33405 Talence cedex, France e-mail:
[email protected]
List of Contributors – Volume VIII
LV
Johann Michler EMPA – Materials Science and Technology Feuerwerker Strasse 39, CH-3602 Thun, Switzerland e-mail:
[email protected] Martin Munz Analytical Science Team, Quality of Life Division, National Physical Laboratory Teddington, Middlesex TW11 0LW, UK e-mail:
[email protected] Cattien V. Nguyen Research Scientist, Eloret Corp./NASA Ames Research Center MS 229-1, Moffett Field, CA 94035-1000, USA e-mail:
[email protected] Jason R. Osborne Eng. Manager, Advanced Development Group 112 Robin Hill Road, Santa Barbara, CA 93117, USA e-mail:
[email protected] Salvatore Patanè Dipartimento di Fisica della Materia e Tecnologie Fisiche Avanzate Università di Messina Salita Sperone n. 31, 98166 Messina, Italy e-mail:
[email protected] Jose Portoles Laboratory of Biophysics and Surface Analysis, University of Nottingham Nottingham, NG72RD, UK e-mail:
[email protected] Yossi Rosenwaks School of Electrical Engineering, Faculty of Engineering, Tel Aviv University Ramat-Aviv, 69978, Israel e-mail:
[email protected] John E. Sader Department of Mathematics and Statistics, The University of Melbourne Victoria 3010, Australia e-mail:
[email protected] Shimon Saraf School of Electrical Engineering, Faculty of Engineering, Tel Aviv University Ramat-Aviv, 69978, Israel e-mail:
[email protected]
LVI
List of Contributors – Volume VIII
Alex Schwarzman School of Electrical Engineering, Faculty of Engineering, Tel Aviv University Ramat-Aviv, 69978, Israel Hayato Sone Department of Production Science and Technology, Gunma University 1-5-1 Tenjin-cho, Kiryu, Gunma, 376-8515, Japan e-mail:
[email protected] Salvatore Spadaro Dipartimento di Fisica della Materia e Tecnologie Fisiche Avanzate Università di Messina Salita Sperone n. 31, 98166 Messina, Italy e-mail:
[email protected] Stephan J. Stranick National Institute of Standards and Technology Surface and Microanalysis Science Division (837) 100 Bureau Drive Gaithersburg, MD 20899-8372, USA e-mail:
[email protected] Oren Tal School of Electrical Engineering, Faculty of Engineering, Tel Aviv University Ramat-Aviv, 69978, Israel e-mail:
[email protected] Ivo Utke EMPA – Materials Science and Technology Feuerwerker Strasse 39, CH-3602 Thun, Switzerland e-mail:
[email protected]
List of Contributors – Volume X
Alexander Alexeev NT-MDT Co., Post Box 158, Building 317, Zelenograd, Moscow 124482, Russia e-mail:
[email protected] Tianming Bao Veeco Instruments Inc, 14990 Blakehill Dr, Frisco, TX 75035, USA e-mail:
[email protected] Bharat Bhushan Nanotribology Lab for Information Storage and MEMS/NEMS (NLIM) The Ohio State University 201 W. 19th Avenue, W390 Scott Laboratory, Columbus, Ohio 43210-1142, USA e-mail:
[email protected] Yasuo Cho Research Institute of Electrical Communication, Tohoku University 2-1-1 Katahira Aoba-ku Sendai, 980-8577, Japan e-mail:
[email protected] Gianluca Ciardelli Department of Mechanics, Politecnico di Torino Corso Duca degli Abruzzi 24, 10129 Torino, Italy e-mail:
[email protected] Mario D’Acunto Department of Chemical Engineering and Materials Science, University of Pisa via Diotisalvi 2, I-56126 Pisa, Italy e-mail:
[email protected] Patrick Fiorenza CNR-IMM, Stradale Primosole 50, I-95121 Catania, Italy e-mail:
[email protected] David Fong Veeco Instruments Inc, 112 Robin Hill Rd, Santa Barbara, CA 93117, USA e-mail:
[email protected]
LVIII
List of Contributors – Volume X
Filippo Giannazzo CNR-IMM, Stradale Primosole 50, I-95121 Catania, Italy e-mail:
[email protected] Paolo Giusti Department of Chemical Engineering and Materials Science, University of Pisa via Diotisalvi 2, I-56126 Pisa, Italy e-mail:
[email protected] Sean Hand Veeco Instruments Inc, 112 Robin Hill Rd, Santa Barbara, CA 93117, USA Yasushi Kadota Ricoh, 1-3-6 Nakamagame, Ohta-ku, Tokyo 143-8555, Japan e-mail:
[email protected] Toshi Kasai Cabot Microelectronics 870 N. Commons Drive, Aurora, IL 60504, USA e-mail:
[email protected] Kazushige Kawabata Division of Biological Sciences, Graduate School of Science, Hokkaido University North 10 West 8, Kita-ku, Sapporo 060-0810, Japan e-mail:
[email protected] Joachim Loos Department of Chemical Engineering and Chemistry Eindhoven University of Technology P.O. Box 513, 5600 MB Eindhoven, The Netherlands e-mail:
[email protected] Franco Montevecchi Department of Mechanics, Politecnico di Torino Corso Duca degli Abruzzi 24, I-10129 Torino, Italy e-mail:
[email protected] Vito Raineri CNR-IMM, Stradale Primosole 50, I-95121 Catania, Italy e-mail:
[email protected] Robert A. Sayer Nanotribology Laboratory for Information Storage and MEMS/NEMS (NLIM) The Ohio State University 201 West 19th Avenue, Columbus, OH 43210-1142, USA
List of Contributors – Volume X
LIX
Ken-ichi Shinohara School of Materials Science, Japan Advanced Institute of Science and Technology Asahi-dai, Nomi, Ishikawa 923-1292, Japan e-mail:
[email protected] Hiroyuki Sugimura Kyoto University, Sakyo, Kyoto 606-8501, Japan e-mail:
[email protected] Tatsuo Ushiki Division of Microscopic Anatomy Graduate School of Medical and Dental Sciences, Niigata University 757 Asahimachi-dori, Chuo-ku, Niigata 951-8510, Japan e-mail:
[email protected] Jiping Ye Nano Analysis Section, Research Department, Nissan ARC 1 Natsushima-cho, Yokosuka 237-0061, Japan e-mail:
[email protected]
13 Ultrathin Fullerene-Based Films via STM and STS Luca Gavioli · Cinzia Cepek
Abstract. This chapter overviews the major results obtained by scanning tunneling microscopy and spectroscopy on systems including the C60 molecule, starting from the free C60 molecule up to C60 in the solid state. The structural and electronic properties of fullerene films are described. Moreover, the behavior of fullerenes in low-dimensional films obtained by molecular deposition on clean metal and semiconductor substrates is reviewed, as well as the influence on the film properties by external doping. Key words: Scanning tunneling microscopy and spectroscopy, Fullerene-based systems, Electronic and structural properties, Alkali doping, Surfaces, Low-dimensional systems, C60 , Surface reconstruction, Surface–molecule interaction
Abbreviations fcc HOMO JT LDA LDOS LUMO SC STM STS
Face-centered cubic Highest occupied molecular orbital Jahn–Teller Local-density approximation Local density of states Lowest unoccupied molecular orbital Simple cubic Scanning tunneling microscopy Scanning tunneling spectroscopy
Dedication This work is dedicated to Prof. Massimo Sancrotti, a scientist who was able to couple scientific experimental research, encouraging young scientists, managing skills, and university teaching. He was a great person, an example of altruism and respect, who we had the opportunity to work with. He prematurely passed away on 22 April 2006.
13.1 Introduction In the last few years, the possibility of producing, controlling, and manipulating atomic and/or molecular blocks at the nanometer scale opened a new era, where
2
L. Gavioli · C. Cepek
novel artificially structured materials with tailored properties can be fabricated. C60 ˚ diameter, may be molecules, with their quasi-spherical shape of approximately 7 A considered as perfect nanoparticles and elemental blocks for nanomaterials. In this framework, the formation of chemically stable C60 monolayer films on many kinds of surfaces has become a subject of great interest for the scientific community [1,2]. Actually fullerene thin films are proving useful to stabilize fragile surface reconstruction and/or are functioning as templates for the formation of novel structures [1, 3–8]. This soccer-ball-shaped molecule differs significantly from the elemental or simple molecular adsorbates because of its three-dimensional character along with the strong modulation of the valence charge distribution over the molecular cage [9,10]. This unique electronic structure leads to several possible interaction mechanisms between C60 and the host surface, while the geometrical constraints imposed by the three-dimensional cage allow only a few carbon atoms per molecule to be in contact with the substrate [11, 12]. Moreover the electron band dispersion depends on the cage orientation [13–45] and the C60 molecules have the very important property that it is possible to change their electronic structure from semiconducting to metallic and even to the superconducting state just by doping [9, 10, 46]. Scanning tunneling microscopy (STM) and scanning tunneling spectroscopy (STS) are unique experimental techniques that allow us to study the spatial distribution of the electron charge all over the cage, and enable us to study individually the doping centers. This work is an overview of the major results obtained on systems including the C60 molecule, and it is divided into two sections. In the first section the basic principles of STM are quickly exposed. The second section gives a survey of the main properties of fullerene-based systems, starting from the free C60 molecule and extending to C60 in the solid state. Different subsections treat the electronic structures of these systems, while Sect. 13.3.4 deals with the main properties of the two-dimensional C60 -based system.
13.2 Basic Principles of STM and STS In this section we briefly review the principles of STM and STS, referring the reader to a complete theoretical treatment for details [47] and to other books for a wide range of STM applications [48]. The most widely used STM operation mode is constant current imaging, with a feedback loop keeping the tunneling current constant at a given bias voltage by adjusting the distance between the tip and the sample. The height (z) adjustment at each rastering point of the tip provides the “topography” z(x, y). However, not only topographic features contribute, but the electron density distribution affects the tunneling current [47, 49–54]. To understand z(x, y), one can start from the simplest case of an electron tunneling through a one-dimensional rectangular potential barrier with height V0 and width s [47]. In the limit of a strongly attenuating barrier the transmission coefficient T , which is the ratio of the transmitted to the incident current density T = jt / ji , is proportional to an exponential factor and is typical of tunneling, independent of the actual barrier shape [47].
13 Ultrathin Fullerene-Based Films via STM and STS
3
The three-dimensional case is treated perturbatively [55], considering the tunneling process between the sample surface and the metal tip. The tunneling current is dependent on the Fermi function, the bias voltage applied to the sample and the tunneling matrix element between the unperturbed electronic states of the sample surface and of the tip. The key part of the expression is the tunneling matrix element, depending explicitly on the sample and tip wave functions [55]. Since the atomic structure of the tip is usually not known, one needs to model the tip wave function to understand the information provided by the tunneling current. Assuming a tip with radius R and s-type only (quantum numbers l = 0 neglected) wave functions [49, 50], and small applied bias voltage, the current is 2 ψν r 2 δ (E ν − E F ) , I ∝ Un t (E F ) e2x R (13.1) 0 ν
where E F is the Fermi energy, r0 is the center of curvature of the tip, n t (E F ) is the density of states at the Fermi level for the tip, and the decay rate x = (2mϕ)1/2 / depends on the effective potential barrier height ϕ. The quantity 2 ψν r 2 δ (E ν − E F ) n s (E F , r0 ) = (13.2) 0 ν
is the charge density of the electronic states at the Fermi level [49, 50], i.e., the surface local density of states (LDOS) at E F , evaluated at the center of curvature of the tip. Owing to the exponential decay of the wave function of the sample in the z direction normal to the surface towards the vacuum, one finds 2 2 ψν r ∝ e−2x(s+R) . (13.3) 0
Therefore the tunneling current of Eq. (13.1) is exponentially dependent on the tip-to-sample distance s: I ∝ e−2xs .
(13.4)
The interpretation of the tunneling current as a contour map of the surface LDOS is limited to small applied bias (less than 0.5 V) and ignores the angular dependence of tip wave functions. Since most of the tips used for STM imaging are made of metals with a density of states at the Fermi level dominated by d states (tungsten or platinum-iridium), it becomes very important to consider the role of such states in the interpretation of STM images [53, 54]. Moreover, the s-like wave function approximation would not explain the atomic resolution experimentally obtained on many metal surfaces. Simulation of tip apex states by first-principles calculations has shown the presence of dz2 states near the Fermi level at the tip apex [56]. Therefore, one has to calculate the tunneling matrix elements considering the higher quantum number states. It is also a limit in the interpretation of STM constant current images as topographic representations of the surface features, since such l = 0 states act as filters for the tunneling current [53, 54], and one should take much care in the assignment of the atomically resolved structures, without a comparison with ab initio calculations of the electronic structure or comparison with other experimental techniques.
4
L. Gavioli · C. Cepek
At higher bias (volts), the tunneling current generally does not show ohmic behavior and the constant current images can critically depend on the applied voltage [51, 52]. A finite bias enters through the summation of electronic states contributing to the tunneling current, and can lead to a distortion of tip and sample wave functions, as well as a modification of the energy eigenvalues [57]. Hence, when a bias voltage U is applied to the sample, the energy levels shift upward (U negative) or downward (U positive) by |eU|, depending on the polarity of the bias. Therefore, one can probe either the occupied or the empty states of the sample surface, selecting the contributing electronic states by varying the amount of applied bias. Since electrons in states with the highest energy “feel” the smallest effective barrier height, most of the tunneling current arises from electrons near the Fermi level. This means that tunneling from the sample to the tip is more sensitive to the shape of tip’s empty states. Moreover, in the case of molecular overlayers, the interpretation of the STS spectra is complicated by localization and polarization effects, already observed in molecular solids [58], and by the variation of the electrostatic potential in the molecular conductor [59].
13.3 Survey of Fullerene-Based Systems 13.3.1 Bulk Properties More than a decade ago, Kratschmer et al. [60] isolated sizable quantities of pure stable C60 molecules, first discovered by Kroto et al. [61] in 1985, giving rise to a new field in solid-state physics. In the neutral state C60 is the highest-symmetry molecule among fullerenes, having the shape of a truncated icosahedron –like a soccer ball– with a diameter of 7.05 Å (Fig. 13.1). It consists, therefore, of 12 pentagons formed by single C–C bonds and 20 hexagons formed by alternate single and double C–C bonds. The hexagonal faces are not regular, reflecting the different bond lengths of single and double bonds. The
Fig. 13.1. Model structure of a C60 buckyball
13 Ultrathin Fullerene-Based Films via STM and STS
5
lowest unoccupied molecular orbital (LUMO) has t1u symmetry and is threefold degenerate. Symmetry-lowering, Jahn–Teller (JT) distortions and lifting of degeneracy are therefore expected when electrons populate the LUMO orbital [62, 63]. In the condensed phase at T > 260 K the C60 molecules form a face-centered cubic (fcc) plastic crystal, with a lattice constant of 14.2 Å, corresponding to about 10-Å separation between the centers of adjacent molecules. The LUMO and the highest occupied molecular orbital (HOMO) broaden into bands that are about 0.5–0.6 eVwide according to calculations [64] (Fig. 13.3). The HOMO-derived band is completely filled by ten electrons per molecule, while the LUMO-derived band is empty and capable of accepting up to six electrons per molecule. Estimates of the on-ball electron phonon coupling [62, 63, 65, 66] and experimental evidence [67–70] suggested that JT distortions (possibly dynamic) may be important also in the solid or at the solid surface. The molecules are held in the fcc lattice mainly by van der Waals interactions and are characterized by a very high degree of orientational disorder. The C60 molecules undergo rapid reorientations and for many purposes they can be considered as free rotors. The form factor obtained when X-rays, neutrons or electrons are scattered by C60 molecules at T > 260 K is readily approximated by spherical harmonics with l = 0, which correspond to the scattering by a sphere. The high symmetry of the truncated icosahedron and the fast molecular rotations allow one to consider in many cases the C60 molecule as a spherical shell. The fine structure of the truncated icosahedron is manifested therefore only when the rotations are suppressed. Below 260 K the C60 crystal adopts an orientationally ordered simple cubic (SC) structure, whose space group is determined to be Pa3, with four inequivalent molecules per unit cell [71–73]. In the SC phase the rotational degrees of freedom are hindered and the molecules can only perform thermally activated jumps between two equilibrium positions (merohedral disorder) and librations around these positions. The quenching of the rotational degrees of freedom is a first-order phase transition and is initiated at the surface. Figure 13.2 shows the orientations of the two cages in the low-temperature phase as seen by STM [74] on the (111) surface of a multilayer C60 film at 78 K. In the ordered (2 × 2) superlattice, each unit cell has one C60 molecule with a three-lobe intramolecular pattern and three C60 molecules with dumbbell-like intramolecular patterns. By comparison of the intramolecular patterns with theo-
Fig. 13.2. Empty-state topography scanning tunneling microscopy (STM) image on a C60 island surface. The (2 × 2) superlattice is clearly observed. The C60 molecules with three-lobe intramolecular patterns are indicated by the white triangle. (Adapted from [74])
6
L. Gavioli · C. Cepek
retical simulated results, it is found that the orientational configurations of the C60 molecules on the surface are similar to those in the bulk. This similarity indicates that it is the interaction of the neighboring C60 molecules that drives the molecules to an orientational ordered state [74]. 13.3.2 Electronic Structure The C60 molecule resembles a graphite layer wrapped around a sphere with 12 pentagonal “defects”, needed by geometrical constraints to close a hexagonal network on itself. The C60 curvature of the bonding network mixes the mostly sp2 hybridization characteristic of graphite with some sp3 bond character. Three of the four valence electrons of each carbon atom are shared in 90 bonds that are along the edges of the truncated icosahedron. The remaining valence electron lies in the states that are formed from radially oriented atomic orbitals. These orbitals are mainly p-type, with some (8%) mixing of s-orbital character. The bonds are delocalized (aromatic), although they are somewhat stronger in the hexagon-shared edges of the C60 cage. The high symmetry of the molecule allows the existence of singly, twofold, fourfold and fivefold degenerate electronic states (a, t, g and h symmetry, respectively). The simplest assumption is that in solid C60 these states broaden into electron bands and are possibly split by the crystal field. Figure 13.3 shows the electronic energy levels of the C60 molecule and the band structure of the fcc C60 crystal as calculated within the local-density approximation (LDA) [64]. The fcc C60 crystal is found to be a semiconductor with a LDA energy gap between the HOMO (hu ) derived bands and the LUMO (t1u ) derived bands of 1.5 eV. Because of the weak van der Waals interaction between the C60 molecules in the solid, the molecular character of the electronic states is retained when these orbitals broaden to form bands. Each molecular orbital of the isolated molecule forms a band with a bandwidth of 0.5–
Fig. 13.3. Electronic energy levels for isolated C60 (a) and for face-centered cubic bulk C60 (b), obtained within the local-density approximation. (Adapted from [64])
13 Ultrathin Fullerene-Based Films via STM and STS
7
0.8 eV for π states and 0.2 eV or less for σ states. That the valence electronic states have strong molecular character is experimentally confirmed by photoemission experiments and STS on isolated C60 molecules and on C60 films or single crystals, showing practically the same spectra in the three different cases [74–76]. A typical valence band photoemission spectrum (filled states) of solid C60 is shown in Fig. 13.4 together with the inverse photoemission spectrum (empty states) [75]. The gap measured comparing the photoemission (N − 1) and inverse photoemission (N + 1) spectra in Fig. 13.4 is 2–2.5 eV, some 0.5–1 eV larger than that obtained for the ground-state N-electron system by LDA calculations. This reflects the presence of strong electron correlation effects in solid C60 . That solid C60 is a strongly correlated system was confirmed by the experimentally estimated Hubbard U of 1.5–1 eV, about 2–2.5 times the HOMO or LUMO band width [75]. The same electronic structures are seen also in STS measurements, as shown, for example, in Fig. 13.5 [77]. This fact has important consequences for the nature and transport properties of doped C60 compounds, and also for the electronic excitations in undoped C60 . The first electronic excitation observed in electron energy loss spectroscopy corresponds to the hu (HOMO)–t1g (LUMO) Frenkel exciton (localized on-ball transition) at 1.55 eV. In addition to strong electron–electron interactions, C60 is also characterized by important electron–phonon interactions. A rich vibronic structure is visible on a side of sharp photoemission [67] and electron energy loss peaks [68, 69]. Moreover, the comparison between photoemission and X-ray absorption spectroscopy [78] data for C60 in the gas phase and in the fcc solid phase indicates that vibronic effects account well for the majority of the measured solid-state bandwidths. This further indicates that in such a molecular solid, where the solid-state effects are weak, on-ball electron localization and electron–phonon interactions are rather important.
Fig. 13.4. Photoemission and inverse photoemission spectra of bulk C60 . (Adapted from [75])
8
L. Gavioli · C. Cepek
Fig. 13.5. Spectrum of the differential conductivity dI/ dV of one C60 layer on Au(887). The tunneling barrier was set at −3 V and 0.1 nA. The full lines represent fits to the experimental data, depicted as dots. LUMO lowest unoccupied molecular orbital, HOMO highest occupied molecular orbital, ML mololayer. (Adapted from [77])
13.3.3 Alkali-Metal-Doped C60 Electron doping of the molecule can be obtained by intercalation of metal atoms (e.g. alkali metals) into the crystal structure of solid or single layers of C60 . The first study of conductivity versus alkali concentration in C60 compounds appeared to be consistent with a very simple one-electron rigid band filling interpretation [79]. The conductivity of Kx C60 increases monotonically with K doping, reaching its maximum at x = 3 (three electrons per molecule) and then decreasing with further uptake of alkali atoms up to x = 6 (six electrons per molecule). There are some problems with the conventional understanding of the electronic properties of alkali-metal-doped C60 compounds. The phase diagram of the alkali metal fullerides turned out to be quite rich, with unexpected metallic or insulating phases. Several experimental studies indicate that at saturation doping (x = 6) all the alkali metals form a semiconducting phase [4, 76, 80–83]. Stable metallic phases occur for K3 C60 [76,82], Rb3 C60 and for some mixed compounds An B3−n C60 (where A is K), Rb, Na and Cs, while B is K and Rb, n = 0, 1, 2) [82, 84]. While these observations were still in agreement with a rigid band filling of the LUMO band, the observed insulating nature of stable phases with two or four electrons per molecule, for example (Na, Cs, Li)2 C60 and (K, Rb, Na, Cs)4 C60 , clearly breaks this picture [1, 76, 82], indicating that correlation effects play an important role [70, 75, 82, 83, 85, 86]. This subject has been recently addressed for a Kx C60 monolayer on Au(111) for 3 < x < 4 [76], showing that the interplay between molecular electronic and lattice degrees of freedom (intramolecular and intermolecular atomic displacements)
13 Ultrathin Fullerene-Based Films via STM and STS
9
is particularly important in fullerenes, because the energy scale of their electron– lattice interaction is comparable to their electronic bandwidths [82, 87]. The triangular lattice of the local structure of the K-doped C60 monolayer reflects a K-induced surface reconstruction for x = 3, with pronounced disorder in some regions (see STM data in Fig. 13.6). At x = 4, the layer undergoes a dramatic restructuring, assuming a nearly rectangular structure with four molecules per unit cell [76]. The difference in electronic properties between the x = 3 and x = 4
Fig. 13.6. a STM image for the mixed-phase Kx C60 ML on Au(111), with x = 0.35 (V = 0.5 V, I = 10 pA, 25 × 25 nm2 ). b Spatially averaged dI/ dV spectrum for K3 C60 showing metallic behavior. c Spatially averaged dI/ dV spectrum for K4 C60 showing metallic behavior for the energy gap around E F , with spatially averaged dI/ dV spectra for a bare C60 ML and a K4 C60 ML over a wider energy range in the inset. d Filled-states STM topograph of K4 C60 (V = 0.10 V, I = 20 pA) and e empty-states STM topograph of K4 C60 (V = 0.10 V, I = 10 pA). Single molecules are marked by circles. (Adapted from [76])
10
L. Gavioli · C. Cepek
phases can be seen by comparing spatially averaged dI/ dV spectra (Fig. 13.6, panels b and c). In the first case, all of the spectra reveal a similar peak near E F , while data for the K4 C60 phase show a 200-mV energy gap. A corresponding major difference is observed in the topographic images associated with filled and empty states only for the K4 C60 phase (Fig. 13.6, panels d and e). The dramatic difference in the filled-state and empty-state topography is a manifestation of the JT distortion [76, 88]. When four extra electrons are added, a C60 molecule is expected to undergo a JT distortion whereby the three degenerate LUMO levels split into a group of twofold degenerate levels at lower energy and one nondegenerate level at higher energy [63, 89, 90]. The four extra electrons are accommodated by the two lower energy levels, making the distortion energetically favorable. The modification of the topography is therefore explained by the change in the molecular levels of the distorted C60 . 13.3.4 Interaction of C60 with Surfaces As described in the previous sections, the discovery of the C60 molecule has triggered massive research on the synthesis and evaluation of fullerene-based compounds, which can even exhibit superconducting character, as is the case for some alkali or rare earth intercalated bulk systems. The transition to superconductivity of these compounds has been established to depend crucially on the charge state of the molecules and on the presence of merohedral disorder. The critical temperature can be tailored by artificially changing the C60 -C60 spacing and in general moves to higher values when this distance is increased [1]. Bulk samples, however, are usually strongly affected by vacancies and chemical disorder that make the interpretation of the physical properties and the comparison between the results of different research groups not straightforward. The possibility of finding systems where doping and structure can be changed in a controlled manner, without the inconvenience occurring in bulk compounds [3– 5, 82], has focused the interest of many researchers on low-dimensional ordered C60 films [6, 91]. As a consequence, the formation of chemically stable C60 monolayer films on many kinds of surfaces has became a subject of great interest for the scientific community [9, 10, 76, 91–94]. A single ordered layer of C60 molecules on a single crystal surface and doped phases are quite well-defined systems in comparison with bulk C60 samples or thicker layers, with perhaps modified crystalline and electronic properties with respect to the corresponding bulk phases. Moreover, fullerene thin films are proving useful to encapsulate chemically reactive surface reconstructions and/or are functioning as templates for the formation of novel structures [9–12]. This soccer-ball-shaped molecule differs significantly from the elemental or simple molecular adsorbates because of its three-dimensional character along with the strong modulation of the valence charge distribution over the molecular cage [46, 76, 93, 95]. This unique electronic structure leads to several possible interaction mechanisms between C60 and the host surface, while the geometrical constraints imposed by the three-dimensional cage allow only a few carbon atoms per molecule to be in contact with the substrate (Table 13.1) [95–98].
Substrate
(a)
Graphite GeS(001) SiO2
(b)
Al(110) Al(111)
(c)
1
2
3
4
Type of bonding
Approximate desorption Temperature (K)
Mobility at room temperature
Equilibrium C60–C60 bond lengths (Å)
500
Mobile
10.01
490
Mobile
10.02
470
Mobile
Weak, predominantly van der Waals Intermediate, predominantly covalent
730 730
9.91, ~10.04
Mobile (islands)
10.07, 10.23i
Intermediate, predominantly ionic
Mobile (steps) Mobile (steps) Mobile (steps)
~10 10.04 m ~10 9.7–11.1
Mobile (steps)
10.1
Dissoc. at 1070 K
Immobile
9.87, 11.58
Dissoc. at 1100 K 970
Immobile Immobile
Dissoc. at 760 K
Reduced mobility
14.4 9.6, 11.52 10, 10.5, ~11.6
Dissoc. at 1050 K
Low mobility
10.0 ± 0.3
770 800 770 730
Cu(111) (d)
Si(100) Si(111) Ge(111) Ge(100) Ni(110) Pt(111)
Strong, predominantly covalent
11
Mobile (steps)
Ag(110) Ag(111) Au(110) Au(111) Cu(110)
9.91, 11.44, 12.13
13 Ultrathin Fullerene-Based Films via STM and STS
Table 13.1. Summary of previous results for C60 monolayer systems. The equilibrium bond lengths refer to the lowest-energy structure for coverage close to a monolayer. The desorption temperature represents the temperature for which the entire first layer is desorbed
12
L. Gavioli · C. Cepek
To enhance the binding, the C60 molecules might choose some optimal orientations, might undergo structural distortions of the cage and in some cases they might also induce structural instability and reconstruction of the host substrate [7, 8, 13, 34, 99–103] (Fig. 13.7). When the substrate reconstruction leads to large modifications of the substrate, STM is able to visualize topographic changes, as, for example, in the case of C60 adsorption on Au [101] (Fig. 13.7), or on Pd [104], while other experimental techniques, like surface X-ray diffraction, are needed, as in the case of C60 on Pt [8] (see the model in Fig. 13.8). As a consequence, a valuable property of these adsorbates is the molecular orientation of the cage with respect to the host substrate [13, 31, 34, 99, 100], which is intimately related to the interfacial interaction between the substrate and the C60 molecules. Fasel et al. [99, 100] have shown by means of photoelectron diffraction experiments that a variety of molecular orientations can be observed on several single crystalline surfaces, while it has been shown that different binding orientations are present at the same time for C60 adsorbed on Si(111)-(7×7) [13,31]. This distribution of molecular orientations, together with the above arguments, highlights the delicate balance between C60 –C60 and C60 –substrate interactions that drive the physics of fullerene adsorption on surfaces. In other words, a very important point in this research area is to understand the nature of the interaction between C60 and the substrate and how the molecules orient themselves onto the crystal surface. Monolayer and submonolayer films based on fullerene molecules deposited on top of metal or semiconductor surfaces, therefore, form a class of systems that could exhibit a wide range of physical behaviors. The properties of these low-dimensional systems depend, for example, on the nature of the chemical bond (ionic versus covalent versus van der Waals), on the partial filling of the LUMO as a result of a charge transfer from the substrate and/or from dopant atoms (influenced by the
Fig. 13.7. Example of C60 -induced reconstruction of the Au(110)-(2 × 1) clean surface structure into the (6 × 5) one. Left: Scanning tunneling microscope image. Right: Hard-sphere model, small circles are Au atoms, large circles C60 molecules. The dashed rectangle indicates the (6 × 5) unit cell. (Adapted from [101])
13 Ultrathin Fullerene-Based Films via STM and STS
13
Fig. 13.8. Adsorption structure of the C60 molecules at the Pt(111) surface as deduced by surface Xray diffraction measurements. The presence of C60 induces the formation of a regular lattice of vacancies on the Pt surface, with the molecules located at the top of the vacancies. Top: Details of the fullerene–Pt bonds. The hexagon orientation of the molecule maximizes the numbers of C–Pt bonds with a length of 0.2 nm, comparable with the sum of the covalent radii. Bottom: Side view. The electron charge density radius of C60 is indicated by the dashed circles: although the C atoms are located above the Pt topmost layer, the electron charge density penetrates inside the first layer, thus supporting the view of a strong interaction between the C60 base atoms and the Pt top layer. (Adapted from [8])
substrate work function, the fullerene–substrate interaction and possible screening effects) [33,105–107], on the geometry of the fullerene film (which could depend on the substrate surface chosen and on the thermal history of the film) [35, 108–110], and, as already mentioned, on possible adlayer-induced substrate reconstruction. An increasing number of works have aimed at studying the interaction of single layers of C60 with metal surfaces. Owing to its large work function (7.6 eV) and electron affinity (2.7 eV) [9, 10], C60 has the tendency to accept electrons from many metallic surface; however, charge transfer from the metallic substrate to C60 is not always observed and the character of the interaction depends mainly on the metal, almost independently of its work function [6, 96]. According to Maxwell et al. [96] the bonding interaction of C60 molecules in direct contact with a metal surface can be divided into three categories: strong predominantly covalent bonding, intermediate predominantly covalent bonding and intermediate predominantly ionic bonding (Table 13.1). Strong predominantly covalent bonding has been observed for a single layer of C60 on Ni(110) [38, 111], Pt(111) [111, 112], Rh(111) [113, 114] and Ta(110) [115, 116],1 where the strength of the interaction is able to catalyze the fullerene cage decomposition prior to the desorption of the monolayer. The C60 substrate interaction is also related to the adsorption-induced substrate reconstruction observed for C60 on Ni(110) [38] and for C60 on Pd(110) [104]. In the latter case, different short-range ordered phases are observed as a function of C60 coverage up to one monolayer. Thermal treatment greatly increases the order, and the buckyball desorbs before decomposition at about 1053 K. The long-range ordered triple-stripe phase observed after annealing at 1003 K is presented in Fig. 13.9. This structure consists of two dark rows in-between bright ones running along the [110] direction, resulting in a single domain. Neighboring dark and bright rows 1
Rudolf and Gensterblum [116] argue that [115] does not contain enough experimental evidence to sustain the fragmentation of C60 deposited onto Ta(110) at 300 K
14
L. Gavioli · C. Cepek
Fig. 13.9. STM image of the triple-stripe phase (multilayer C60 annealed to 1000 K) consisting of alternating bright and two dark rows oriented along the [110] direction. Bottom: Suggested real-space model in which the bright molecular rows seen in STM correspond to C60 molecules adsorbed in a vacancy formed by Pd atoms released out of the first layer. The dark molecular rows correspond to the C60 molecules adsorbed in a vacancy formed by Pd atoms released out of the first two layers. (Adapted from [104])
are shifted by half a molecular width. The height difference between dark and bright rows (independent of the tunneling parameters) is approximately the height of a monatomic step of the Pd(110), and is thus interpreted as topographic difference. The X-ray photoelectron diffraction analysis shows that C60 molecules are oriented in the same way with a 5–6 bond facing towards the substrate and attributes the height difference to a thermally activated bonding transition, where the C60 molecules are accommodated in microscopic pits, resulting in a higher C–Pd coordination. It is hence concluded that the different heights of the bright and the dark rows are due to a surface reconstruction [104]. Predominantly covalent bonding with an intermediate strength has been observed on Al(110) and Al(111) [96] where the fullerene monolayer desorbs at about 730 K. Interaction of intermediate strength with bonds of predominantly ionic character has been observed for the C60 monolayer on Au(110) [32, 33], Au(111) [34– 37], Cu(110) [38], Cu(111) [39, 40], Ag(110) [41, 42], Ag(111) [34, 35, 43, 44], W(100) [45] and polycrystalline Ag [117–119]. These predominantly ionic bondings are characterized by charge transfer from the metal to the LUMO of the C60 molecule. On the C60 /Au(111) interface, STS data indicate that the molecular orbitals are not strongly perturbed by the interaction with the substrate and reported a HOMO– LUMO gap of 2.3 eV [36]. However, the STS maps of the HOMO and LUMO spatial charge distribution (Fig. 13.10) and a comparison of peak width and energy for different molecules suggest that for the brighter molecules a lower amount of electronic charge is transferred from the substrate and they are identified as adsorbed on top sites. Moreover, the narrower shape of the LUMO peaks of these brighter molecules indicates that their interaction with the metal surface is weaker, i.e., their electronic structure is less perturbed by the substrate on these particular sites. Consequently, their adsorption energy is expected to be also smaller than that of molecules adsorbed on bridge sites [36]. However, the slightly different interaction with the substrate does not result in a preferred molecular orientation on this surface. This is explained as a consequence of the predominantly ionic character of the C60 –Au(111) interaction.
13 Ultrathin Fullerene-Based Films via STM and STS
15
Fig. 13.10. a Topography image on a C60 island aligned parallel to the main directions of the substrate. b A comparison of the normalized conductance spectra measured on the two types of molecules forming the periodic pattern shown in c. c Image of the normalized differential conductance at Vs = 0.6 V. The ordered pattern reveals the (2 × 2) periodic adlayer structure, not visible in a. (Adapted from [36])
In the C60 /Cu(111) system the adsorbate–substrate bonding is supposed to be stronger than in the C60 /Au(111) system [121]. In fact preferred orientations of the molecules are observed on different C60 phases [120]. However, the morphology and molecular orientation are driven by the balance between competing forces: C60 and Cu lattice matching, C60 -induced reconstruction of substrate, and the C60 –C60 interaction. The various metastable phases observed as a function of substrate deposition temperature [120] are explained by taking into account the three competing forces. In particular, the substrate reconstruction taking place for the different phases is the tuning factor determining the relative strength of the adsorbate–adsorbate and adsorbate–substrate interactions. The charge transfer mechanism responsible for the origin of the predominantly ionic bonding has been also observed for C60 monolayers deposited on polycrystalline Au [117] and Cu [117] samples, but the stability of the C60 monolayer and the strength of the interaction with these surfaces was not investigated in detail. A number of works concentrated on the Ag(100) surface [93, 122–124] where a particular kind of bonding interaction that falls outside the above classification, namely, strongly predominantly ionic, has been observed, where there is a charge transfer from Ag to C60 and the strength of the interaction is able to catalyze the C60 decomposition at 860 K prior to the complete desorption of the adsorbed molecular film [119]. On this interface, bright and dim molecules are clearly evidenced by STM, as reported in Fig. 13.11a, while the corresponding STS spectra are shown in Fig. 13.11b. The spectroscopic behavior of dim molecules is qualitatively very similar to that of bright molecules, except for a slight shift in peak locations and a broadening of the resonance widths. By comparing the C60 monolayer with individual C60 molecules, one observes ˚ irrespective of that the isolated C60 monomers have an apparent height of 5.7 A, orientation [93]. This implies that differences in electronic structure due to C60 –Ag orientation are an unlikely cause for apparent height differences between bright and dim molecules in the monolayer. Moreover, the fact that bright monolayer,
16
L. Gavioli · C. Cepek
Fig. 13.11. a A 20 nm × 11 nm topographic image of a C60 /Ag(001) monolayer (T = 7 K, V = +2.0 V, I = 1 nA). b Spatially averaged spectra for bright and dim C60 molecules in the C60 /Ag(100) monolayer, as well as for an isolated C60 molecule on the Ag(001) surface. Inset: Low-bias spectrum of a C60 /Ag(001) monolayer, showing no features near E F at T = 7 K. (Adapted from [122])
dim monolayer and isolated monomer spectra are all nearly identical implies that there is no significant difference in C60 –Ag bonding character between the three molecule types and hence that differences in the C60 –Ag chemical bond are unlikely to be the source of bright/dim contrast. This points towards the explanation of the contrast from C60 -induced modification of the underlying Ag surface. The interaction of C60 with elemental semiconductor surfaces [13–31], in particular with Si, has been the subject of intense experimental work because of the potential industrial applications of these systems. The interaction of C60 with Si, after annealing above 850 K, is predominantly covalent and strong [13–21, 31]: the substrate–adsorbate interactions tend to dominate over the weak C60 –C60 interactions and the molecules are immobile at room temperature, forming structures with no long-range order. Moreover, STM data show the reduction of the molecular icosahedral symmetry after interaction with the surface (Fig. 13.12), though the relevant electronic structure remains fairly unchanged upon adsorption [31]. Many different adsorption sites are observed on Si(111)-(7×7), and the images reveal that fullerenes at each site have multiple orientations and, consequently, multiple different bonding configurations with the surrounding silicon adatoms. Moreover, the fullerene molecules do not desorb from the low-index Si surfaces even at 1000 K, while above this temperature they fragment to form SiC films. In the case of the Ge(100) and Ge(111) surfaces, the interface formation with C60 molecules has been largely studied via microscopic [22–28] and diffraction
13 Ultrathin Fullerene-Based Films via STM and STS
17
Fig. 13.12. Comparison of experimental STM images of individual molecules (a–c) with calculated shapes of C60 molecular orbitals (d–f). The applied voltages and tunnel currents are −1.5 V, 0.2 nA (a), 2 V, 1 nA (b) and −2 V, 1 nA (d), respectively. The calculations are based on density-functional theory, in the local-density approximation. In all cases, the isosurfaces of the squared HOMO orbital are plotted. (Adapted from [31])
techniques that provide much information on the growth mode, morphology, and stability of the C60 film, but unsatisfactorily address the nature of the C60 –substrate interaction. However, an apparently strong bond mediated by charge transfer (i.e., prevalently ionic) from the Ge to the C60 molecules has been proposed for the C60 /Ge(111) interface. This conclusion has been inferred mainly on the basis of the observed structural flexibility of the one-monolayer C60 /Ge(111) system, which exhibits a number of surface orderings temperature. √ √as a function√of the annealing √ Two stable surface phases, 3 3 × 3 3R30◦ and 13 × 13R14◦ , form upon annealing at 723–773 K and above 773 K, respectively, √ √ and a localized metastable 5×5 phase forms by suddenly quenching the 3 3×3 3R30◦ phase heated at 773 K. On the basis of their low energy electron diffraction patterns and STM data, Xu et al. [24] concluded that the formation of these three phases requires reconstructions of the virgin Ge(111)-c(2 × 8) surface structure.
13.4 Summary This short review on C60 -based systems provides just a little taste of the variety of interfaces and structures that can be constructed with this molecule. The subjects that can be studied on this playground range from fundamental physics, such as molecule–substrate interactions, electron correlation effects, phase transitions and superconductivity, to more applicative fields like nanotechnology. We hope that this will stimulat interest to pursue and deepen understanding of fullerene-based structures, which might have a profound impact on various fields of science and technology.
18
L. Gavioli · C. Cepek
References 1. Yildrim T (1998) Electrochem Soc Proc 98:360 2. Rafii-Tabar H, Ghafoori-Tabrizi K (2001) Prog Surf Sci 67:217 3. Merkel M, Knupfer M, Golden MS, Fink J, Seemann R, Johnson (1993) Phys Rev B 47:11470 4. Benning PJ, Stepniak F, Poirier DM, Martins JL, Weaver JH, Chiabante LFP, Smalley RE (1993) Phys Rev B 47:13843 5. Morikawa T, Takahashi T (1993) Solid State Commun 87:1017 6. Rudolf P (1996) Fullerenes and fullerene nanostructures. World Scientific, Singapore 7. Pedio M, Felici R, Torrelles X, Rudolf P, Capozi M, Rius J, Ferrer S (2000) Phys Rev Lett 85:1040 8. Felici R, Pedio M, Borgatti F, Iannotta S, Capozi M, Ciullo G, Stierle A (2005) Nat Mater 4:688 9. Dresselhaus MS, Dresselhaus G, Eklund PC (1996) Science of fullerenes and carbon nanotubes. Academic, San Diego 10. Klyachko D, Chen DM (1997) J Vac Sci Technol B 15:1295 11. Hong H, McMahon WE, Zschack P, Lin DS, Aburano RD, Chen H, Chiang TC (1992) Appl Phys Lett 61:3127 12. Ma YR, Moriarty P, Upward MD, Beton PH (1998) Surf Sci 397:421 13. Hou JG, Jinlong Y, Haiqian W, Qunxiang L, Changgan Z, Hai L, Wang B, Chen DM, Qingshi Z (1999) Phys Rev Lett 83:3001 14. Hamza AV, Balooch M (1993) Chem Phys Lett 201:404 15. Balooch M, Hamza AV (1993)Appl Phys Lett 63:150 16. Moriarty P, Upward D, Dunn AV, Ma YR, Beton PH, Teehan D (1998) Phys Rev B 57:362 17. Moriarty P, Ma YR, Upward MD, Beton PH (1998) Surf Sci 407:27 18. Sakamoto K, Kondo D, Ushimi Y, Harada M, Kimura A, Kakizaki A, Suto S (1999) Phys Rev B 602579 19. Suto S, Sakamoto K, Kondo D, Wakita T, Kimura A, Kakizaki A, Hu CW, Kasuya A (1999) Surf Sci 438:242 20. Chen D, Sarid D (1994) Phys Rev B 49:7612 21. Chen D, Sarid D (1995) Surf Sci 329:206 22. Klyachko D, Chen DM (1996) J Vac Sci Technol B 14:974 23. Dunphy JC, Klyachko D, Xu H, Chen DM (1997) Surf Sci 383:L760 24. Xu H, Chen DM, Creager WN (1994) Phys Rev B 50:8454 25. Wirth KR, Zegenhagen J (1996) Surf Sci 351:13 26. Wirth KR, Zegenhagen J (1997) Phys Rev B 56:9864 27. Klyachko D, Lopez-Castillo JM, Jay-Gerin JP, Chen DM (1999) Phys Rev B 60:9026 28. Kidd T, Aburano RD, Hong H, Gog T, Chiang TC (1998) Surf Sci 397:185 29. Goldoni A, Cepek C, De Seta M, Avila J, Asensio MC, Sancrotti M (2000) Surf Sci 454:514 30. Goldoni A, Cepek C, De Seta M, Avila J, Asensio MC, Sancrotti M (2000) Phys Rev B 61:10411 31. Pascual JI, Gomez-Herrero J, Rogero C, Baro AM, Sanchez-Portal D, Artacho E, Ordejon P, Soler JM (2000) Chem Phys Lett 321:78 32. Modesti S, Cerasari S, Rudolf P (1993) Phys Rev Lett 71:2469 33. Maxwell AJ, Brühwiler PA, Nilsson A, Mårtensson N, Rudolf P (1994) Phys Rev B 49:10717 34. Altman EI, Colton RJ (1993) Phys Rev B 48:18244 35. Altman EI, Colton RJ (1993) Surf Sci 295:13 36. Rogero C, Pascual JI, Gomez-Herrero J, Barò AM (2002) J Chem Phys 116:832
13 Ultrathin Fullerene-Based Films via STM and STS
19
37. Lu X, Grobis M, Khoo KH, Louie SG, Crommie MF (2004) Phys Rev B 70:115418 38. Murray PW, Pedersen MØ, Lægsgaard E, Stensgaard I, Besenbacher F (1997) Phys Rev B 55:9360 39. Hashizume T, Sakurai T (1994) J Vac Sci Thechnol B 121992 40. Tsuei KD, Yuh JY, Tzeng CT, Chu RY, Chung SC, Tsang KL (1997) Phys Rev B 56:15412 41. Purdie D, Bernhoff H, Reihl B (1996) Surf Sci 364:279 42. Magnano E, Vandre S, Cepek C, Goldoni A, Laine AD, Curro GM, Santaniello A, Sancrotti M (1997) Surf Sci 377-379:1066 43. Tjeng LH, Hesper R, Heessels ACL, Heeres A, Jonkmann HT, Sawatzky GA (1997) Solid State Commun 103:31 44. Wertheim GK, Buchanan DNE (1994) Phys Rev B 50:11070 45. Wertheim GK, Buchanan DNE (1993) Solid State Commun 88:97 46. Weaver J, Poirier DM (1994) Solid state physics: advances in research and applications, vol 48. Academic, Boston 47. Wiesendanger R (1994) Scanning probe microscopy and spectroscopy. Cambridge University Press, Cambridge. 48. Bhushan B, Fuchs H, Hosaka,S (eds) (2004) Applied scanning probe methods. Springer, Berlin 49. Tersoff J, Hamann DR (1983) Phys Rev Lett 50:1998 50. Tersoff J, Hamann DR (1985) Phys Rev B 31:805 51. Lang ND (1987) Phys Rev Lett 58:45 52. Lang ND (1987) Phys Rev B 36:8173 53. Chen CJ (1990) Phys Rev Lett 65:448 54. Chen CJ (1990) Phys Rev B 42:8841 55. Bardeen J (1961) Phys Rev Lett 6:57 56. Ohnishi S, Tsukada M (1990) J Vac Sci Technol A 8:174 57. Chen CJ (1988) J Vac Sci Technol A 6:319 58. Hill I.G, Kahn A, Soos Z.G, Pascal RA Jr (2000) Chem Phys Lett 327:181 59. Datta S, Tian W, Hong S, Reifenberge R, Henderson JI, Kubiak CP (1997) Phys Rev Lett 79:2530 60. Krätschmer W, Lamb LD, Fostiforopoulos K, Huffman DR (1990) Nature 347:354 61. Kroto H, Heat JR, Brien SCO, Curl RF, Smalley RE (1985) Nature 318:162 62. Auerbach A, Mannini N, Tosatti E (1994) Phys Rev B 49:12998 63. Mannini N, Tosatti E, Auerbach A (1994) Phys Rev B 49:13008 64. Saito S, Oshiyama A (1991) Phys Rev Lett 66:2637 65. Gunnarsson O (1997) Rev Mod Phys 69:5754 66. Fabrizio M, Tosatti E(1997) Phys Rev B 55:13456 67. O Gunnarsson, Handschuh H, Bechthold PS, Kessler B, Gantefor G, Eberhardt W (1995) Phys Rev Lett 74:1875 68. Goldoni A, Cepek C, Modesti S (1996) Synth Met 77:189 69. Cepek C, Goldoni A, Modesti S, Negri F, Orlandi G, Zerbetto F (1996) Chem Phys Lett 250:537 68:3924 70. Knupfer M, Fink J (1997) Phys Rev Lett 79:2714 71. Savin S, Harris AB, T Yildirim T (1997) Phys Rev B 55:14182 72. David WIF, Ibberson RM, Dennis TJS, Hare JP, Prassides K(1992) Europhys Lett 18:219 73. David WIF, Ibberson RM, Matthewman JC, Prassides K, Dennis TJS, Hare JP, Kroto HW, Taylor R, Walton DRM (1991) Nature 353:147 74. Wang H, Zeng C, Wang B, Hou JG, Li Q, Yang J (2001) Phys Rev B 63:085417 75. Lof RW, van Veenendaal MA, Koopmans B, Jonkman HAT, Sawatzky GA (1992) Phys Rev Lett 68:3924
20
L. Gavioli · C. Cepek
76. Wachowiak A, Yamachika R, Khoo KH, Wang Y, Grobis M, Lee DH, Louie SG, Crommie MF(2005) Science 310:468 77. Schiller F, Ruiz-Osés M, Ortega JE, Segovia P, Martínez-Blanco J, Doyle BP, Pérez-Dieste V, Lobo J, Néel N, Berndt R, Kröger J (2006) J Chem Phys 125:144719 78. Ehrenreich H, Spaepen F (Eds) (1994) Solid state physics, vol 48. Academic, San Diego 79. Haddon RC, Hebard AF, Rosseinsky MJ, Murphy DW, Duclos SJ, Lyons KB, Miller B, Rosamilia JM, Fleming RM, Kortan AR, Glarum SH, Makhija AV, Muller AJ, Eick RH, Zahurak SM, Tycko R, Dabbag G, Thiel FA (1991) Nature 350:320 80. Tycko R, Dabbagh G, Rosseinsky MJ, Murphy DW, Fleming R, Ramirez AP, Tully JC (1991) Science 253:884 81. Benning PJ, Stepniak F, Weaver JH (1993) Phys Rev B 48:9086 82. Gunnarsson O (2004) Alkali-doped fullerides: narrow-band solids with unusual properties. World Scientific, Singapore 83. Durand P, Darling GR Dubitsky Y, Zaopo A, Rosseinsky MJ (2003) Nat Mater 2:605 84. Holczer K, Klein O, Huang SM, Kaner RB, Fu KJ, Whetten RL, Diederich F Science (1991) 252:1154 85. Bruhwiler PA, Maxwell AJ, Nilsson A, Martensson N, Gunnarsson O (1993) Phys Rev B 48:18296 86. Knupfer M, Fink J, Armbruster JF (1996) Z Phys B 101:57 87. Yang WL, Brouet V, Zhou XJ, Choi HJ, Louie SG, Cohen ML, Kellar SA, Bogdanov PV, Lanzara A, Goldoni A, Parmigiani F, Hussain Z, Shen ZX (2003) Science 300:303 88. Jahn HA, Teller E (1937) Proc R Soc Lond Ser A 161:220 89. Ceulemans A (1987) J Chem Phys 87 :5374 90. Koga N, Morokuma K (1992) Chem Phys Lett 196:191 91. Rudolf P, Golden MS, Brühwailer PA, (1999) J Electron Spectrosc Relat Phenom 100:409 92. Hou JG, Jinlong Y, Haiqian W, Qunxiang L, Changgan Z, Lanfeng Y, Bing W, Chen DM, Qingshi Z (2001) Nature 409:304 93. Lu X, Grobis M, Khoo KH, Louie SG, Crommie MF (2003) Phys Rev Lett 90:096802 94. Nazin GV, Qiu XH, Ho W (2005) J Chem Phys 122:181105 95. Wang LL, Cheng HP (2004) Phys Rev B 69:045404 96. Maxwell AJ, Brühwiler PA, Arvanitis D, Hasselstöm J, Johansson MKJ, Mårtensson N (1998) Phys Rev B 57:7312 97. Silien C, Pradhan NA, Ho W, Thiry PA (2004) Phys Rev B 69:115434 98. Ohno TR, Chen Y, Harvey SE, Kroll GH, Weaver JH, Haufler,RE, Smalley RE (1991) Phys Rev B 44:13747 99. Fasel R, Aebi P, Agostino RG, Naumovic D, Osterwalder J, Santaniello A, Schlapbach L (1998) Phys Rev B 57:7312 100. Fasel R, Agostino RG, Aebi P, Schlapbach L (1991) Phys Rev B 60:4517 101. Gimzewski JK, Modesti S, Gerber Ch, Schlittler RR (1993) Chem Phys Lett 213:401 102. Gimzewski JK, Modesti S, Schlittler RR (1994) Phys Rev Lett 72:1036 103. Modesti S, Gimzewski JK, Schlittler RR (1995) Surf Sci 331–333:1129 104. Weckesser J, Cepek C, Fasel R, Barth JV, Baumberger F, Greber T, Kern K (2001) J Chem Phys 115:9001 105. Rowe JE, Rudolf P, Tjeng LH, Malic RA, Meigs G, Chen CT, Chen J, Plummer E (1992) Int J Mod Phys B 6:3909 106. Ohno TR, Chen Y, Harvey SE, Kroll GH, Benning PJ, Weaver JH, Chibante LPF, Smalley RE (1993) Phys Rev B 47:2389 107. Maxwell AJ, Bruhwiler PA, Andersson S, Arvantis D, Hernnas B, Karis O, Mancini DC, Mårtensson N, Gray SM, Johansson MKJ, Johansson LSO (1995) Phys Rev B 52:R5546 108. Li YZ, Chander M, Patrin JC, Weaver JH, Chibante LPF, Smalley RE (1992) Phys Rev B 45:13837
13 Ultrathin Fullerene-Based Films via STM and STS
21
109. David T, Gimzewsky JK, Purdie D, Reihl B, Schlittler RR (1994) Phys Rev B 50:5810 110. Kuk Y, Kim DK, Suh YD, Park KH, Noh HP, Oh SJ, Kim SK (1993) Phys Rev Lett 70:1948 111. Cepek C, Goldoni A, Modesti S (1996) Phys Rev B 53:7466 112. Pedio M, Hevesi K, Zema N, Capozi M, Perfetti P, Gottebaron R, Pireaux JJ, Caudano R (1999) Surf Sci 437:249 113. Sellidj A, Koel BE (1993) J Phys Chem 97:10076 114. Jiang LQ, Koel BE (1994) Chem Phys Lett 223:69 115. Ruckman MW, Xia B, Qiu SL (1993) Phys Rev B 48:15457 116. Rudolf P, Gensterblum G (1994) Phys Rev B 50:12215 117. Hoogenboom BW, Hesper R, Tjeng L, Sawatzky GA (1998) Phys Rev B 57:11939 118. Hunt MRC, Modesti S, Rudolf P, Palmer RE (1995) Phys Rev B 51:10039 119. Goldoni A, Paolucci G (1999) Surf Sci 437:353 120. Pai WW, Hsu CL, Lin MC, Lin KC, Tang TB (2004) Phys Rev B 69:125405 121. Tsuei KD, Johnson PD (1997) Sol State Commun 101:337 122. Grobis M, Lu X, Crommie MF (2002) Phys Rev B 66:161408 123. Pai WW, Hsu CL (2003) Phys Rev B 68:121403 124. Hsu CL, Pai WW (2003) Phys Rev B 68:245414
14 Quantitative Measurement of Materials Properties with the (Digital) Pulsed Force Mode Alexander M. Gigler · Othmar Marti
Abstract. In this contribution we present the potential of digital pulsed force mode (DPFM) atomic force microscopy (AFM) for mapping the local mechanical properties to surface topology. In the first part, we review the pulsed force mode principles and compare this mode with other common modes of dynamic AFM operation. Then, we present an overview of the different model for tip–sample interaction based on continuum mechanics theories. In a general view, we then show how the huge amount of data that is acquired during DPFM imaging is stored and how it can be processed in order to provide insight into local mechanics of the sample surfaces. Applications of this measurement technique show the possible impact of DPFM on condensed soft matter samples as well as possible extensions towards measurements of the in-plane mechanics of the tip and the surface. All mechanical DPFM data were analyzed on the basis of continuum mechanical models. We conclude our contribution with an overview of this technique and its benefits for lateral compositional mapping. Key words: Digital pulsed force mode, polymer, HeLa cells, Material properties, Local elasticity mapping, Atomic force microscopy data analysis, Atomic force microscopy calibration
Abbreviations AFM DMT DPFM JKR MD PFM PMMA SBR STM
Atomic force microscopy Derjaguin–Muller–Toporov Digital pulsed force mode Johnson–Kendall–Roberts Maugis–Dugdale Pulsed force mode Poly(methyl methacrylate) Styrene–butadiene rubber Scanning tunneling microscope
14.1 Introduction Atomic force microscopy (AFM) nowadays belongs to the standard toolset of surface science. Its ability to not only probe the topography of samples but also to provide a wealth of other information on local properties led to the application of the technique in fields such as life science, materials science and quality control. The original atomic force microscope [1] operated in contact mode. This operation
24
A.M. Gigler · O. Marti
provides information on the averaged atom position for atomically flat surfaces [2,3]. It was soon realized that contact mode imaging can be disruptive to molecules adsorbed on a surface [4]. The tip in contact with the surface and sliding laterally can exert forces large enough to displace molecules. It was later found that contact mode was one way to lithographically modify and structure surfaces [5, 6]. If one compares the operation of a scanning tunneling microscope (STM) [7,8] with that of an atomic force microscope, one notices that the tip of an STM is about 1 nm above the surface, whereas the atomic force microscope tip is in contact, meaning that there is a possibility that chemical bonds can be formed between an atom of the tip and the sample surface. The STM tip can be positioned at an arbitrary distance to the sample, because it is rigid. On the other hand, the atomic force microscope tip is mounted on a rather soft spring. The combination of the spring potential and the tip–sample interaction potential causes instabilities, which prevents the operation of an atomic force microscope at certain distances. To overcome this problem, one could use much stiffer cantilevers, at the expense of force sensitivity. Modulation techniques can then be used to get more sensitivity [9]. Noncontact AFM gives useful information on gradients of interaction potentials, and on energy dissipation [10]. Force-modulation AFM works best in ultrahigh vacuum. Therefore, several alternatives have been proposed for operation in air and liquids, such as tapping mode [11], pulsed force mode (PFM) imaging [12], jumping mode [13], CODY mode [14] or force volume mode [15]. All these operating modes are summarized by the term “intermittent contact mode”. In this chapter we will first give an overview of the intermittent contact modes and then concentrate on the physics and the applications of PFM imaging.
14.2 Modes of Intermittent Operation Intermittent contact modes are the method of choice for compositional imaging of surfaces. They offer minimum mechanical interaction for reduced destructiveness together with a maximum material sensitivity. Resonant imaging is the most widely distributed method of the dynamic AFM modes. Here, the cantilever is oscillated near or at its fundamental resonance and the damping or phase shift of the oscillatory movement in close proximity to the surface is observed [11]. More recent approaches also focus on the higher harmonics of the cantilever’s fundamental mode or the higher-order eigenmodes of the cantilever beam while imaging [16–19]. To access the mechanical properties of the sample materials directly, however, subresonance methods are most promising. Oscillating a cantilever far below its resonance frequency, one can assume a direct transfer of every movement of its base directly to the movement of the tip; thus, one is able to directly control the interaction of the tip and the surface. In essence, there are two different approaches to subresonant dynamic imaging. One is to conduct a so-called force volume imaging or jumping mode [13], i. e., to capture one force–distance curve for each pixel of the image. To do so, a ramped approach signal is applied to the positioning of the cantilever base. The other subresonance method is PFM. Here, the linearly ramped modulation is replaced by a sinusoidal positioning of the cantilever. In a linearly ramped positioning, the modulation signal (a triangle) contains high-frequency com-
14 Quantitative Measurement of Material Properties using DPFM
25
ponents in its Fourier spectrum that also have to be applied to the positioning system. Compared with this, a solely sinusoidal modulation requires only a bandwidth in the range of the modulation frequency itself. Thus, the sinusoidal modulation of the positioning system allows higher repeat rates than the triangular one. Both PFM and jumping mode allow for 3D imaging by following the topography under a predetermined stress by keeping the applied maximum force constant. In addition to this 3D image, both allow for the correlation with material properties deduced from the indentation measurements and are therefore adding a fourth dimension to the topographical measurements. 14.2.1 Destructive Versus Nondestructive Measurements Delicate samples are easily modified by scanning with an atomic force microscope tip. Figure 14.1 shows this behavior on a sample of polyethylene [20]. The sample consists of a highly swollen polyethylene network spin-coated to the surface. It was first imaged in tapping mode [11]. Then the atomic force microscope was switched
Fig. 14.1. Effect of scanning in contact mode on a polyethylene sample [20]. a The sample before scanning in contact mode. b The sample after one single line scan. c The normal and friction forces occurring in the scan
26
A.M. Gigler · O. Marti
to contact mode and moved across the sample. The resulting destruction was again measured in tapping mode. 14.2.2 The Pulsed Force Mode In order to measure the local material properties of sample surfaces at the nanoscale, indentation curves can be investigated. One can learn about the local mechanical stiffness and adhesion of samples as well as the viscous damping and creep during indentation processes. Since mechanical interactions always occur at the surface of the bodies involved in the contact, these quantities are highly interesting to the entire microelectromechanical/nanoelectromechanical systems or nanotribology community. In the following we focus on the quantitative differentiation of such material properties and describe both the experimental as well as the theoretical approaches. 14.2.3 Operating Principle of the Pulsed Force Mode Various properties, measured simultaneously at the same point of the sample, facilitate the characterization of the behavior of the sample material. An overview of the theoretical description of the tip–sample contact will be given in the next section. To keep the measurement process as nondestructive as possible, the tip must only be in contact for the shortest time possible. This can be achieved in PFM [12, 21–23] or jumping mode. Here, the tip follows a predefined approach and retract trajectory. Such a controlled positioning is not possible in a resonant mode of operation as discussed earlier. Even though in PFM as well as in jumping mode most of the lateral scanning movement of the tip is out of contact, the shear forces acting between the tip and the sample cannot be fully neglected. An elaborate study for the jumping mode was conducted by Moreno-Herrero et al. [24]. On the other hand, a controlled lateral excitation with simultaneous PFM operation can also be utilized to learn about further frictional properties of the sample surface. An example of CODY mode operation [22] is described later. At low speeds and triangular modulation the approach–retract cycle described above is known under the name of “force–distance curve” as seen in Fig. 14.2. In the literature force–distance curves have been well examined by many sources [25–30]. The pull-off force, i. e., the peak attractive force on retraction, provides an easy way to measure local adhesion. By application of a triangular modulation to the base of the cantilever, the tip is brought periodically into contact with the sample surface. Local adhesion, elastic, viscoelastic and plastic properties of the sample can be derived from the measured forces and positions; however, the range of accessible deformation rates is quite limited. In force–distance curves only repeat rates of a few hertz are used, since the bandwidth of the excitation electronics in general is not sufficiently wide for a triangular signal. Fourier decomposition of a triangular wave shows that quite a few higher harmonics have to be taken into account in order to synthesize a real
14 Quantitative Measurement of Material Properties using DPFM
27
triangular wave; hence, the low-pass behavior of amplifiers and other components affects the transmitted curve shape and, thus, cantilever positioning. For sinusoidal drive signals this dispersion does not occur; therefore, a purely sinusoidal modulation will not suffer from this bandwidth limitation. Nevertheless phase shifts between the drive signal and the cantilever movement are present. In addition, sinusoidal drive eliminates the excitation of spurious resonances in the atomic force microscope. A setup designed for minimal disturbance of the drive signals is the PFM. The equivalence between this mode and force–distance curves can be shown by simple transformations. Examples of force–distance curves and force curves acquired using PFM are shown in Fig. 14.3. Both curves were measured with the same AFM setup at the same frequency of 100 Hz. It is obvious that at the turning point of the triangular drive either the triangular wave shape is not transmitted correctly to the cantilever or the cantilever is excited at its resonance, or both. Yet, it is clear that for repeat rates of more than 100 Hz of a triangular modulation, higher harmonics from the wave shape couple to the cantilever. The interpretation of the results becomes very intricate. If the mechanical processes could be modeled analytically, it would be possible to simulate force–distance curves at those high repeat rates. However, PFM offers a compromise, since it can be operated in a frequency range from 50 Hz to 20 kHz, which has been verified experimentally. It has the advantage that the content of higher harmonics is negligible in the spectrum of the wave form, avoiding possible problems. Given that each material has its own characteristic temporal response as shown in Fig. 14.3, one might conclude that the type of excitation function could affect the results. The coordinate transformation removes the trajectory from the signal. The resulting data turn out to be the equivalent to those obtained by triangular excitation. This can be seen in the sinusoidally driven trace in Fig. 14.1.
Fig. 14.2. Force versus distance curve as it is acquired by atomic force microscopy (AFM). The measured force is plotted versus the position of the base of the cantilever. Coming closer to the surface, the tip snaps into contact with the surface. On further approach, the cantilever is forced to bend backwards depending on the elastic, viscoelastic and plastic behavior of the sample material. A certain maximum force is reached, and then the base is retracted, causing relaxation of the cantilever. On the retrace, a material-dependent negative force has to be exceeded to pull off the tip from the surface. All forces have to be measured relative to the “free” position of the cantilever at the rightmost point
28
A.M. Gigler · O. Marti
Fig. 14.3. A view of the development of the force within the time domain. Trace a shows the response of the system upon modulation with a triangular trajectory. As in Fig. 14.1, snapin, maximum force, the repulsive regime, pull-off force and the baseline can be assigned to the curve. In trace b, which corresponds to a sinusoidal movement of the cantilever, the same characteristics can be found as in trace a. However, in order to sustain the steep turns, the triangular wave is bandwidth-limited to a certain maximum velocity (or repetition rate). By coordinate transformation (assigning the trajectory of the base of the cantilever to the horizontal axis of the force plot), the modulation-specific curve shape is removed from the curve shape. What remains is the classic force–distance curve plot (Fig. 14.1)
14.2.3.1 Topography—Measurement and Reality Topography in AFM measurements always comprises crosstalk from the sample compliance. Since the average force applied to the sample is kept constant, the tip follows equiforce lines rather than the real sample morphology; thus, the interpretation of topography from AFM measurements may be misleading. However, from knowledge of the interaction of the tip and the sample, one is able to reconstruct the “real” topography of the surface by simply adding the maps of measured topography and indentation depth. Further corrections of the topography can be made by calculating the additional deformation of the sample due to the error of the control loop. 14.2.3.2 Adhesion Force The adhesive behavior of surfaces is accessible by the attractive force that has to be overcome to retract the tip completely from the sample. In PFM, the detachment force is measured as the minimum force of an indentation loop while the tip is retracted from the sample (Fig. 14.4a). In ambient conditions, the interpretation of adhesive properties is often problematic, because of capillary condensation. The force needed to rupture the liquid neck can easily exceed and, thus, hide the actual tip–sample detachment from being observed. The detachment force can be mapped to the surface features as shown in Fig. 14.4b.
14 Quantitative Measurement of Material Properties using DPFM
29
Fig. 14.4. Before the tip–sample contact is released, a certain attractive force has to be overcome. Measurement of the minimum force upon retraction of the tip for each force curve (a) results in a map of the local adhesion of the sample (b). PMMA poly(methyl methacrylate)
14.2.3.3 Sample Compliance The slope of the indentation curve depends on the compliance of the sample. For hard reference surfaces like silicon or elastic materials such as poly(methyl methacrylate) (PMMA), the approach and retract curve is essentially the same (Fig. 14.5a). Softer materials reveal a much flatter trace in the repulsive regime. Following the descriptions of contact mechanical models such as Derjaguin–Muller–Toporov (DMT) or Johnson–Kendall–Roberts (JKR), the indentation curvature can be analyzed by a fit to the force trace. As shown in Fig. 14.5b, maps of the local elasticity of the sample can be created. In this composition, styrene–butadiene rubber (SBR) is the softest of the three components (dark regions), silicon is the hardest (bright color) and PMMA shows an intermediate stiffness.
Fig. 14.5. Local stiffness of the sample is calculated from a fit of a Johnson–Kendall–Roberts (JKR) based model to the force curve. For the elastic PMMA, the slope of the force curve is lower than for the hard silicon (a). Y map of the local elasticity of the sample can be created by evaluating all traces recorded in the measurement (b)
30
A.M. Gigler · O. Marti
14.2.3.4 Electrical Fields On charged samples, PFM is also capable of the determination of electrostatic interactions between the cantilever and the sample surface. Interaction forces between the tip and the sample such as van der Waals forces [31], capillary forces [32] or repulsive forces [33] are short-ranged. Electrostatic forces in nonpolarizable environments and magnetic forces are long-ranged. Near the sample surface the long-range forces are small compared with the short-range ones; therefore, the measured adhesion forces, the stiffness and the topography are almost not affected by electrostatic and magnetic forces. However, the baseline, the level of the force far away from the sample surface, strongly depends on the long-range forces [12]. Figure 14.6 shows a simultaneous measurement of the topography, the electrostatic forces and the adhesion on a glass surface partially covered by gold electrodes. The left and the right electrode were oppositely charged. The voltage between the two electrodes was modulated. The atomic force microscope tip was made of silicon nitride and was not conducting. The measurements show that the topography and the adhesion are not affected by the charge on the electrodes. If the charge had been static, one would not be able to discriminate the influence of the electrical field on the PFM measurement. Alternatively one can detect local electrical fields at the jump-in on the force curves [34].
Fig. 14.6. Mapping the electric fields relies on the attractive or repulsive interaction before the snap into contact of the tip. Close to the surface, electric forces can be measured [12]
14.2.3.5 Energies After conversion of the force traces, areas enclosed by the force curves correspond to work. A very interesting quantity is the hysteretic loss during a single experimental cycle, which is equal to the enclosed area in the repulsive regime. It gives an idea of the energy deposited in the surface by the experiment. Hysteretic losses might be due
14 Quantitative Measurement of Material Properties using DPFM
31
Fig. 14.7. Areas enclosed by force curves correspond to energies. In a, the red area corresponds to hysteretic loss, while in b, the detachment energy is colored in red
to thermal dissipation as well as viscous damping. For SBR, the energy deposited in the sample is equal to the red area in Fig. 14.7a. Another interesting energy is the work needed to detach the tip from the sample. It can be determined as the area enclosed by the zero-force line and the force trace in the attractive regime. This is colored yellow in Fig. 14.7b. 14.2.3.6 Postflow Since experiments in PFM are conducted at rather high repeat rates (approximately 1 kHz), the interaction timescale for the release of the mechanical load in the repulsive part (upon retraction of the tip) is on the order of microseconds. For rubber-elastic or viscoelastic materials, relaxation times are also within this time scale. For deformations, this means that the material follows the loading with a temporal delay. In force versus indentation plots as shown for SBR in Fig. 14.8a, this effect can be observed
Fig. 14.8. Viscous creep can be observed at the turning point of the indentation. Elastic materials such as PMMA respond instantaneously on the release. Viscoelastic or rubber-elastic materials such as styrene–butadiene rubber (SBR) are deformed with a temporal delay. While the force on the cantilever is already reduced, the tip is still continuing to indent into the material (a). b A map of the postflow distance
32
A.M. Gigler · O. Marti
as a postflow of the material, since the force is already decreasing, while the tip still indents further into the sample. Again, the distance of this postflow can be used to distinguish between materials at the sample surface. A map of the postflow is shown in Fig. 14.8b. 14.2.4 Analog Pulsed Force Mode The setup of a PFM is rather straightforward. What is needed is the possibility to modulate the tip–sample distance in order to rupture the point of contact before a new approach curve is acquired. Actually, the contact time should be kept as short as possible to avoid sample destruction during the lateral scanning movement. This z-modulation is usually conducted by a separate piezo within the cantilever holder. Sampling and recording of the measurement data is another demanding topic. On the one hand, data acquisition has to be sophisticated enough to allow the reduction of the amount of data to a reasonable scale, while, on the other hand, the data should only be compressed to a predetermined error margin. 14.2.5 Digital Pulsed Force Mode In the analog implementation of the measurement system, this is accomplished by peak-pickers and sample-and-hold circuitry, while with the new digital version, it is possible to capture and store the entire data of an experiment. The digitization is therefore a very demanding topic and has been set up as a dedicated transfer and storage connection. The I/O system, for example, has to be able to constantly handle about 10 MB/s for a 1-kHz experiment. For a typical experimental run with a scan rate of about one line per second, this adds up to a file size of several gigabytes. Anyway, the data also have to be processed. Here, the analog setup has a clear advantage over the digital version, since only the most important points and features of the interaction are investigated in this case. On the other hand, all the information contained in the remaining force curve is neglected. To also record this data means that a more powerful CPU has to process the streaming data. This is done by a field programmable gate array, which mimics the analog behavior by operating simple algorithms such as peak-picking and sample and hold in real time, while being able to analyze the data in an off-line data evaluation more thoroughly. Another advantage of the digital implementation is that a synthesized curve can be used as the modulation signal. Thus, the curve shape is interpolated between 5000 support points. Triangular and sinusoidal curves as well as, for example, trapezoidal excitations can be realized and material responses to changing excitations can be investigated. By simply switching to the one or the other modulation function inside the software, one can even do this at the exact same position of the sample, allowing direct comparison of the results. With the analog setup this has not been possible without major modification.
14 Quantitative Measurement of Material Properties using DPFM
33
14.3 Contact Mechanics Relevant for Pulsed Force Mode Investigations The interaction of the tip with the sample dominates the response of the cantilever in contact with the sample [35–40]. We summarize here the most relevant models. 14.3.1 Hertz Model The contact mechanics of rigid bodies was first described by Hertz [41]. Assuming two perfectly elastic spheres made of an isotropic material (Rm , E m , νm , being the radius, Young’s modulus and Poisson’s ratio of each sphere of index m, respectively), whose centers are being pushed towards each other (Fig. 14.9) by a load P and neglecting adhesion forces, he was able to calculate the analytical solution to the problem. Assuming that the spheres are only deformed inside the area of contact, one can calculate the geometric changes. Two spheres of radius R1 and R2 are given by R12 = r 2 + [R1 − z 1 (r)]2 , R22 = r 2 + [R2 − z 2 (r)]2 , where z 1 and z 2 are the distances from the x-axis. For small distances r from the main axis z, z 1 (r) and z 2 (r) can be written as z 1 (r) =
r 1 , 2 R1
z 2 (r) =
r 1 . 2 R2
Fig. 14.9. The situation of the impact of two fully elastic spheres of radius R1 and R2 as used for the derivation of the Hertzian contact model. At a given force, the centers of the spheres will be closer by a distance δ, resulting in a deformed contact area of radius a. Before the impact two points on the spheres at the same distance r from the z-axis were separated by z 1 (r) + z 2 (r). The displacement δ moves sphere 1 by δ1 and sphere 2 by δ2 , while the distance between the points only changes by δ1 + δ2 − [z 1 (r) + z 2 (r)]
34
A.M. Gigler · O. Marti
The sum equals the distance of two opposing points of the spheres (M, N). Thus, pushing together the two spheres by a distance δ leads to a change in distance along each sphere’s boarder line. The deformation of both spheres along the x-axis then is Δ(r) = δ1 + δ2 − (z 1 + z 2 ) 1 r2 1 + =δ− 2 R1 R2 2 r 1 . =δ− 2 R Per definition R is the effective radius: 1 1 1 = + . R R1 R2 On the other hand, this deformation also has to be the result of an elastic deformation by pushing the spheres together. The pressure distribution for this problem is to first order r2 p(r) = p0 1 − 2 , a where r is the actual position within the area A = πa2 of radius a and p0 is the maximum pressure in the center of this area (Fig. 14.10a). Within the loaded circle, the local displacement of a sphere with the index m along the z-axis is given by [42, 43] u zm (r) =
1 − υm2 π p0 2 2a − r 2 Em 4 a
r≤a.
Thus, the total deformation is—again—the sum of these displacements and has to equal Δ(r) as calculated above: 1 − υ22 π p0 2 1 − υ12 2a − r 2 . + Δ(r) = E1 E2 4 a The total load P (the interaction force) in this area can be calculated as the integral of the pressure distribution over the entire area: P=
2π a 0
p(r)r dr dφ = 2π
0
a 0
a r2 1 − 2 r dr . p(r)r dr = 2π p0 a 0
This integral can be evaluated by using the substitution z 2 = 1 − r 2 /a2 , or z dz = −r dr/a2 . Using the substitution of the boundaries 0 → 1 and a → 0, we get for the load P = −2π p0
0 1
2π p0 2 3 1 2π p0 a2 a z dz = a z = . 3 3 0 2 2
14 Quantitative Measurement of Material Properties using DPFM
35
Fig. 14.10. Indentation of a sphere into an elastic half-space (a). It is valid for all contact models. The basic difference between the Derjaguin–Muller–Toporov (DMT), JKR and Maugis models is the definition of the area that is involved in adhesive forces (b). In the DMT model (top) it is only an annular zone around the Hertzian contact area which causes adhesion. The JKR model (middle) assumes that only the contact area contributes to the adhesive force, but the surface is compliant enough to adhere to the tip even beyond point zero. Both models are special cases of the more general Maugis model (bottom)
On the other hand, we have at the center of the contact (r = 0) Δ(r) = δ . If we substitute at r = 0 the second equation for Δ(r) into the first one we obtain 1 − υ12 1 − υ22 πa δ= + p0 . E1 E2 2 Now we use the equation r2 = 2R
r2 2R
= δ − Δ(r) = Δ(0) − Δ(r) and get
1 − υ12 1 − υ22 + E1 E2
p0
π 2 r 4a
36
A.M. Gigler · O. Marti
or 1 = R
1 − υ12 1 − υ22 + E1 E2
p0
π . 2a
If we now divide the equation for δ by the equation for 1/R we get the contact radius √ a = Rδ . Defining an effective modulus K as 1 − ν22 3 1 − ν12 1 = + K 4 E1 E2 2 and using P = 2π 3 p0 a , we can also write this contact radius as PR 1/3 . a= K √ Also, by solving for P and substituting a = Rδ, we get the dependency of the force on the deformation of the spheres: √ P = K Rδ3 .
Given this load, the deformation can also be written as 2 31 P . δ= K2R For the calculation we used the contact of two spheres with different radii. The AFM geometry can be handled by letting R2 → ∞. Then R1 equals the effective radius R. This describes a fully elastic sphere indenting into a fully elastic half-space. Furthermore, if one sphere is incompressible, e. g., meaning E 1 tends to infinity, only the other one will be deformed. 14.3.2 Sneddon’s Extensions to the Hertz Model In 1965, based on the nineteenth century calculations of Hertz [41], Sneddon [44] went on to extend the theoretical description to nonspherical, but also axisymmetric, indenters. Later, many groups [25, 26, 28, 45–49] extended the theories to virtually every sufficiently smooth indenter shape with rotational symmetry. A brief summary will be given in the following. The basic assumption for the calculations is that for a purely elastic material, the stiffness S (or the change in force upon a given change in indentation) is directly proportional to the root of the contact area of the indenter and the surface at this position, i. e., S=
dP √ ∝ A. dδ
14 Quantitative Measurement of Material Properties using DPFM
37
For some of the most important boundaries of indenters, we will now show the consequences. One of the easiest shapes for an indenter is a cylinder that confers a “flatpunch.” In this case, the radius of the indenter is constant for all depths r(δ) ≡ R∀δ. Clearly, the area of the punch is thus a constant too: A ≡ πR2 . This implies S ∝ const., which results in a linear dependency of the force P =
S dδ ∝ δ. √ 2 2 A spherical indenter with √ a boundary line of r(δ) = R − (R − δ) ≈ 2Rδ implies a stiffness S ∝ π2Rδ, which means that the force will be P =
S dδ ∝ δ3/2 , which exactly reproduces the Hertz behavior. Owing to the approximation of the contact radius, this description is correct only for indentation depths small compared to the radius of the sphere (δ = R3 already results in an error of 10%). √ Using√a parabolic curve to describe the surface of the indenter, we find r(δ) = δ, S ∝ δ and, hence, the force P ∝ δ3/2 , which is exactly the same as the result for a sphere for small indentations (near the apex of a sphere it can be approximated as a parabola). Carrying out the same calculation for a cone which has a straight line as its boundary results in a radius r(δ) = δ and a stiffness S ∝ δ. This leads to the force law P = S dδ ∝ δ2 . For a curve of arbitrary shape r(δ) = δ1/α , the stiffness is proportional to δ1/α as well, while the force on the contact will increase like P = S dδ ∝ δ1+1/α . This last equation with all its simplicity has the power to explain measured indentation data. One often finds exponents greater than 2, which cannot be explained by standard indenter shapes. However, if the indenter were formed as a root function, it would even allow for an exponent of 3. The only problem with this formulation is that one of the basic assumptions was to have a smooth surface, in order to avoid singularities in the pressure distribution at point zero; thus, this theory would only be applicable at large indentations. The small indentation depths could be modeled with an indenter made of a small spherical part attached to the root shape; thus, this would be a clever way to circumvent the problem of the singularity, while having an exponent greater than 2. Anyway, this correction has been neglected throughout the published literature for the conical indenter too. Only the shape of indentation curves at the initial contact is affected by this problem. From the analytical point of view, one simply has to be aware that basic assumptions are violated at this point. Recently, there have been several attempts at a more general description of an indenter by Borodich et al. [50, 51] or for more specialized, more realistic shapes, e. g., a Vicker’s indenter which is a pyramidal tip, by Franco et al. [52].
14.4 Models Incorporating Adhesion One of the shortcomings of the Hertz model and, thus, the Sneddon extensions is that adhesion is not part of the model. A more complete approach was chosen by Maugis [42, 53] based on an earlier theory of Dugdale [54]. Their theory contains an adjustment parameter for the systematic inclusion of adhesion effects
38
A.M. Gigler · O. Marti
2 −1/3 λ = 2σ0 π̟K . It is assumed that the region in which adhesive forces are R acting is larger than the contact area derived in the Hertz model; m = ac being the parameter determining the actual region of influence of adhesion (Fig. 14.10b). The formulae describing the depth of indentation δ and the load P are √ 8aσ0 m 2 − 1 a2 − , δ= R 3K
a3 K − 2a2 σ0 P= m 2 − 1 + m 2 arctan m2 − 1 , R
where σ0 is the surface stress at the edge of the contact. There are two limiting cases where the Maugis–Dugdale (MD) model is reduced to much simpler theories: the JKR model [35, 43, 55, 56] and the DMT model [31, 36, 57, 58]. The JKR model holds for the special case when adhesive forces are only acting in the region of immediate contact, while the DMT model describes the situation when these forces are acting along the circumference of the contact only. In addition, the JKR model assumes that the sample is compliant enough to adjust to the detaching sphere, while the DMT model claims that the surface will not deform on detachment. To derive the JKR model, one has to assume that the maximum indentation force leads to the same radius of contact as the Hertz model (a3 = PR K ), while at the same time at zero load, the sphere is still in contact with the surface. For this situation
2 1/3 a2 the radius of contact is a0 = 6π̟R and δ0 = 3R0 , where ̟ is Dupré’s work K of adhesion to separate two solidsreversibly andisothermally. Generally, the radius 2 3π̟R 3π̟R . It is obvious 1 + 3π̟R of contact is given by a3 = PR K P + 2 P + P
that for the nonadhesive case (̟ = 0) this equals the Hertzian contact. Pulling the indenter and the surface further apart, one reaches an unstable position where the indenter finally detaches. The force at which this occurs is Fadh = − 23 π̟R with
2 1/3 a contact radius of a = 32 π̟R or a = 0.63a0 . The “indentation depth” of this K
1/3 2̟ 2 R a2 = − π12K . position is δ = − 3R 2 In the DMT model describing an adhesive contact of an elastic sphere with an elastic surface, it is assumed that adhesive forces originate from a ring-shaped zone around the contact radius and that the surface does not deform owing to adhesion. The deformation profile is therefore identical to the one derived in the Hertzian 2π̟R2 model. The contact radius is given by a3 = PR K + K , where the nonadhesive case yields the Hertzian contact again. The depth of indentation is given by √ 1/3 2 2π̟ R P + δ = aR = K √ . Under zero load the contact radius is given by K R 2
a2
at a indentation of δ0 = R0 . The pull-off force is Fadh = −2π̟R at a03 = 2π̟R K depth δ = 0 and a = 0. The behavior of a DMT-like indenter is simply that of a Hertz sphere, but offset by the adhesive force. Introducing normalization parameters, we can compare the theoretical descriptions more easily:
14 Quantitative Measurement of Material Properties using DPFM
A=
a π̟R2 K
1/3 ,
P , π̟R δ Δ=
1/3 , 2 2 P=
39
the reduced contact area;
the reduced force on the contact; the reduced indentation depth; and
π ̟ R K2
λ=
2σ0
, 2 1/3
the degree of adhesion according to Maugis.
π̟K R
Thus, the equations describing the contacts are simplified to: Model
Reduced load and indentation depth
Hertzian
P = A3 Δ = A2
DMT
P = A3 − 2 Δ = A2
JKR MD
√ P = A3 − A√ 6A Δ = A2 − 23 6A
√
√ m 2 − 1 + m 2 arctan m 2 − 1 √ Δ = A2 − 43 Aλ m 2 − 1 √ √ 1 = 12 A2 λ m 2 − 1 + m 2 − 2 arctan m 2 − 1 √ √ + 34 Aλ2 1 − m + m 2 − 1 arctan m 2 − 1
P adh = A3 − λA2
The reduced force curves versus indentation depth are shown in Fig. 14.11 for comparison. 14.4.1 Data Processing When indentation data are recorded in digital PFM (DPFM), the storage protocol is also capturing the information of the search windows as well as the data evaluated by on-line algorithms such as maximum force, detachment force or the linear stiffness as recorded during the measurement. In postprocessing, the search windows can be accessed in order to estimate regions of interest instead of analyzing the entire data. Before the data are analyzed with respect to continuum-mechanical models, the data have to be preconditioned. For data recorded in a viscous environment like buffer solution or water, the drag force has to be subtracted from the force signal. In a first-order approach, the drag has to follow a cosine function for a sinusoidal modulation, since the drag is velocity-dependent. Fitting such a cosine function to
40
A.M. Gigler · O. Marti
Fig. 14.11. With normalized parameters, all models describing contact mechanics can be plotted on one graph. It is obvious that the DMT curve differs from the Hertzian curve only by a constant offset, while the JKR plot has a steeper increase. However, JKR and Maugis–Dugdale are the only theories describing negative indentation depths which explain the pull-off behavior of very compliant materials. Thus, these explain hysteresis in detachment processes. (From [42])
the data requires knowledge of the repulsive regime, since the indentation information is contained here. Using the information from the search windows, one can easily do this by rejecting the information inside the snap-in as a starting point and the end of the adhesion window as the stop point. This approach works well as demonstrated in [59, 60]. Afterwards, the baseline has to be subtracted in order to measure forces free of long-range interaction offsets. Therefore, the average value of the force signal outside the repulsive regime has to be subtracted. To enter the physics of the indentation measurements, the data have to be converted from detected signals into forces measured by the atomic force microscope. To do so, calibration data have to be obtained from force curves on a hard reference material. In the case of such a hard material, the cantilever is assumed to compensate each nanometer of approach by 1 nm of deflection.Owing to the sinusoidal modulation, the x-axis has to be transferred from the time domain (Fig. 14.12a) to the z-position of the lever (Fig. 14.12b). In such a plot, the deflection signal has to be a linear function of the z-modulation. Since there are time delays in the modulation and detection signals, which cannot be avoided, the linear fit of the data can only be optimized for a certain phase shift. This phase shift, however, is constant for a given measurement setting. The slope of the linear fit then allows the calibration of the measurement data. Before calculation of elastic properties or energies involved in the indentation process, another conversion has to be done. The x-axis has to be transferred to the indentation depth instead of absolute z-position. In general, the deflection of the
14 Quantitative Measurement of Material Properties using DPFM
41
Fig. 14.12. Three materials were probed in this measurement: SBR (circles), PMMA (diamonds) and silicon (solid line). From the force versus time data as acquired by the digital pulsed force mode (a), the known trajectory of the base of the cantilever can be removed, resulting in plot of force versus position of the cantilever support (b). Silicon is taken as the reference material. The slope of the calibration curve is subtracted from all other data in order to rescale the horizontal axis. Finally in c, all three materials are plotted versus the indentation depth (the real physical quantity). Areas now correspond to work applied to the sample. The theories of continuum mechanics can be used to analyze the data, revealing Young’s moduli or surface energies of the sample materials
Table 14.1. Characteristics of the sample materials as used for the measurements shown in Fig. 14.12 Silicon wafer Young’s modulus 60–150 GPa (bulk, static) Expected behavior Hard; marginal deformation Glass temperature – Molecular mass – Polydispersity – Radius of gyration – Specialty Sonication cleaning: 15 min each toluene/ethanol/toluene
PMMA
SBR
3.3 GPa
50–150 MPa
Elastic; glassy polymer structure 118 ◦ C 100 kg/mol 2.15 7 nm –
Rubber-elastic; clear relaxations −20 ◦ C 390 kg/mol 2.6 14 nm Not cross-linked; 30% S + 70% Br; statistically arranged
PMMA poly(methyl methacrylate), SBR styrene–butadiene rubber
42
A.M. Gigler · O. Marti
cantilever is caused by two components. First, the deflection of the cantilever due to a given force. This is the same deflection for all sample materials. Second, the deformation of the sample until this force is reached. This is intrinsic to the sample itself and allows the determination of the material stiffness. Now, to calculate the indentation depth coordinate, the deflection of the cantilever at a given force on the reference material has to be subtracted from the deflection of the cantilever on the (softer) sample material. Finally, the data are ready for analysis. A set of exemplary force versus indentation curves for silicon, PMMA and SBR are shown in Fig. 14.12c. The indentation measurements shown in Fig. 14.12 were carried out on three different sample materials: silicon, PMMA and SBR. The material properties for the bulk can be found in Table 14.1. 14.4.1.1 Search Methods In DPFM, the entire force data are recorded instead of just saving the most interesting features of the scans. Together with the advanced hardware of the system, it is possible to apply certain search algorithms to the data. The algorithms work within the range of preset search windows. 14.4.1.2 Compressing the Data Data can be reduced in two fundamentally different ways. One way is to reduce the sampling rate, i. e., to collect fewer data points for each curve; the other way is to compress the amount of data that has to be stored without loss of information. The latter can be realized by storing the first curve in full depth (16 bit) and storing the following traces only differentially at a lower bit rate. 14.4.1.3 Creating Images from Data Analysis Every single pixel of each image carries at least the full information of an entire force curve. Evaluation of all the data allows one to determine the local mechanical properties of samples. Maps representing the physical quantities can be created from this analysis. 14.4.2 Polymers To test the capabilities of a system like PFM, a sample system that contains all different mechanical behaviors, elastic, viscoelastic, adhesive, etc., at once is the most desirable. For this purpose, a dewetting polymer mixture consisting of PMMA and SBR has been used. A binary mixture of two parts by volume of PMMA and one part by volume SBR will be discussed in the following. In general, polymers do not
14 Quantitative Measurement of Material Properties using DPFM
43
mix, since the mixing entropy too small, owing to the huge number of monomers – of low molecular weight – combined in long chains. However, mixing can be mediated by short molecules diffusing into the long-chain polymers. In this case, toluene was used as the solvent. In solution, the polymers are more or less homogeneously dispersed. Spincasting the solution onto a flat solid substrate such as mica, glass or silicon and evaporating the solvent then breaks the mixing equilibrium and the polymers try to separate again in order to minimize the surface free energy. Together with the substrate, there is a four-component system: substrate, air and two polymers. The last of these components is the only one that is able to change conformation and shape. Thus, the affinity of the polymers towards the substrate decides what amount of area is occupied by which polymer. If both polymers show the same wetting behavior towards the substrate, both will extend down to the substrate, while if one component has a minor tendency to wet the substrate, a stratified surface coverage will be observed. The polymer–polymer interface remains. Here, the polymers dewet owing to the low mixing entropy, again. Minimizing the surface energy causes separation of the polymer mixture into different phases with minimum contact areas. Finally, there is the polymer–air interface, which will be a purely flat surface. An example of such a sample is shown in Fig. 14.13. There have been studies and publications [61–64] where the mechanism of dewetting and phase separation of mixtures of polystyrene and PMMA has been described in great detail. For thin films as produced by spin-casting, these structures are accessible by AFM. Mechanical properties of such surfaces can be nicely measured in PFM. As common in sciences on the nanoscale, this expectedly leads to results deviating from literature values for material properties of bulk samples. In order to probe the local mechanical properties of the sample, the maximum indentation depth should be one fifth of the film thickness for AFM experiments. Otherwise, the results can be strongly influenced by the properties of the substrate. All measurements were conducted using the same experimental parameters: modulation amplitude:250 nm; modulation frequency 1.0 kHz; maximum normal load 1.2 mN. The experiments were conducted in ambient conditions with cantilevers of 51–62 N/m force constant at a fundamental frequency of 328–340 kHz.
Fig. 14.13. Solutions of polymers phase-separate when spin-cast onto a favorable substrate. The morphology that forms depends on the ratio of the polymers in the solution and their tendency to wet the substrate. The measurement shown here is the topographic image of one of the experiments discussed in the text
44
A.M. Gigler · O. Marti
Most important in AFM experiments is to depict the sample topography correctly. Most of the contact or intermittent-contact AFM modes, however, rely on an average force that has to be kept constant in order to ensure stable imaging conditions. However, materials of different compliance result in the same force being reached only for a certain deformation of the sample. Thus, the real topography and the measured topography for a given interaction force are commonly not the same. The measured topography is shown in Fig. 14.14a. The cross section in Fig. 14.14b (blue line in Fig. 14.14a) implies a curved surface of the SBR compared with a rather rectangular shape of the PMMA islands. The average height of the SBR is about 30 nm, while the PMMA is leveled at about 55 nm above the silicon wafer. The indentation depth at which the maximum force is reached is shown in Fig. 14.15a. The inverse effect of the measured topography is observed in these measurements. The SBR is deformed by about 25 nm until the set-point force is reached (Fig. 14.15b). The PMMA regions, on the other hand, are only compressed by about 3 nm.
Fig. 14.14. Topography of the polymer mixture as recorded for a given maximum set-point force. The blue line in the map (a) marks the position of the cross section (b). The same cross-section coordinates were also used in the following figures
Fig. 14.15. Indentation depth of the tip into the material. a The strong deformation of SBR that is needed to reach the set-point force, while on PMMA and silicon this force is reached after only a few nanometers. b The cross section of the indentation depth
14 Quantitative Measurement of Material Properties using DPFM
45
Adding these measurements allows the “real” surface topography to be restored (Fig. 14.16). It corresponds to the surface at the initial contact of the tip and the surface, before compression. Yet, there is still a height difference between the PMMA islands and the SBR background, but the polymer surfaces are almost equally flat. This complies well with what is known from polymers dewetting towards the air interface. Besides a small corrugation by the polymer coils at the surface, the interface has to be very smooth. The ratio of indentation depth and original sample thickness gives a measure for the deformation of the sample. A map is shown in Fig. 14.17a. For the silicon region ξ = 0, since the initial thickness is the thickness of the silicon wafer. Compared with this, all nanoscale deformations are negligibly small. For SBR and PMMA, however, the results are clear: with an original height of (55 ± 3) nm, the PMMA areas are indented by about 6–8% of their initial film thickness, while SBR, starting at an average height of about (47±3) nm, is extremely
Fig. 14.16. Topography of the sample after correction of the indentation deformation. a The corrected surface morphology. As required for polymers dewetting towards the air interface, the surface is very flat. The same can be found in the cross section (b)
Fig. 14.17. By division of the deformation, a map (a) of the deformation ratio can be created. A cross section (b) shows the strong deformation of SBR at the same force, while PMMA is only deformed very little
46
A.M. Gigler · O. Marti
deformed, with x ranging from 20 to 65%. This behavior can also be found in the cross section along the same line as before. Applying an elastic model to the data, one can obtain the elastic properties of the sample. Here, we used a JKR-based approach. The map of local elasticity is given in Fig. 14.18. As a reference material, the silicon region in the left part of each image was used. With assuming an elastic modulus of 160 GPa for silicon, and νSi = 0.15, νPMMA = 0.4 and νSBR = 0.5 for the Poisson’s ratios, the Young’s moduli calculated are (110.1 ± 7.5), (42.6 ± 10.5) and (0.84 ± 0.64) GPa for silicon, PMMA and SBR, respectively. In comparison with literature values, these results appear by far too high to describe the materials present on the sample. Literature values for a bulk sample under static conditions are: 3.3 GPa for PMMA and 150 MPa for SBR. However, this is exactly the benefit of the AFM technique. The contact between any two bodies always occurs on the surface; thus, the mechanical properties of the surface are important. One underestimates the elastic behavior if only the lower literature values for bulk samples are used instead of the higher values obtained by a specialized surface-sensitive technique. PMMA will be discussed first. The reason why the calculation of the Young’s modulus yields a value that is too high is that at a surface the PMMA chain molecules are not free to align as coils as in a bulk phase, but rather have to gain a flat arrangement owing to the confined geometry. This increases the initial resistance towards indentation. Thus, indenting into a material like PMMA, which might have a coil structure in the bulk but a network or at least a stretched filament structure at its surface, must yield values that are higher than those from literature that were obtained for a macroscopic experiment on a bulk material. The fact that the values for SBR are too big by about an order of magnitude is backed by the theory of heat capacity. As in thermodynamics, isothermal and adiabatic processes occur. An isothermal process is characterized by energetic equilibration and an adiabatic process occurs in the absence of balancing. Since the indentation experiment happens at nonzero velocities, one has to move from an isothermal view that would be valid for a quasi static evolution of the forces applied, to an adiabatic model that is too fast for the exchange of energy. Energy diffusion can be described by the random walk. The average distance is given by 2 x = 6Dt , where t is the interaction time and D the diffusion constant. For the DPFM experiments the interaction time is about 0.1 ms and D is about 10−7 m2 /s; thus, the
Fig. 14.18. Map of the Young’s modulus of the materials of the polymer mixture
14 Quantitative Measurement of Material Properties using DPFM
47
average diffusion length scale is about 250 nm, which is about the scale of the deformations in use. Therefore, the modulus of SBR or PMMA cannot be calculated by an isothermal approach. This is why the Young’s modulus has to be higher than for a purely static experiment. In the case of the large deformations of SBR on the order of the initial film thickness, an influence of the substrate on the measurement of the Young’s modulus also cannot be ruled out. 14.4.3 Other Applications The ions in electrolytes cause potential gradients at surfaces [65]. The electrical double layer shields surface potentials and often gives rise to an ideally polarized layer [66]. This double layer exists on both surfaces, the tip and the sample. When the tip approaches the sample, the electrical double layer modifies the short-range interaction forces. The snap-in in the force–distance curve is changed depending on the configuration of the electrical double layers. Both the position of the snap-in and the size of the force-jump will vary. With analog PFM [67, 68] it is only possible to detect the size of the force jum; however, DPFM can give all quantities, when a postanalysis is performed. 14.4.4 Pulsed Force Mode and Friction Measurements At first sight it seems impossible to measure friction in PFM. Friction measurements [69,70] require an intimate contact between the tip and the sample, whereas the PFM measurement scheme tries to avoid that. However, it was realized, that a rapid lateral movement of the sample induces signals in the friction channel [22, 71]. The combined measurement of topography, adhesion, stiffness and friction is achieved with a PFM atomic force microscope where the sample is mounted on a shear piezo inducing movement lateral to the axis of a dashboard-type cantilever (Fig. 14.19). The PFM brings the tip 100–5000 times per second in contact with the sample. It is important that this frequency is far below the resonance frequency of the cantilever. Every time the tip comes into contact with the sample, the permanent highfrequency modulation of the lateral position induces the friction (lateral force) signal. This signals frequency ranges from several kilohertz up to several hundred kilohertz with amplitudes ranging from 0.1 nm to several nanometers. Either the frequency of the fast modulation is sufficiently far from the resonances of both the cantilever and the system, or it is exactly at one of these resonances. In the latter case, thorough knowledge of the entire measurement process is necessary. In most cases a frequency sweep on a test location is required to identify the resonances. If the lateral movement were induced with the scanning piezo, one would have more or less severe crosstalk of the modulation to the scanning. Furthermore, the power amplifiers for the scanning axis would need a much larger bandwidth than they had otherwise. As usual, the bending and deflection of the cantilever is detected with an optical lever detection scheme [2, 72, 73].
48
A.M. Gigler · O. Marti
Fig. 14.19. The experimental setup of the CODY mode. The sample is modulated along the horizontal axis by a high-frequency piezoelectric transducer while scanning under the principle of pulsed force mode scanning force microscopy
The amplitude and phase-sensitive detection of the cantilever response must be carried out by a two-phase lock-in amplifier (e. g., SR844 RF, Stanford Research Systems, CA, USA). Best suited are cantilever probes with a force constant of about 1 N/m and rather high resonance frequencies. Figure 14.20 shows a schematic overview of the principle of the combined dynamic mode AFM technique. In the force versus time diagrams, signal 1 is the normal force signal, signal 2 is the lateral force signal as detected and signals 3 and 4 are the amplitude and phase output from the lock-in amplifier, respectively. At the starting point of the pulsed-force cycle the tip is well above the sample surface. The PFM amplitude usually ranges from 10 to 500 nm and is kept con-
Fig. 14.20. An overview of the principle of the new scanning force microscopy mode during one pulsed force mode period. 1 The force signal, 2 the modulation induced by the oscillating sample, 3 the real and 4 the imaginary component of the lock-in amplifier output. The sample can be oscillated horizontally to measure friction or vertically to measure the dynamical compliance
14 Quantitative Measurement of Material Properties using DPFM
49
stant during the measurement. At this point no forces are acting on the cantilever (Fig. 14.20, left). Owing to the sinusoidal modulation, the cantilever moves closer to the surface. The tip snaps into contact because of the attractive force between the tip and the sample surface. A characteristic snap-on peak can be seen in the PFM signal (1) while the lateral force signal (2) and the lock-in signals (3 and 4) are still
Fig.14.21.CODY mode image of a micro-contact-printed silicon surface. A polydimethylsiloxane stamp was inked by aminosilanes—3-(aminopropyl)triethoxysilane. The molecules were then transferred to the silicon surface. The lateral modulation amplitude was 5.5 nm. The left column shows the results acquired at 150-kHz lateral modulation frequency. The center column was measured at 400-kHz lateral modulation frequency and the right column was measured at 700 kHz. The top row shows the pull-off force, mostly given by the adhesion. The middle row and the bottom row show the friction amplitude and phase, respectively
50
A.M. Gigler · O. Marti
at zero. Lateral forces due to the high-frequency modulation begin to act on the cantilever. The tip is pushed further into the surface, increasing the applied load, as observed in the PFM signal (1). Responding to the continuous lateral high-frequency excitation of the sample (2), the lock-in amplifier begins to output the values of the real (3) and imaginary (4) parts (Fig. 14.20 right). After reaching the user-selectable maximum of the applied normal force to the sample surface, the piezo moves back again, resulting in a decrease in the PFM signal (1) and the lateral force signal (2). Because of the time delays inherent in the lock-in-amplifier circuit, the real (3) and imaginary (4) portions of the detected lock-in signals reach a maximum and subsequently begin to decrease. As a consequence of the adhesive forces between the tip and sample, the PFM signal changes from repulsive to attractive. Finally, the tip loses contact and a free damped oscillation of the probe begins. As the resonance frequency decays, a new period begins. Figure 14.21 shows the correlation between adhesion and lateral force signals. The sample is micro-contact-printed with aminosilanes to create a surface with little topography and an adhesion and friction force contrast. All scan fields are 12.5 µm by 12.5 µm. While the adhesion signal is similar for all three frequencies, the imaginary part of the lateral force peaks at the circles. This indicates an increased loss at that frequency. The technique can be applied to measure dynamic stiffness too [22, 74–76]. 14.4.5 Cell Mechanics The measurement of local mechanical properties of cells utilizing an atomic force microscope [1,2,72,73,77] requires the use of small and controlled interaction forces while imaging the cell surface or acquiring force–distance curves [78–80]. To obtain images of entire living cells the intermittent contact mode [81,82] has been preferred. The phase lag in resonant intermittent mode imaging can provide information on the dissipative processes on the sample surfaces. On the other hand, the determination of surface force gradients by measuring force–distance curves while the cantilever is oscillating at its resonance [81] can provide data on the mechanical response of cells. The mechanical properties of cartilage hve been determined by force–distance curves and indentation [83]. However, the calibration of intermittent contact mode measurements in a liquid environment is possible, but quite demanding. As an alternative, nonoscillating force–distance curves [84] with repeat rates far below the resonance frequency of the cantilever can be performed, also known as force volume imaging. The evaluation and calibration of these force–distance curves is a commonly known procedure. Yet, to be practical for measuring living cells, the entire data acquisition has to be quick in order to capture the entire set of data before the cell may move or detach. Thus, the measurements need to be carried out at rather high repeat rates to minimize these unwanted effects. Increasing this rate to reasonable imaging speeds makes it necessary to switch from a triangular modulation of the cantilever position to a sinusoidal trajectory. This is done in PFM [12] or DPFM [23, 85].
14 Quantitative Measurement of Material Properties using DPFM
51
Fig. 14.22. Pulsed force mode image of a cell. The left image shows the topography and the right image shows the stiffness signal
Figure 14.22 shows an example of a cell measured by PFM. Shown here are some basic measurements which can be done by analog PFM or by DPFM measurements. DPFM allows a more detailed sample analysis [59, 60].
14.5 Summary PFM is one of the operating modes of a scanning force microscope which provides a virtually destruction-free imaging capability for delicate samples. PFM works similar to the jumping mode. Owing to the nonresonant measurement, one can analyze the force–distance curves in detail. This is either done by measuring at selected times during one period of the measurement cycle or by recording the entire curve at high speed. PFM imaging was and is successfully applied to polymer surfaces. It is rather gentle on the surface and gives data on stiffness, adhesion and similar quantities. Since the relevant time scale is between that of contact mode imaging and tapping mode imaging, one gains access to another range of the frequency-dependent viscoelastic properties of polymers. The nonresonant measurement principle makes PFM ideally suited for measurements in liquids. Electrical double layers and surface charges have been imaged successfully. Recently the technique was applied to the measurement of cells in buffer solution. The quantitative analysis of PFM data strongly depends on the thorough understanding of the tip–sample interaction. Therefore, we presented here a recipe for how to calibrate the tip and the microscope to measure reproducible values of the stiffness and the adhesion, where tips can be changed. Finally we gave some examples of results obtained with PFM. Acknowledgements. The authors gratefully acknowledge funding by the Landesstiftung BadenWürttemberg (Project B8, priority program “Hierarchical nanostructures” and the partial funding by the German Science Foundation (SFB 569, project C4). Some of the measurements would not have been possible without the support of WITec. The micro-contact-printed data were acquired by Peter Barth. The measurement of electrostatic forces was done by A. Rosa-Zeiser. Monika
52
A.M. Gigler · O. Marti
Asbach and Anne-Marie Saier prepared the samples. Some structures were made in collaboration with Stefan Walheim and Thomas Schimmel (Karlsruhe). The authors gratefully acknowledge discussions with Martin Pietralla, Bernd Heise, Gerd-Ingo Asbach, Bharat Bhushan, Armin Rosa-Zeiser, Peter Barth, Hanns-Georg Kilian, Olaf Hollricher and Joachim Koenen.
References 1. Binnig G, Quate CF, Gerber C (1986) Phys Rev Lett 56:930 2. Alexander S, Hellemans L, Marti O, Schneir J, Elings V, Hansma PK, Longmire M, Gurley J (1989) J Appl Phys 65:164 3. Binnig G, Gerber C, Stoll E, Albrecht TR, Quate CF (1987) Europhys Lett 3:1281 4. Blackman GS, Mate CM, Philpott MR (1990) Phys Rev Lett 65:2270 5. Majumdar A, Oden PI, Carrejo JP, Nagahara LA, Graham JJ, Alexander J (1992) Appl Phys Lett 61:2293 6. Jung TA, Moser A, Hug HJ, Brodbeck D, Hofer R, Hidber HR, Schwarz UD (1992) Ultramicroscopy 1446:42–44 7. Binnig G, Rohrer H (1982) Helv Phys Acta 55:726 8. Binnig G, Rohrer H, Gerber C, Weibel E (1982) Phys Rev Lett 49:57 9. Hartmann U (1991) Ultramicroscopy 59:42–44 10. Erlandsson R, Olsson L (1998) Appl Phys A 66:S879 11. Hansma PK, Cleveland JP, Radmacher M, Walters DA, Hillner PE, Bezanilla M, Fritz M, Vie D, Hansma HG, Prater CB, Massie J, Fukunaga L, Gurley J, Elings V (1994) Appl Phys Lett 64:1738 12. Rosa-Zeiser A, Weilandt E, Hild S, Marti O (1997) Meas Sci Technol 8:1333 13. De Pablo PJ, Colchero J, Gomez-Herrero J, Baro AM (1998) Appl Phys Lett 73:3300 14. Krotil H-U, Stifter T, Marti O (2000) Rev Sci Instrum 71:2765 15. Reynaud C, Sommer F, Quet C, El Bounia N, Tran Minh D (2000) Quantitative determination of Young’s modulus on a biphase polyamer system using atomic force microscopy. Surface and Interface Analysis, Vol. 30, Issue 1, pp 185–189, doi:10.1002/10969918(200008)30:1<185::AID-SIA862>3.0.CO;2-D 16. Stark RW, Drobek T, Heckl WM (1999) Appl Phys Lett 74:3296 17. Stark RW, Heckl WM (2003) Rev Sci Instrum 74:5111 18. Rodriguez T, Garcia R (2004) Appl Phys Lett 84:449 19. Proksch R (2006) Appl Phys Lett 89:113121 20. Marti O, Meinhard D, Rieger B, Schmatulla A (2007) Nanotechnology 18:044013(7pp) 21. Luginbühl R, Garrison MD, Overney RM, Weiss L, Schieferdecker H, Hild S, Ratner BD (2001) In: Castner DG, Grainger DW (eds) Fluorinated surfaces, coatings and films. American Chemical Society, Washington, p 187 22. Krotil H-U, Stifter T, Marti O (2000) Appl Phys Lett 77:3857 23. Spizig PM (2002) Dynamische Rasterkraftmikroskopie, in Experimental Physics. PhD Thesis, Ulm University, Ulm 24. Moreno-Herrero F, de Pablo PJ, Colchero J, Gomez-Herrero J, Baro AM (2000) Surf Sci 453:152 25. Bulychev SI, Alekhin VP, Shorshorov MK, Ternovskii AP, Shnyrev GD (1975) Ind Lab 41:1409 26. Bulychev SI, Alekhin VP, Shorshorov MK, Ternovskii AP (1976) Strength Mater:1084 27. Mizes HA, Loh K-G, Miller RJD, Ahuja SK, Grabowski EF (1991) Appl Phys Lett 59:2901 28. Pharr GM, Oliver WC, Brotzen FR (1992) J Mater Res 7:613 29. Boehnke U-C, Brodowsky HM, Groothues H, Kremer F (1999) J Phys Chem B 103:6741 30. Prokhorenko P, Rudnitsky V, Kren A (2004) Rev Quant Nondestruct Eval 23:1135
14 Quantitative Measurement of Material Properties using DPFM
53
31. Derjaguin BV, Landau L (1941) Acta Physicochim URSS 14:633 32. Stifter T, Marti O, Bhushan B (2000) Phys Rev B 62:13667 33. Grigg DA, Russell PE, Griffith JE (1992) Tip–sample forces in scanning probe microscopy in air and vacuum. J Vac Sci Technol A: Vacuum, Surfaces, and Films. July 1992, Vol. 10, Issue 4, pp 680–683, doi:10.1116/1.577709 34. Miyatani T, Okamoto S, Rosa A, Marti O, Fujihira M (1998) Appl Phys A 66:349 35. Johnson KL, Kendall K, Roberts AD (1971) Proc R Soc Lond Ser A 324:301 36. Derjaguin BV, Muller VM, Toporov YP (1980) J Colloid Interface Sci 73:293 37. Cohen SR, Neubauer G, McClelland GM (1990) J Vac Sci Technol A 8:3449 38. Landman U, Luedtke WD (1990) Science 248:454 39. Johnson KL (1992) In: Singer IL, Pollock HM (eds) Fundamentals of friction: macroscopic and microscopic processes. Kluwer, Dordrecht, p 589 40. Landman U, Luedtke WD, Ringer EM (1992) Wear 153:3 41. Hertz H (1881) J Reine Angew Math 92:156 42. Maugis D (2000) Contact, adhesion, and rupture of elastic solids. Springer, Berlin 43. Johnson KL (1996) Contact mechanics. Cambridge University Press, Cambridge 44. Sneddon IN (1965) Int J Eng Sci 3:47 45. Shorshorov MK, Bulychev SI, Aleklin VP (1981) Sov Phys Dokl 26:769 46. King RB (1987) Int J Solids Structures 23:1657 47. Oliver WC, Pharr GM (1992) Mater Res 7:1564 48. Lankarani HM, Nikravesh PE (1994) Continuous contact force models for impact analysis in multibody systems. Nonlinear Dynamics, Vol. 5, Number 2, März 1994, pp 193–207, doi:10.1007/BF00045676 49. Carpick RW, Ogletree DF, Salmeron M (1999) Colloid Interface Sci 211:395 50. Borodich FM, Keer LM, Korach CS (2003) Nanotechnology 14:803 51. Borodich FM, Keer LM (2004) Evaluation of elastic modulus of materials by adhesive (no-slip) nano-intendation. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 460, Issue 2042, pp 507–514, doi:10.1098/rspa.2003.1224 52. Franco AR Jr, Pintaúde G, Sinatora A, Pinedo CE (2004) Mater Res 7:483 53. Maugis D (1992) J Colloid Interface Sci 150:243 54. Dugdale DS (1960) J Mech Phys Solids 8:100 55. Johnson KL (1996) In: Bhushan B (ed) Micro/nanotribology and its applications. Kluwer, Dordrecht, pp 151–168 56. Johnson KL (1997) Proc R Soc Lond Ser A 453:163 57. Derjaguin BV, Muller VM, Toporov YP (1975) J Colloid Interface Sci 53:314 58. Muller VM, Yushenko VS, Derjaguin BV (1980) J Colloid Interface Sci 77:91 59. Holzwarth MJ, Gigler AM, Marti O (2006) Imaging Microsc 8:37 60. Gigler A, Holzwarth M, Marti O (2007) J Phys Conf Ser (in press) 61. Krausch G (1997) J Phys Condens Matter 9:7741 62. Ludwigs S, Krausch G, Reiter G, Losik M, Antonietti M, Schlaad H (2005) Macromolecules 38:7532 63. Walheim S, Boeltau M, Mlynek J, Krausch G, Steiner U (1997) Macromolecules 30:4995 64. Wang C, Krausch G, Geoghegan M (2001) Langmuir 17:6269 65. Israelachvili JN (1991) Intermolecular and surface forces. 2 edn. Academic, London 66. Zoltowski P (1998) J Electroanal Chem 443:149 67. Miyati T, Horii M, Rosa A, Fujihira M, Marti O (1997) Appl Phys Lett 71:2632 68. Miyatani T, Okamoto S, Rosa A, Marti O, Fujihira M (1998) Appl Phys A 66:S349 69. Mate CM, McClelland GM, Erlandsson R, Chiang S (1987) Phys Rev Lett 59:1942 70. Mate CM, McClelland GM, Erlandsson R, Chiang S (1987) Phys Rev Lett 59:1942 71. Krotil H-U, Stifter T, Marti O (2001) Rev Sci Instrum 72:150
54
A.M. Gigler · O. Marti
72. Marti O, Colchero J, Mlynek J (1990) Nanotechnology 1:141 73. Meyer G, Amer NM (1990) Appl Phys Lett 57:2089 74. Maivald P, Butt HJ, Gould SAC, Prater CB, Drake B, Gurley JA, Elings VB, Hansma PK (1991) Nanotechnology 2:103 75. Krotil H-U, Weilandt E, Stifter T, Marti O, Hild S (1999) Surf Interface Anal 27:341 76. Troyon M, Wang Z, Pastre D, Lei HN, Hazotte A (1997) Nanotechnology 8:163 77. Meyer G, Amer NM (1988) Appl Phys Lett 53:1045 78. Banerjea A, Smith JR, Ferrante J (1990) J Phys Condens Matter 2:8841 79. Weisenhorn AL, Khorsandi M, Kasas S, Gotzos V, Butt H-J (1992) Nature 80. Weisenhorn AL, Maivald P, Butt H-J, Hansma PK (1992) Phys Rev B 45:11226 81. Putman CAJ, Vanderwerf KO, Degrooth BG, Vanhulst NF, Greve J (1994) Biophys J 67:1749 82. Putman CAJ, van der Werf KO, de Grooth BG, van Hulst NF (1994) Appl Phys Lett 64:2454 83. Stolz M, Raiteri R, Daniels AU, VanLandingham MR, Baschong W, Aebi U (2004) Biophys J 86:3269 84. Nagao E, Dvorak JA (1998) Am Lab 30:36 85. Jacobs EP, Roux SP, van Reenen AJ, Morkel C, Meincken M (2006) J Membr Sci 276:8
15 Advances in SPMs for Investigation and Modification of Solid-Supported Monolayers Bruno Pignataro
Abstract. Improvement in nanotechnology research has been strongly driven by the advancement that nanoscale characterization and manipulation tools made from the beginning. Solidsupported monolayers (SSMs) represent one of the most important systems in nanotechnology. High-quality SSMs may be fabricated at low cost by solution processes employing techniques like self-assembly and Langmuir–Blodgett methods. In the last decade, advances in scanning probe methods provided more and more efficient tools able both to probe with largely improved resolution the morphological features of SSMs and to modify selectively their functional groups. The modified self-assembled monolayers can be used as nanotemplates for 3D nanofabrication. Key words: Self-assembled monolayers, Langmuir–Blodgett monolayers, Dynamic Scanning Force Microscopy, Scanning Probe Lithography
15.1 Introduction The research community is spending much effort in nanotechnology, since it is more and more revolutionizing our everyday life by providing solutions for the environment, health, energy and materials. Nanosystems are conventionally considered those that at least in one dimension measure roughly less than 100 nm [1]. Systems in this size regime have been found to show unique physical and chemical properties, leading to new laws, behaviour and phenomena which in turn open up the possibility of novel technological opportunities. Thus, for instance, in spite of their inertness noble metals may show important catalytic activity when nanodimensioned [2]; nanoclusters and nanowires as well as conjugated single molecules have shown singular electrical behaviour, including quantized conductance [1, 3–7]; nanosized systems may show have different colours [1] and lower melting points [1] than the bulk materials as well as new magnetic properties [8]; nanotribology in lubricated junctions [9] and the physics of nanofluids or nanojets [ [10] are other fields where the behaviour of “small” systems is different. Furthermore, in considering a nanosystem not only the size but also the shape is a fundamental factor as, for instance, the closed-double-shell atomic structure of Au55 clusters confers to the system significant inertness [11] or also the allotropic nanoscopic forms of carbon (nanotubes and fullerenes) exhibit remarkable hardness as well as tuneable electrical and optical properties [1, 12].
56
B. Pignataro
Among nanosystems ultrathin films aroused great interest, since they show unique properties owing to their confined thickness. For instance, two immiscible polymers may completely mix if restricted within a nanoscopic layer [13]; an adsorbed nanoscopic film may show unconventional wetting, like in the case of a few layers of water which have been observed to spread completely on hydrophobic surfaces in spite of the conventional nonwetting behaviour [14]; electrical transport through an ultrathin film may show quantized behaviour, including negative resonant tunnelling and Coulomb blockade effects [15, 16]. A special class of ultrathin films consists of solid-supported monolayers (SSMs). These systems are of outstanding fundamental interest but also they have a significant technological impact. For instance, the integration of organic monolayers with existing microelectronic technology, based on materials like silicon, silicon oxide, silicon nitride, aluminium and gold, opens up opportunities in a wide panorama of applications such as single-molecule electronics, optoelectronics, sensors and biosensors, nonlinear optics, molecular computers, microdevices or nanodevices such as microelectromechanical systems (MEMS) or nanoelectromechanical systems (NEMS) as well as in the area of scanning force microscopy (SFM) probes [17–20]. A more complete view of applications and technological needs related to SSMs is reported in Sect. 15.3. SSMs can be prepared in well-ordered forms using Langmuir–Blodgett (LB) and self-assembly methods. These allow for the fabrication of molecular surfaces with tailored terminal groups along with predesigned properties [17]. Once the SSM has been fabricated, its application requires methods able to characterize and in some cases for it to be able to modify itself with high spatial resolution. In the last decade, there have been important advances in scanning probe microscopies (SPMs) which notably improved the means available to investigate and modify SSMs. As to investigation, gentle and/or really noncontact dynamic SFM modes have been developed [21–25]. Dynamic SFM in the attractive regime (DSFMA) improved notably both the vertical and the lateral resolution of SSMs and individual molecules as well [21–25]. As to modification, SPMs have been used to inscribe structures inside or at the top of an organic monolayer, in a way similar to that we use to write with a pencil on a sheet of paper. Different nanopatterning SPM tools have been developed, including elimination, addition, substitution and constructive nanolithography methods [26– 29]. This chapter is intended just to give a focus on recent applications of SPMs on SSMs. In order to make more fluent the reading, first the structural characteristics, preparation methods, applications as well as the characterization and modification tools of SSMs are described. Then the latest advances of SPMs for imaging and modification of SSMs are reported in some detail. Conclusions with perspectives close the chapter.
15 Advances in SPMs for SSMs
57
15.2 SSMs and Their Preparation SSMs may be divided essentially into two classes depending on their interaction with the solid support: 1. Monolayers which are chemisorbed or covalently linked 2. Monolayers which are physisorbed or loosely bonded By confining the discussion to ordered SSMs the first class is mainly represented by the so-called self-assembled monolayers (SAMs), while LB monolayers are widely studied examples of the second class. 15.2.1 Self-Assembled Monolayers Chemisorption is commonly used to generate well-defined, tightly packed molecular structures that coat a surface uniformly. The resulting film is a SAM, i. e. a monolayer consisting of an ordered lattice of molecules on the surface of a solid substrate [17]. Both the structure of a single molecule and that of a SAM bound to a solid surface are schematized in Fig. 15.1. The structure of a molecule at a solid surface can be divided into three main regions: 1. The head group, whose nature is responsible for the sorption process. The bond formed between the molecule and the solid surface site is essentially covalent (tens of kilocalories per mole) in the case of organosilanes on silicon oxide or glass or hydroxilated silicon (–O–Si), alkylthiols on gold or other metals (–S–Au), and 1-alkenes or 1-alkynes on hydrogenated silicon (–C–Si). It is essentially ionic in nature in the case of carboxylic acids on silver or silver + oxide (–CO− 2 Ag ). 2. The chain, which is the skeleton of the molecule. The most common chains are those consisting of ten to 20 methylene units, but many studies (especially in the field of electronics) concern chains with conjugated units and/or etheroatoms. The attractive chain-to-chain van der Waals interactions (a few kilocalories per mole) play a major role in maintaining molecules highly packed within the monolayer (see also later for other important forces). 3. The terminal group of the molecule, which constitutes the real interface with the environment. For instance, highly hydrophobic and highly hydrophilic surfaces result, respectively, from methyl (–CH3 ) and hydroxyl (–OH) terminal SAMs. The preparation of SAMs exhibiting different chemistry is in principle not limited and SAMs with terminal groups like –(C=O)OCH3 , –O(C=O)CH3 , –O(C=O)CF3 , –O(C=O)C6 H5 , –COOH and –OSO3 H can be obtained. High-quality SAMs may be obtained spontaneously at low cost by solution processes (Fig. 15.2a), i. e. by the immersion of the substrate in a dilute solution of the active molecule. More expensive processes are those employing sublimation of the molecular system under vacuum with successive spontaneous adsorption at surface.
58
B. Pignataro
Fig. 15.1. a A molecule at a solid surface with its head, skeleton and terminal groups. b A selfassembled monolayer (SAM) represented with molecules in an all-trans conformation.
Depending on the systems involved in the process (molecule, substrate, solvent) the time scale of self-assembly can vary from a few minutes to some days. Widespread examples of SAMs are those described below in some detail, including systems which showed the highest stability and quality of packing. As a first class of systems, thiolated molecules have been used to realize SAMs on surfaces of gold or other metals like silver and copper. Depending on the chemical nature of the molecule and of the substrate, these systems experience a wide variety of packing phenomena and surface ordering as a result of the interplay between different potentials, including those of steric hindrance, dispersive forces, chemisorption and intermolecular elasticity. Thiol head groups (S–H) and gold substrates have been observed to work excellently, leading to highly compact monolayers with the tails
15 Advances in SPMs for SSMs
59
Fig. 15.2. Preparation methods of solid-supported monolayers (SSMs). a Self-assembly leads to the spontaneous formation of an ordered and compact monolayer covalently linked at a solid surface by a solution process. b The Langmuir–Blodgett (LB) method gives physisorbed monolayers with controlled pressure
pointing outwards from the surface. By using thiols with different tails and terminal groups, one can vary the resulting SAM function within a broad range. Alternatively, it is also possible to derivatize the SAM after self-assembly both by chemical and by physical methods (Sect. 15.3). By using a mixture of two differently terminated thiols, one can also prepare mixed SAMs, whose relative amount of functionalities can be modulated depending on parameters like mixing ratio, chain lengths, molecular solubility and nature of the surface or of terminal and head groups. The strong Au–S bond (about 44 kcal mol−1 ) together with the van der Waals forces between adjacent chains (1.4–1.8 kcal mol−1 for alkyl thiols) confers to the SAM a highly packed structure (for alkylthiols a molecular area of about 20 Å2 ) and longterm stability [30–32], the whole chemisorption energy being about −5 kcal mol−1 . Alkylthiols form typically an angle of about 30◦ with the surface normal and appear aligned with the alkyl chains parallel to each other in a nearly all-trans configuration. The order is nearly perfect at temperatures as low as −173 ◦ C, although even at room
60
B. Pignataro
temperature only a few gauche defects have been found [17]. Just for comparison with a different metal substrate, SAMs of alkylthiols on Ag(111) are more densely packed with the molecular chains almost perpendicular to the surface. A different class of molecules which form high-quality SAMs is constituted by organosilicon derivates. These systems have the general formulae RSiX3, R2SiX2 and R3SiX, with X typically a chlorine and R a chain that can bear different groups (methylene, amine, pyridyl, etc.). The Si-based head group reacts with hydroxylated surfaces like SiO2 or TiO2 via the Si–Cl bonds and the –OH groups of the surface. In the presence of traces of water a network of Si–O–Si is formed, with the molecules connected both to the surface and to each other by strong chemical bonds. For instance, a widely used alkyltrichlorosilane is octadecyltrichlorosilane (OTS), which shows a highly packed and highly hydrophobic (water contact angle 112°) structure with a small molecular tilt (less then 15°) with respect to the surface normal. Because of these features, SAMs of organosilicon derivatives are generally quite stable towards ageing, thermal treatment and chemicals. At variance with the abovementioned organosilanes, high-quality alkyl SAMs can be fabricated directly on silicon surfaces via a Si–C molecule–surface linkage, i. e. without involving oxide functions. This was first achieved using 1-alkenes (or 1-alkynes) and HF-etched substrates, as proposed by Linford and Chidsey [33]. These authors built up high-quality alkyl monolayers through radical-initiated reactions at a temperature of nearly 180 ◦ C. Then such a strategy was used for the fabrication of different organic monolayers on differently hydrogenated silicon faces and silicon-based materials [33–36]. The Si–C link leads to robust monolayers along with a large number of advantages: large and homogeneous atomically flat surfaces with densely packed molecules can be routinely prepared with reproducibility, and terminal variability; the true silicon–organic monolayer interface exhibits improved electrical contacts; high thermal and chemical stability even towards acids as strong as HF. The direct attachment of conjugated alkenes, alkynes and aromatics with silicon (Si–C bond) is intriguing and potentially very useful to avoid the insulating and electronically defective native silicon oxide interface. Just to give one example, by using 1-octadecene and hydrogenated Si(111), one can get a highly hydrophobic monolayer (water contact angle of 110 ◦ C) with molecules highly packed and tilted about 30° with respect to the substrate normal [33–36]. 15.2.2 LB Monolayers At variance with chemisorption, physisorption of molecules at surfaces arises from a small enthalpy change (a few kilocalories per mole). So a physisorbed molecule retains its identity, but it could be more easily distorted by the interaction with the surface. Physisorbed ordered monolayers can be prepared by using the LB technique [17]. This technique is generally limited to water-insoluble molecules since molecular monolayers are usually transferred from the air–water interfaces to a solid substrate (Fig. 15.2b). Typical systems are those of amphiphiles, although different compounds, including proteins, calixarenes, porphyrins and polymers, have been widely used.
15 Advances in SPMs for SSMs
61
In order to prepare a LB monolyer, first a known amount of molecules is diluted in a volatile solvent and spread drop by drop at the air–water interface of a Langmuir bath. After a few minutes the solvent evaporates. Then the monolayer is preorganized at the air–water interface by compressing two mobile barriers generally made of Teflon. The physical state of the monolayer at the air-water interface may be controlled by recording surface pressure versus moleculer area (π/A) isotherms. Indeed, the pressure–area isotherm provides information on the stability of the monolayer at water–air interface, the orientation of the molecules in the 2D system, and the phases and phase transitions involving expanded (low pressure, gaslike or liquid-like) or compact (high pressure, solid-like) systems [17,37]. Thus, at low pressure molecules typically are loosely arranged in a disordered or gaslike state. In this state one can apply the general rule πA = kT , which is reminiscent of the well-known equation of the perfect gas in three dimensions. The compact phase is generally characterized by a predominant attractive interaction between the molecular skeletons along with a short- and long-range order. The rupture or collapse of the monolayer occurs at high pressure when the attractive interactions are overcome by head-to-head repulsions and/or steric hindrance. Since the intermolecular and surface forces are a function of the experimental conditions, it is easily predictable that the state of a monolayer depends strongly on different parameters, including pressure, temperature, subphase pH and ionic strength [17,37,38]. The LB transfer generally occurs by transferring the monolayer onto a solid substrate which is oriented perpendicularly to the air–water interface (vertical deposition) [17], but several papers have been published reporting use of the Langmuir–Schaefer method with the substrate oriented parallel to the air–water interface (horizontal deposition) [37, 39]. As to the vertical transfer, the molecules can be deposited either with a upstroke (Z-type) or downstroke (X-type) operation by using a hydrophilic or a hydrophobic substrate, respectively. The transfer process can be repeated several times, thus leading to multilayer structures or 3D ordered thin films.
15.3 Fundamental and Technological Applications of SSMs SSMs allow modification of just the real surface of a material, tailoring its properties in a controlled fashion, thus providing solutions for interface problems. It should be quite easy to understand that interest in SSMs derives both from fundamental and technological purposes. As to the fundamental issues, the study of such systems allows the ability of molecules to self-organize to be inspected, the structure of the monolayer to be related to the properties of the interface, membrane behaviour to be mimicked, and phenomena connected to the molecule–substrate and molecule–molecule interactions to be discovered and studied, like wetting or dewetting, adhesion, friction and corrosion. On the other hand, applications of solid-supported organic monolayers cover a broad spectrum (Fig. 15.3). In particular, technological needs related to SSMs are in the fields of coatings for optical, chemical, electrical, adhesion, antismudge, antifouling, wear resistance, low
62
B. Pignataro
Fig. 15.3. Different technological applications which can involve SSMs. MEMS microelectromechanical systems, NEMS nanoelectromechanical systems
friction, antibacterial, and culture heritage applications; adaptive, sensitive and multifunctional systems; surfaces and processes for printed electronics, nanoelectronics and photonics; new advanced fabrication methods including patterning; nanotechnology for improving tools and efficient energy transfer; surfactants involving synthesis of polymeric nanoparticles; biomimetic surfaces; advanced probes for characterization of nanomaterials; stiction in NEMS and MEMS; chemical force microscopy, etc. Just to focus on the importance that an organic monolayer can have for the development of devices, Fig. 15.4 reports some representative cases. Figure 15.4 a shows the scheme of a typical organic thin-film transistor in the bottom-gate top-contacts configuration. The arrows mark the interfaces which typically are responsible for the good performance of organic thin-film transistors. SAMs could be used to improve the electric contact between the metal electrodes and the organic layer as well as to act as adhesion promoters or electrical screening agents for dipoles at the dielectric–organic interface. In the field of electronics the possibility to generate organic monolayers by solution processes is a further advantage, since it is possible to develop processes at low cost and in most cases for large-scale production. As an example, plastic electronics is an emerging field which is expected to conquer in the near future 30% of the whole market of electronics. In order to fabricate a plastic-supported device, it is necessary to develop a process at low temperature (below about 300 ◦ C) and organic coatings; an organic interface as well as functional organic layers would provide the best solutions [40]. Figure 15.4b shows a mechanical sensor made of multiple levers functionalized by SAMs of thiolated single-strand DNA segments [41]. Each lever deflects following the hybridization
15 Advances in SPMs for SSMs
63
Fig. 15.4. Specific examples of technological applications involving SSMs: a An organic fieldeffect transistor and the interfaces improved by SSMs. b A cantilever sensor array covered by thiolated-DNA SAMs for evaluating DNA hybridization along with treatment response efficacy for personalized medical diagnostics. c A multifunctional nanoparticle supporting different organic and biological ligands. (b Reprinted from Zhang et al. [41] with permission from Macmillan Publishers, copyright 2006)
process. Finally Fig. 15.4c gives a sketch of a multifunctional nanoparticle. This is typically composed of an inorganic core with its peculiar properties and of a shell of different organic or biological self-assembled molecules [42]. As a further very important issue, these systems are among the best candidates for miniaturization at the nanometre level. The miniaturization of the active and passive components for the construction of useful devices and machines is currently pursued by the so-called top-down approach which includes lithography techniques using beams of photons, neutrals, ions or electrons. These tools have become available for structuring substrates at the 10-nm level [43–45]. The bottom-up strategy to fabricate devices is an alternative and promising strategy towards the technology on the nanometre scale. Such a strategy starts from atoms or molecules to build up nanostructures like in the case of molecular monolayers which are typically systems from a fraction of a nanometre to a few nanometres thick. The combination of the two approaches, bottom-up for monolayer fabrication and top-down for lateral miniaturization, has led to nanostructured laterally ordered monolayers. The achievement of laterally ordered monolayers has been obtained only in few cases by exploiting self-organization processes alone, such as, for instance, by using instability phenomena during the transfer of a LB film [46, 47]. The use of SSMs and especially of SAMs as ultrathin resists for patterning surfaces with sub-10-nm lateral dimensions
64
B. Pignataro
is extremely convenient since these systems exhibit low defect density and their terminal functionality can be controlled in a precise fashion. Patterned SAMs can be employed as biomolecule templates for biochip fabrication: by arranging SAMs terminated with different groups, biomolecules can spontaneously organize on preferential sites of the pattern [26]. Furthermore, nanometric biomolecular arrays may enable high-throughput screening of biomolecules at a single-molecule level. Such arrays with a precise control of position and orientation of individual molecules may become a powerful tool for studying multivalent and multicomponent molecular interactions in biological systems, which are of fundamental interest for developing biosensors, DNA chips, etc. Seemingly nanopatterned SAMs could constitute one way for interfacing molecular wires, conducting polymers and metal nanoparticles directly with materials/surfaces typically used in current technology which make use mainly of silicon-based materials. This would be very important to solve a main challenge in nanoelectronics which consists of providing a means of hardwiring with integrated circuitry, which is the planned assembly of surface-immobilized active nanocomponents and their effective electrical connection to addressable contact electrodes. In the frame of emerging technologies the preparation of SSMs may constitute also a first step for the fabrication of complex 3D supramolecular architectures as layers, membranes and multilayer films with predeterminated physicochemical properties [47] which again can be used as key components for the building up of devices.
15.4 Characterization and Modification of SSMs In parallel to the preparation techniques, more and more sophisticated characterization and modification techniques have been developed, allowing for the resolution of a number of problems connected with their preparation, structure, properties and applications. SPMs are today among the most important methods to characterize SSMs. In addition SPM-related techniques are continuously being developed to modify SSMs once they have been prepared. 15.4.1 Characterization of SSMs Today, a large number of techniques allow the investigation of the structure and the properties of SSMs. Among those one can distinguish techniques (conventional spectroscopies and microscopies) which “make use of the sense of sight” and techniques which “make use of the sense of feeling” (Fig. 15.5). The first have been classified in analogy with the human eye, which reveals the light coming from an illuminated sample, as accordingly by these techniques the sample is irradiated with a source (light, electrons, ions, etc.) and the radiation or particles coming from it are revealed by a detector. Depending on the nature of the particle detected, the sampling depth generally varies from some angstroms (ions and neutrals) to some nanometres (electrons)
15 Advances in SPMs for SSMs
65
Fig. 15.5. Pictorial sketches as an analogy of how spectroscopies and microscopies “which make use of eyes” work in comparison with techniques “which make use of feeling” like scanning probe microscopies (SPMs)
up to hundreds of nanometres or microns (photons). Figure 15.6 reports a picture where sampling depths, relevant surface properties and emitted secondary particles are compared. Very sensitive and valuable for studying the elemental and molecular composition of SSMs are real surface tools like X-ray photoelectron spectroscopy (XPS) [50] and time of flight (TOF) secondary ion mass spectrometry (SIMS) [51]. Thanks to their scanning capacity and low sampling depth (electrons for XPS and ions for SIMS), these techniques allow for surface chemical imaging with submicrometric resolution. Also ion scattering techniques or thermal atom scattering [52, 53] have been used to get spatially resolved chemical information. There are anyway a number of experimental approaches with reduced sampling depth using photon-sensitive techniques like, for instance, surface-enhanched Raman scattering [54], scanning near field optical microscopy [55] and grazing-angle spectroscopies and diffractometers [17]. Among those, Fourier transform IR spectroscopy, in both grazing-angle and attenuated total reflection [56] modes, have been used to learn about the direction of transition dipoles, and to evaluate dichroic ratios, molecular orientation, packing and coverage of SSMs. For instance, in the case of alkyl chains the C–H stretching vibrations are very sensitive to packing density and to the presence of gauche defects. The related signals are then ideally suited as probes to determine the SAM quality. Also, the antisymmetric CH2 stretching vibration at approximately 2918 cm−1 is a useful probe, since its position varies from 2916 cm−1 for SAMs cooled below room temperature to 2918 cm−1 forhigh-quality SAMs, up to about 2926 cm−1 for heavily disordered “spaghetti-like” systems. Ellipsometry [1] is used to measure the film thickness and uniformity. For example, in the case of ˚ can be evaluated, which is in hexadecanethiolate on gold a thickness of 21±1 A ˚ length are tilted 30◦ . agreement with a picture where all-trans molecules of 24.5-A The use of lasers [53] as well as the increasing availability of synchrotron radiation [57] increased considerably the characterization capabilities. Many other very useful techniques are electron paramagnetic resonance [58], electron Auger spectroscopy [59], surface plasmon microscopy [60], scanning electron microscopy [61] and surface second-harmonic-generation spectroscopy [62].
66
B. Pignataro
Fig. 15.6. Comparison of surface properties and sampling depths of different probes, including electrons, ions, photons and scanning probe microscope tips. SIMS secondary ion mass spectrometry, XPS X-ray photoelectron spectroscopy
At variance with the techniques reported above, the techniques which “make use of the sense of feeling” probe the ultimate surface and the surface properties. Surface potential measurements [63] provide information on the coherence of a molecular surface on metal surfaces. The contact-angle technique [17] involves the direct contact of a drop of liquid with the surface. Depending on adhesion forces, the drop can wet or dewet the surface. By using different liquids, one can obtain quantitative information on the wetting properties and the surface free energy, although the lateral resolution is obviously limited at the microscale.
15 Advances in SPMs for SSMs
67
Finally, SPMs make use of feeling with a lateral nanoscopic resolution, providing direct and impressive information to investigate SSMs. In this case the analogy may be made with a man who closes his eyes and puts his finger on the sample. SPMs use a nanoscopic finger (tip) to probe the morphological, mechanical,and also electrical and magnetic properties of the surface of a sample. The applications of SPMs to SSMs are reported in Sect. 15.5. 15.4.2 Modification of SSMs As mentioned earlier in this chapter, once a SSM has been prepared, one sometimes wants to modify it. A distinction similar to that we made for the characterization techniques can be made in the case of modification techniques and in particular for patterning of SSMs. This in fact can be performed by radiative sourcing or by mechanical touching. Radiative methods include UV and electron-beam sources which can modify the chemical functionalities of a SAM [64,65]. A different strategy consists of generating a prepattern with lithographic techniques (UV, ions, electrons, etc.) of a substrate followed by self-assembly of the monolayer [66]. Otherwise one can make bottomup patterning by using SAM-based multilayers to bridge gaps between previously constructed systems [67]. Mechanical touching is referred to “soft lithography” [68, 69]. Among the soft lithography techniques one can cite microcontact printing. This technique generally allows SAMs to be patterned at the micrometre scale, but several extensions of it offered also nanometre-scale resolution. Nanoscale patterns have been obtained more easily by employing nanoimprint lithography [70, 71]. In this scenario, the use of SPM-related modification tools, more generally called scanning probe lithographies (SPLs), provides important means to modify a SSM. SPLs offer more general approaches and in most cases the best resolution. In Sect. 15.6) we discuss how a SAM can be modified with molecular-group definition by SPLs.
15.5 Latest Advances in SPMs and Applications for Imaging of SSMs Scanning tunnelling microscopy (STM), invented in 1982, was the first near-field probe microscopy to be used [72, 73]. It makes use of a metal tip which scans the surface to be probed. The tunnelling current is set as a feedback signal to maintain the tip–sample distance constant. In order to scan the sample surface, the instrument moves the tip over the sample (or vice versa) along the x- and y-axes typically by a piezoelectric scanner. The scanner reacts along the z-direction in order to maintain constant the tunnelling current throughout the tip and the sample. By the scanner movement, the computer reconstructs a false-colour map which reproduces the topography with a resolution that can be as high the angstrom scale. If the feedback is switched off the microscope works in the constantdistance mode, thus allowing point the electrical properties of the surface to be
68
B. Pignataro
inspected point by point and the map of the local density of states to be reconstructed. The application of STM to SSMs is limited to conductive substrates which ensure the charge flux. Typical examples of SSMs investigated by STM are those of thiols on Au(111) [73]. For this system the lateral resolution is generally very high. Thus, ˚ scale in mixed for instance, STM allowed the formation of islands on the 20–50-A CH3 /CO2 CH3 SAMs to be revealed or the observation of phase separation on the shell of a ligand-coated nanoparticle [42, 73]. This technique has also been widely used to study in situ adsorbed molecules prepared by evaporation under a vacuum [73]. On the other hand, for these systems STM provides important information beyond topography, and, for instance, several studies have been done dealing with electrical transport through metal-supported molecules [73, 74]. However, although STM has been miniaturized, multiplexed and adapted to many environments, there have been few major innovations since its invention and the restriction to metal substrates limits its applicability to SSMs. The same cannot be said for atomic force microscopy (AFM), also called scanning force microscopy (SFM), invented in 1986 [75]. It was rapidly introduced in the research community as a breakthrough for investigating and manipulating organic and biological samples. The principle which drives SFM is quite different from that for STM. SFM works by employing as feedback the tip-sample force and thus does require the tip or the sample to be conductive; therefore, it allows virtually every kind of material to be imaged. A major problem of this technique is that in the case of soft samples like organic monolayers and biological samples damage can be induced by the tip–sample contact. This problem was crucial for the first generation of scanning force microscopes working in contact mode. Indeed, this mode operates by putting the tip in static contact with the sample, and thus soft samples or samples loosely bound to the surface can be manipulated by the scanning tip, resulting in a failed image. This, even if in some cases contact mode is used, can improve information as in the cases of chemical force microscopy [76] or friction [77] and force-distance curve maps [78]. Nevertheless, owing to the abovementioned problems, contact mode has been overtaken by dynamic modes like the so-called tapping mode [21–25], at least as far as imaging of soft samples is concerned. Indeed, tapping mode operates by putting an oscillating tip in intermittent contact with the sample, and this limits the time of contact and lateral dragging. Tapping mode exercises in air normal forces of tens of nanonewtons and this can still cause contact effects [24, 25]. Such tip pressure mainly arises from capillary forces which are set up between the tip and the sample. By working under water it is possible to avoid the capillary effects, thus operating with forces in the piconewton scale [79] and improving image quality as well as resolution. Also better resolution has been obtained by changing the chemical nature of tips along with their aspect ratio. Innovative probes like carbon nanotubes [80] and electron beam deposited carbon supertips [81] have been introduced and are still being continuously developed. However, working under a liquid and using special tips is not always easy and convenient. Imaging in air with conventional tips is cheaper, easier and speedier. In addition an air environment could be necessary when small quantities of samples are available, or the
15 Advances in SPMs for SSMs
69
molecule–substrate adhesion is weak [82], as well as for different technological needs [18, 83]. Confining the discussion to SSMs and/or individual molecules, as mentioned above, tapping mode in air may result in low image quality. A widely observed feature is that molecules exhibit lower heights than expected, giving a scarce contrast. This can be due to different factors such as different probe damping/energy dissipations, since the tip oscillates on the monolayer or otherwise on the substrate [25,84], and/or tip indentation inducing sample deformation [24,85]. In some cases height anomalies can be ascribed to molecular adhesion with the substrate, as well as contamination, dehydration or humidity effects [86]. It may be intuitive that by reducing the tip–sample forces height anomalies can be minimized. This in particular would limit damping/energy dissipation effects and tip-induced deformation as well. Following this strategy it is just in this last decade that a few groups have had success in improving the resolution of the dynamic modes for soft samples, including SSMs [21–25]. By operating in the conventional dynamic mode (i.e tapping mode), the tip propagates at high amplitude through large parts of the tip–sample interaction potential in a single oscillation cycle, i. e., through the attractive and repulsive interaction regions. The level of tip–sample interaction force can be minimized by changing the driving amplitude (amplitude of the piezoelectric actuator vibrating the cantilever) or the set-point ratio rsp = Asp /A0 (Asp is the cantilever oscillation amplitude to be maintained during imaging by the feedback loop; A0 is the cantilever free oscillation amplitude in air). In particular by increasing rsp , the tip works on average further from the surface and as a consequence the interaction force is reduced. The correlation between rsp and the level of interaction force cannot be easily quantitatively estimated since it depends on several difficult-to-control factors, including tip–sample effective elastic modulus, tip structure and sample contamination. Thus, in discussing the level of interaction forces, the rsp has to be considered as a relative parameter in the particular experiment. 15.5.1 Dynamic SFM in the Attractive Regime It has been demonstrated [21–25] that it is possible to limit the tip–sample interaction. The instrument was in some cases put exclusively and stably in the attractive regime. This allowed working in a real noncontact mode by dynamic SFM in the attractive regime (DSFMA) [21–25]. In principle this would be possible by working with every SFM instrument, but it could, however, require a lot of effort and time along with the choice of the best tips. A way to reach the noncontact condition straightforwardly consists of increasing the quality factor Q of the oscillating probe (Q = f/Δ f , with f denoting the cantilever resonance frequency and Δ f the full width at half maximum of the resonance peak). Some years ago, Ancycowski et al. [22] invented a special feedback circuit (called Q-control) which allows the range of the attractive interaction regime to be controlled and enhanced by increasing Q up to 1 order of magnitude.
70
B. Pignataro
The whole electronic scheme proposed by these authors is reported in Fig. 15.7 including the extra Q-control circuit. This basically includes a variable phase shifter and an amplifier with adjustable gain. Such a modification of the electronics implies that the equation of motion of the free cantilever–tip ensemble is represented by the formula m eff z(t) ¨ −
m eff w0 z(t) ˙ − Kz(t) = F0 cos(wt) + Geiφ z(t) , Q eff
(15.1)
where G is the gain of the amplifier, φ is the phase shift, m eff is the effective mass, K is the force constant, F0 is the amplitude and Q eff and w0 are the effective quality factor and the resonance frequency of the cantilever–tip ensemble, respectively. The cantilever motion can be described as almost sinusoidal even in hard tapping and z(t) ≈ A sin(wt) .
(15.2)
Therefore, by choosing an apposite phase shift (φ = ±π/2), we get e±iπ/2 qz(t) = 1/wz(t) ˙
(15.3)
Fig. 15.7. Dynamic scanning force microscopy (SFM) modes. a The main features of the instrument are sketched together with the Q-control feedback which allows for the achievement of the net-attractive regime. b Phase-lag signals induced by contact effects in conventional tapping mode
15 Advances in SPMs for SSMs
71
and the equation of motion is m eff z(t) ¨ − αeff z(t) ˙ − Kz(t) = F0 cos(wt) + Geiφ z(t) ,
(15.4)
where αeff = mQeffeffw0 = α ± 1ω G is the effective damping constant. As a consequence, by tuning the gain it is possible to vary the effective damping and thus to regulate the effective quality factor [22]. With use of this special feedback improved resolution and image quality have been obtained in comparison with what was obtianed the conventional tapping mode by using common substrates, tips and structurally well defined soft nanostructures [21–25, 79]. 15.5.2 Dynamic SFM at Different Level of Interaction Forces As the attractive condition is reached, it is then possible to switch the system deliberately from the net-attractive to the repulsive regime and vice versa by properly modulating either the drive amplitude or set-point, or both (see earlier). Phase images can help to monitor the level of tip–sample interaction forces since (1) in a real noncontact condition (net-attractive regime) it does not show any contrast and (b) in intermitted contact phase lag is sensitive to different properties, including sample surface free energy and hardness, whose relative contribution depends on the level of tip pressure [25]. According to the above considerations, high-amplitude dynamic modes allow one to work within a wide spectrum of tip–sample interaction forces by switching between: 1. The high-amplitude mode running in the repulsive regime. This is the tapping mode, i. e. the tip oscillates with high amplitude (10–100 nm) through large parts of the interaction potential, i. e., through the attractive and repulsive interaction regions. 2. The high-amplitude mode in the attractive regime (DSFMA). At variance with tapping, the tip can be oscillated at high amplitude (10–100 nm) without touching the sample. By combining these two modes, one can furthermore get mechanical information with single-molecule resolution. One can first image the molecules in the attractive regime, then deform them by switching to intermittent repulsive contact, and finally observe the damage by scanning in the attractive regime again. By doing this it was suggested that in the case of DNA fragments and streptavidin proteins the singlemolecule mechanical response under the pressure of the scanning force microscope tip depends on the intra-molecular organization and that molecular “squeezing” is crucial in considering the observed height reduction. In particular Fig. 15.8 reports on two loading–unloading curves performed on single DNA and single streptavidin molecules, respectively. The molecules were first imaged in the attractive regime, then in the repulsive regime at increasing loads (see on the right of the dashed line), and finally again in the attractive regime. The switching from the attractive
72
B. Pignataro
Fig. 15.8. Single-molecule mechanics experiments on a double-strand DNA segment and a streptavidin (STV) protein by exploiting dynamic SFM working both in the attractive and the repulsive regime. DNA behaves as a rod-like pseudo-elastic system, while STV is plastically deformed by the pressure of the tip pressure. (Adapted with permission from Pignataro et al. [24], copyright Springer-Verlag 2002)
to the repulsive regime caused for both molecules a reduction in the height, which had a major extent at larger forces. Coming back to the attractive regime (noncontact) only DNA showed its original height (pseudo-elastic behaviour), whereas streptavidin remained plastically deformed [24]. The above scenario could stimulate the reader to address different questions: What do we really observe by using the cantilever amplitude to sense the surface: topography or mechanical properties (including adhesion)? How important is the level of tip–sample interaction? Do we modify the molecular structure during scanning, avoiding correct information? Chi [21] showed that the introduction of dynamic SFM modes increases the range of detectable surfaces and surface properties of soft materials. This author highlighted that it is fundamental to choose the appropriate experimental conditions. Thus, force modulation microscopy (low-amplitude oscillating tip in constant contact with the surface) of LB monolayers was found to provide information on local viscoelasticity in addition to molecular packing and long-range orientation order. Also in the same paper it was shown that the damage to a gold cluster monolayer is effectively prevented and increased resolution is achieved by employing DSFMA [21]. High-amplitude dynamic modes have been successively applied to LB monolayers with laterally defined nanostructures, which constitute an optimal framework to explore the advantages/limitations of employing different levels of tip–sample interaction forces. In particular thermodynamics conditions by mixing quercetin-3-Opalmitate (QP) [87] with dimyristoylphosphatidylcholine (DMPC), a characteristic
15 Advances in SPMs for SSMs
73
phase separation process leads to the formation of QP fibrelike structures. These fibres wrap up, forming self-organized spiral domains within the DMPC-rich matrix [25, 87]. These structures have been imaged by three dynamic regimes [25]: (1) netattractive regime; (2) light tapping (rsp = 0.9–0.8); (3) moderate tapping (rsp = 0.5-0.7). The application of the net-attractive regime resulted in the observation of QP wrapped fibres about 15 nm in lateral size and 0.6 ± 0.2 nm raised with respect to the DMPC-rich plane. Light tapping gave a reduced image quality and contrast since the domains appear only 0.3 ± 0.1 nm raised with respect to the plane. Moderate repulsive tapping results in a poor contrast because of a further reduced height of the structures (0.2 ± 0.1 nm). The better contrast and resolution obtained by means of the attractive regime have been evidenced by the possibility to explore more structural details than those observed in tapping mode. Figure 15.9 shows the dynamic SFM images of circular structures made of fibres wrapped in circular assemblies. DSFMA allowed fibre fractures (about 20 nm in size) with an angular periodicity of 60° to observed. Such features cannot generally be observed in tapping mode. The different resolution achieved was related to the level of interaction of the SFM tip with the two phases. At variance with the work from the abovementioned experiments [21, 24] in this case the tip–sample interaction did not alter the LB structures, as proved by sequentially recording in the same sample region several images with the system working alternately in DSFMA and in tapping mode. Figures 15.10 and 15.11 report on image experiments made on the same sample at different levels of interaction forces. By means of hard tapping (driving amplitude
Fig. 15.9. A circular feature obtained by phase separation in LB quercetin-3-O-palmitate (QP)– dimyristoylphosphatidylcholine (DMPC) mixed monolayers. The ring consists of wrapped QP fibres and dynamic SFM in the attractive regime allows the visualization of ruptures every 60◦ , which otherwise cannot be clearly observed by tapping mode. (Reprinted with permission from Pignataro et al. [87], copyright 2003 American Chemical Society)
74
B. Pignataro
Fig.15.10.Investigation of spiral-like domains in mixed QP–DMPC LB monolayers by employing dynamic SFM at different levels of interaction forces (z-scale 4 nm). a Light tapping (rsp = 0.8). b Moderate tapping (rsp = 0.65). c Net-attractive regime. d Zoom-in of the circular structure in c (upper part) and section analysis along the segment in c (lower part). (Reprinted with permission from Pignataro et al. [25], copyright IOP Publishing 2003)
120 mV, response about 0.3) a reverse contrast was obtained, with the QP domains lowered by about 0.4 nm. In Fig. 15.11b the same region has been imaged immediately after in the net-attractive regime (40 mV, response about 0.3, no contrast in phase mode). The QP domains are well visible, and 0.5 nm taller than the surrounding plane. The characteristic ruptures along the fibres are also visible. By increasing the driving amplitude (200 mV) hard tapping gave again lower fibres with the fine structures completely lost. A reversible behaviour was obtained by switching once more to the attractive regime (40 mV). After each experimental “loading–unloading” trial the structures were unaltered without any damage.
15 Advances in SPMs for SSMs
75
Fig. 15.11. Investigation of spiral-like domains in mixed QP–DMPC LB monolayers by using dynamic SFM operating alternately in hard tapping and in the net-attractive regime by the Qcontrol system. a, c Hard tapping using rsp ≈ 0.3 and drive amplitudes of 120 and 200 mV, respectively. b, d Net-attractive regime using rsp ≈ 0.3 and a drive amplitude of 40 mV. Z-scale 2 nm. (Reprinted with permission from Pignataro et al. [25], copyright IOP Publishing 2003)
This fact indicated that at variance with proteins [23, 24] (see earlier), gold clusters [21], etc., in these samples the height lowering and the loss of resolution cannot be ascribed to plastic deformation, but more likely to local tip–sample response (adhesion, indentation), which differs by moving the tip from the QP domains to the surrounding matrix. Furthermore, the extent of the contribution of factors such as adhesion or indentation depends on the tip force. The reverse contrast observed in the height images by hard tapping was interpreted by considering a change of the effective resonance frequency of the cantilever weff . This would be lowered by attractive forces and raised by repulsive forces [88]. Thus, assuming that in light tapping mode the attraction forces prevail, the driving frequency w0 would be larger than weff either on the spiral domains or on the surrounding phase; thus, the cantilever amplitude measured at w0 would be larger on the stiffer (or less adhesive) sample owing to the larger operative repulsion (or smaller adhesion). Also, the softer (or more adhesive) regions would look brighter, given that the feedback requires a larger downward sample displacement to maintain constant the set point. The situation is reversed if the microscope operates in hard
76
B. Pignataro
tapping mode. This hypothesis has been corroborated by the images obtained in force modulation and phase-lag modes and with the results obtained by Magonov et al. [88]. Recently the Q-control system was applied under a liquid [79]. Improved resolution along with higher features were observed in comparison with conventional instrumentation by investigating LB dipalmitoylphosphatidylcholine (DPPC) stripes supported by mica in decane. The authors observed that the topographical height differences of DPPC features measured with conventional amplitude-modulated AFM were lowered by several angstroms compared with those measured with the Q-control. In the same wotk, similar features were observed by looking at solidsupported DNA fragments. It was certainly interesting to understand whether or not imaging was performed in the attractive or the repulsive regime. In contrast to experiments done in ambient conditions, the authors did not observe an instability which marks the transition between the “repulsive” and the “attractive” regime. Nonetheless, since adhesion forces are typically significantly reduced in liquids, they assumed that the net tip–sample interaction forces were repulsive, but significantly reduced with Q-control.
15.6 Nanopatterning by SPMs and SSMs Scanning a tip to create a pattern over a SSM has generally been applied to SAMs. LB films are indeed scarcely used as nanopatterning “resists” owing to their mechanical lability from tip scanning and their instability upon ageing. The LB method offers, however, the possibility to realize laterally ordered features only by exploiting self-organization processes upon transfer [46, 47]. SAMs have more attractive features for nanopatterning by SPMs since they have and hold strongly anchored functional molecules generally highly packed in planelike structures with tailored head groups. It was natural to marry the manipulation capabilities of scanning probe microscopes with the chemical definition and prospect for realizing patterns offered by SAMs. In the last decade, different nanopatterning methods have been developed which use SPM and SSMs. These include “dip-pen nanolithography” (DPN; addition nanolithography), STM-based lithography (elimination nanolithography) and SFM nanoelectrochemistry methods (constructive nanolithography) [28,29,47,89–115]. These techniques allow for patterning at nanometric resolution and are, in general, grouped under the family of SPL. 15.6.1 Addition Nanolithography The term “addition nanolithography” was introduced to indicate the possibility of adding molecules to a surface in a controlled matter and with nanoscale resolution. There are essentially two different ways to perform addition nanolithography. The first one consists of using a nanopipet. A broad range of materials have been deposited by nanopipets [116], including DNA, enzymes, proteins, photoresists and gases. The
15 Advances in SPMs for SSMs
77
second one consists of using an atomic force microscope tip. This method is called dip-pen nanolithography (DPN). It allows the transfer of molecules on a substrate in a similar way that a pencil transfers the ink onpaper. Figure 15.12 illustrates the general procedure of DPN. In a typical experiment an atomic force microscope tip is inked with a molecular system which is known to self-assemble on the solid substrate. The molecules are then transferred from the tip to the solid surface via the water meniscus formed between the tip and the sample. The tip is put in contact with the substrate at a point or moved at a certain speed to write a pattern. By consideration of this scheme, it could be intuitive that the transfer process is dependent on different factors, including the humidity conditions, the dynamics of the tip motion as well as the nature of the substrate, the tip and
Fig. 15.12. The dip-pen nanolithography method illustrated as an addition-lithography technique exploiting a water meniscus to deposit molecules from the tip to the substrate (a, b). Dip-pen nanolithography is followed by molecular self-assemblies (c, d) to build up a 3D nanopatterned system
78
B. Pignataro
molecules. In particular by increasing humidity, one expects the features size i to generally increase because of an enlargement of the water meniscus. Also, the rate of scanning or the time of tip–sample contact is a fundamental factor. Longer contact times or slower scan rates generate larger features. Of course, the nature of the tip, molecules and samples is also crucial since it should allow for the easy detachment of molecules from the tip along with the molecular self-assembly at the surface. This technique was pioneered by using thiols on gold since these systems spontaneously self-assemble on gold, thus giving rise to stable structures (Sect 15.2). As an example, 16-mercaptohexadecanoic acid (MHA) was written on gold and successively used as a mask for assembling other systems, including molecules and nanostructures. To make this the patterned surface was immersed in a solution of 1-octadecanethiol (ODT) for 30 s. ODT self-assembled on the free gold region, thus filling in the empty spaces. By employing the MHA–ODT pattern, Fe3 O4 magnetic nanostructures adsorbed preferentially on MHA, leading to a coherent feature of nanosystems with the patterned surface [117]. This is just an example to show the power of combining SPL with SAM for the construction of ordered 3D systems (Sect. 15.6.4). Several other molecular systems and substrates have been used with DPN allowing for the realization of nanostructures like dots and lines with a lateral resolution of some tens of nanometres and a height of about 10 nm. For instance, disilazane molecules have been patterned on semiconductor (Si, AsGa) substrates [118]. Since these systems do not polymerize in the water meniscus they were preferred to trichlorosilane, which is largely used for self-assembly on these substrates (Sect. 15.2). However, these systems showed a slower patterning rate. Oligonucleotides have been patterned by DPN on Si and on Au (for the latter case various systems have been functionalized by a thiol linker). Such patterns showed biological activity by using complementary fluorophore-labelled DNA [119]. Also proteins including collagen, immunoglobulin G, lysozome, etc. could be inscribed on Au and Si/SiOx by DPN. The possibility to write with these systems indicates that DPN can also be applied by exploiting molecule–substrate noncovalent electrostatic interactions. This allowed a series of systems to be deposited by DPN, including dyes [120], dendrimers [121], negatively charged and positively charged polymers [122], metal nanostructures [123], porphyrins [124] and conjugated molecules [125, 126]. 15.6.2 Elimination and Substitution Nanolithography The term “elimination nanolithography” derives from the idea of detaching molecules from a surface to generate a pattern. The molecules that are removed can be in a second step replaced by a different molecular system which binds to the surface, thus leading to so-called substitution nanolithography. In principle one can perform elimination and substitution in situ or ex situ by a number of techniques and systems, but in order to achieve the best lateral and vertical resolution this strategy is very successful by using SPMs and SAMs. By using this last combination, one can follow two strategies: (1) the use of a scanning tunnelling microscope which can induce a process reminiscent of electron-beam lithography by low-energy electron currents;
15 Advances in SPMs for SSMs
79
Fig. 15.13. Scanning tunnelling microscopy based self-replaced method: elimination is followed by substitution of a different molecular system
(2) the use of an atomic force microscope that can induce molecular removal by tip mechanical scratching under high load. Figure 15.13 sketches the self-replacement method which combines in situ elimination by STM and substitution by spontaneous self-assembly. The empty spaces induced by STM elimination are refilled by a different molecular system dissolved in the apolar medium, which binds covalently and self-assembles to the surface. This strategy has been largely applied to thiols on gold. Depending on the system studied, the elimination processes have been induced by employing biases of 3–10 V and currents from 100 pN to 10 nN. By using low bias and high tunnelling current, one essentially induces removal as in AFM by mechanical scratching since in this condition the tip is in close contact with the sample. By the elimination and substitution strategies, the resolution can achieve lines about 15 nm wide and 1–3 nm high, i. e. the molecular length. A wide range of molecular systems and substrates have been used [27]. 15.6.3 Nanoelectrochemical Lithography The local modification of a surface with nanoscopic control can be achieved by using an atomic force microscope tip which induces local chemical reactions by applying an external bias. Nanoelectrochemical techniques have been applied to inorganic substrates (oxidation of silicon) [127] but also to molecular surfaces, mainly SAMs [26, 29, 48, 89, 90]. By this method both lateral and vertical resolution can be largely increased with respect to the other above-reported SPLs. In effect there is no linear relation between tip size and lateral resolution since this latter depends mainly on the electric field distribution, which in turn affects exponentially the kinetics of the reactive process [128]. For this reason structures about 10 nm wide have been obtained with a coated atomic force microscope tip which typically has diameters 3–5 times larger. The use of thinner tips made, for instance, of carbon nanotubes should allow forbetter resolution. As to the vertical resolution, the selec˚ tive modification of the terminal group of a SAM (Angstrom resolution) has been demonstrated [29]. This method works well especially in the case of oxidation pro-
80
B. Pignataro
cesses, which typically occur in water-rich environments. Thus, crucial factors are writing rate or time of contact, tip size, applied biases, tip–SAM–substrate nature, and humidity. Larger patterned features have been observed by increasing the tip size, time of contact, applied bias and humidity within a certain range. This strategy was first used to modify SAMs of organosilanes [28, 111]. In particular, spatially defined high-friction nanometric features have been successfully obtained on top of SAMs of different alkyltrichlorosilanes on silicon oxide by locally inducing electrochemical modification via a biased conductive scanning force microscope tip. The characterization of the chemical nature of the probe-induced modification of the organic layer was, however, a difficult task owing to the inherent difficulty in achieving a nanometric-scale chemical characterization. The authors gave realistic hypotheses on the basis of macroscale simulation experiments through the use of a tip-mimicking biased copper grid pressed against the SAM. IR spectroscopy of this system showed that the bands of the –CH2 groups were unchanged, thus indicating no modification of the molecular tails. In contrast, the intensity of –CH3 signals decreased and a new signal for the C=O stretching appeared. The IR picture was interpreted by considering the oxidation of the terminal –CH3 groups to COOH moieties [28]. In our laboratory we used 1-terminated alkenes self-assembled on hydrogenated silicon (see earlier). It was demonstrated that the true silicon–organic monolayer interface provides improved electrical contacts and that the threshold bias for the electrochemical oxidation is lowered. Figure 15.14 reports on the experiment performed by using 1-octadecene, a HFetched silicon surface and a Pt/Ir tip in a water-saturated environment. We first performed the nanoelectrochemical modification with a conductive atomic force microscope tip of a 1-octadecene SAM obtained on hydrogenated silicon. SPM was employed to characterize the topographical and frictional changes of the monolayer upon tip modification. As a relevant feature, no modification of the organic layer was observed in low-humidity conditions, thus suggesting that the electrochemical reaction occurs through the formation of reactive oxygen-rich radicals at the tip–SAM interface. The application of a negative bias to the tip generally led to a much stronger modification, indicating that the reaction is preferably initiated by an electron injection process (from the tip to the sample). SPM friction and height
Fig. 15.14. A tip-writing oxidation experiment performed at the top of a 1-alkene SAM on Si. AFM atomic force microscope
15 Advances in SPMs for SSMs
81
images showed that at absolute bias values smaller than 6 V, the tip–sample friction force was increased by about a factor of 2, whereas the height of the monolayer was ˚ Higher biases led to an almost linear height increase owing to the raised only 3–6 A. formation of silicon oxide underneath the organic layer. The direct investigation of the chemical information written in the patterned areas has been done by using TOF-SIMS [29]. This was accomplished by performing chemical imaging and the retrospective spectral analysis of rectangular patterns, inscribed with various biases applied to the tip and large enough (some micrometres) to be individualized on the surface for the TOF-SIMS analysis. Figure 15.15 shows a possible strategy to combine a chemical surface-sensitive tool (low lateral resolution) with scanning probe written features. Figure 15.16 reports high-resolution TOF-SIMS chemical images of SAMs of 1-octadecene patterned by a negatively biased (from −6 to −9 V) Pt/Ir tip. The AFM friction (retrace) image of the pattern is reported as a reference. The TOFSIMS chemical maps of Six H y , Cx H y O, SiCx H y , Cx H y and Cx H y N related peaks are reported. The images show that in the modified regions the intensity of Six H y and SiCx H y peaks decreases, while that of Cx H y , Cx H y O and Cx H y N peaks increases. The spectra did not show evidence of an increase of the COOH and SiOCx H y signals. These findings suggest that at low bias (−6 V) there is formation of high-friction polar moieties (mainly CHO and CHN) at the top layer, with no large morphological modification. The observation of nitrogen-rich carbon fragments suggests that upon the tip bias possible reactions occur between activated carbon moieties and the nitrogen water carrier, possibly mediated by surface and oxygen-containing radicals.
Fig. 15.15. A prepattern performed on a Si surface by ion beams (top), just in order to write using an AFM tip the modifications in a selected area easily accessible for the successive time-of-flight secondary ion mass spectrometry (TOF-SIMS) analysis. At the bottom the total ions TOF-SIMS image from the region circled in top image is shown
82
B. Pignataro
Fig. 15.16. Tip-written oxidation at the top layer of a 1-octadecene SAM on Si as observed in a retrace direction friction force microscopy image (A) and TOF-SIMS positive ion maps + + obtained by adding different series of ions: B Si+ , Si2+ , SiH+ , Si2 H+ 5 and Si2 H2 ; C Cn H2n−3 O , + + + + + + + + Cn H2n−1 O , Cn H2n O , Cn H2n+1 O , C3 H3 O , and C3 H4 O ; D SiCH , SiC2 H2n , SiCn H2n+2 , + + + + + + + + SiCn H+ 2n+3 , and SiCn H2n ; E CH , C3 H3 , CnH2n−1 , Cn H2n , Cn H2n+1 , Cn H2n+2 and C7 H11 ; + + + F Cn H2n N, Cn H2n+2 N , C14 H33 N and Cn H2n+4 N . The colour scale intensity maximum (Max) is reported below the images. (Reprinted from Pignataro et al. [29], copyright 2003, with permission from Elsevier)
15.6.4 3D Nanolithography The possibility to use 2D nanopatterns as templates for the realization of 3D ordered nanosystems has already been mentioned (Sect. 15.4.1). This can be done by combining the power of SPL with the self-organization capabilities of molecular systems. The concept is simple and its power has already been addressed in a series of recent publications [26, 111, 112, 114]. A template-guided 3D self-assembly picture is sketched in Fig. 15.17. The lateral accuracy achievable by SPLs is of the order of several nanometres, but its combination with molecular self-assembly (a few angstrom resolution) offers even greater capabilities in terms of constructive tools and resolution. SPL nanotemplates offer remarkable chemical versatility and flexibility in the selection and handling of a large variety of different organic, inorganic and biological systems. Just to give an example, Fig. 15.18 shows a demonstration from Liu et al. [114] of the hierarchical self-assembly approach to the fabrication of ordered gold nanostructures on silicon. They first defined the nanotemplate via a nanoelectrochemical oxidation of the terminal –CH3 moiety of a OTS SAM on Si. This guided the
15 Advances in SPMs for SSMs
83
Fig. 15.17. A 3D nanolithography process performed by combining scanning probe lithographies (first step) and self-assembly (successive steps)
Fig. 15.18. AFM image of 17-nm gold nanostructures at the top of a 3D nanopattern fabricated as sketched on the left. (Reprinted with permission from Liu et al. [114], copyright 2004 American Chemical Society)
second step consisting of the spontaneous bilayer assembly of amino-terminated silanes. These last systems were in situ modified to generate the top amine functions constituting the specific sites for anchoring in a third step gold nanoparticles from a colloidal solution.
84
B. Pignataro
15.7 Conclusions and Perspectives In addition to fundamental purposes, the fabrication of self-assembled species at surfaces is a prerequisite for the realization of working device prototypes, in particular regarding cost-effective upscaling. Thus, it is crucial to achieve a full control over the interplay between intramolecular, intermolecular and interfacial interactions in order to tailor complex 2D and 3D architectures. It is possible to tailor SSMs by means of easy-to-control methods like LB and SA. In the recent period advances in SPMs allowed for an improved resolution both for imaging and modification. In particular the application of dynamic modes at different levels of tip–sample interaction forces allowed for increasing capability to control surface properties from the morphological to the mechanical ones. The development of DSFMA extended the available characterization tools, providing for a real noncontact technique which constitutes a breakthrough in imaging of soft samples. In light of this new technique, not only those employing SSMs, but all experiments on soft sample performed in tapping mode with molecular resolution should be reconsidered as far as they concern the size of molecules. Scanning probe microscopes have been developed also as nanoscopic manipulator, providing a new class of nanochemistry tools called SPLs. By combining SPLs with SSMs nanopatterned features and nanotemplates for 3D nanofabrication can now be achieved. The main drawback of these techniques is constituted by the inherently slow nature of serial patterning. Efforts have been made to build up parallel tip arrays. These techniques may serve to create functional patterns to make, for instance, sensor element arrays and interconnects of micrometric contacts with nanoscopic metal lines. Low-cost methods are needed for the fabrication of connector features bridging the dimensional gap needed for communication with the external world, including materials and functions. Many questions remain unsolved on the mechanisms of SPLs. Some of these concern the ink transport process in DPN and the reactive processes in nanoelectrochemical processes. Also, the limit in resolution of these techniques is still indefinite. Can, for instance, other probes like carbon nanotubes make smaller structures? In addition, the reproducibility and stability of the patterned nanostructures with time and under various environments has to be better investigated. To conclude, combining self-organisation with methods for controlled structuring like SPLs represents part of the exploration of concepts for future application. These kinds of methods and concepts might give in a long-term period innovation in industrial processes, more added value and lower production costs, as well as innovative ways for contributing to a sustainable development. The combination of SPLs, dynamic modes and SSMs has, however, great promise in the light of scientific knowledge. Nanotemplates and real nondestructive imaging tools can facilitate the study of fundamental issues like molecular recognitions, laws of nanoconfined molecular clusters, self-organization of biomolecules and so on.
15 Advances in SPMs for SSMs
85
Acknowledgements. Claudia Cascio, Giuseppe Francesco Indelli, Chiara Musumeci and Antonio Scandurra are gratefully acknowledged for their useful help and discussions. The work was supported financially by the University of Palermo (Fondi d’Ateneo ex 60%, Fondi CORI – 2005) and Italian MUR (PRIN-2006 prot. 2006037708).
References 1. Poole CP, Owens FJ (2003) Introduction to nanotechnology. Wiley, Hoboken 2. Sanchez A, Abbet S, Heiz U, Schneider WD, Hakkinen H, Barnett RN, Landman U (1999) J Phys Chem A 103:9573 3. Wang SC, Lu WC, Zhao QZ, Bernholc J (2006) Phys Rev B 7419:5430 4. Fink H-W, Schönenberger C (1999) Nature 398:407 5. Giese B, Amaudrut J, Köhler A-K, Spormann M, Wessely S (2001) Nature 412 318 6. Kasumov A, Kociak M, Guéron S, Reulet B, Volkov VT, Klinov DV, Bouchiat H (2001) Science 291:280 7. Porath D, Bezryadin A, De Vries S, Dekker C (2000) Nature 403:635 8. Moseler M, Hakkinen H, Barnett RN, Landman U (2001) Phys Rev Lett 86:2545 9. Bhushan B, Israelachvili JN, Landman U (1995) Nature 374:607 10. Moseler M, Landman U (2000) Science 289:1165 11. Boyen HG, Kästle G, Weigl F, Koslowski B, Dietrich C, Ziemann P, Spatz JP, Riethmüller S, Hartmann C, Möller M, Schmid G, Garnier MG, Oelhafen P (2002) Science 297:1533 12. Cohen ML (2005) Phys E Low Dimen Syst Nanostruct 29:447 13. Zhu S, Liu Y, Rafailovich MH, Sokolov J, Gersoppe D, Winesett DW, Ade H (1999) Nature 400:49 14. Salmeron M, Xu L, Hu J, Dai Q (1997) MRS Bull Aug 36 15. Guisinger NP, Yoder NL, Hersam MC (2005) Proc Natl Acad Sci USA 102: 8838 16. Suzuki T (2006) J Surf Sci Nanotech 4:588 17. Ulman A (1991) An introduction to ultrathin organic films from Langmuir–Blodgett to self assembly. Academic, New York 18. Reed MA, Zhou C, Muller CJ, Burgin TP, Tour JM (1997) Science 278:252 19. Edelstein RL, Tamanaha CR, Sheehan PE, Miller MM, Baselt DR, Whitman LJ, Colton RJ (2000) Biosens Bioelectron 14:805 20. Aviram A, Ratner M (1998) Ann N Y Acad Sci 852 21. Chi LF (1999) Appl Phys A 68:203 22. Anczykowski B, Cleveland JP, Krüger D, Elings V, Fuchs H (1998) Appl Phys A 66:885 23. Niemeyer CM, Adler M, Pignataro B, Lenhert S, Gao S, Chi LF, Fuchs H, Blohm D (1999) Nucleic Acids Res 27:4553 24. Pignataro B, Chi LF, Gao S, Anczykowski B, Niemeyer CM, Adler M, Blohm D, Fuchs H (2002) Appl Phys A 74:447 25. Pignataro B, Sardone L, Marletta G (2003) Nanotechnology 14:245 26. Krämer S, Fuierer RR, Gorman CB (2003) Chem Rev 103:4367 27. Gorman B, Carroll RL, He Y, Tian F, Fuierer R (2000) Langmuir 16:6312 28. Maoz R, Frydman E, Cohen SR, Sagiv J (2000) Adv Mater 12:725 29. Pignataro B, Licciardello A, Cataldo S, Marletta G (2003) Mater Sci Eng C 23:7 30. Strong L, Whitsides GM (1988) Langmuir 4:546 31. Dubois LH, Nuzzo RG (1992) Ann Phys Chem 43:437 32. Ulman A (1989) J Mat Ed 11:205 33. Linford MR, Fenter P, Eisenberger PM, Chidsey CED (1995) J Am Chem Soc 117:3145 34. Terry J, Linford MR, Wigren C, Cao RY, Pianetta P, Chidsey CED (1999) J Appl Phys 85:213
86
B. Pignataro
35. Sieval AB, Demirel AL, Nissink JWM, Linford MR, Vandermaas JH, Dejeu WH, Zuilhof H, Sudholter EJR (1998) Langmuir 14:1759 36. Pignataro B, Grasso G, Renna L, Marletta G (2002) Surf Interface Anal 33:54 37. Schwartz DK, (1997) Surf Sci Rep 27:241 38. Israelachvili J (1995) Intermolecular and surface forces. Academic, London 39. Sortino S, Petralia S, Pignataro B, Marletta G, Conoci S, Valli L (2003) New J Chem 27:615 40. Dodabalapur A (2006) Mater Today 9: 24 41. Zhang J, Lang HP, Huber F, Bietsch A, Grange W, Certa U, Mckendry R, Guntherodt H-J, Hegner M, Gerber C (2006) Nat Nanotechnol 1:214 42. Jackson AM, Myerson JW, Stellacci F (2004) Nat Mater 3:330 43. Bezryadin A, Dekker C (1997) J Vac Sci Technol B 15:793 44. Bezryadin A, Dekker C (1997) Appl Phys Lett 71:1273 45. Hersam MC, Hoole ACF, O’Shea SJ, Welland ME, (1998) Appl Phys Lett 72:915 46. Gleiche M, Chi LF, Fuchs H, (2000) Nature 403:173 47. Raudino A, Pignataro B (2007) J Phys Chem B 111:9189 48. Hoeppener S, Schubert U (2005) Small 1:628 49. Chen X, Lu N, Zhang H, Hirtz M, Wu L, Fuchs H, Chi L (2006) J Phys Chem B 110:8039 50. Briggs D, Seah MF (1992) (eds) Practical surfaces analysis. Wiley, Chichester 51. Benninghoven A (1994) Angew Chem Int Ed Engl 33:1023 52. Poelsema B, Comsa G (1989) Springer tracts in modern physics 115. Scattering of thermal energy atoms from disordered surfaces. Springer, Berlin 53. Zewail A (1992) The chemical bond, structures and dynamics. Academic, London 54. Compagnini G, Pelligra B, Pignataro B (1998) Chem Phys Lett 285:266 55. Kirstein S (1999) Curr Opin Colloid Interface Sci 4:256 56. Maoz R, Sagiv J (1984) J Colloid Interface Sci 100:465 57. De Padova P, Larciprete R, Ottaviani C, Priori S, Quaresima C, Reginelli A, Ressel B, Perfetti P (1999) Surf Sci 433-435:362 58. Messier J, Marc G (1971) J Phys (Paris) 32:799 59. Somorjai GA, Szalkowski F (1971) J Adv High Temp Chem 4:137 60. Rothenhäusher B, Duschl C, Knoll W (1988) Thin Solid Films 159:323 61. Ruska H (1993) In: Nobel lectures physics 1981-1990. World Scientific, Singapore 62. Shen YR (1989) Liq Cryst 5:635 63. Kinloch CD, McMullen AI (1959) J Sci Instrum 36:347 64. Lercel MJ, Craighead HG, Parikh AN, Seshadri K, Allara DL (1996) Appl Phys Lett 68:1504 65. Golzhauser A, Eck W, Geyer W, Stadler V, Weimann T, Hinze P, Grunze M (2001) Adv Mater 13:806 66. Inoue A, Ishida T, Choi N, Mizutani W, Tokumoto H (1998) Appl Phys Lett 73:1976 67. Hatzor A, Weiss PS (2001) Science 291:1019 68. Xia YN, Whitesides GM (1998) Angew Chem Int Ed Engl 37:551 69. Xia YN, Rogers JA, Paul KE, Whitesides GM (1999) Chem Rev 99:1823 70. Chou SY (2001) MRS Bull 26:512 71. Torres CMS, Zankovych S, Seekamp J, Kam AP, Cedeno CC, Hoffmann T, Ahopelto J, Reuther F, Pfeiffer K, Bleidiessel G, Gruetzner G, Maximov MV, Heidari B (2003) Mater Sci Eng C 23:23 72. Binning G, Rohrer H, Gerber C, Weibel E (1982) Phys Rev Lett 49:57 73. De Feyter S, De Schryver FC (2003) Chem Soc Rev 32:139 74. Ikai A (1996) Surf Sci Rep 26:261 75. Binnig G, Quate CF, Gerber C (1986) Phys Rev Lett 56:930 76. Frisbie CD, Rozsnyai LF, Noy A, Wrighton MS; Lieber CM (1994) Science 265: 2071
15 Advances in SPMs for SSMs
87
77. Bhushan B (1999) Handbook of micro/nano tribology. CRC, Boca Raton 78. Butt HJ, Cappella B, Kappl M (2005) Surf Sci Rep 59:1 79. Ebeling D, Holscher H, Fuchs H, Anczykowski B, Schwarz UD (2006) Nanotechnology 17:S221 80. Hansma HG, Vesenka J, Siegerist C, Kelderman G, Morrett H, Sinsheimer RL, Elings V, Bustamante C, Hansma PK (1992) Science 256:1180 81. Wong SS, Joselevich E, Woolley AT, Cheung C-L, Lieber CM (1998) Nature 394:52 82. Lyubchenko YL, Oden PI, Lampner D, Linsday SM, Dunker KA (1993) Nucleic Acids Res 21:1117 83. Niemeyer CM (1999) Appl Phys A 68:119 84. Quist AP, Björck LP, Reimann CT, Oscarsson SO, Sundqvist BUR (1995) Surf Sci 325:406 85. Sarid D, Hunt JP, Workman RK, Yao X, Peterson CA (1998) Appl Phys A 66:283 86. Bustamante C, Vesenka J, Tang CL, Rees W, Guthold M, Keller R (1992) Biochemistry 31:22 87. Pignataro B, Sardone L, Marletta G (2003) Langmuir 19:5912 88. Magonov SN, Elings V, Whangbo MH (1997) Surf Sci 375:385 89. Ozin GA, Arsenault AC (eds) (2005) Nanochemistry: a chemical approach to nanomaterials. Royal Society of Chemistry, Northampton 90. Wouters D, Willems R, Hoeppener S, Flipse CFJ, Schubert US (2005) Adv Funct Mater 15:938 91. Zheng J, Zhu Z, Chen H, Liu Z (2000) Langmuir 16:4409 92. Zheng J, Chen Z, Liu Z (2000) Langmuir 16:9673 93. Li Q, Zheng J, Liu Z (2003) Langmuir 19:166 94. Ling X, Zhu X, Zhang J, Zhu T, Liu M, Tong L, Liu Z (2005) J Phys Chem B 109:2657 95. He M, Ling X, Zhang J, Liu Z (2005) J Phys Chem B 109:10946 96. Snow ES, Campbell PM (1994) Appl Phys Lett 64:1932 97. Kinser CR, Schmitz MJ, Hersam MC (2005) Nano Lett 5:91 98. Day HC, Allee DR (1993) Appl Phys Lett, 62:2691 99. Kim Y, Kang SK, Choi I, Lee J, Yi J (2005) J Am Chem Soc 127:9380 100. Sugimura H, Hanji T, Hayashi K, Takai O (2002) Ultamicroscopy 91:221 101. Sugimura H, Hanji T, Hayashi K, Takai O (2002) Adv Mater 14:524 102. Sugimura H, Takai O, Nakagiri N (1999) J Electroanal Chem 473:230 103. Sugimura H, Nakagiri N (1997) Nanotechnology 8:A15 104. Kim B, Pyrgiotakis G, Sauers J, Sigmund WM (2005) Colloids Surf A 253:23 105. Xie XN, Chung HJ, Sow CH, Wee ATS (2004) Chem Phys Lett 388:446 106. Park J, Lee H (2004) Mater Sci Eng C 24:311 107. Wouters D, Schubert US (2003) Langmuir 19:9033 108. Ahn SJ, Jang YK, Lee H, Lee H (2002) Appl Phys Lett 80:2592 109. Lee H, Kim SA, Ahn SJ, Lee H (2002) Appl Phys Lett 81:138 110. Lee W, Kim ER, Lee H (2002) Langmuir 18:8375 111. Maoz R, Cohen SR, Sagiv J (1999) Adv Mater 11:55 112. Liu S, Maoz R, Schmid G, Sagiv J (2002) Nano Lett 2:1055 113. Hoeppener S, Maoz R, Cohen SR, Chi LF, Fuchs H, Sagiv J (2002) Adv Mater 14:1036 114. Liu S, Maoz R, Sagiv J (2004) Nano Lett 4:845 115. Maoz R, Frydman E, Cohen SR, Sagiv J (2000) Adv Mater 12:424 116. Rodolfa KT, Bruckbauer A, Zhou DJ, Schevchuk AI, Korchev YE, Klenerman D (2006) Nano Lett 6:25 117. Liu XG, Fu L, Hong SH, Dravid VP, Mirkin CA (2002) Adv Mater 14:231 118. Ivanisevic A, Mirkin CA (2001) J Am Chem Soc123:7887 119. Demers LM, Ginger DS, Park SJ, Li Z, Chung SW, Mirkin CA (2002) Science 296:1836 120. Su M, Dravid VP (2002) Appl Phys Lett 80:4434
88
B. Pignataro
121. McKendry R. Huck WTS, Weeks B, Florini M, Abell C, Rayment T (2002) Nano Lett 2:713 122. Lim JH, Mirkin CA (2002) Adv Mater 14:1474 123. Liao JH, Huang L, Gu N (2002) Chin Phys Lett 19:134 124. Agarwal G, Naik RR, Stone MO (2003) J Am Chem Soc 125:7408 125. Wilson DL, Martin R, Hong S, Cronin-Golomb M, Mirkin CA, Kaplan DL (2001) Proc Natl Acad Sci USA 98:13660 126. Noy A, Miller AE, Klare JE, Weeks BL, Woods BW, DeYoreo JJ (2002) Nano Lett 2:109 127. Zhang YY, Zhang J, Luo G, Zhou X, Xie GY, Zhu T, Liu ZF (2005) Nanotechnology 16:422 128. Yang M, Zheng Z, Liu Y, Zhang B (2006) Nanotechnology 17:330
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells Félix Rico · Ewa P. Wojcikiewicz · Vincent T. Moy
Abstract. The study of viscoelastic and adhesive properties of cells falls under the discipline of cellular mechanics. Our knowledge of cell mechanics has been limited until the development of tools that allow us to measure minute forces (piconewton) and small dimensions (nanometer) at the cellular level. The atomic force microscope (AFM) is a very useful tool for determining the mechanical properties of cells as it allows experiments to be conducted under physiological conditions. The focus of this review is to provide an overview of the AFM as a tool for measuring the mechanical (viscoelastic and adhesive) properties of living cells. First, the different modes of operation of the AFM are briefly introduced, drawing special attention to the force mode. Then, the methods used to determine the viscoelasticity of cells are described, followed by a review of the available methods for determining cell adhesion at both single-molecule and whole-cell levels.
Key words: Atomic force microscopy, Atomic force microscope, Cell adhesion, Cell mechanics, Viscoelasticity, Young’s modulus, Dynamic force spectroscopy, Scanning probe microscopy, Cantilever, Tips, Indentation, Force, Shear modulus, Elasticity, Single molecules
16.1 Introduction Mechanics is the branch of physics that studies forces. As such, mechanics describes the relations between the strain and the applied stress to obtain the elastic and viscous properties of bodies. But mechanics also describes the forces required to maintain two bodies in contact when pulled apart. Thus, mechanics involves both viscoelasticity and adhesion, two properties that are very difficult to study independently, as we will see in the following sections [1]. Cells are living systems that are constantly subjected to or exerting forces. That is the reason why the mechanics of living cells is proving to be as important to cells’ functions as their biochemical counterparts [2–5]. The structure, shape and function of living cells are regulated in part by their mechanical properties, and vice versa [6–13]. Until recently, little was known about single cell mechanical properties owing to the lack of adequate tools to control displacements and forces at the cellular level. Our understanding of the mechanics of the cell has increased thanks to the emergence of nanotechnology. The currently available nanotools enable us to manipulate biomolecules and cells with subnanometer resolution and simultaneously measure forces at the piconewton scale. One of the most versatile nanotools available today is the atomic force microscope (AFM). The AFM is capable of acquiring both topographical images and mechanical measurements under physiological conditions,
90
F. Rico · E.P. Wojcikiewicz · V.T. Moy
making it a very powerful tool for studying living systems at the piconewton and nanometer scales. The AFM was first introduced in 1986 as a tool to investigate the surfaces of insulators in air [8, 10–12, 14, 15]. It was then rapidly applied to measure forces and nanomechanical properties of different material surfaces [16,17]. Its application for imaging biological systems came about once measurements could be conducted in liquids [18–20]. In 1992, the AFM was first used to probe the mechanics of biological samples. Tao et al. [21] applied the AFM to measure the elasticity of sectioned biological material, and Hoh and Schoenenberger [22] used the AFM to measure the mechanical properties of living cells. Following these pioneering studies, there has been a continuous increase in studies assessing the mechanical properties of living cells under different physiological conditions as well as other novel methods to improve AFM measurements of biological samples, especially of living cells [23–46]. The focus of this review is to provide an overview of the AFM as a tool for measuring the mechanical (viscoelastic and adhesive) properties of living cells. The first part of the review will focus on the basic principles of atomic force microscopy operation, and will be followed by descriptions of the methods used to determine the viscoelastic and adhesive properties of cells.
16.2 Principle of Operation The AFM and its precursor, the scanning tunneling microscope, were first applied to obtain topographical information of hard surfaces in air [14, 47–49]. For that purpose a sharp tip was used to probe the sample while scanning its surface in the horizontal plane. Atomic force microscopy is still mostly used as an imaging technique, especially in the field of material sciences. However, its capability of imaging nonconducting surfaces in fluids led to its immediate application to biological specimens [50–54].
Fig. 16.1. A typical atomic force microscope (AFM)
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
91
Fig. 16.2. a Au-coated Si3 N4 cantilever chip compared with a regular paper clip. The upperright image shows a magnification of the front of the chip where five different cantilevers can be observed, each one has a different spring constant (the longest cantilever is 320 µm long). The lower-right image is a scanning electron microscopy lateral view of the end of one of the cantilevers. b Commercially available AFM cantilever tips include, e. g., four- and eight-sided pyramidal tips, and spherical tips. Bars stand for 1, 5 and 2 µm, respectively
The AFM makes use of a silicon or silicone nitride flexible cantilever with a tip at its end for probing the sample surface (Fig. 16.1). The cantilever is the most important part of the AFM, since it determines the applied and measured forces and the lateral resolution. There are different tips and cantilevers commercially available, some of which are shown in Fig. 16.2. The probe is moved relative to the sample by means of piezoelectric elements, ceramic materials that expand or contract when different voltages are applied. Owing to the inherent nonlinearity of piezoelectric materials, newer AFM models use position sensors to detect the actual extension, and correct the nonlinearities using electrical feedback circuits. When the probe tip makes contact with the surface, the cantilever bends or deflects. This deflection is usually measured using the optical lever method [55, 56]. A laser beam is focused on the cantilever surface and the position of the reflected beam is detected using a segmented photodiode. The difference in the photocurrents of the photodiode segments provides a measure of the deflection change. AFM cantilevers are very flexible and act as springs with a characteristic constant, k. Thus, the applied force of the cantilever (F) can be obtained using Hooke’s law, F = −kd, where d is the cantilever’s deflection. An illustrative sketch of an AFM is shown in Fig. 16.1.
92
F. Rico · E.P. Wojcikiewicz · V.T. Moy
16.2.1 AFM Imaging The operation of an imaging AFM is similar to that of a stylus profiler or a record player but with superior resolution. AFM images are obtained in contact mode by scanning the sample surface with the cantilever tip at a constant force. The force level is kept constant by using feedback electronics (usually digital proportional integral circuits) that continuously correct the vertical position of the cantilever as the deflection signal varies as a result of topographical changes. The topographical images are then obtained from the vertical height corrections used to maintain the force constant. 16.2.2 Force Measurements Apart from imaging purposes, the AFM is widely used to apply and measure forces on materials and to determine their mechanical properties [57, 58]. To achieve this, the AFM cantilever is moved in the vertical direction at a constant speed until the tip contacts the sample, and the cantilever undergoes deflection before being retracted away from the sample. The result of this complete cycle is termed a force–distance curve (F–z) or, simply, a force curve. The approach or the loading trace of the force– distance curve is usually analyzed to obtain the elastic properties of the sample, while the retract or the unloading trace of the curve is mainly used to measure adhesion. Figure 16.3 shows an example of a force curve obtained with a pyramidal tip on a living cell, together with an illustration of the position of the cantilever at each stage of the cycle. When the cantilever tip makes contact with a soft material, such as a living cell, the resulting deformation or indentation of the sample is recorded in the loading curve. If, in addition, the sample adheres to the cantilever, the retraction curve may present negative force values, as shown in the figure. As we mentioned before, cantilever deflection is measured using the laser beam that is reflected onto the photodiode. To obtain a quantitative measure of such a deflection, it is necessary to calibrate the photodiode signal before any measurement. To do so, a force curve is obtained on a hard surface where the cantilever will deflect as much as the piezoelement displaces it. Thus, knowing the applied displacement and recording the photodiode signal, one obtains the optical lever sensitivity (usually reported in units of volts per nanometer), enabling the conversion of the deflection from voltage to distance. In practice, this is obtained by computing the slope of the F–z curve obtained on a hard surface. In addition, to obtain mechanical measurements, it is necessary to convert the deflection of the cantilever into force. To do so, the spring constant of the cantilever has to be determined. Several methods have been reported to obtain the spring constant. These range from calculated values derived from the cantilever’s geometry and material properties to measured values using a reference lever [59–71]. There are a number of reviews which provide an excellent overview of these methods [72, 73]. However, the most commonly used method to calibrate the spring constant of the cantilevers is the thermal fluctuation method introduced by Hutter and Bechhoefer [65]. It consists of modeling the behavior of the cantilever to a simple harmonic oscillator. Then, by
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
93
measuring the thermal fluctuations ( d 2 ) of the cantilever away from the surface and applying the equipartition theorem, one can relate d 2 to the thermal energy (kB T , kB being the Boltzmann constant and T the absolute temperature) to obtain the cantilever spring constant. As suggested by Butt and Jaschke, the method may require correcting for the secondary modes of oscillation which are not considered in the Hutter and Bechhoefer method but are measured in practice [66, 69]. Typically, there are several cantilevers with different spring constants on an AFM cantilever chip (Fig. 16.2a). They are intended for different applications. Hard samples will require stiffer cantilevers, with spring constants of a few thousand piconewtons per nanometer, while more delicate and malleable samples, such as living cells, will call for softer cantilevers. The softer cantilevers commercially available have a spring constant of approximately 6 pN/nm (Biolever, Olympus, Tokyo, Japan). Two of the most widely used chips of cantilevers for cellular use are the Microlever series commercialized by Veeco (Santa Barbara, CA, USA) and the OMCL-TR series by Olympus (Tokyo, Japan). All these cantilevers are made of silicon nitride since this material allows for the fabrication of thinner cantilevers, which results in lower spring constants.
16.3 Cell Viscoelasticity The approach trace of the force curve shown in Fig. 16.3 (dotted line) has a zone of relatively constant force when the tip is away from the cell surface, followed by a nonlinear increase when the tip makes contact with and indents the surface. In this case, the curve was obtained using a pyramidal tip, and the nonlinear response does not come from the nonlinear behavior of the cell, but from the increasing contact area between the tip and the sample. The shape of the curve will thus depend on the
Fig. 16.3. Force–displacement (deflection– displacement) curve acquired at approximately 1.8 µm/s on a living cell using a pyramidal tip. Approach (dotted line) and retract (solid line) traces are plotted. The various steps of the tip–sample interaction are depicted
94
F. Rico · E.P. Wojcikiewicz · V.T. Moy
tip geometry and on the elastic properties of the sample, its stiffness. On the other hand, the retract trace of the curve follows a similar shape, although slightly shifted, i. e., the complete cycle has hysteresis. This hysteresis is mainly due to the viscosity of the sample. Taken together, these behaviors reflect the viscoelastic nature of living cells. This section will review the most-used methods to quantitatively determine the elasticity and viscoelasticity of cells with atomic force microscopy. 16.3.1 AFM Tip Geometries As mentioned in Sect. 16.1, cantilever tips can have a variety of geometries, and the choice of tip will depend on its final application. The cantilever tips most widely used in the biological field are silicon nitride four-sided pyramidal tips (Fig. 16.2a). These cantilever tips are very suitable since they allow both mechanical measurements and imaging with high lateral resolution [74]. Sharper tips are also being used for imaging and probing single molecules, where the interaction is limited to the very end of the tip, typically with a curvature radii of approximately 10 nm (Fig. 16.2b) [75–77]. In contrast, when a smooth contact with the sample surface is required, a tip with dull geometry is a better choice. Thus, some authors have suggested the use of spherical beads of a few microns in diameter attached to the end of conventional cantilevers (Fig. 16.2b) [30, 39, 78]. This procedure will provide a more global estimate of the mechanical properties and may render the simultaneous imaging mode inoperative. As we will see in the following section, the geometry of the tip greatly affects force measurement analysis. 16.3.2 Elasticity: Young’s Modulus All commercially available tip geometries (e. g., conical, pyramidal and spherical1 ) lead to nonlinear force–indentation responses. The force increases as the tip indents the sample as a result of the increasing area of contact during the indentation process. The force-indentation relationship (F(δ)) of two elastic bodies in contact was first described by Hertz [79] in 1881 for two spheres. Later on, Sneddon [80] solved the problem for various tip geometries. These so-called elastic models relate the force as a function of the indentation and the sample elastic modulus, taking tip geometry into consideration. Because of his pioneering work, elastic models are also commonly referred to as Hertz models. The most frequently used elastic models are summarized in Table 16.1. Note that the only model that leads to a linear force response is that of the flat-ended cylinder, the only geometry which maintains a constant area of contact independently of the indentation depth. The most common approach used to obtain the elastic or Young’s modulus (E) of living cells involves acquiring force curves (i. e., indenting the cell surface) and then fitting the suitable elastic model to the loading trace [32, 37, 81–84]. In principle, the unloading trace could also be used; however, using the loading trace minimizes 1
Novascan Technologies (Ames, IA, USA) commercializes cantilevers with spherical beads of various materials attached at their end.
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
95
Table 16.1. Contact elastic models for common rigid-tip geometries indenting an elastic half space. E and ν are Young’s modulus and the Poisson ratio of the sample, respectively Tip geometry
Contact models, F(δ)
Sphere of radius R
F=
Cone of semi-included angle θ
F=
Flat-ended cylinder of radius a
F=
Four-sided regular pyramid of semi-included angle θ
F=
√ 4E R 3/2 δ [79, 85] 3(1−ν2 ) 2E tan θ 2 δ [80] π(1−ν2 ) 2Ea δ [80] (1−ν2 ) 0.7453E tan θ 2 δ [168] (1−ν2 )
artifacts due to adhesion between the tip and the sample. The elastic modulus of living cells has been measured using different techniques, revealing a wide range of values from a few pascals to hundreds of kilopascals, depending on cell type and the region of the cell being probed. Compared with everyday material such as rubber and jelly, cells are relatively soft2 and a very soft cantilever is required to determine their Young’s modulus. A reliable estimation of a cell’s elastic modulus requires acquisition of experimental force curves obtained at an adequate indentation range. It has been shown that the elastic modulus is difficult to determine from curves with indentations lower than approximately 200 nm [74, 85]. The heterogeneity of the cell’s glycocalix and membrane as well as the poorly defined geometry of the tip at such low indentations make it difficult to obtain a good estimation of E. In addition, the use of very high indentation forces which result in the probing of the hard substrate on which cells are plated is also not recommended for obtaining accurate Young’s modulus values [83, 86]. It should also be noted that under very high indentation forces cells can behave nonlinearly, i. e., cells can have a nonelastic response [87–94]. Thus, it is important to limit the indentation in order to work within the linear regime of cellular elasticity. The actual indentation range may vary between cell types and regions of the cell being probed. However, as a rule of thumb, the hard surface will not affect the E values as long as the maximum indentation does not exceed 10–20% of the cell’s thickness [83]. Secondly, two main issues are crucial for a reliable estimation of E during data analysis: accurate determination of the point of initial contact and suitability of the elastic model to the tip geometry being used [26, 39, 43, 74, 95–98]. The point of contact coordinates are defined by the force (or deflection) offset (Foff or doff ) and the vertical position point at which the cantilever starts to deflect (z c ) and are used to determine the absolute indentation [δ = z − z c − (d − doff )]. As noted in Fig. 16.4, it is difficult to visually determine the point of contact for living cells since the deflection increases very smoothly following contact owing to the cells’ softness. Thus, an objective method to determine Foff and z c should be used. The inclusion of the noncontact region of the loading trace in the fitting process allows for the simultaneous determination of the point of contact coordinates and the Young’s modulus [26, 74]. Figure 16.4 shows a representative force curve overlaid with the fit of the appropriate contact model. In addition, the 2
Rubber has a Young’s modulus in the megapascal range, while that of jelly falls in the kilopascal range.
96
F. Rico · E.P. Wojcikiewicz · V.T. Moy
Fig. 16.4. Example of a representative approach trace of a force curve (dotted line) used to determine the cell elasticity. The fit of the contact model (Table 16.1) for a four-sided pyramid with a 35◦ semiopen angle is also shown (solid line). The Young’s modulus obtained was 4.5 kPa. The arrow indicates the estimated point of contact. The retract trace (dashed line) presents hysteresis owing to the viscoelastic nature of living cells. The curve was acquired at a speed of approximately 1.8 µm/s by indenting an epithelial cell using a pyramidal tip
suitability of the contact elastic model depends on the similarity of the tip geometry to the idealized geometry and on the accuracy of the actual geometry parameters (radii, semiopen angles, etc.). As an example, it has been shown that using the common pyramidal tips (Fig. 16.2b) leads to a spherical-like F–δ response at low indentation, followed by a conical-like response at higher indentation3 [81, 99]. This behavior was attributed to the actual bluntness of pyramidal tips, which lead to a region in the approach trace of tens of nanometers with a spherical-like geometry. There are a number of publications that take these more realistic blunted geometries of actual tips into account [74, 97]. In summary, a reliable estimation of the Young’s modulus of living cells requires appropriate experimental conditions (indentation range), an accurate and objective determination of the point of contact, and the application of an elastic model that describes the actual tip geometry. 16.3.3 Viscoelasticity: Complex Shear Modulus Even if we obtain a reliable estimation of cellular elasticity, a great number of studies have shown that cells are not only elastic but also viscous: they have a viscoelastic nature [18, 27, 81, 89, 100–106]. Viscoelasticity is roughly deduced from the hysteresis between the approach and retract traces of the curves shown in Figs. 16.3 3
Pyramidal tips were first modeled as cones. Indeed, the indentation dependence of the force is the same for a pyramid and for a cone, apart from a scaling factor (Table 16.1).
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
97
and 16.4. However, the viscoelastic dynamic response of living cells has been shown to be more complex [104, 107–112]. Several authors have measured the viscoelasticity of living cells by applying a force or deformation step and measuring the resulting response [27, 101, 113, 114]. The principal limitation of these approaches is that the estimated viscous parameters are limited by the time window used during the measurements [85]. A more elegant method to probe dynamic viscoelastic cell behavior involves the application of low-amplitude oscillations at a wide frequency range and subsequent derivation of the complex shear modulus G ∗ (ω). The complex shear modulus is divided into a real or in-phase component, which stands for the elasticity or the stored energy, and an imaginary or out-of-phase component, which stands for the viscosity or lost energy. The real part is known as the elastic modulus (G ′ ), while the imaginary part is called the loss modulus (G ′′ ). This technique has been successfully applied to a wide variety of living cells studied under different conditions [39, 86, 88, 103, 104, 108, 110, 111, 114]. In AFM viscoelastic measurements conducted on living cells, the oscillatory technique involves applying low-amplitude oscillations (50–100 nm) at an operating indentation (δ0 ). Under these conditions, a contact elastic model (Table 16.1) described by F(δ) can be approximated by taking the first two terms of its Taylor expansion around the fixed point (δ0 ) ∂F (δ − δ0 ) + . . . . (16.1) F∼ = F(δ0 ) + ∂δ For the sake of simplicity, we will follow the derivation using the elastic model for a flat-ended cylinder of radius a, whose expansion leads to F = F0 +
2Ea (δ − δ0 ) . 1 − ν2
(16.2)
Expressing (16.2) in terms of the shear modulus, using E = 2(1 + ν)G, and solving for G, we obtain G=
1 − ν F − F0 . 4a δ − δ0
(16.3)
By applying the Parseval theorem, we can transform (16.3) into the frequency domain, resulting in an expression for the complex shear modulus G ∗ (ω) =
1 − ν F(ω) , 4a δ(ω)
(16.4)
where ω is the angular frequency (ω = 2π f ) and F(ω) and δ(ω) are the Fourier transforms of force and indentation, respectively. When samples are probed in liquid conditions, (16.4) has to be corrected for the hydrodynamic drag force exerted on the cantilever by the surrounding liquid. It has been determined that the drag contribution is not negligible for oscillatory frequencies above 0.3 Hz [115]. The drag force (Fd ) depends on the geometry and dimensions of the cantilever, which give us the drag factor b, and on the velocity of the cantilever’s free end v = dδ/ dt [i. e., Fd = b(0)v]. As described by Alcaraz et al. [115], the drag factor increases as the tip approaches the surface. Thus, it is important to measure b(h) experimentally by oscillating the cantilever at various
98
F. Rico · E.P. Wojcikiewicz · V.T. Moy
tip–surface distances (h) and extrapolate its value to zero separation [84, 110, 115]. The hydrodynamic drag force contribution in the frequency domain has the form Fd (ω)/δ(ω) = iωb(0) (i being the imaginary unit). To correct for the drag force, the last expression should be included in (16.4), resulting in the final form of the complex shear modulus 1 − ν F(ω) ∗ − iωb(0) . (16.5) G (ω) = 4a δ(ω) Notice, again, that the cylindrical punch has no dependency on the absolute indentation (the operating indentation, δ0 ), since its force response is linear. That feature makes it ideal to probe cell mechanics at different indentation depths. The corresponding shear modulus for spherical or pyramidal models does depend on the absolute indentation [74]. A recent work proposed another viscoelastic approach that took into account the possible contribution of the hard substrate in oscillatory measurements. The authors used this procedure to determine the viscoelastic properties of thin regions of living cells [86]. Even if its application is quite a bit more complicated than the one presented here, it might be useful when probing very thin samples. Experimentally, the procedure to carry out oscillatory measurements would first involve the acquisition of a high-amplitude force curve ( a few microns) to determine the point of contact. Then the probe would be moved to an operating indentation in the linear regime (approximately 0.5 µm), and low-amplitude oscillations approximately (50 nm) would be applied at a given frequency or at a multifrequency, recording a few tens of cycles. The resulting data would be analyzed in the frequency domain according to (16.5). The oscillatory method is well suited for experiments conducted on living cells since it enables us to measure their dynamic viscoelastic response with a single measurement by applying a multifrequency wave as the oscillatory signal [85]. This procedure provides faster and less invasive measurements, convenient for living systems. 16.3.4 Cell Adhesion Cell adhesion plays a major role in many biological processes, including developmental and immune system functions. It involves many cell types, from nonadherent cells, such as platelets or leukocytes, to adherent cells that comprise tissues, such as endothelial or epithelial cells [116–119]. Cell adhesion is mediated by cell adhesion receptors that adhere specifically to ligands on other cells or to the extracellular matrix [120, 121]. The AFM can be used to study cell adhesion at the singlemolecule level as well as at the whole-cell level, where the contribution of multiple receptor–ligand interactions can be examined. The principle of operation of adhesion measurements by atomic force microscopy and other techniques is the same for both types of experiments. The method involves measuring the forces necessary to rupture the ligand–receptor bonds. We have seen that the approach trace of a force curve can provide us with an estimate of the cellular elasticity. In adhesion measurements, the unloading curve, or the retract trace, is usually used to acquire adhesive information at both molecular and cellular levels (Fig. 16.3). Different approaches have been used for that purpose. One method involves the functionalization or coating
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
99
of the cantilever tip with receptors and acquiring force measurements by bringing the cantilever tip into contact with a cell or ligand-coated surface below. This is the most common approach when measuring adhesion between cell ligands that specifically attach to extracellular matrix proteins [122, 123]. A second technique involves attaching a cell to the end of a cantilever and bringing it into contact with a surface functionalized with the ligand of interest or another cell [40, 45, 124]. These two approaches are equivalent and their use is more or less appropriate depending on the system being studied. In both types of experiments, the first step is the coating of the cantilever tip with the desired molecules. The most widely used procedures for studying cell adhesion are explained in this section. 16.3.4.1 Functionalization of AFM Tips There are many different protocols for coating surfaces, since surface chemistry modification is an intensive field of study [125]. We will describe two commonly used protocols, one involves covalently attaching receptors to AFM cantilevers, and the second is used to immobilize living cells on the cantilevers. 16.3.4.2 Attachment of Protein – Clean cantilevers in acetone and irradiate them with UV light for 10 min. – Immerse cantilevers in a glass dish containing 2% 3-aminopropyltriethoxysilane in acetone and shake them for 10 min at room temperature. – Rinse extensively with nanopure water. – Incubate for 30 min in 0.1% glutaraldehyde in phosphate-buffered saline (PBS) at room temperature. – Rinse extensively with PBS. – Air-dry cantilevers. – Cover dried cantilevers with the protein solution and incubate for at least 1 h at room temperature or overnight at 4 ◦ C. – Rinse with tris(hydroxymethyl)aminomethane (Tris) buffered saline (TBS). – Store coated cantilevers in TBS at 4 ◦ C until use. 16.3.4.3 Attachment of Living Cells – Clean cantilevers in acetone and irradiate them with UV light for 10 min. – Incubate cantilevers in biotinylated bovine serum albumin (0.5 mg/ml in NaCO3 , overnight at 4 ◦ C). – Rinse cantilevers with PBS. – Incubate cantilevers with streptavidin (0.5 mg/ml) for 10 min. – Rinse again with PBS. – Incubate cantilevers in a solution of biotinylated concanavalin-A (0.5 mg/ml). – Rinse with PBS. – Store coated cantilevers in PBS at 4 ◦ C until use.
100
F. Rico · E.P. Wojcikiewicz · V.T. Moy
16.3.4.4 Dynamic Force Spectroscopy The study of single-molecule interactions shed light on the ultimate biophysical determinants of cell adhesion. Measurements can be carried out on isolated receptor– ligand complexes attached to synthetic surfaces or directly on living cells [23, 75, 126–134]. An updated review on single-molecule measurements using atomic force microscopy has recently been published [77]. Single-molecule measurements are performed by allowing the cantilever tip coated with receptors or a living cell to make contact with a ligand-coated substrate or another cell. The resulting force curve is used to measure the rupture forces of the adhesion events. The bottom curve in Fig. 16.5 shows the rupture force ( f u ) of a single bond. The forces necessary to break single bonds range from tens to hundreds of piconewtons. All force curves in Fig. 16.5 reveal fluctuations, especially visible in the noncontact region. This noise is mainly due to thermal fluctuations, although mechanical perturbations of the system and turbulence of the liquid also have an effect. The noise level resulting from thermal fluctuations is on the order of tens of piconewtons; thus, rupture events involving forces below 10–20 pN are difficult to resolve and are usually rejected [135,136]. In order to obtain single-molecule measurements on living cells, force curves are obtained at low indentation forces to minimize the contact area and with a brief contact time (on the order of fractions of a second), thus reducing the binding probability. Under these conditions, most of the force curves will show no adhesion, as illustrated in Fig. 16.5. This effect is desirable, since, assuming Poisson statistics, an adhesion frequency of less than 30% in the force measurements ensures that more than 83% of the events are caused by single bonds [137, 138]; therefore, thousands of measurements need to be acquired, from which only curves presenting rupture events will be analyzed. Automatic procedures for detecting and
Fig. 16.5. Unbinding curves obtained from single-molecule measurements. Measurements were acquired at speeds of 5 µm/s and conditions that minimized contact (approximately 50 ms contact time and approximately 100 pN indentation force) ensuring an adhesion frequency of less than 30%. Adhesion was observed in the third and sixth traces. f u is the unbinding force of the interaction, and rf is the actual loading rate derived from the slope of the force–time graph
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
Fig. 16.6. a Rupture force histograms obtained from measurements at different loading rates of the interaction between a T-cell integrin and one of its ligands. Rupture events were grouped according to loading rates. The most probable unbinding forces were taken from the Gaussian fit of each histogram (solid lines). b Dynamic force spectrum resulting from the histograms in a. Measurements were acquired at a loading rate range from 50 to 60,000 pN/s, which was achieved by varying the retraction rate of the cantilever from 0.1 to 26 µm/s. Forces acquired at cantilever retraction speeds exceeding 1 µm/s were corrected for hydrodynamic drag [150]. The fits were obtained using (16.6) on each linear regime (solid lines). The error bars represent the standard error of the mean. Two linear regimes can be differentiated corresponding to inner (fast loading rates) and outer (slow loading rates) barriers
101
102
F. Rico · E.P. Wojcikiewicz · V.T. Moy
measuring the rupture events have been developed to accelerate the data analysis process [139]. Weak molecular interactions involved in biology can be characterized by their free-energy landscape, which, in the simplest case, can be represented by a potential well with a single barrier. This barrier has to be overcome in order to break the molecular interaction. Thermal motion determines the kinetics of unbinding when no force is applied. However, AFM measurements involve the application of a force that distorts the energy landscape by lowering its barrier height. This theory was first proposed by Bell [120] in 1978 for a constant applied force. Evans and Ritchie [140, 141] developed it for an increasing ramp of force, which better describes common force measurements. According to this theory, the rupture force will depend on the logarithm of the loading rate, i. e., the rate at which force is applied ( dF/ dt or rf ). A large number of studies have tested this prediction experimentally [24, 25, 41, 122, 132, 141–146]. Thus, to completely characterize the dynamic strength of molecular bonds, rupture forces have to be measured ranging over several orders of magnitude of the loading rate. In AFM experiments, this parameter is controlled by varying the spring constant of the cantilever and the retraction speed, from a few namometers per second to tens of microns per second. Owing to the elasticity of cells and adhesion molecules, the actual loading rate is determined by the effective spring constant of the cantilever–sample system in series. The loading rate is then obtained from the slope immediately before rupture in a force–time curve (Fig. 16.5). The loading rates measured in experiments using living cells usually cover 3 or 4 orders of magnitude, from tens of piconewtons per second to hundreds of nanonewtons per second.Owing to the stochastic nature of single-molecule bond rupture, hundreds of rupture events have to be grouped for each fixed loading rate and organized into histograms. A Gaussian fit is applied to each histogram and provides the mean rupture force for each loading rate (Fig. 16.6a) [122, 127, 142, 147, 148]. The resulting force versus loading rate plot is called the dynamic force spectrum of the interaction (Fig. 16.6b). According to Bell’s theory, each linear regime in the semilogarithmic plot corresponds to an energy barrier whose width (xβ ) and dissociation rate (ko ) under no force can be obtained from the following equation: xβ kB T kB T + ln o ln rf . (16.6) F= xβ k kB T xβ Fitting the above equation for each linear regime to the experimental data will provide us with the unknown variables that characterize the molecular interaction. More elaborate models have been proposed, which may also allow us to determine the potential barrier height, although their practical application appears to be more complex [149]. 16.3.4.5 Whole-Cell Adhesion Measurements Even if single-molecule interactions are the ultimate determinants of cell adhesion, cells usually attach to each other and to the extracellular matrix by forming clusters of
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
103
receptors which may cooperate and break in a more intricate manner [151–156]. To explain the rupture of multiple-bond attachments, the three simplest configurations are N identical bonds loaded in series, zipperlike attachment of N identical bonds, and N identical bonds loaded in parallel. The first configuration is common in protein unfolding measurements, in which each bond (or protein) experiences the same force and ruptures (or unfolds) in random order [157]. In a zipper, the same force propagates from one bond to the next after rupture of the lead bond has taken place. Bond rupture occurs in sequence from first to last. Finally, in the parallel-bonds arrangement, the force is shared by all the bonds that are involved in the attachment and rupture may occur cooperatively. The last of these is probably the configuration more physiologically relevant and the one which better describes multiple adhesion measurements obtained with atomic force microscopy on living cells. Whole-cell adhesion is usually measured by obtaining high-amplitude curves (tens of microns) and keeping the cell in contact with the ligand-coated surface for a longer time (a few seconds) than in the single-molecule measurements. The retract trace is then analyzed by measuring the maximum force and the total work required to completely detach the two surfaces [40, 41, 132]. Owing to the large area of contact and the long contact time, multiple bonds are typically formed, leading to multiple rupture events [124]. Retract traces of the force–distance curves usually present a peak of force followed by a cascade of rupture events until complete detachment is achieved (Fig. 16.7). In living cells, long tethers may also form. Curves then present a long and constant offset of force, whose value determines the force needed to extract the tether from the cell membrane [144]. The maximum peak of force in the retract trace of the curve determines the detachment force, which is usually hundreds of piconewtons or a few nanonewtons. The work of adhesion (or deadhesion) is computed from the area under the unloading curve on a force–distance plot (shaded region in Fig. 16.7). The dependence of the single-bond rupture force on the loading rate may have an influence on the measured adhesion strength; thus, dynamic measurements at different loading rates might be required. The slope just before initial rupture in a force–time plot provides an estimate of the loading force. This loading rate, however, is not usually determined since the force increase before
Fig. 16.7. Unbinding curves involving the rupture of multiple bonds. a A T cell, which expresses integrins on its surface, was attached to the cantilever and brought into contact with a surface coated with integrin ligands. b The same cell in the presence of ligand-blocking antibody. The retraction curves were obtained at a cantilever retraction speed of 5 µm/s
104
F. Rico · E.P. Wojcikiewicz · V.T. Moy
rupture is usually nonlinear; therefore, the cantilever retraction speed is the parameter used. The forces involved in whole-cell adhesion depend on the contact area, the time of contact, the loading rate, the density of receptors and the strength of the specific interaction [124]. As an illustrative example, Fig. 16.7 shows the retraction curves obtained with living cells under different conditions. The decrease in both force and work of de-adhesion is noticeable when blocking the ligand with antibody and reflects the suitability of these measurements to quantify cell adhesion strength. Both force and work of deadhesion are good estimates of cell adhesion strength, but they depend on the specific experimental conditions. Therefore, in order to be able to compare data from different cell systems and methods, such as the biomembrane force probe, it is necessary to estimate a parameter which describes intrinsic cell adhesion properties such as force or energy density. This may require knowledge of the contact area, which is not always possible. In both single-bond and multiple-bond measurements, the hydrodynamic drag force introduces a systematic bias in the force that has to be corrected. As mentioned before, the drag force is proportional to the tip velocity and to the drag coefficient b. Both rupture forces are corrected by adding the product of vb(h) to the measured peak force. Since the drag coefficient depends on the tip–sample distance (h), a previous calibration may be obtained to determine the correction factor taking h into account. Typically a constant drag coefficient is estimated and used to correct for the drag force. At high retraction speeds (tens of microns per second) and when the interaction lasts several hundreds of nanometers, this simplified approach may introduce important errors and should be avoided, especially in single-molecule measurements where the drag force is of the same order of magnitude as the actual rupture forces [158]. Therefore, measurements at high retraction speeds may be inaccurate and should be avoided. High loading rates can be achieved by using stiffer and smaller cantilevers and lower retraction speeds, thus minimizing the drag forces. Finally, an important issue at either single-molecule or multiple-molecule level concerns the specificity of the interaction. Negative controls are required to ensure that the probed ligand–receptor interaction is specific. There are different procedures available. Similar measurements using uncoated surfaces are the simplest. More sophisticated and effective measures may require cell mutagenesis to obtain cells without the specific protein, or the use of specific antibodies directed against the molecules involved to block the interaction. In practice, it is recommended to carry out complementary negative controls.
16.4 Concluding Remarks and Future Directions We have seen that the AFM is a very versatile tool to measure cellular mechanics in living cells, a feature that complements its ability to image cells. Recent articles also show the ability of the AFM to map the viscoelastic and adhesive properties of the whole cell surface with nanometer resolution and single-molecule recognition [77, 95, 123, 134]. The development of models describing the behavior of living cell mechanics is the next area of research to advance. There has been intense progress in this area in recent years [9, 15, 92, 93, 104, 107, 149, 151, 153, 154, 156, 159–167].
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
105
The next technological step would be the miniaturization of the AFM to enable us to measure the mechanical properties of living cells in their native environment, the human body.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.
Johnson KL (1985) Contact mechanics. Cambridge University Press, Cambridge Zhu C, Bao G, Wang N (2000) Annu Rev Biomed Eng 2:189 Heidemann SR, Wirtz D (2004) Trends Cell Biol 14:160 Janmey PA, Weitz DA (2004) Trends Biochem Sci 29:364 Sheetz MP, Sable JE, Dobereiner H-G (2006) Annu Rev Biophys Biomol Struct 35:417 Chicurel ME, Chen CS, Ingber DE (1998) Curr Opin Cell Biol 10:232 Goldmann WH, Ingber DE (2002) Biochem Biophys Res Commun 290:749 Huang S, Ingber DE (1999) Nat Cell Biol 1:E131 Ingber DE (1993) J Cell Sci 104:613 Ingber DE, Prusty D, Zhengqi S, Betensky H, Ning W (1995) J Biomech 28:1471 Maniotis AJ, Chen CS, Ingber DE (1997) Proc Natl Acad Sci USA 94:849 Pourati J, Maniotis A, Spiegel D, Schaffer JL, Butler JP, Fredberg JJ, Ingber DE, Stamenovic D, Wang N (1998) Am J Physiol 274(5 Pt 1):C1283 Wang N, Butler JP, Ingber DE (1993) Science 260:1124 Binnig G, Quate CF, Gerber C (1986) Phys Rev Lett 56:930 Ingber DE, Dike L, Hansen L, Karp S, Liley H, Maniotis A, McNamee H, Mooney D, Plopper G, Sims E (1994) Int Rev Cytol 150:173 Burnham NA, Colton RJ (1989) J Vac Sci Technol A 7:2906 Weisenhorn AL, Hansma PK, Albrecht TR, Quate CF (1989) Appl Phys Lett 54:2651 Butt HJ, Wolff EK, Gould SA, Dixon NB, Peterson CM, Hansma PK (1990) J Struct Biol 105:54 Haberle W, Horber JKH, Binnig G (1991) J Vac Sci Technol B 9:1210 Henderson E, Haydon PG, Sakaguchi DS (1992) Science 257:1944 Tao NJ, Lindsay SM, Lees S (1992) Biophys J63:1165 Hoh JH, Schoenenberger CA (1994) J Cell Sci 107:1105 Florin EL, Rief M, Lehmann H, Ludwig M, Dornmair C, Moy VT, Gaub HE (1995) Biosensors Bioelectron 10:895 Baumgartner W, Hinterdorfer P, Ness W, Raab A, Vestweber D, Schindler H, Drenckhahn D (2000) Proc Natl Acad Sci USA 97:4005 Baumgartner W, Hinterdorfer P, Schindler H, Drenckhahn D (2000) Biophys J 78:446A Crick SL, Yin FC (2007) Biomech Model Mechanobiol 6:199 Wu HW, Kuhn T, Moy VT (1998) Scanning 20:389 Goldmann WH, Galneder R, Ludwig M, Kromm A, Ezzell RM (1998) FEBS Letters 424:139 Goldmann WH, Galneder R, Ludwig M, Xu W, Adamson ED, Wang N, Ezzell RM (1998) Exp Cell Res 239:235 Dimitriadis EK, Horkay F, Maresca J, Kachar B, Chadwick RS (2002) Biophys J 82:2798 Dvorak JA (2003) Methods 29:86 Domke J, Dannohl S, Parak WJ, Muller O, Aicher WK, Radmacher M (2000) Colloids Surf B 19:367 Charras GT, Horton MA (2002) Biophys J 83:858 Horton MA, Charras GT, Ballestrem G, Lehenkari PP (2000) Single Mol 1:135 Lehenkari PP, Charras GT, Nykanen A, Horton MA (2000) Ultramicroscopy 82:289
106
F. Rico · E.P. Wojcikiewicz · V.T. Moy
36. Mathur AB, Collinsworth AM, Reichert WM, Kraus WE, Truskey GA (2001) J Biomech 34:1545 37. Rotsch C, Radmacher M (2000) Biophys J 78:520 38. Obataya I, Nakamura C, Han S, Nakamura N, Miyake J (2005) Nano Lett 5:27 39. Mahaffy RE, Shih CK, MacKintosh FC, Kas J (2000) Phys Rev Lett 85:880 40. Wojcikiewicz EP, Zhang X, Chen A, Moy VT (2003) J Cell Sci 116:2531 41. Zhang XH, Chen A, De Leon D, Li H, Noiri E, Moy VT, Goligorsky MS (2004) Am J Physiology Heart Circ Physiol 286:H359 42. Lee I, Marchant RE (2000) Colloids Surf B 19:357 43. A-Hassan E, Heinz W, Antonik MD, D’Costa NP, Nageswaran S, Schoenenberger CA, Hoh JH (1998) Biophys J 74:1564 44. Han SW, Nakamura C, Obataya I, Nakamura N, Miyake J (2005) Biosensors Bioelectron 20:2120 45. Benoit M, Gaub HE (2002) Cells Tissues Organs 172:174 46. You HX, Lau JM, Zhang S, Yu L (2000) Ultramicroscopy 82:297 47. Binnig G, Rohrer H (1982) Helv Phys Acta 55:726 48. Bustamante C, Dunlap D (1991) Semin Cell Biol 2:179 49. Dunlap DD, Bustamante C (1989) Nature 342:204 50. Muller DJ, Sapra KT, Scheuring S, Kedrov A, Frederix PL, Fotiadis D, Engel A (2006) Curr Opin Struct Biol 16:489 51. Hansma H, Pietrasanta L (1998) Curr Opin Chem Biol 2:767 52. Hansma HG, Bezanilla M, Zenhausern F, Adrian M, Sinsheimer RL (1993) Nucleic Acids Res 21:505 53. Poggi MA, Gadsby ED, Bottomley LA, King WP, Oroudjev E, Hansma H (2004) Anal Chem 76:3429 54. Simon A, Durrieu M-C (2006) Micron 37:13 55. Hansma PK, Drake B, Grigg D, Prater CB, Yashar F, Gurley G, Eling SV, Feinstein S, Lal R (1994) J Appl Phys 76:796 56. Putman CAJ, Degrooth BG, Vanhulst NF, Greve J (1992) J Appl Phys 72:6 57. Butt HJ, Cappella B, Kappl M (2005) Surf Sci Rep 59:1 58. Cappella B, Dietler G (1999) Surf Sci Rep 34:1 59. Gibson CT, Watson GS, Myhra S (1996) Nanotechnology 7:259 60. Chon JWM, Mulvaney P, Sader JE (2000) J Appl Phys 87:3978 61. Sader JE (1995) Rev Sci Instrum 66:4583 62. Sader JE (1998) J Appl Phys 84:64 63. Sader JE, Larson I, Mulvaney P, White L (1995) Rev Sci Instrum 66:3789 64. Sader JE, White L (1993) J Appl Phys 74:1 65. Hutter JL, Bechhoefer J (1993) Rev Sci Instrum 64:1868 66. Stark RW, Drobek T, Heckl WM (2001) Ultramicroscopy 86:207 67. Cleveland JP, Manne S, Bocek D, Hansma PK (1993) Rev Sci Instrum 64:403 68. Walters DA, Cleveland JP, Thomson NH, Hansma PK, Wendman MA, Gurley G, Elings V (1996) Rev Sci Instrum 67:3583 69. Butt HJ, Jaschke M (1995) Nanotechnology 6:1 70. Gibson CT, Watson GS, Myhra S (1997) Scanning 19:564 71. Gibson CT, Weeks BL, Abell C, Rayment T, Myhra S (2003) Ultramicroscopy 97:113 72. Burnham NA, Chen X, Hodges CS, Matei GA, Thoreson EJ, Roberts CJ, Davis MC, Tendler SJB (2003) Nanotechnology 14:1 73. Gibson CT, Smith DA, Roberts CJ (2005) Nanotechnology 16:234 74. Rico F, Roca-Cusachs P, Gavara N, Farre R, Rotger M, Navajas D (2005) Phys Rev e72 75. Rief M, Grubmuller H (2002) Chemphyschem 3:255
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
107
76. Frederix P, Akiyama T, Staufer U, Gerber C, Fotiadis D, Muller DJ, Engel A (2003) Cur Opin Chem Biol 7:647 77. Hinterdorfer P, Dufrene YF (2006) Nat Methods 3:347 78. Stolz M, Raiteri R, Daniels AU, VanLandingham MR, Baschong W, Aebi U (2004) Biophys J 86:3269 79. Hertz H (1881) Hertz’s miscellaneous papers (miscellaneous papers). Macmillan, London, p 146 80. Sneddon IN (1965) Int J Eng Sci 3:47 81. Radmacher M, Fritz M, Kacher CM, Cleveland JP, Hansma PK (1996) Biophys J 70:556 82. Radmacher M (1997) IEEE Eng Med Biol Mag 16:47 83. Domke J, Radmacher M (1998) Langmuir 14:3320 84. Rico F, Alcaraz J, Fredberg JJ, Navajas D (2005) Int J Nanotechnol 2:180 85. Roca-Cusachs P, Almendros I, Sunyer R, Gavara N, Farre R, Navajas D (2006) Biophys J 91:3508 86. Mahaffy RE, Park S, Gerde E, Kas J, Shih CK (2004) Biophys J 86:1777 87. Coughlin MF, Stamenovic D (2003) Biophys J 84:1328 88. Trepat X, Grabulosa M, Puig F, Maksym GN, Navajas D, Farre R (2004) Am J Physiol Lung Cell Mol Physiol 31:L1025 89. Fernandez P, Pullarkat PA, Ott A (2006) Biophys J 90:3796 90. Stamenovic D, Coughlin MF (1999) J Theor Biol 201:63 91. Stamenovic D, Coughlin MF (2000) J Biomech Eng 122:39 92. Stamenovic D, Ingber DE (2002) Biomech Model Mechanobiol 1:95 93. Wang N, Naruse K, Stamenovic D, Fredberg JJ, Mijailovich SM, Tolic-Norrelykke IM, Polte T, Mannix R, Ingber DE (2001) Proc Natl Acad Sci USA 98:7765 94. Wang N, Tolic-Norrelykke IM, Chen J, Mijailovich SM, Butler JP, Fredberg JJ, Stamenovic D (2002) Am J Physiol Cell Physiol 282:C606 95. Touhami A, Nysten B, Dufrene YF (2003) Langmuir 19:4539 96. Costa KD, Yin FCP (1999) J Biomech Eng 121:462 97. Briscoe BJ, Sebastian KS, Adams MJ (1994) J Phys D Appl Phys 27:1156 98. Costa KD, Sim AJ, Yin FC (2006) J Biomech Eng 128:176 99. Radmacher M, Fritz M, Hansma PK (1995) Biophys J 69:264 100. Hansma PK, Cleveland JP, Radmacher M, Walters DA, Hillner PE, Bezanilla M, Fritz M, Vie D, Hansma HG, Prater CB, Massie J, Fukunaga L, Gurley J, Elings V (1994) Appl Phys Lett 64:1738 101. Thoumine O, Ott A (1997) J Cell Sci 110:2109 102. Thoumine O, Ott A, Cardoso O, Meister JJ (1999) J Biochem Biophys Methods 39:47 103. Maksym GN, Fabry B, Butler JP, Navajas D, Tschumperlin DJ, Laporte JD, Fredberg JJ (2000) J Appl Physiol 89:1619 104. Fabry B, Maksym GN, Butler J, Glogauer M, Navajas D, Fredberg JJ (2001) Phys Rev Lett 87:148102 105. Puig-de-Morales M, Grabulosa M, Alcaraz J, Mullol J, Maksym GN, Fredberg JJ, Navajas D (2001) J Appl Physiol 91:1152 106. Puig-de-Morales M, Millet E, Fabry B, Navajas D, Wang N, Butler JP, Fredberg JJ (2004) Am J Physiol Cell Physiol 287:C643 107. Fabry B, Maksym GN, Butler JP, Glogauer M, Navajas D, Taback NA, Millet EJ, Fredberg JJ (2003) Phys Rev e68 108. Alcaraz J, Buscemi L, Grabulosa M, Trepat X, Fabry B, Farre R, Navajas D (2003) Biophys J 84:2071 109. Stamenovic D, Suki B, Fabry B, Wang N, Fredberg JJ (2004) J Appl Physiol 96:1600 110. Smith BA, Tolloczko B, Martin JG, Grutter P (2005) Biophys J 88:2994
108
F. Rico · E.P. Wojcikiewicz · V.T. Moy
111. Trepat X, Grabulosa M, Buscemi L, Rico F, Farre R, Navajas D (2005) J Appl Physiol 98:1567 112. Bursac P, Lenormand G, Fabry B, Oliver M, Weitz DA, Viasnoff V, Butler JP, Fredberg JJ (2005) Nat Mater 4:557 113. Bausch AR, Hellerer U, Essler M, Aepfelbacher M, Sackmann E (2001) Biophys J 80:2649 114. Desprat N, Richert A, Simeon J, Asnacios A (2005) Biophys J 88:2224 115. Alcaraz J, Buscemi L, Puig-de-Morales M, Colchero J, Baro A, Navajas D (2002) Langmuir 18:716 116. Burridge K, Chrzanowska-Wodnicka M (1996) Annu Rev Cell Dev Biol 12:463 117. Burridge K, Tsukita S (2001) Curr Opin Cell Biol 13:528 118. Vogel V, Sheetz M (2006) Nat Rev Mol Cell Biol 7:265 119. Chen CS, Tan J, Tien J (2004) Annu Rev Biomed Eng 6:275 120. Bell GI (1978) Science 200:618 121. Springer TA (1990) Nature 346:425 122. Lehenkari PP, Horton MA (1999) Biochem Biophys Res Commun 259:645 123. Hyonchol K, Arakawa H, Osada T, Ikai A (2002) Colloids Surf B 25:33 124. Zhang X, Wojcikiewicz EP, Moy VT (2006) Exp Biol Med (Maywood) 231:1306 125. Wagner P (1998) FEBS Lett 430:112 126. Florin EL, Moy VT, Gaub HE (1994) Science 264:415 127. Moy VT, Florin EL, Gaub HE (1994) Biophys J 66:A340 128. Hinterdorfer P, Baumgartner W, Gruber HJ, Schilcher K, Schindler H (1996) Proc Natl Acad Sci USA 93:3477 129. Schutz GJ, Sonnleitner M, Hinterdorfer P, Schindler H (2000) Mol Membr Biol 17:17 130. Moy VT, Chen A (2000) Biophys J 78:447A 131. Hinterdorfer P, Schutz G, Kienberger F, Schindler H (2001) J Biotechnol 82:25 132. Zhang XH, Wojcikiewicz E, Moy VT (2002) Biophys J 83:2270 133. Zhang XH, Bogorin DF, Moy VT (2004) Chemphyschem 5:175 134. Stroh CM, Ebner A, Geretschlager M, Freudenthaler G, Kienberger F, Kamruzzahan AS, Smith-Gill SJ, Gruber HJ, Hinterdorfer P (2004) Biophys J 87:1981 135. Rief M, Junker JP, Schlierf M, Hell K, Neupert W (2006) Biophys. J 91:2011 136. Ainavarapu SRK, Li L, Fernandez JM (2006) Biophys J 91:2009 137. Tees DFJ, Waugh RE, Hammer DA (2001) Biophys J 80:668 138. Evans E (1998) Faraday Discus 111:1 139. Baumgartner W, Hinterdorfer P, Schindler H (2000) Ultramicroscopy 82:85 140. Evans E, Ritchie K (1997) Biophys J 72:1541 141. Merkel R, Nassoy P, Leung A, Ritchie K, Evans E (1999) Nature 397:50 142. Benoit M, Gabriel D, Gerisch G, Gaub HE (2000) Nat Cell Biol 2:313 143. Zhang XH, Craig SE, Kirby H, Humphries MJ, Moy VT (2004) Biophys J 87:3470 144. Sun M, Graham JS, Hegedus B, Marga F, Zhang Y, Forgacs G, Grandbois M (2005) Biophys J 89:4320 145. Li FY, Redick SD, Erickson HP, Moy VT (2003) Biophys J 84:1252 146. Wojcikiewicz EP, Abdulreda MH, Zhang X, Moy VT (2006) Biomacromolecules 7:3188 147. Li YQ, Tao NJ, Pan J, Garcia AA, Lindsay SM (1993) Langmuir 9:637 148. Kokkoli E, Ochsenhirt SE, Tirrell M (2004) Langmuir 20:2397 149. Dudko OK, Hummer G, Szabo A (2006) Phys Rev Lett 96 150. Evans E, Leung A, Hammer D, Simon S (2001) Proc Natl Acad Sci USA 98:3784 151. Seifert U (2000) Phys Rev Lett 84:2750 152. van Kooyk Y, Figdor CG (2000) Curr Opin Cell Biol 12:542 153. Evans E (2001) Annu Rev Biophys Biomol Struct 30:105 154. Seifert U (2002) Europhys Lett 58:792
16 Atomic Force Microscopy Studies of the Mechanical Properties of Living Cells
109
155. Prechtel K, Bausch AR, Marchi-Artzner V, Kantlehner M, Kessler H, Merkel R (2002) Phys Rev Lett 89:028101 156. Williams PM (2003) Anal Chim Acta 479:107 157. Rief M, Gautel M, Oesterhelt F, Fernandez JM, Gaub HE (1997) Science 276:1109 158. Janovjak HJ, Struckmeier J, Muller DJ (2005) Eur Biophys J 34:91 159. Stamenovic D, Wang N (2000) J Appl Physiol 89:2085 160. Hummer G, Szabo A (2005) Acc Chem Res 38:504 161. Hummer G, Szabo A (2003) Biophys J 85:5 162. Hummer G, Szabo A (2001) Proc Natl Acad Sci USA 98:3658 163. Dudko OK, Filippov AE, Klafter J, Urbakh M (2003) Proc Natl Acad Sci USA 100:11378 164. Stamenovic D, Fredberg JJ, Wang N, Butler JP, Ingber DE (1996) J Theor Biol 181:125 165. Erdmann T, Schwarz US (2004) J Chem Phys 121:8997 166. Erdmann T, Schwarz US (2004) Europhys Lett 66:603 167. Erdmann T, Schwarz US (2004) Phys Rev Lett 92:108102 168. Bilodeau G (1992) J Appl Mech Trans ASME 59:519
17 Towards a Nanoscale View of Microbial Surfaces Using the Atomic Force Microscope Claire Verbelen · Guillaume Andre · Xavier Haulot · Yann Gilbert · David Alsteens · Etienne Dague · Yves F. Dufrêne
Abstract. In recent years, the atomic force microscope (AFM) has greatly improved our understanding of microbial surfaces. AFM imaging has proved to be a powerful tool for visualizing membrane proteins and live cells at high resolution and in physiological conditions. In addition, AFM force spectroscopy has enabled us to probe and map a variety of properties, including the elasticity of cell walls and cell surface molecules, and the unfolding forces of single proteins, and has allowed the detection and functional analysis of molecular recognition sites. These unique capabilities allow researchers to answer a number of questions that were inaccessible before, such as how does the surface architecture of microbes change as they grow or as they interact with an antibiotic, what are the conformational changes in single membrane proteins, and what are the molecular forces responsible for the interaction between host cells or specific molecules? Key words: Atomic force microscopy, Cell surfaces, Living cells, Imaging, Force spectroscopy, Ultrastructure, Molecular recognition, Single molecule, Surface properties, Membrane proteins
17.1 Introduction Characterization of microbial surfaces is a key to gain insight into their structure and function and is particularly useful in the biomedical context for understanding microbial infection processes such as biofilm formation on medical implants and pathogen–host interactions. Traditionally, elucidation of the cell surface architecture has relied on transmission electron microscopy (TEM) and scanning electron microscopy techniques [1–5]. Remarkably, the development of cryo-TEM of frozen-hydrated thin sections has allowed researchers to get more natural views of bacterial cell envelopes. Unfortunately, these cryogenic methods are very demanding in terms of sample preparation and analysis and are only applied in a few laboratories worldwide. Valuable information on the composition, properties and interactions of cell surfaces can also be gained using electron microscopy approaches, biochemical analysis, biophysical techniques and surface analysis methods [4, 5]. Although very useful, these techniques generally require cell manipulation prior to examination and often provide averaged information obtained on large ensembles of cells. Consequently, there is clearly a need to develop ultrasensitive, highresolution probes for mapping the structure, properties and interactions of microbial surfaces. Recent studies have revealed that atomic force microscopy (AFM) can provide information on microbial surfaces that cannot be obtained with classical
112
C. Verbelen et al.
methods [5–9]. Basically, AFM offers two unique advantages over other microscopies. First, it provides surface images of the sample with nanometer (even subnanometer) resolution, in real time and under physiological conditions. Second, the force sensitivity of the instrument can be exploited to measure physical properties and interactions (e. g., elasticity, adhesion) on a local scale, and to detect and analyze single molecular recognition events. The aim of this chapter is to provide a flavor of the different possibilities offered by AFM in microbiology.
17.2 Imaging 17.2.1 Sample Preparation An important prerequisite for successful AFM experiments is sample preparation, i. e., the specimen must be firmly attached to a solid support in order to withstand the lateral forces exerted by the scanning tip. While isolated membranes are generally well immobilized by simple adsorption on mica [10], stronger attachment is often needed for whole cells. This can be achieved by drying the sample or by bonding the cells covalently to the support. However, these treatments may cause significant rearrangement, denaturation or contamination of the cell walls, yielding information that may no longer be representative of the native state. More appropriate approaches include treating the support with a polycation to strengthen cell adhesion or immobilizing the cells mechanically in an agar gel or in a porous membrane. In the agar method, the gel is used as a soft, deformable immobilization matrix, thereby allowing direct visualization of growth processes [11]. This enabled in situ investigation of the growth process of Saccharomyces cerevisiae cells and direct visualization of micrometer-sized bud scars. In the porous membrane method, spherical cells are trapped in a polymer membrane with a pore size comparable to the dimensions of the cell, allowing repeated imaging without cell detachment or cell damage [12, 13]. 17.2.2 Visualizing Membrane Proteins at Subnanometer Resolution During the past decade, the use of AFM has afforded researchers a glimpse of many different membrane proteins, of which Bacillus S-layers [14], the hexagonally packed intermediate (HPI) layer of Deinococcus radiodurans [15], purple membrane from the archeon Halobacterium [16] and porin 2D crystals of Escherichia coli [17, 18]. AFM yields structural information to a resolution of 0.5–1 nm under physiological conditions, which makes it a complementary tool to X-ray and electron crystallography. Remarkably, function-related conformational changes can be monitored on single proteins, as nicely demonstrated for porin OmpF in which voltage-induced and pH-induced channel closure was observed [19]. In addition, there is clear evidence that the instrument is evolving from an imaging tool to a “lab
17 AFM of Microbial Surface
113
Fig. 17.1. High-resolution imaging of native membrane proteins. High-resolution atomic force microscopy (AFM) image of a high-light adapted native photosynthetic membrane of Rhodospirillum photometricum revealing the complex mixture of LH2 (small rings) and core complexes (large rings). (Reproduced with permission from S. Scheuring)
on a tip” multifunctional tool [8, 20], enabling the study of the unfolding pathways, the assembly, oligomeric states and molecular interactions of membrane proteins. In the first such study, Müller et al. [21] used the atomic force microscope tip as a molecular tweezer to reversibly push away single loops of bacteriorhodopsin proteins, thereby accessing structures otherwise hidden. More recent work has shown that the combined use of AFM imaging and single-molecule force spectroscopy makes it possible to pull on single membrane proteins, thereby providing novel molecular insight into their unfolding pathways and assembly forces [22–26]. Recently, a number of studies have focused on noncrystalline native membranes of different photosynthetic bacteria, including Rhodopseudomonas viridis [27], Rhodospirillum photometricum [28, 29], Rhodobacter sphaeroides [30], Rhodobacter blasticus [31] and Rhodopseudomonas palustris [32], providing new insight into the supramolecular organization of membrane protein complexes. In an elegant study, AFM was used to investigate how the composition and architecture of photosynthetic membranes of R. photometricum change in response to light [29] (Fig. 17.1). Despite large modifications in the membrane composition, the local environment of core complexes remained unaltered, whereas specialized paracrystalline lightharvesting antenna domains grew under low-light conditions. It was concluded that such structural adaptation ensures efficient photon capture under low-light conditions and prevents photodamage under high-light conditions. 17.2.3 Live-Cell Imaging AFM is being used increasingly for imaging living cells, revealing structural details of hydrated cell walls with unprecedented resolution. Examples of microbial species that have been explored by AFM are S. cerevisiae [33], Phanerochaete chrysosporium [13], Aspergillus oryzae [34], Aspergillus nidulans [35], Pinnularia viridis [36],
114
C. Verbelen et al.
lactic acid bacteria [37,38], Bacillus spores [39,40], Staphylococcus aureus [41,42], E. coli [41], Mycobacterium bovis [43] and vaccinia virus [44]. Figure 17.2 illustrates the power of high-resolution AFM imaging in mycological research [34]. As can be seen, the surface of A. oryzae spores is covered with a crystalline-like array of rodlets that are 10 nm in diameter. These “nanobio-arrays” are known to cover the surface of many fungal spores and are thought to have different biological functions, such as spore dissemination and protection. Dramatic changes of cell surface structure are observed upon spore germination, the rodlet layer changing into a layer of soft material within a few hours. On close examination, this material shows streaks oriented in the scanning direction, suggesting that soft, loosely bound material is interacting strongly with the scanning tip. These direct observations are in good agreement with previous structural and chemical studies which showed that germination of Aspergillus spores results in the disruption of the proteinaceous rodlet layer and reveals inner spore walls which are essentially composed of polysaccharides.
Fig. 17.2. Live-cell imaging. Optical micrographs of fungal spores from Aspergillus oryzae in the dormant state (a) and after germination for a few hours in a liquid medium (b). AFM deflection images recorded in aqueous solution for the surface of dormant (c) and germinating (d) spores revealing that, upon germination, the crystalline rodlet layer changes into a layer of soft material. (Reproduced with permission from [34]. Copyright 2001 American Chemical Society)
17 AFM of Microbial Surface
115
AFM can be combined with thin-section TEM to monitor cell growth and division events, as nicely demonstrated for S. aureus [42]. Nanoscale holes were seen around the septal annulus at the onset of division and were attributed to cell wall structures possessing high autolytic activity. After cell separation, concentric rings were observed on the surface of the new cell wall and were suggested to reflect newly formed peptidoglycan. Remarkably, real-time imaging enables researchers to follow cell surfaces as they interact with enzymes and antibiotics. For example, the effect of protease on the cell surface architecture of S. cerevisiae cells was monitored in real time [33]. Progressive
Fig. 17.3. Nanoscale imaging of drug-induced alterations. a Deflection image recorded in aqueous solution showing native mycobacteria immobilized on a polymer surface. b High-resolution image recorded on top of a native cell revealing a smooth, homogeneous surface. c Highresolution image of the cell surface following incubation for 24 h with 4 µg/ml ethambutol showing that the drug created concentric striations. d High-resolution image of the cell surface following incubation for 24 h with 10 µg/ml ethambutol revealing alteration of cell wall layers of 8- and 12-nm thicknesses, attributed to the outer electron-opaque and thick electron-transparent layers usually seen by electron microscopy. (Reproduced with permission from [43])
116
C. Verbelen et al.
changes of the cell surface topography were clearly observed after protease addition. With time, the surface became eroded and showed large depressions, about 500 nm in diameter, surrounded by protruding edges, about 50 nm in height. By contrast, no modification of the cell surface was noted upon addition of amyloglucosidase, which was consistent with the cell wall biochemical composition. In a biomedical context, in situ AFM was used to investigate the effect of the drug ethambutol on the cell wall architecture of mycobacteria [43] (Fig. 17.3). Consistent with the mode of action of ethambutol, which primarily inhibits arabinan polymerization, the drug was found to dramatically alter the fine-surface architecture of the cells, by removing outer cell wall layers to an extent that depends on the drug concentration and by revealing concentric striations on top of the remaining layers. In the future, these noninvasive ultrastructural investigations will have a great potential to provide new insight into the mode of action of antibiotics as well as into the mechanisms of cell wall assembly.
17.3 Force Spectroscopy 17.3.1 Customized Tips Probing surface properties and biomolecular interactions by force spectroscopy often requires the use of customized tips bearing well-defined chemical or biological moieties [45]. Chemically modified tips can be obtained by using organic monolayers terminated by specific functional groups (e. g., OH, COOH or CH3 ), the most common method relying on the formation of self-assembled monolayers (SAMs) of alkanethiols on gold surfaces. The procedure involves coating microfabricated cantilevers with a thin adhesive layer of Cr or Ti, followed by another thin layer of Au, and immersing the coated cantilevers in dilute ethanol solutions of the selected alkanethiol [46, 47]. Molecular recognition studies imply functionalizing the tips (and supports) with relevant ligands (and receptors) [45]. Here, several important issues must be considered: (1) the forces which immobilize the molecules should be stronger than the intermolecular force being studied; (2) the attached biomolecules should have enough mobility so that they can freely interact with complementary molecules; (3) the contribution of nonspecific adhesion to the measured forces should be minimized; (4) attaching biomolecules at low surface density is recommended in order to ensure single-molecule detection; (5) site-directed coupling may be desired to orientate all the interacting molecules in the same way. Immobilization strategies commonly used for making biofunctionalized tips are illustrated in Fig. 17.4 [45]. A first approach consists of modifying tips by adsorption of biotinylated bovine serum albumin and further reaction with avidin (or streptavidin) to attach biotinylated molecules [48–50]. However, this approach is not ideally suited for single molecular recognition studies because (1) parallel breakage of multiple bonds is often observed and (2) the biotin–avidin system has a fairly low binding strength compared with that of covalent bonds. To solve this problem, biomolecules can be firmly attached on atomic force microscope tips using either thiol or silane chemistry. In the first approach, biomolecules
17 AFM of Microbial Surface
117
Fig. 17.4. Strategies commonly used for modifying atomic force microscope tips for singlemolecule recognition studies: physisorption of proteins such as biotinylated bovine serum albumin (BSA), chemisorption of alkanethiols on gold and covalent coupling of silanes on silicon oxide. EG ethylene glycol, NTA nitrilotriacetate, PEG poly(ethylene glycol) (Adapted with permission from [45])
are covalently attached onto SAMs of functionalized alkanethiols on gold tips, using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and N-hydroxysuccinimide (NHS) [51]. Mixing long-chain alkanethiols with COOH terminal functions in a matrix of shorter OH-terminated alkanethiols ensures a certain mobility of the attached biomolecules and minimizes nonspecific adsorption [52]. In the second approach, biomolecules are covalently attached onto silicon tips using various aminefunctionalization procedures [53]. The amino-terminated surfaces are reacted with a cross-linker (spacer) which provides the ligands with motional freedom and prevents their denaturation. The use of long spacer molecules is also advantageous for data treatment, for separating specific from nonspecific adhesion forces. Crosslinkers typically carry two different functional ends, e. g., an amine reactive NHS group on one end and 2-pyridyldithiopropionyl or vinyl sulfone groups, which can be covalently bound to thiols, on the other. Importantly, in the abovementioned approaches it is possible to orientate all biomolecules in the same way, thereby allowing an optimal exposure of the C-terminal or N-terminal domains of proteins, by attaching recombinant histidine-tagged proteins onto an atomic force microscope tip terminated with nitrilotriacetate groups [54, 55]. Finally, it is worth mentioning that for probing cellular interactions, cell probes can be created by attaching microbial cells onto atomic force microscope cantilevers, using glutaraldehyde [56], glue [57], amino groups [58] or lectins [59]. 17.3.2 Probing Nanoscale Elasticity and Surface Properties Because of the small size of microbes, measuring the physical properties of the cell wall at the subcellular level has always been very challenging. For the first
118
C. Verbelen et al.
time, AFM researchers can measure the nanoscale elasticity and surface properties (charge, hydrophobicity) of individual cells. The nanomechanical properties of cell walls can be estimated by pressing the atomic force microscope tip onto the wall components. In pioneering measurements, the elastic modulus of the sheath of the archeon Methanospirillum hungatei GP1 was found to be in the 20–40-GPa range, indicating that this layer of unusual strength could withstand an internal pressure of 400 atm [60]. So-called nanoindentation experiments showed the turgor pressure of Enterococcus hirae and Pseudomonas aeruginosae to be in the 105 -Pa range [61]. Similar nanomechanical analyses revealed that the elasticity of the S. cerevisiae cell wall varies significantly across a single cell, i. e., the Young’s modulus is 6 MPa on the bud scar and 0.6 MPa on the surrounding cell surface, in agreement with the accumulation of chitin in the bud scar area [62]. In this nanomechanical context, a crucial issue is to distinguish true sample deformation from repulsive surface forces, which can be difficult, particularly when working in aqueous solution. One way to clarify these two contributions is to vary the ionic strength or pH, as done by Gaboriaud et al. [63], who quantified the nanomechanical properties (bacterial spring constant, Young’s modulus) of Shewanella putrefaciens at different pH values. The data demonstrated a swelling effect induced by raising the pH that was attributed to a change of water mobility due to intermolecular repulsive forces inside the polymeric fringe. Beside force spectroscopy, it is worth emphasizing that atomic force microscope cantilevers have also been used to measure local nanomechanical motion of S. cerevisiae cell walls, in relation to metabolic processes within the cell [64]. Hydrophobicity and charge are two crucial cell surface properties that are involved in a variety of interfacial events, including cell–support, cell–cell, antigen– antibody, cell–drug and cell–ions interactions. Various experimental methods have been developed to assess these properties, among which are microelectrophoresis, colloid titration, isoelectric focusing of cells, ion-exchange chromatography and surface conductivity, water contact-angle measurements on cell lawns, adhesion to hydrocarbons, partitioning in aqueous two-phase systems and hydrophobic interaction chromatography [5]. Because these assays only provide averaged information obtained on a large ensemble of cells, developing complementary tools for probing properties at the subcellular level is an important challenge. For the first time, functionalized atomic force microscope tips enable researchers to measure local surface charge and surface hydrophobicity. Ahimou et al. [47] used tips functionalized with ionizable carboxyl groups (COO–/COOH) to probe the surface charges of S. cerevisiae at the nanometer level. Force–distance curves were strongly influenced by pH: while no adhesion was measured at neutral/alkaline pH, multiple adhesion forces were recorded at acidic pH. Adhesion maps always showed fairly homogeneous contrast, indicating that the cell surface properties were homogeneous. The change of adhesion forces as a function of pH was interpreted as resulting from a change of cell surface electrostatic properties, a claim which was supported by the correlation obtained between the AFM titration curve constructed by plotting the adhesion force versus pH and the electrophoretic mobility versus pH curve obtained by classical microelectrophoresis. With use of a similar approach, it has been shown that hydrophilic (OH) and hydrophobic (CH3 ) tips can be used to map cell surface hydrophobicity [46].
17 AFM of Microbial Surface
119
17.3.3 Stretching Cell Surface Polysaccharides and Proteins By pulling on single proteins and polysaccharides, researchers can probe their intramolecular and intermolecular forces, thereby providing new insight into their folding/unfolding forces, stability and elasticity [23, 65]. In particular, single-molecule AFM has allowed researchers to visualize single subunits of S-layer proteins and to probe their molecular interactions. In the first such study, force–extension curves recorded for D. radiodurans S-layers showed sawtooth patterns with six force peaks attributed to the sequential pulling out of multiple protomers of the S-layer. After the force curve had been recorded, a molecular defect the size of a hexameric complex was visualized by means of high-resolution imaging [22]. A similar approach was
Fig. 17.5. Measuring the molecular forces driving the folding and assembly of S-layer proteins. A The surface chemistry used to functionalize atomic force microscope tips and supports with CbsA proteins at low surface density. The entire mature CbsA protein as well as CbsA peptides truncated in their N-terminal and C-terminal regions were engineered with a His-tag for sitedirected immobilization onto NTA/EG terminated surfaces. B Typical force–extension curve obtained in phosphate-buffered saline (PBS) between tips and supports modified with CbsA proteins. The well-pronounced force peaks forming a sawtooth pattern are well-described using the wormlike chain model and are attributed to the unfolding of individual α-helical structures. C Interaction forces of the CbsA protein (triangles) and of CbsA peptides truncated in their N-terminal (circles) and C-terminal (squares) regions as a function of the pulling speed
120
C. Verbelen et al.
Fig. 17.6. Detection and mapping of individual adhesins. a Representative force curves and adhesion force histogram obtained in PBS between a heparin-binding hemagglutinin adhesin (HBHA) tip and a heparin surface. The adhesion force histogram revealed a bimodal distribution reflecting the binding strength of one and two adhesins. b Mapping cell-adhesion nanodomains on living bacteria. Adhesion force maps (gray scale 100 pN) recorded in PBS with heparinmodified atomic force microscope tips over Mycobacterium bovis BCG bacteria expressing HBHA adhesins. Adhesion events (clear pixels) were attributed to the detection of single HBHA molecules. Notably, the HBHA distribution was not homogeneous, but apparently concentrated into nanodomains which may play important biological functions. Scale bars 100 nm. (Reprinted with permission from [55]) ◮
used to unzip the S-layer of Corynebacterium glutamicum, providing new insight into the stability of this protective bacterial surface coat [66]. Recently, single-molecule force spectroscopy was used to gain insight into the molecular forces driving the folding and assembly of the CbsA S-layer protein from Lactobacillus crispatus, which is known to have important functions, including promoting binding to collagen, triggering self-assembly of the protein subunits and mediating attachment to teichoic acids of the cell wall [67] (Fig. 17.5). Force curves recorded between tips and supports modified with CbsA proteins (Fig. 17.5a) showed sawtooth patterns with multiple force peaks of 58 ± 26 pN that were attributed to the unfolding of α-helices, in agreement with earlier secondary structure predictions (Fig. 17.5b). The average unfolding force increased with the pulling speed but was independent of the interaction time (Fig. 17.5c). Force curves obtained for CbsA peptides truncated in their C-terminal region showed similar periodic features, except that fewer force peaks were seen. Furthermore, the average unfolding force was 83 ± 45 pN, suggesting the domains were more stable. By contrast, cationic peptides truncated in their N-terminal region showed single force peaks of 366 ± 149 pN, presumably reflecting intermolecular electrostatic bridges rather than unfolding events. Interestingly, these large intermolecular forces increased not only with pulling speed but also with interaction time. It was concluded that the measured intramolecular and intermolecular forces may play a significant role in controlling the stability and assembly of the CbsA protein. AFM has also been used to pull on microbial cell surface polysaccharides. For instance, force curves recorded on germinating A. oryzae spores showed large attractive forces along with elongation forces that were attributed to the stretching of single cell surface polysaccharides [34]. Indeed, fitting the curves with a theoretical model from statistical mechanics yielded parameters similar to those reported for individual amylose and dextran molecules [65]. It was concluded that the stickiness and flexibility of cell surface polysaccharide chains may promote spore aggregation via macromolecular bridging interactions between opposing cells. In another work, mannan polymers were stretched at the surface of S. cerevisiae cells using tips that were functionalized with the lectin concanavalin A [68]. 17.3.4 Nanoscale Mapping and Functional Analysis of Molecular Recognition Sites Functionalization of atomic force microscope tips with ligands (Sect. 17.3.1) enables researchers to detect complementary receptor molecules either on model surfaces or
17 AFM of Microbial Surface
121
122
C. Verbelen et al.
on cellular surfaces [45]. Molecular recognition is of major importance in bacterial pathogenesis, where the infectious process is generally initiated by the interaction between cell-adhesion molecules on the bacterial surface, referred to as adhesins, and specific ligands on the host cell surface. An example of this is found in tuberculosis, which is caused by the infectious agent Mycobacterium tuberculosis. Mycobacteria adhere to heparan sulfates on epithelial cells via the heparin-binding hemagglutinin adhesin (HBHA). To explore the binding forces of individual HBHA molecules, force curves were recorded between HBHA-modified tips and heparin-coated supports [55] (Fig. 17.6a). The adhesion force histogram obtained using a loading rate of 10,000 pN/s revealed a bimodal distribution with average rupture forces of 50 and 117 pN, attributed to one and two binding events between HBHA and heparin. The specificity of the measured interaction was confirmed by showing a dramatic reduction in the number of adhesion events when working in the presence of free heparin or with a bovine serum albumin coated tip instead of a HBHA tip. This work shows that single-molecule force spectroscopy is a powerful tool to gain insight into the molecular mechanisms of cell-adhesion processes. Next, AFM affinity imaging was used to map the distribution of single HBHA molecules on bacterial surfaces with nanoscale resolution [55] (Fig. 17.6b). Affinity maps (300 nm × 300 nm) recorded with a heparin-modified tip revealed adhesion events in about half of the locations, with a magnitude of approximately 50 pN, thus very similar to the value of the model HBHA–heparin interaction. This finding strongly suggests that the unbinding forces correspond to the detection of single HBHA molecules, which was further supported by the observation that a mutant strain lacking HBHA did not bind the heparin tip. Interestingly, the HBHA distribution was not homogeneous, but apparently concentrated into nanodomains which may promote adhesion to target cells by inducing the recruitment of receptors within membrane rafts. In summary, these nanoscale functional analyses contribute to shedding new light onto the molecular mechanisms of bacterial pathogenicity and may ultimately lead to new antiadhesion approaches for therapy.
Fig. 17.7. Imaging vancomycin binding sites on living Streptococcus thermophilus. a The experimental setup. b AFM image showing a single cell during the course of the division process. c Affinity map recorded with a vancomycin tip on the septum region highlighted by the white box in b. Single adhesion events are detected in the septum region and are attributed to the specific recognition of individual terminal d-Ala–d-Ala of nascent peptidoglycan
17 AFM of Microbial Surface
123
Molecular recognition is also involved in the action mode of important antibiotics such as vancomycin, which binds with high affinity and specificity to the terminal d-Ala–d-Ala of peptidoglycan precursors in Gram-positive bacteria, thereby preventing their incorporation into the bacterial cell wall. The use of vancomycinterminated atomic force microscope tips to detect single d-Ala–d-Ala sites on living bacteria was recently established (Fig. 17.7). Topographic images of living Streptococcus thermophilus cells revealed a smooth morphology as well as a well-defined division septum. Affinity maps recorded with vancomycin tips revealed that d-Ala– d-Ala sites were essentially located in the septum region, suggesting that newly formed peptidoglycan is inserted in these regions. Accordingly, these novel antibiotic nanoprobes offer exciting prospects for understanding the assembly of bacterial cell walls and for elucidating the mode of action of antibiotics that target cell wall constituents.
17.4 Conclusions AFM has recently opened a variety of novel possibilities for imaging and manipulating microbial surfaces in their native environment. While AFM imaging is useful for visualizing surface structures at high resolution and in physiological conditions, force spectroscopy enables researchers to probe a variety of cell surface properties, including the elasticity of cell walls and cell surface molecules, the unfolding forces of single proteins, and the detection and functional analysis of molecular recognition sites. These nanoscale analyses allow researchers to answer a number of questions that were difficult to address before, such as how does the surface architecture of microbes change as they grow or interact with antibiotics; what is the force required to unfold a membrane protein; and what are the molecular forces driving the interaction between a pathogen and a host surface or between an antibiotic and a bacterial surface? There are several technological issues to address in the future. First, the relatively slow imaging speed of commercially available atomic force microscopes (30–60 s per image) represents a crucial bottleneck for exploring cell surface dynamic processes like conformational changes. Fortunately, fast instruments have been introduced in the few last years, which enable recording up to 200 frames per second and open up fascinating new perspectives to explore molecular and cellular dynamics [69,70]. Another key issue when dealing with living cells is that the image resolution is often limited because the extremely soft cell surface is deformed by the scanning tip. The use of novel dynamic imaging modes which minimize the applied force [71], combined with improved sample preparation methods, may help solve this problem, thereby providing reliable images at much higher resolution than currently available. With such further technological developments (faster imaging, lower interaction forces), it should become possible to observe, for example, the binding of individual antibodies or drugs to distinct locations of the cell surface, function-related conformational changes of surface proteins, or local alterations of the cell surface topography associated with cellular processes such as cell growth and division.
124
C. Verbelen et al.
Advances in single-molecule force spectroscopy will allow elucidation of the elasticity, conformational changes and unfolding forces of an increasing number of membrane proteins and other cell surface constituents. Progress in developing reliable protocols for attaching biomolecules and cells to atomic force microscope cantilevers will enable the measurement of molecular interactions on microbial surfaces in the daily routine, providing novel insight into the molecular mechanisms of microbial infection processes (biofilms on medical implants, pathogen–host interactions) and new ideas for antiadhesion strategies. Acknowledgements. This work was supported by the National Foundation for Scientific Research (FNRS), the Foundation for Training in Industrial and Agricultural Research (FRIA), the Université Catholique de Louvain (Fonds Spéciaux de Recherche) and the Federal Office for Scientific, Technical and Cultural Affairs (Interuniversity Poles of Attraction Programme). Y.F.D. is a Research Associate of the FNRS.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
Beveridge TJ (1999) J Bacteriol 181:4725 Matias VRF, Beveridge TJ (2005) Mol Microbiol 56:240 Matias VRF, Beveridge TJ (2006) J Bacteriol 188:1011 Beveridge TJ, Graham LL (1991) Microbiol Rev 55:684 Ubbink J, Schar-Zammaretti P (2005) Micron 36:293 Engel A, Müller DJ (2000) Nat Struct Biol 7:715 Dufrêne YF (2004) Nat Rev Microbiol 2:451 Müller DJ, Sapra KT, Scheuring S, Kedrov A, Frederix PL, Fotiadis D, Engel A (2006) Curr Opin Struct Biol 16:489 Gaboriaud F, Dufrêne YF (2007) Colloids Surf B 54:10 Müller DJ, Amrein M, Engel A (1997) J Struct Biol 119:172 Gad M, Ikai A (1995) Biophys J 69:2226 Kasas S, Ikai A (1995) Biophys J 68:1678 Dufrêne YF, Boonaert CJP, Gerin PA, Asther M, Rouxhet PG (1999) J Bacteriol 181:5350 Györvary ES, Stein O, Pum D, Sleytr UB (2003) J Microsc 212:300 Müller DJ, Baumeister W, Engel A (1996) J Bacteriol 178:3025 Müller DJ, Schabert FA, Büldt G, Engel A (1995) Biophys J 68:1681 Schabert FA, Henn C, Engel A (1995) Science 268:92 Scheuring S, Ringler P, Borgnia M, Stahlberg H, Müller DJ, Agre P, Engel A (1999) EMBO J 18:4981 Müller DJ, Engel A (1999) J Mol Biol 285:1347 Janovjak H, Kedrov A, Cisneros DA, Sapra KT, Struckmeier J, Müller DJ (2006) Neurobiol Aging 27:546 Müller DJ, Büldt G, Engel A (1995) J Mol Biol 249:239 Müller DJ, Baumeister W, Engel A (1999) Proc Natl Acad Sci USA 96:13170 Oesterhelt F, Oesterhelt D, Pfeiffer M, Engel A, Gaub HE, Müller DJ (2000) Science 288:143 Janovjak H, Kessler M, Oesterhelt D, Gaub H, Müller DJ (2003) EMBO J 22:5220 Kedrov A, Ziegler C, Janovjak H, Kühlbrandt W, Müller DJ (2004) J Mol Biol 340:1143 Sapra KT, Besir H, Oesterhelt D, Müller DJ (2006) J Mol Biol 355:640 Scheuring S, Seguin J, Marco S, Lévy D, Robert B, Rigaud JL (2003) Proc Natl Acad Sci USA 100:1690
17 AFM of Microbial Surface
125
28. Scheuring S, Rigaud JL, Sturgis JN (2004) EMBO J 23:4127 29. Scheuring S, Sturgis JN (2005) Science 309:484 30. Bahatyrova S, Frese RN, Siebert CA, Olsen JD, van der Werf KO, van Grondelle R, Niederman RA, Bullough PA, Otto C, Hunter CN (2004) Nature 430:1058 31. Scheuring S, Busselez J, Lévy D (2005) J Biol Chem 280:1426 32. Scheuring S, Gonçalves RP, Prima V, Sturgis JN (2006) J Mol Biol 358:83 33. Ahimou F, Touhami A, Dufrêne YF (2003) Yeast 20:25 34. van der Aa BC, Michel RM, Asther M, Zamora MT, Rouxhet PG, Dufrêne YF (2001) Langmuir 17:3116 35. Ma H, Snook LA, Kaminskyj SGW, Dahms TES (2005) Microbiol SGM 151:3679 36. Crawford SA, Higgins MJ, Mulvaney P, Wetherbee R (2001) J Phycol 37:543 37. Schaer-Zammaretti P, Ubbink J (2003) Ultramicroscopy 97:199 38. Schaer-Zammaretti P, Ubbink J (2003) Biophys J 85:4076 39. Chada VGR, Sanstad EA, Wang R, Driks A (2003) J Bacteriol 185:6255 40. Plomp M, Leighton TJ, Wheeler KE, Malkin AJ (2005) Biophys J 88:603 41. Doktycz MJ, Sullivan CJ, Hoyt PR, Pelletier DA, Wu S, Allison DP (2003) Ultramicroscopy 97:209 42. Touhami A, Jericho M, Beveridge TJ (2004) J Bacteriol 186:3286 43. Verbelen C, Dupres V, Menozzi FD, Raze D, Baulard AR, Hols P, Dufrêne YF (2006) FEMS Microbiol Lett 264:192 44. Malkin AJ, McPherson A, Gershon PD (2003) J Virol 77:6332 45. Hinterdorfer P, Dufrêne YF (2006) Nat Methods 3:347 46. Dufrêne YF (2000) Biophys J 78:3286 47. Ahimou F, Denis FA, Touhami A, Dufrêne YF (2002) Langmuir 18:9937 48. Lee GU, Kidwell DA, Colton RJ (1994) Langmuir 10:354 49. Florin EL, Moy VT, Gaub HE (1994) Science 264:415 50. Fritz J, Katopodis AG, Kolbinger F, Anselmetti D (1998) Proc Natl Acad Sci USA 95:12283 51. Dammer U, Popescu O, Wagner P, Anselmetti D, Güntherodt HJ, Misevic GN (1995) Science 267:1173 52. Berquand A, Xia N, Castner DG, Clare BH, Abbott NL, Dupres V, Adriaensen Y, Dufrêne YF (2005) Langmuir 21:5517 53. Hinterdorfer P, Baumgartner W, Gruber HJ, Schilcher K, Schindler H (1996) Proc Natl Acad Sci USA 93:3477 54. Kienberger F, Kada G, Gruber HJ, Pastushenko VPh, Riener C, Trieb M, Knaus HG, Schindler H, Hinterdorfer P (2000) Single Mol 1:59 55. Dupres V, Menozzi FD, Locht C, Clare BH, Abbott NL, Cuenot S, Bompard C, Raze D, Dufrêne YF (2005) Nat Methods 2:515 56. Ong YL, Razatos A, Georgiou G, Sharma MM (1999) Langmuir 15:2719 57. Bowen WR, Hilal N, Lovitt RW, Wright CJ (1998) Colloid Surf A 136:231 58. Lower SK, Hochella MF, Beveridge TJ (2001) Science 292:1360 59. Benoit M, Gabriel D, Gerisch G, Gaub HE (2000) Nat Cell Biol 2:313 60. Xu W, Mulhern PJ, Blackford BL, Jericho MH, Firtel M, Beveridge TJ (1996) J Bacteriol 178:3106 61. Yao X, Walter J, Burke S, Stewart S, Jericho MH, Pink D, Hunter R, Beveridge TJ (2002) Colloids Surf B 23:213 62. Touhami A, Nysten B, Dufrêne YF (2003) Langmuir 19:4539 63. Gaboriaud F, Bailet S, Dague E, Jorand F (2005) J Bacteriol 187:3864 64. Pelling AE, Sehati S, Gralla EB, Valentine JS, Gimzewski JK (2004) Science 305:1147 65. Rief M, Oesterhelt F, Heymann B, Gaub HE (1997) Science 275:1295 66. Scheuring S, Stahlberg H, Chami M, Houssin C, Rigaud J-L, Engel A (2002) Mol Microbiol 44:675
126
C. Verbelen et al.
67. Verbelen C, Antikainen J, Korhonen TK, Dufrêne YF (2007) Ultramicroscopy 107:1004 68. Gad M, Itoh A, Ikai A (1997) Cell Biol Int 21:697 69. Ando T, Kodera N, Takai E, Maruyama D, Saito K, Toda A (2001) Proc Natl Acad Sci USA 98:12468 70. Humphris ADL, Hobbs JK, Miles MJ (2003) Appl Phys Lett 83:6 71. Humphris ADL, Miles MJ (2002) Methods Cell Biol 68:337
18 Cellular Physiology of Epithelium and Endothelium Christoph Riethmüller · Hans Oberleithner
Abstract. Epithelial and endothelial cell layers have many essential functions for keeping higher organisms alive. One main task which they both share is the regulation of water balance in the body. Regulation of water permeability can be performed by altering the conductivity between the cells (paracellularly) or through the cells (transcellularly). Here, four examples are shown of how atomic force microscopy (AFM) can indirectly assess the effects of these physiological functions on primary cell preparations under buffer conditions. (1) assessing the tightness of paracellular junctions via their height profile, (2) softening of the cytoskeleton in response to thirst hormone, (3) probing intracellular vacuoles and (4) exploring the invagination formed by a transmigrating immune cell after nanomanipulation. It can be concluded that AFM can help us to examine mechanisms regulating the permeability of cell layers. Key words: Physiology, Epithelium, Endothelium, Cell Junctions, Permeability, Cell Elasticity
18.1 Introduction Physiology is a discipline of biology that investigates phenomena of organisms with respect to their function in keeping the individual alive. The spectrum of organisms covers plants, animals and humans. Not only whole organisms, but also parts of them, like isolated organs, cell ensembles, single cells or even organelles, are used to study their respective role in enabling a normal life. An important aim of physiological investigation of isolated organs or at the cellular level is to discover regulatory cycles that allow the organism adapt to the ever-changing environmental conditions. Classic examples of physiological models are isolated beating hearts kept in buffer solution to study the effects of flow resistance or prepared single nephrons to investigate fluid absorption. The search for mechanisms can make it necessary to dissect a regulatory cycle down to the molecular level, i. e., single metabolic steps. Especially in human physiology, out-of-tune states can be of considerable interest for understanding the development of diseases, the knowledge of which is comprised in the subdiscipline pathophysiology. The latter is intimately intertwined with pharmacology, because the impact of chemical compounds on cellular and organ function forms the basis for exploring their therapeutic potential. Major categories of physiological interest are the uptake and transport of nutrients, circulation, secretion, hormonal regulation, muscle contraction and bone mechanics as well as biological data acquisition (eye, ear, skin) or information processing through neurons. Since a physiological measurement focuses on entities like pressure, flow, concentration gradients and electric potential or current, it
128
C. Riethmüller · H. Oberleithner
mainly relies on biophysical and biochemical methods. Reversely, biological questions sometimes drive the technical improvement of measuring devices, for instance, the patch-clamp technique promoted the design of ultra-low-noise amplifiers [65,89]. For the atomic force microscope (AFM) as a rather young “tactile” instrument, new aspects have been gained in the past decade when it was applied to old questions and more are expected in the years to come. The fact that the AFM can not only be run in a high vacuum like its predecessor the scanning tunneling microscope but it can also be run under atmospheric pressure made it feasible for many more applications. And after it had surprisingly been proven to work even under fluid buffer conditions, the biosciences started to recognize its huge potential. Since then, the AFM has experienced a continuously growing acceptance in the life sciences [15, 17, 28, 32, 38, 44, 52, 63, 67, 75, 99, 111]. New insights have been obtained into lipid biochemistry [4, 36, 49, 61], structural biology [24, 31, 43, 51, 81, 91], cell biology [14, 19, 41, 44, 67], and pharmacology [14, 22, 79, 88] to name just a few fields. In cellular physiology of higher animals, two closely related cell types, namely epithelium and endothelium, have diverse functions of outstanding importance. They form cellular monolayers separating different body compartments and thus build pseudo-two-dimensional organs with a maximal surface. As such they are ideally suited specimens for investigation by the surface probing tool the AFM. The results promise to be of physiological relevance, because the surface maximization follows their function [34,35,52,69]. Here, we focus on these two types of cells in mammals. This chapter will spotlight that specific area of AFM application of the past decade and provide some examples of ongoing experiments with a clear perspective for physiological implications in the future.
18.2 Epithelium 18.2.1 Transport Through a Septum The mammalian body is separated from its environment by at least one epithelial cell layer. This layer primarily forms a barrier between inside and outside to control any kind of transport in both directions. A regulation is very important, because literally all of the substances like water, oxygen, salt and nutrients or waste products going either into or out of the body have to cross epithelial layers. The asymmetry between up and down faces of the cells is termed “polarization”. Basically, the outside (with respect to the organism) is called “apical” and the inside is called “basolateral.” For in vitro experiments, isolated cells are usually cultivated in a petri dish coated with basal membrane proteins for better growth. For this reason, the cells interpret the interface to the culture dish as “inside” the organism (basolateral) face. Hence, only the apical surface of polarized cells is accessible to the AFM tip, unless a special membrane preparation is used [3, 23, 46]. A well-differentiated layer develops cell-to-cell borders consisting of adherent transmembrane proteins interacting with their counterparts in the neighboring cell to form a zipper-like structure. Fibrillary actin is found in a higher density in this
18 Cellular Physiology of Epithelium and Endothelium
129
Fig. 18.1. Epithelial cells. The environment of an epithelial cell is asymmetric: the apical side is oriented to the outside of the organism, the basolateral side to the inside. The tight junctions accomplish and regulate the barrier function of an intact epithelium. In the basolateral plasma membrane, the Na/K exchanger, a transmembrane protein, builds up concentration gradients and membrane potential. Microvilli and cilia on the apical membrane increase the interaction area available for the transport of substances
belt, offering mechanical support from inside the cell as well as a better coupling of barrier properties to an intracellular regulation. The junctions are tight enough to build up an electrical resistance in the range of kiloohm square centimeters. Accordingly, the conductivity for solutes between the cells (paracellularly) is reduced down to transport rates comparable to the minute amounts passing through the cells (transcellularly). The belt of tight-junction proteins effectively hinders the diffusion of membrane proteins within the lipid bilayer from the basolateral to the apical side or in the reverse direction. As a consequence, transport proteins, ion channels and receptors are differently partitioned between apical and basolateral faces, reflecting the asymmetric nature of epithelial function [21,104,108]. For the physiologist, only a well-polarized epithelium is a good one, because it can be expected to represent the in vivo situation. In functionally different regions, the epithelial layers are specialized according to their respective tasks. The skin barrier protects against entrance of particles and bacteria as well as against the loss of water and solutes. Alveolar epithelium of the lung enables an effective exchange of gaseous oxygen and carbon dioxide. Epithelium of the guts absorbs sugar and peptides, in the stomach it secretes protons, and the epithelia of secretory glands produce saliva, sweat or tears. The process of epithelial secretion including the formation of pores has been described in pancreatic acinar cells [44, 92]. Another secretion process of high physiological importance is that of the proteoglycan “surfactant” of alveolar type II cells in the lung, which reduces surface tension at the gas–fluid interface. Although the protein has been well studied [56, 100] and alveolar cells can be subjected to AFM imaging [1], the
130
C. Riethmüller · H. Oberleithner
process of secretion could not successfully be addressed yet, probably owing to its tenside-like properties of the surfactant. In the following, we will focus on the kidney epithelium as an example. It separates urine from interstitial space and serves as a good model for different types of transport. Typical AFM applications can be transferred to other epithelia. Technical details of methods beyond normal contact imaging are given at the end of the chapter. 18.2.2 In the Kidney The main task of the kidney is to control the water and salt balance of the organism. It consists of a complex system with around one million capillaries called nephrons. A nephron is segmented into four functional domains: (1) in the glomerulus, blood is filtrated to give the primary urine; (2) then, the vast majority of nutrients and water are reabsorbed by the proximal tubulus epithelium; (3) salt is actively pumped out of the loop of Henle into the interstitial space, setting up an osmotic gradient; (4) the fine-tuning of water reabsorption is performed via regulating the water permeability of the cell membranes. The first reports on kidney cells viewed with the AFM date back to the early 1990s [40, 50, 68]. The studies used immmortalized epithelial cells from canine kidney collecting duct (MDCK), which has established as the standard cell line for AFM studies on epithelium. Since then, numerous reports have extended the spectrum of cells to various types of freshly isolated specimens, so-called primary culture, which can be regarded as better representing the in vivo characteristics [20,75,77,109]. Following the functional segmentation of nephrons, glomerular epithelial cells have very recently been imaged [103]. They are termed “podocytes” owing to their feet(toe)like interdigitations that have a role in the filtration process. However, efforts to get such a morphologically differentiated cell culture exhibiting this special feature have not been successful yet and are still ongoing. Cells from the proximal tubule have been immortalized to give the LLCPK-1 line, which has been demonstrated to offer the functional characteristics and is convenient to work with [82]. While cells from the loop of Henle have not been addressed by AFM studies, cells of the collecting duct, especially the MDCK line, have been serving as a model for a vast number physiological and biophysical studies since the mid-1990s [68, 95, 96]. 18.2.2.1 Protrusions The luminal surface of epithelia is covered by (micro-)villi, which are small protrusions of the plasma membrane (in the micron range). Many membrane proteins, like transporters, receptors and ion channels, are preferentially localized in the villi. The AFM is a good tool to measure the height of microvilli accurately. This is of significance, since villi are the physiological means to increase the interfacial area between cells and the environment. This area can be a limiting factor, as seen from
18 Cellular Physiology of Epithelium and Endothelium
131
highly active transporting epithelia like in the ileum of the intestine or in proximal tubules of the kidney, which exhibit a higher density and length of villi. In other words, the transport capacity correlates to the density and height of villi. The number and correct size of these microstructures is obtained rather accurately and easily with atomic force microscopy as compared to transmission electron microscopy. In Fig. 18.2, typical examples from freshly isolated nasal and kidney cells are given. The data from atomic force microscopy are in good accordance with those from scanning ion conductance microscopy [29], a method that uses the resistance of a patch pipette as a feedback signal instead of the deflection of the cantilever in the AFM [33, 47]. Combination of atomic force microscopy with fluorescence
Fig. 18.2. Fixed epithelial cells. Epithelial cells freshly isolated from a and b human nasal biopsies or c and d rat inner medulla collecting duct (IMCD) were fixed with glutardialdehyde and scanned in physiological buffer in contact mode. a At 70-µm2 scan area, the cell borders can be seen. b A zoom into 10 µm resolves the microvilli protruding into the lumen by around 0.5 µm. c An area of (100 µm)2 covers around six cell bodies, four of which exhibit a ciliary structure in their center. d Zooming to 25 µm resolves the microvillar pattern and shows the seam of a cell border (in the upper third of the image)
132
C. Riethmüller · H. Oberleithner
microscopy demonstrated that these villi contain a high density of fibrillary actin, probably in order to keep them in an upright position to avoid a loss of interaction surface [80]. Occasionally, the villi do not only protrude, but they also seem to build long tubes lying flat on the surface [30]. Moreover, well-differentiated epithelial cells show a central cilium on their apical side. This prominent structural feature carries many transmembrane proteins and signaling molecules [101]. Since bending of this cilium activates intracellular messenger cascades, it has been interpreted as a sensor antenna, which reports via its bending state on the velocity of urine flow in the nephron. Its loss is associated with a severe illness, polycystic kidney disease [53, 76]. This prominent structure is known primarily from ex vivo preparations for electron microscopy, but could recently be imaged by atomic force microscopy in a cell-culture model of inner medulla epithelial cells from rat collecting duct [9, 57] (Fig. 18.3). Another functionally important element is the cell–cell junctions. Epithelial junctions vary in terms of their tightness. Apart from measuring transepithelial electrical resistance or solute permeabilities, it might be useful to align morphological data to these functional characteristics. In contrast to the sketches in textbooks on cell biology, AFM contact imaging clearly demonstrates the cell-to-cell border to
Fig. 18.3. Living epithelial cells. Cells isolated from rat collecting duct were imaged in physiological buffer at 37 ◦ C. a In a 50-µm-height image, central cilia show up in the middle of each cell body. b The corresponding deflection image better depicts the hexagonal “lattice” of this well differentiated cell layer. c From a section along the white line in a the height of the central cilia can be taken (1 µm). d The application of an image-filtering routine can resolve the microvillar pattern even in living cells
18 Cellular Physiology of Epithelium and Endothelium
133
Fig. 18.4. Stiffness map of living epithelial cells. a In the so-called force volume mode, the tip is run like the needle in a sewing machine, thus recording an array of force–distance curves b. From the raw data, maps of c true height and d stiffness can be obtained. The 10-µm maps on a border area of four IMCD cells show analogously elevated values (represented by light gray) for height and stiffness along the tight junctions. e ,f From sections along the horizontal lines in c and d, the values can be deduced to resemble e 150-nm elevation and f a factor of 3 in stiffness (log scale)
134
C. Riethmüller · H. Oberleithner
be higher than the normal cell level by more than 100 nm (Fig. 18.4). This elevation reflects the belt of actin fibers surrounding the cell which constitutes the barrier properties. Correspondingly, the elasticity map yields 3 times higher mechanical rigidity (Young’s modulus) on these junctional areas. These values speak in favor of primary cells to be taken for physiological investigations, because the MDCK cell line does not form comparably prominent junctions [96]. Measuring the height elevation and the stiffness increase might indicate the degree of cell differentiation and thus could replace an electrical or functional assessment of barrier properties in cases when cells have to grow on optically impermeable substrates, e. g., biomaterials for implants. 18.2.2.2 Volume Accurate regulation of its volume is very important for a cell to survive [13,25,30,37, 45, 66, 71, 93, 102, 106]. This is difficult to fulfill by actively transporting epithelia, because only a slight imbalance between apical uptake and basolateral loss rates leads to massive volume change and cell damage. Taken out of the organism context, cell volume changes in vitro hint at alterations of transport rates or permeabilities. Measuring the cellular volume can be performed via exchange of radioactively labeled water, for instance. Apart from the disadvantage of radioactive hazard, this method allows only for end-point determinations, without any possibility to monitor volume changes in living cells. Advantageously, the AFM has been made use of in a study on volume dynamics in migrating epithelial cells [72, 94]. It turned out that volume dynamics are aligned to the direction of migration in a population of MDCK-F cells. At the lamellipodium, where the cell front searches for its way, a volume increase can be detected, whereas the rear end shrinks. Accordingly, the stiffness exhibits a gradient too [48]. The study of cell migration may have some relevance for cancer research, since it is regarded to take part in tumor metastasis. An important player in terms of cell and blood volume regulation is the steroid hormone aldosterone. Being secreted by the suprarenal glands, it increases blood volume and pressure. The mechanism involves cell swelling via activation of the epithelial sodium channel, which is highly expressed in epithelia [74]. Narrowing of the nephron lumen with concomitant increase of flow resistance is among the responsible effects discussed, because this would lead to subsequent counterregulation via angiotensin [54, 55, 85]. This mechanism would complete the “vicious cycle” of raising blood pressure. Endothelial aspects of this hormone action are discussed later. 18.2.2.3 Invaginations Tightly connected with the volume regulation is a cell’s ability to take up or secrete fluid. Secretion is the main physiological function in glandular epithelia that produce saliva, sweat or pancreatic enzymes. Secretion of enzymes is usually performed via an exocytotic process. Vesicles store the protein product of the cell and upon appropriate signals fuse with the plasma membrane. The secretion process has been addressed by various AFM studies on pancreatic cells [44, 92].
18 Cellular Physiology of Epithelium and Endothelium
135
A process closely related to secretion is the translocation of membrane proteins which are stored in the vesicle membrane and are inserted into the plasma membrane upon the respective stimulus. The process itself is more or less the same, but the focus is on the membrane protein instead of the vesicle containment. A well-described model of membrane insertion is the translocation of aquaporin-2 to the apical side of collecting duct cells. When cells which have been stimulated with arginine vaso-
Fig. 18.5. Stiffness map of fixed epithelial cells. IMCD cells were stimulated with arginine vasopressin (AVP), fixed with glutardialdehyde and imaged in contact mode (A). B From a section along the white line in A, shallow circular depressions of the plasma membrane can be found to measure around 200 nm in diameter. C, E A 10-µm force volume map was recorded for this sample, showing deviants from mean values in C height and E stiffness mainly at the microvilli. D, F Performing a 2-µm force volume exhibits D shallow (5–10-nm) depressions between microvilli, which F do not correspond to soft spots
136
C. Riethmüller · H. Oberleithner
pressin (AVP) are fixed with glutardialdehyde, some of the vesicles should be caught in the process of fusion. Dimples of the expected diameter (200 nm) could be found in such a preparation, which might represent such cases (Fig. 18.5). From a geometric point of view, any intense interaction of a vesicle with the plasma membrane, be fusion or fission, must produce alterations of the actual cell height during the process. The AFM is ideally suited to monitor height changes with nanometer accuracy on living cells. Such height changes have been reported for bacteria as well as for single proteins [39, 70]. While yeast cells seem to prefer a certain frequency in the kilohertz region [78], this has not been reported for other cells. We have tested the inner medulla collecting duct (IMCD) cell model, which was proven to translocate vesicles upon a stimulation with AVP. Indeed, the height changes determined on a single spot as a function of time gave less than 1 nm mean square z-displacement (Rms ) for the control and 1.3 nm after addition of AVP. It remains to be clarified how much actomyosin interaction contributes to the effect. However, no possible result would contradict an interpretation of the z-vibrations as a sign of membrane traffic.
18.3 Endothelium The endothelium is a specialized sort of epithelium, which is inwardly bound and does not have any contact with the environment. It covers the inside of blood and lymph vessels and shares the main characteristics with the epithelium. Apart from strictly forming not more than one layer, it has the ability to communicate in a highly sensitive manner. One extreme example in a long row may be given by the diatomic radical of nitric oxide, which mediates blood vessel dilation via diffusion in the nanomolar concentration range. Endothelial cells organize the internal workflow of the organism on a humoral (non-neuronal) basis. They are central in controlling any kind of transport from blood
Fig. 18.6. Height fluctuations of living epithelial cells. The height of living IMCD cells at one spot was monitored in physiological buffer at 37 ◦ C A before and B after stimulation with AVP. AVP elevates intracellular cyclic AMP and leads to membrane translocation of aquaporin-containing vesicles. The height fluctuations correlate to the intracellular activities
18 Cellular Physiology of Epithelium and Endothelium
137
Fig. 18.7. Endothelial cells. Analogous to epithelial cells, endothelial cells are asymmetrically constructed. The apical side is oriented to the bloodstream and the basolateral side to the interstitial space (tissue). Most of the fluid that crosses the cell layer passes through the paracellular clefts between the cells, but in special cases also transendothelial channels (TEC) form a shortcut between the compartments
to intercellular space and vice versa. They regulate the bloodstream and the healing of wounds and they modulate the immune response. Endothelial dysfunction leads to disorders as severe as bleeding, dementia, hypertension and arteriosclerosis—most of which develop slowly, but tend to be irreversible and lethal. Therefore, improved knowledge of endothelial cell biology deserves the interest of the physician. The first studies on endothelium date from a little later than those on on epithelium: images of fixed cells were reported in 1995, when a comparison with electron microscopy was performed [10, 11]. (Atomic force microscopy has been applied to endothelial cells of vascular origin only; lymphatic endothelial cells are virtually nonexistent in cell-culture laboratories.) Not only cell morphology has been addressed, but also cytoskeletal mechanics [58, 62, 90]. Moreover, atomic force microscopy has demonstrated a shape-dependence of cells from the biological coating of the culture support. This effect is caused by different degrees of adhesion forces between the cell and the support [8]. 18.3.1 Paracellular Gaps One of the permanent tasks of endothelia is to let pass a suitable amount of liquid and solutes from the blood plasma to the diverse tissues and backwards. The endothelium in different regions is morphologically adapted to the requirements of the respective tissues. Its shape ranges from a discontinuous appearance in liver sinusoids, which is highly permeable, to a tight layer in the brain with characteristics close to those of epithelium. Rather than transporting small inorganic ions via membrane transporters through apical and basolateral membranes like epithelia do, the typical passage across endothelial layers is a rather the unspecific flow of fluid between the cells, i. e., paracellularly. An imbalance in regulating fluid balance may lead to tissue swelling (edema), as happens in the course of an inflammation. Transport of
138
C. Riethmüller · H. Oberleithner
fluid is experimentally assessed by physiologists using cells grown on permeable filter membranes. When the layer reaches confluence, a cargo is added to the apical side and its appearance on the other side is monitored. Not that atomic force microscopy could replace this measurement, but since the majority of the passage takes a paracellular route, it can inspect cell-to-cell contacts. Although a direct access to
Fig. 18.8. Paracellular gap formation. Endothelial cells of human umbilical vein are isolated, fixed with glutardialdehyde and imaged in contact mode (70 µm). A, C Deflection and height images of a control sample. B, D Deflection and height images of a sample having been treated with ionomycin (1 µM) 5 min before fixation. Ionomycin elevates intracellular calcium concentration and induces cell contraction. This leads to the formation of paracellular gaps (red area in D), that perforate the layer. From the sections below C and D, it can be seen that the bottom of the culture dish is exposed in D the stimulated sample (straight horizontal part of the line between 40 and 50 µm), as opposed to C the control
18 Cellular Physiology of Epithelium and Endothelium
139
the intercellular cleft (normally only few nanometers wide) is not possible owing to steric hindrance, the formation of paracellular gaps can be observed by the AFM. An interplay of two counteracting factors balances the opening of junctions. On the one hand, transmembrane proteins interact with their counterparts in the neighboring cells to define a certain adherent force. On the other hand, intracellular contractile force is generated from actomyosin interaction. Increasing tension as a consequence of elevated intracellular calcium concentrations leads to a local detachment of cells at sites of weakest adherence. Eventually, these “pores”—when big enough—allow the AFM tip to access the culture dish (Fig. 18.8). This state of cell activation reflects pathophysiological conditions in vivo. The paracellular gaps demonstrate a clear contribution of intracellular tension in the induction of endothelial hyperpermeability on a time scale of minutes, which is definitely too short for downregulation of transmembrane proteins like cadherins [5, 98, 105]. These proteins have been shown to regulate the barrier properties of endothelial cells over long range. It is of considerable interest to the biologist that filter membranes can be subjected to AFM imaging as seen from Fig. 18.8. There is not a big difference in appearance compared with cells grown on glass, but the development of internal signaling pathways can be expected to better represent the in vivo situation. The various degree of barrier function can be inspected from cells of different origin. The most intimate cell—cell contacts are built up by endothelial cells in brain capillaries (blood–brain barrier). A morphological correlate to this function is called “marginal folds”, which are cell lobes that overlap neighboring cells and are easily measured by the AFM technique as heightened lines [97, 107]. 18.3.2 Cellular Drinking In special cases of proinflammatory activation, a contribution of transcellular transport could be indirectly shown [86]. The study compared the diffusion of a fluid marker and the paracellular electrical resistance. Upon the proinflammatory stimulus bradykinin, which is known to cause edema in vivo, both detection channels should alter in parallel, but they surprisingly pointed in different directions, i. e., closing of the paracellular gaps coincided with increased dextran diffusion. This paradox could only be solved by assuming transcellular transport of fluid [86]. Transcellular transport of fluid was hypothesized earlier, not at least owing to diverging results between in vitro and in vivo experiments. Transendothelial channels have been found which resemble a tunnel made out of plasma membrane spanning from the apical side directly to the basolateral side [59,60]. These have been reported to increase in number as determined by transmission electron microscopy, when intracellular calcium concentration is raised [64]. However, the mouth of such a channel has not been observed by atomic force microscopy yet, although observation should be possible from a geometric point of view. One reason might be that it is inaccessible to the AFM tip owing to coverage by the glycocalix. The term describes a layer of long, branched and hydrophilic proteoglycans anchored at membrane proteins. Proteoglycans can be imagined as a kind of microscopic seaweed that dampens turbulence in plasma flow and mediates the adhesion of immune cells. Their role in AFM image convolution has not been clarified yet.
140
C. Riethmüller · H. Oberleithner
Fig. 18.9. Comparison of living and fixed endothelial cells. Human endothelial cells (HUVEC) were imaged in contact mode either A before or B after fixation with glutardialdehyde. A In living cells, the cytoskeletal fibers are predominant. B After fixation, they almost disappear, generating dimples as a new structural feature. C Three-dimensional view with three of the characteristic depressions marked by arrows
An unspecific uptake of fluid by cells is termed pinocytosis (“cellular drinking”). From electron-microscopic studies it is known that their size ranges from 200 to more than 1000 nm. The lack of specificity does not only concern the cargo, but also reflects the lack of specific markers for these organelles. Fluorescence microscopy can easily identify those, that were internalized while being exposed to a solution containing a fluorescently labeled and biologically inert compound like fluorescein isothiocyanate–dextran [42]. But the fraction of pinocytotic organelles that were formed before exposition to fluid marker remain dark. Atomic force microscopy can add a completely different approach to this question. The technical “trick” is that the cross-linking of intracellular proteins elevates the cellular stiffness by roughly an order of magnitude [88]. This leads to an embedding of the actin fibers that have dominated the picture of living cells. These prominent structures are mechanically buried in the polymeric biocement made from glutardialdehyde and protein-bound amino groups. But organelles devoid of (cytosolic) proteins like pinosomes or vacuoles remain as soft as they were before; hence, in an elasticity map they appear as soft spots like bubbles of air in a pudding [87].
18 Cellular Physiology of Epithelium and Endothelium
141
Fig. 18.10. Stiffness map of fixed endothelial cells. Isolated human endothelial cells were fixed with glutardialdehyde and subjected to force volume recording. By calculation from the raw data, maps of A indented height, B true height and C Young’s modulus are obtained. The indented height map is reconstructed from all points at a force load of 0.5 nN and the true height is at zero force, as illustrated by the sketches in D and E, respectively. The stiffness map (C) shows a colocalization of soft spots with the depressions of A. Consistently, they disappear (B) concomitant to vanishing force load
Fig. 18.11. von Willebrandt fibers. Purified glycoprotein of von Willebrandt factor are deposited on mica, dried and imaged in contact mode under air. A Long fibers can be found in samples of this adherent protein. B A zoom reveals a treelike structure
Mapping the elasticity overcomes the limitation of the AFM to the surface of a sample and enables intracellular organelles to be detected via their lack of mechanical support. It compares to the palpation of intraabdominal organs by a physician,
142
C. Riethmüller · H. Oberleithner
which is the classic approach to first diagnosis without cutting. The vacuolar organelles have been implicated in angiogenesis (forming of new blood vessels), a critical step in tumor nutrition and growth [18, 26]. Typical angiogenesis stimulating factors concomitantly increase the vascular permeability and the formation of intracellular vacuoles [6, 7, 16]. Up to now, it has not been investigated,whether these organelles have a role in the volume regulation of endothelial cells in response to the mineralocorticoid hormone aldosterone, which was mentioned in Sect. 18.2.2.2 [67, 72, 73]. Therefore, a more detailed study of vacuolization may offer some therapeutic aspects in the future. 18.3.3 Wound Healing In the context of wound healing, the von Willebrand factor (vWF) has a central role. This fibrous glycoprotein is stored in endothelial cells and is secreted upon cell damage. It builds long sticky fibers which platelets adhere to, in order to kind of plumb the lesion site of a blood vessel. Technical difficulties in trying to image the purified protein by atomic force microscopy led to the notion of vWF being a molecular spring that can be unrolled through the action of fluid shear stress [2]. The activation of platelets as the next step downstream of vWF-network formation has also been imaged by atomic force microscopy [27, 84].Owing to the small size of around 1 µm, classic assays for determining the aggregation state have to rely on
Fig. 18.12. TRECTM imaging of living endothelial cells. TRECTM mode has been developed by P. Hinterdorfer, University of Linz, Austria. A cantilever is functionalized with an antibody (here: anti-β1 integrin) and with use of a modified AFM controller, the simultaneous recording of A topography and B molecular recognition is possible. Integrins are membrane proteins that are important for cell adhesion. The recognition events are black spots marked by arrows. From C and D, it can be seen that there are C no depressions of the surface height that correlate to D the recognition events
18 Cellular Physiology of Epithelium and Endothelium
143
Fig. 18.13. Immune cells on endothelium. Endothelial cells were freshly isolated from porcine brain and treated with a proinflammatory stimulus. Isolated polymorphonuclear neutrophiles (PMN), a sort of leukocyte, are added and the sample is fixed. a Height and b deflection images of one PMN (arrow) in close vicinity of an endothelial cell nucleus. The zoom (c, d) reveals the interendothelial cell border (arrow) lying underneath the PMN. (The image was gratefully obtained from H.-J. Galla, University of Münster, Germany; S. Schrot, K. Psathaki and H.-J. Galla, unpublished results)
macroscopic turbidity measurement via light scattering methods which need a high number of platelets; to get an image of single platelets, electron microscopy was necessary. Nowadays, atomic force microscopy has taken advantage of working in buffer solution to investigate a single platelet’s activation state via its morphology [84]. 18.3.4 Transmigration of Leukocytes The concept of transcellular transport in a proinflammatory context has very recently been shown to apply even to intact cells of the immune system. For some decades the notion of paracellular pathways exclusively dominated the notion of leukocyte transmigration. The term describes the puzzling fact that blood-born white immune
144
C. Riethmüller · H. Oberleithner
Fig. 18.14. Footprint of a leukocyte in endothelium. Endothelial cell line bEND5 was added to freshly isolated PMNs and fixed (samples were gratefully obtained from D. Vestweber, Max Planck Institute, Münster, Germany). A Contact image. B Nanomanipulation of the sample by application of high force. C Image of the same region as in A after nanomanipulation; a transmigratory cup remains imprinted in the endothelial cell. D It has a diameter of around 10 µm and a depth of 800 nm
cells of 10 µm in size get across the endothelial barrier into the tissues, while the plasma fluid is mainly kept in the blood vessel. This paradox has been solved by the (experimentally proven) assumption that leukocytes squeeze through the zipper of interendothelial clefts. But recently, it was shown by video microscopy that leukocytes can even invade and cross endothelial cells without even touching a cell border [110]. In the course of this transmigration, the endothelial cell forms a complementary entry channel, named a “transmigratory cup” [12]. Atomic force microscopy is not only able to image and characterize leukocytes adhering to endothelium in both fixed and unfixed states, but also can be used as a nanomanipulator to scratch off the adhering leukocyte, uncovering the transmigratory cup lying underneath (Fig. 18.12). Atomic force microscopy uniquely offers the possibility to image and manipulate cells. The high-resolution capabilities render it a valuable tool for endothelial cell biology and physiology.
18.4 Technical Remarks The data shown here were acquired under physiological buffer conditions (except those in Fig. 18.11) using scanning microscopes built by Digital Instruments
18 Cellular Physiology of Epithelium and Endothelium
145
(Santa Barbara, USA). Mostly, standard contact imaging in the low-nanometer range was applied using DNP or MLCT cantilever substrates (purchased through Veeco, Mannheim, Germany). The force–volume recordings were performed in 64 × 64 matrices with the softest cantilevers available (0.01 N/m) at very low indentation (100–200 nm) and piezo travel (less than 3 µm/s). The data were processed by a software routine for IGOR ProTM based on the Hertz model for elasticity [83].
18.5 Summary Atomic force microscopy throws light from a different angle onto the fascinating scenery of cells whose physiological architecture is nearly two-dimensional. Regulatory cycles in physiological contexts can uniquely be addressed by this technique on living specimens with unprecedented resolution. Nanometer movements of cells can be monitored to get insight into their intracellular machinery. The future will see the AFM providing the basis of physiological assays for pharmacological testing on single cells. Acknowledgements. We would like to thank Rainer Matzke (Munich), Philippe Carl (Mannheim) and Tilman Schaeffer (Münster) for advice with the force curve analysis. Primary cultures were gratefully obtained from Kristina Kusche (human nasal airway), Bayram Edemir (rat collecting duct) and Joachim Wegener (porcine brain endothelial cells) (all University of Münster).
References 1. Alcaraz J, Buscemi L, Grabulosa M, Trepat X, Fabry B, Farre R, Navajas D (2003) Biophys J 84:2071 2. Alexander-Katz A, Schneider MF, Schneider SW, Wixforth A, Netz RR (2006) Phys Rev Lett 97:138101 3. Allen S, Rigby-Singleton SM, Harris H, Davies MC, O’Shea P (2003) Biochem Soc Trans 31:1052 4. Almofti MR, Harashima H, Shinohara Y, Almofti A, Baba Y, Kiwada H (2003) Arch Biochem Biophys 410:246 5. Baumgartner W, Hinterdorfer P, Ness W, Raab A, Vestweber D, Schindler H, Drenckhahn D (2000) Proc Natl Acad Sci USA 97:4005 6. Bayless KJ, Davis GE (2002) J Cell Sci 115:1123 7. Berlin RD, Oliver JM (1980) J Cell Biol 85:660 8. Bhadriraju K, Hansen LK (2002) Exp Cell Res 278:92 9. Birnbaumer M, Seibold A, Gilbert S, Ishido M, Barberis C, Antaramian A, Brabet P, Rosenthal W (1992) Nature 357:333 10. Braet F, de Zanger R, Kalle W, Raap A, Tanke H, Wisse E (1996) Scanning Microsc Suppl 10:225 11. Braet F, de Zanger R, Seynaeve C, Baekeland M, Wisse E (2001) J Electron Microsc (Tokyo) 50:283 12. Carman CV, Springer TA (2004) J Cell Biol 167:377 13. Charles CJ, Nicholls MG, Rademaker MT, Richards AM (2002) J Endocrinol 173:123 14. Costa KD (2003) Dis Markers 19:139
146
C. Riethmüller · H. Oberleithner
15. 16. 17. 18. 19. 20. 21. 22. 23.
Costa KD (2006) Methods Mol Biol 319:331 Davis GE, Bayless KJ (2003) Microcirculation 10:27 Dufrene YF (2003) Curr Opin Microbiol 6:317 Dvorak AM, Feng D (2001) J Histochem Cytochem 49:419 Dvorak JA (2003) Methods 29:86 Dvorak JA, Nagao E (1998) Exp Cell Res 242:69 Ebnet K, Suzuki A, Ohno S, Vestweber D (2004) J Cell Sci 117:19 Edwardson JM, Henderson RM (2004) Drug Discov Today 9:64 Ehrenhofer U, Rakowska A, Schneider SW, Schwab A, Oberleithner H (1997) Cell Biol Int 21:737 Engel A, Gaub HE, Muller DJ (1999) Curr Biol 9:R133 Evans EA (1989) Methods Enzymol 173:3 Feng D, Nagy JA, Pyne K, Hammel I, Dvorak HF, Dvorak AM (1999) Microcirculation 6:23 Fritz M, Radmacher M, Gaub HE (1993) Exp Cell Res 205:187 Fritz J, Katopodis AG, Kolbinger F, Anselmetti D (1998) Proc Natl Acad Sci USA 95:12283 Gorelik J, Zhang Y, Shevchuk AI, Frolenkov GI, Sanchez D, Lab MJ, Vodyanoy I, Edwards CR, Klenerman D, Korchev YE (2004) Mol Cell Endocrinol 217:101 Gorelik J, Zhang Y, Sanchez D, Shevchuk A, Frolenkov G, Lab M, Klenerman D, Edwards C, Korchev Y (2005) Proc Natl Acad Sci USA 102:15000 Graham JS, Vomund AN, Phillips CL, Grandbois M (2004) Exp Cell Res 299:335 Hansma PK (2006) Structure 14:390 Hansma PK, Drake B, Marti O, Gould SA, Prater CB (1989) Science 243:641 Henderson RM (2003) Methods Mol Med 86:139 Henderson RM, Oberleithner H (2000) Am J Physiol Renal Physiol 278:F689 Henderson RM, Edwardson JM, Geisse NA, Saslowsky DE (2004) News Physiol Sci 19:39 Hillebrand U, Hausberg M, Stock C, Shahin V, Nikova D, Riethmuller C, Kliche K, Ludwig T, Schillers H, Schneider SW, Oberleithner H (2006) Cardiovasc Res 69:916 Hinterdorfer P, Dufrene YF (2006) Nat Methods 3:347 Hoh JH (1998) Proteins 32:223 Hoh JH, Schoenenberger CA (1994) J Cell Sci 107(Pt 5):1105 Horton M, Charras G, Lehenkari P (2002) J Recept Signal Transduct Res 22:169 Iwasa Y, Hirono C, Sugita M, Takemoto K, Shiba Y (2001) Cell Physiol Biochem 11:311 Janicijevic A, Ristic D, Wyman C (2003) J Microsc 212:264 Jena BP (2006) Ultramicroscopy 106:663 Jiao Y, Schaffer TE (2004) Langmuir 20:10038 Kienberger F, Stroh C, Kada G, Moser R, Baumgartner W, Pastushenko V, Rankl C, Schmidt U, Muller H, Orlova E, LeGrimellec C, Drenckhahn D, Blaas D, Hinterdorfer P (2003) Ultramicroscopy 97:229 Korchev YE, Gorelik J, Lab MJ, Sviderskaya EV, Johnston CL, Coombes CR, Vodyanoy I, Edwards CR (2000) Biophys J 78:451 Laurent VM, Kasas S, Yersin A, Schaffer TE, Catsicas S, Dietler G, Verkhovsky AB, Meister JJ (2005) Biophys J 89:667 Lawrence JC, Saslowsky DE, Edwardson JM, Henderson RM (2003) Biophys J 84:1827 Le GC, Lesniewska E, Cachia C, Schreiber JP, de FF, Goudonnet JP (1994) Biophys J 67:36 Lehto T, Miaczynska M, Zerial M, Muller DJ, Severin F (2003) FEBS Lett 551:25 Lesniewska E, Giocondi MC, Vie V, Finot E, Goudonnet JP, Le GC (1998) Kidney Int Suppl 65:S42 Liu W, Murcia NS, Duan Y, Weinbaum S, Yoder BK, Schwiebert E, Satlin LM (2005) Am J Physiol Renal Physiol 289:F978
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.
18 Cellular Physiology of Epithelium and Endothelium
147
54. Mackie FE, Campbell DJ, Meyer TW (2001) Kidney Int 59:1458 55. Magaldi AJ, Cesar KR, de Araujo M, Simoes e Silva AC, Santos RA (2003) Pflugers Arch 447:223 56. Malcharek S, Hinz A, Hilterhaus L, Galla HJ (2005) Biophys J 88:2638 57. Maric K, Oksche A, Rosenthal W (1998) Am J Physiol 275:F796 58. Mathur AB, Truskey GA, Reichert WM (2000) Biophys J 78:1725 59. Michel CC, Curry FE (1999) Physiol Rev 79:703 60. Michel CC, Neal CR (1999) Microcirculation 6:45 61. Milhiet PE, Giocondi MC, Baghdadi O, Ronzon F, Roux B, Le Grimellec C (2002) EMBO Rep 3:485 62. Miyazaki H, Hayashi K (1999) Med Biol Eng Comput 37:530 63. Muller DJ, Heymann JB, Oesterhelt F, Moller C, Gaub H, Buldt G, Engel A (2000) Biochim Biophys Acta 1460:27 64. Neal CR, Michel CC (1995) J Physiol 488(Pt 2):427 65. Neher E, Sakmann B, Steinbach JH (1978) Pflugers Arch 375:219 66. Nilius B, Prenen J, Voets T, Eggermont J, Droogmans G (1998) J Physiol 506(Pt 2):353 67. Oberleithner H (2007) Pflugers Arch 454:187 68. Oberleithner H, Schwab A, Wang W, Giebisch G, Hume F, Geibel J (1994) Nephron 66:8 69. Oberleithner H, Schneider S, Larmer J, Henderson RM (1996) Kidney Blood Press Res 19:142 70. Oberleithner H, Schneider SW, Henderson RM (1997) Proc Natl Acad Sci USA 94:14144 71. Oberleithner H, Reinhardt J, Schillers H, Pagel P, Schneider SW (2000) Cell Physiol Biochem 10:429 72. Oberleithner H, Schneider SW, Albermann L, Hillebrand U, Ludwig T, Riethmuller C, Shahin V, Schafer C, Schillers H (2003) J Membr Biol 196:163 73. Oberleithner H, Ludwig T, Riethmuller C, Hillebrand U, Albermann L, Schafer C, Shahin V, Schillers H (2004) Hypertension 43:952 74. Oberleithner H, Riethmuller C, Ludwig T, Shahin V, Stock C, Schwab A, Hausberg M, Kusche K, Schillers H (2006) J Cell Sci 119:1926 75. Ohnesorge FM, Horber JK, Haberle W, Czerny CP, Smith DP, Binnig G (1997) Biophys J 73:2183 76. Olteanu D, Yoder BK, Liu W, Croyle MJ, Welty EA, Rosborough K, Wyss JM, Bell PD, Guay-Woodford LM, Bevensee MO, Satlin LM, Schwiebert EM (2006) Am J Physiol Cell Physiol 290:C952 77. Osada T, Takezawa S, Itoh A, Arakawa H, Ichikawa M, Ikai A (1999) Chem Senses 24:1 78. Pelling AE, Sehati S, Gralla EB, Valentine JS, Gimzewski JK (2004) Science 305:1147 79. Pereira RS (2001) Biochem Pharmacol 62:975 80. Puech PH, Poole K, Knebel D, Muller DJ (2006) Ultramicroscopy 106:637 81. Puntheeranurak T, Stroh C, Zhu R, Angsuthanasombat C, Hinterdorfer P (2005) Ultramicroscopy 82. Rabinovich Y, Esayanur M, Daosukho S, Byer K, El-Shall H, Khan S (2005) J Colloid Interface Sci 285:125 83. Radmacher M (1997) IEEE Eng Med Biol Mag 16:47 84. Radmacher M, Fritz M, Kacher CM, Cleveland JP, Hansma PK (1996) Biophys J 70:556 85. Reddy HK, Campbell SE, Janicki JS, Zhou G, Weber KT (1993) J Lab Clin Med 121:510 86. Riethmuller C, Jungmann P, Wegener J, Oberleithner H (2006) Pflugers Arch 453:157 87. Riethmuller C, Schaeffer TE, Kienberger F, Stracke W, Oberleithner H (2007) Ultramicroscopy 107:895 88. Rotsch C, Radmacher M (2000) Biophys J 78:520 89. Sakmann B, Neher E (1984) Annu Rev Physiol 46:455 90. Sato M, Nagayama K, Kataoka N, Sasaki M, Hane K (2000) J Biomech 33:127
148
C. Riethmüller · H. Oberleithner
91. Scheuring S, Levy D, Rigaud JL (2005) Biochim Biophys Acta 1712:109 92. Schneider SW, Sritharan KC, Geibel JP, Oberleithner H, Jena BP (1997) Proc Natl Acad Sci USA 94:316 93. Schneider SW, Yano Y, Sumpio BE, Jena BP, Geibel JP, Gekle M, Oberleithner H (1997) Cell Biol Int 21:759 94. Schneider SW, Pagel P, Rotsch C, Danker T, Oberleithner H, Radmacher M, Schwab A (2000) Pflugers Arch 439:297 95. Schneider SW, Matzke R, Radmacher M, Oberleithner H (2004) Methods Mol Biol 242:255 96. Schoenenberger CA, Hoh JH (1994) Biophys J 67:929 97. Schrot S, Weidenfeller C, Schaffer TE, Robenek H, Galla HJ (2005) Biophys J 89:3904 98. Seebach J, Dieterich P, Luo F, Schillers H, Vestweber D, Oberleithner H, Galla HJ, Schnittler HJ (2000) Lab Invest 80:1819 99. Shahin V (2006) Bioessays 28:935 100. Sibug-Aga R, Dunn RC (2004) Photochem Photobiol 80:471 101. Singla V, Reiter JF (2006) Science 313:629 102. Stock C, Allen RD, Naitoh Y (2001) J Exp Biol 204:291 103. Tandon R, Levental I, Huang C, Byfield FJ, Ziembicki J, Schelling JR, Bruggeman LA, Sedor JR, Janmey PA, Miller RT (2007) Am J Physiol Renal Physiol 292:F701 104. Turksen K, Troy TC (2004) J Cell Sci 117:2435 105. Vestweber D (2000) J Pathol 190:281 106. Wagner CA, Huber SM, Warntges S, Zempel G, Kaba NK, Fux R, Orth N, Busch GL, Waldegger S, Lambert I, Nilius B, Heinle H, Lang F (2000) Pflugers Arch 440:295 107. Weidenfeller C, Schrot S, Zozulya A, Galla HJ (2005) Brain Res 1053:162 108. Wolburg H, Lippoldt A (2002) Vascul Pharmacol 38:323 109. Yamane Y, Shiga H, Haga H, Kawabata K, Abe K, Ito E (2000) J Electron Microsc (Tokyo) 49:463 110. Yang L, Froio RM, Sciuto TE, Dvorak AM, Alon R, Luscinskas FW (2005) Blood 106:584 111. Zhang X, Wojcikiewicz EP, Moy VT (2006) Exp Biol Med (Maywood) 231:1306
19 Application of Atomic Force Microscopy to the Study of Expressed Molecules in or on a Single Living Cell Hyonchol Kim · Hironori Uehara · Rehana Afrin · Hiroshi Sekiguchi · Hideo Arakawa · Toshiya Osada · Atsushi Ikai
Abstract. One of the most useful ways to study the physical and biochemical properties of expressed molecules in or on living cells is to directly manipulate the molecules in question. In this chapter, we describe how the atomic force microscope (AFM) was used as a nanomanipulator, and membrane receptors or expressed messenger RNA (mRNAs) were touched, pulled and picked up to study the expressed information and biophysical properties of these molecules. First, the chemical and physical modifications of the AFM tip are introduced. For the study of membrane receptors over a large area of cell surface using an AFM, the colloidal probe method is useful. Details of probe preparation, chemical modification of colloids and data analysis are introduced. Furthermore, a useful method to measure the interaction force between a single molecular pair on a soft material is also introduced. Second, some examples of expressed receptor mapping are given. Ligand molecules were immobilized on a microbead attached to the AFM tip and force measurements were performed one by one over a large area of living cell surface. Distributions and amounts of expressed receptors on the cell surface were studied by measuring ligand–receptor interactions. Third, extractions of both membrane receptors and mRNAs are introduced. For the extraction of membrane receptors, the AFM tip was treated with cross-linkers. The tip was moved close to the cell surface and covalent bonds were formed between the tip and the receptors through the cross-linkers. Extraction of membrane proteins was confirmed by monitoring the force in each assay. For the extraction of mRNAs, the AFM tip was inserted into the cell body, and incubated for a few minutes. After incubation, the tip was withdrawn and inserted into a PCR tube. Collection of mRNAs was then confirmed with PCR analysis. By application of this method, the time course expression of specific mRNAs was successfully measured. The methods introduced in this chapter represent new and useful applications of the AFM to the study of the properties of single cells. Key words: Single cell, Nanomanipulation, Membrane receptor, Messenger RNA, Colloidal probe
Abbreviations AFM CHO CHX ECM FBS FN GFP mRNA PCR
Atomic force microscope Chinese hamster ovary Cycloheximide Extracellular matrix Fetal bovine serum Fibronectin Green fluorescent protein Messenger RNA Polymerase chain reaction
150
PG SD VN
H. Kim et al.
Prostaglandin Standard deviation Vitronectin
19.1 Introduction A recent trend in drug discovery is to find molecular markers which determine cellular characteristics and functions. For example, finding specific molecular markers for the determination of stem cell characteristics has critical importance not only for the development of regenerative medicine but also for the understanding of the occurrence and metastasis of cancers [1–3]. Recently, several methods have been developed to identify expressed molecules in or on cells. One of the promising methods is the application of array-based technologies, such as DNA microarray [4–8]. This technology is suitable for the simultaneous detection of a large number of expressed molecules; however, it is difficult to detect the expression of molecules in a single cell without amplification such as by polymerase chain reaction (PCR) and hence the result generally gives the average value of the properties of many cells. Another suitable method is the application of in situ hybridization technology [9–12]. By application of this method, expressed molecules in each cell can be visualized; however, as this method is usually based on fluorescent dye, spatial resolution of target detection is restricted by optical limits, and quantification of the target number is difficult because of the photobleaching of fluorescent dyes. Thus, to study the cellular properties and characteristics of each cell at the molecular level, a new technology should be developed. As examples of molecular markers representing cellular characteristics and hence good candidates for detection, two types should be mentioned: receptors on a cell surface, and messenger RNAs (mRNAs) in a cell. Cellular condition is usually controlled by outer environmental stimulation, and those signals are recognized at receptors on the cell surface. The cellular condition is also usually related to the amount of expressed mRNAs in a cell. Taking these points into account, direct observation and quantification of both cellular receptors and mRNAs are key factors to understand each cellular characteristic and to discover key biomarkers. In this chapter, methods to identify target receptors and mRNAs using atomic force microscopy force measurement mode are introduced. To determine the cellular conditions and responses of cells to environmental stimulations, the amount and the distribution of expressed receptors are good parameters, because the expression of receptors is usually controlled by the environment. To study the distribution of receptors using an atomic force microscope (AFM), it is useful to employ a microbead as a probe to increase the area of contact with the cell. With use of this colloidal probe tip, the distribution of molecules such as integrins has been determined on a living cell [13, 14]. Membrane receptors were also harvested using an AFM tip modified with bifunctional cross-linkers [15–17]. This technology is especially remarkable for the establishment of a method to identify expressed characteristic marker proteins on the cell surface at the single-molecule level. Moreover, we can harvest expressed
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
151
mRNAs from a cell using an AFM tip as a nanomanipulator without serious damage to the cell, and time-course developments of mRNA expressions have been studied using this method [18–20]. One of the advantages of using the AFM is that observation and manipulation can be made simultaneously in a liquid environment. From this benefit, all of the abovementioned experiments could be performed using living cells under physiological conditions. Another advantage is that quantitative force measurement is possible. Both observation of the distribution of receptors and harvest of mRNA were carried out by monitoring the interaction force between the AFM tip and the sample. By using force as a parameter, we can carry out quantitative nanomanipulations for a single living cell.
19.2 Methods of Manipulation To Study Molecules in or on a Living Cell Using an AFM 19.2.1 AFM Tip Preparation To Manipulate Receptors on a Cell Surface For the observation and manipulation of receptors on a cell surface, several kinds of chemical treatments or physical modifications have been applied to the AFM tip. In comparison with the size of a cell, the usual AFM tip is too sharp and small; thus, the colloidal probe method is useful for observation of the distribution of receptors. The colloidal probe was first introduced by Ducker et al. [21, 22]. Silica particles are a popular material for colloidal probes, because the chemistry of silica is well developed [23, 24]. By changing the material and the size of the colloidal particles, one can easily control the chemical properties or the contact area of the colloid. As the contact area of the colloid is larger than that of the usual AFM tip, many molecular pairs make bonds during an assay. When ligand molecules are immobilized on a colloidal particle and force measurements are performed on a living cell, the strength of interaction forces depends on the number density of receptors under the contact area at the position of each measurement. With use of this method, the distribution of several specific kinds of receptors over a large area of a cell surface has been studied [13, 14, 25]. The method of preparing the colloidal AFM probe is as follows: a suitable size and material for microbeads should be chosen to increase the contact area between the sample and the tip of the AFM. A suitable colloid for the modification of protein is carboxylated polystyrene microbeads, which are ideal for the immobilization of proteins or as a size reference, being a series of monodispersed polystyrene particles, and carboxyl groups are exposed on the bead surface. We can wash and gather dispersed colloids in solution by centrifugation. After they have been washed, beads are suspended in suitable solvent such as ethanol. The procedure to mount a microbead onto the AFM cantilever is as follows (Fig. 19.1). First, an AFM cantilever, epoxy resin, and a drop of microbead solution are put onto the warmed hot plate, respectively. Ethanol in the bead suspension immediately evaporates and the epoxy resin melts. A small amount of melted epoxy resin is picked up by the glass pipette of the micromanipulator, and placed on the
152
H. Kim et al.
end of the cantilever. A microbead is also picked up using another glass rod by physisorption and is placed on the epoxy resin coated spot on the cantilever, and the temperature is allowed to fall to room temperature. Figure 19.2 shows scanning electron microscopy images of an AFM cantilever to which a microbead has been fixed with epoxy resin. The microbead is tightly attached to the cantilever, while its surface facing the sample is not covered with epoxy resin. Figure 19.3 shows a schematic image of the experiment to mechanically locate receptors on
Fig. 19.1. A method to make a colloidal probe using a manipulator. a An atomic force microscope (AFM) cantilever, epoxy resin, and a drop of microbead solution were put onto warmed glass plates, respectively. Ethanol in the bead solution immediately evaporated and epoxy resin melted. b A small amount of molten epoxy resin was picked up by immersing a glass pipette in melted epoxy resin, and was placed on the end of the cantilever. c A microbead was picked up using another glass rod and placed on the epoxy resin coated spot on the cantilever
Fig. 19.2. Scanning electron microscopy images of an AFM cantilever with a polystyrene microbead. A microbead was placed at the end of the cantilever using a micromanipulator. a Overhead view of the cantilever. The length of the cantilever is 200 µm and the diameter of the microbead is about 10 µm. Scale bar 50 µm. b Side view of the cantilever. Scale bar 10 µm
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
153
Fig. 19.3. The experimental system. A microbead was attached to an AFM cantilever and several kinds of cell-adhesive molecules or antibodies were covalently immobilized on the bead surface. Interactions between these molecules and target receptors were measured in a liquid. Distributions of receptors were obtained from measurement results
the cell surface. As shown in this figure, ligand molecules are covalently immobilized to the bead surface and AFM force measurements are carried out. When using carboxylated polystyrene beads, ligand molecules are covalently immobilized to microbeads with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride and N-hydroxysuccinimide, which are well-known reagents to connect the amino group of the former with the carboxyl group of the latter. Measurements of interaction between ligand molecules and target receptors are then performed. The distribution of receptors on a cell surface is obtained from these measurements. 19.2.2 Analysis of Molecular Interactions Where Multiple Bonds Formed The interaction force is usually analyzed using a “force curve” (also usually referred to as a “force–distance curve”). Figure 19.4 shows a schematic image of the force curve. The abscissa indicates the vertical piezo movement relative to the sample surface, while the ordinate indicates the deflection of the cantilever. Cantilever deflection can be converted to applied force by multiplying it with the cantilever spring constant based on Hooke’s law. The downward deflection of the cantilever on the retraction sequence reflects complex adhesive interactions between the tip and the sample. Such adhesive interactions are terminated with a rupture force indicated by a sudden jump of the cantilever to the free level. Figure 19.5 shows a schematic image of the force curve, where a colloidal probe was used. Because of the large contact area, multiple point adhesions and detachments were involved in the measurement and the shape of the force curve was complicated. To evaluate the number of receptors under the contact area, the forces of all peaks must be added and calculated; however, it was almost impossible to distinguish between small force peaks and noise. Thus, force curves were converted to force–extension curves (Fig. 19.5b), where the abscissa indicates the
154
H. Kim et al.
Fig. 19.4. Force curves. a When the AFM tip does not interact with the sample surface, a retraction period follows in the same way as the approach period. b When the tip binds with the sample surface, a retraction period does not follow in the same way as the approach period, because the tip bends in the direction of the sample surface. This bend is suddenly recovered by detachment of the tip from the surface, and the binding force is calculated from the cantilever deflection, Δz. Extension of the sample is also calculated from the difference between cantilever deflection and piezo movement, which corresponds to Δl
Fig. 19.5. The method used to calculate both the separation work, W, and the indentation depth. a The force curve at which a microbead was used as a probe. b Conversion of the force curve to a force–extension curve. The retraction period of the force curve was transformed to a force– extension curve, and W, which corresponds to the area as indicated in the image, was calculated. c Method to calculate indentation depth. The broken line shows linear tip deflection calculated from piezo movement. The depth of indentation of the cell surface, d, corresponds to the length of the arrow, and was calculated from the compression part of the force curve
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
155
elongation of molecules on a sample surface and the ordinate indicates the interaction force calculated from the cantilever deflection. It is expected that the area under the force–extension curve is a better measure of interaction [26], and the force–extension curve was integrated to obtain the quantity with the dimension of energy (joules or piconewton nanometers) referred to as “separation work” (denoted as W, hereafter) [13, 14, 25, 27, 28]. To quantify the number of receptors, the contact area of the colloidal probe on the cell surface should be calculated. The contact area between a microbead and the cell is estimated according to the Hertzian contact model [29–34] as explained below. Figure 19.5c shows a schematic image of force curves obtained during separation of the bead from the cell surface. The broken line in the figure corresponds to tip deflection, which is proportional to piezo movement. The depth of indentation of the cell surface corresponds to the length of the arrow in Fig. 19.5c, and is calculated from the extension of the force curve. According to the Hertzian contact model for a spherical indenter, the depth of indentation, d, is given as d = {[3F(1 − ν2 )]/4E R1/2 }2/3 ,
(19.1)
with the tip radius, R, loading force, F, Young’s modulus, E, and Poisson ratio, ν, for the cell. The contact radius, a, is given according to the Hertzian contact model as a = [3FR(1 − ν2 )/4E]1/3 .
(19.2)
From (19.1) and (19.2), the relation between d and a is obtained as a = (Rd)1/2 .
(19.3)
The contact area, S, corresponds to the spherical cup of the radius R and height d, and is given as θ 2 sin θ dθ , (19.4) S = 2πR 0
where θ is defined as sin θ = a/R = (d/R)1/2 .
(19.5)
We can therefore calculate the contact area S based on d obtained from a force curve. In their interesting trial, Charras et al. [34] estimated the contact area for living cells from the analysis of fluorescence images, but optical errors in their analysis were large. The calculation of the exact contact area remains as an important problem in the studies of cellular properties with an AFM, and a theoretical approach using the Hertzian contact model may be one of the useful approaches, as indicated by the result of Charras et al. 19.2.3 Measurement of Single-Molecule Interaction Strength on Soft Materials For the measurement of interaction between a single ligand molecule and its receptor on the cell surface, the usual sharp AFM tip was used. In particular, a gold-coated AFM tip is convenient for chemical modification. One of the procedures of tip
156
H. Kim et al.
Fig. 19.6. “Compressionfree force spectroscopy.” An AFM tip on which a ligand molecule was immobilized is falseengaged, gradually approaches the cell surface, and is retracted before pushing into the cell
modification is as follows [25]. The AFM tip is cleaned with a UV–ozone cleaner and reacted with 10-carboxy-1-decanethiol. After the washing and activation procedure, the amino group of the ligand molecule is covalently cross-linked to the carboxylated tip in the same way as microbeads. To measure the interaction of a single receptor–ligand on a cell surface, “compression-free force spectroscopy”, where piezo movement is reversed from approaching the tip to retraction immediately before the start of the upward deflection of the cantilever, is used to reduce the area of contact (Fig. 19.6) [35, 36]. An AFM tip on which a ligand molecule has been immobilized is false-engaged and set into the force curve mode. The parameters of the force curve are the same as for microbeads. The tip gradually approaches the cell surface using a step motor. When the tip approaches the cell surface and retracts, we can observe interactions without pushing the cell surface.
19.3 Observation of the Distribution of Specific Receptors on a Living Cell Surface 19.3.1 Distribution of Fibronectin Receptors on a Living Fibroblast Cell As the first example of receptor mapping, fibronectin (FN) receptor on a living cell surface was observed using an AFM and its colloidal tip [13]. FN is a large, extracellular matrix (ECM) glycoprotein that exists in two major states, as a soluble
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
157
dimeric form in plasma and as an insoluble component within ECMs [37]. FN is composed of three types of repeating modules, termed type I, II, and III repeats, which are organized into functional domains [38, 39]. FN mediates its biological effects through binding to the heterodimeric transmembrane glycoproteins, integrins, which physically couple the cytoskeleton to the ECM [40–43]. The majority of integrinmediated interactions of FN with cells occur through the Arg–Gly–Asp (RGD) cell binding sequence. Through these interactions, integrins play a critical role in the regulation of a number of cellular functions, including adhesion and migration, cytoskeletal reorganizations, and the cell cycle progression. Disruption of the FN gene results in an embryonic lethal phenotype, confirming the importance of FN in cellular development [43]. By taking this clinical importance into account, the distribution of FN receptors has been studied on the single-cell level. In detail, FN molecules were immobilized onto the surface of the AFM colloidal tip, and force measurements were performed at several places on a single living cell. Figure 19.7 shows a histogram of W obtained between FN-coated beads and living cells. Because of the softness of the cell surface, the contact area changed depending on the position of the cell, so the contact area was calculated from the extension realm of the force curve according to the Hertzian contact model, and W was normalized with this value. The mean and the standard deviation (SD) of normalized W between FN-coated beads and living cells was
Fig. 19.7. Histogram of W obtained between a fibronectin (FN) coated bead and a living cell, between an unmodified bead and a cell, and between a FN-coated bead and a cell in the presence of 0.1 mg/mL FN
158
H. Kim et al.
0.69 ± 0.45 × 10−17 J/µm2 . In the presence of 0.1 mg/mL FN, the values decreased to 0.10 ± 0.08 × 10−17 J/µm2 . Unmodified microbeads gave a mean and a SD of 0.15 ± 0.13 × 10−17 J/µm2 . Because of the observed difference of cell adhesion strength from place to place on a cell surface, the data distribution was far from the normal distribution and the SD was large for the data given in Fig. 19.7. However, the mean and the SD were calculated to emphasize local differences of cell adhesion strength on a cell surface and to compare cell adhesion strength numerically with control experiments. From these results, it was concluded that W represented specific interactions between FN and the cell surface. Figure 19.8 shows the results of measurement of W in 16 subdivided areas on a living cell, a process we call W-mapping. All W-mappings were done in the same region. It took about 1 min for one cycle of mapping. The pattern of W-mapping was reasonably reproducible for six consecutive runs. To check the reproducibility of interaction between FN and a living cell, Student’s t test was carried out. The average P value of results during each pattern was 0.45, indicating that there were no clear differences between all possible pairs of the patterned data. Since the diffusion coefficient of the integrin β1 subunit in the membrane is reported to be in the range 0.01–0.1 µm2 /s [44–46], the receptor could migrate a distance covering a few subdivisions within a map between two consecutive measurements. The observed reproducibility of this results can be interpreted as follows: first, there was no large-
Fig. 19.8. The results of measurement of W in 16 subdivided areas on a living cell. All six maps were obtained in the same region. It took about 1 min for one mapping
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
159
Fig. 19.9. W-mapping over a large area of the cell surface. The area surrounded with a solid line was subdivided into 1-µm2 units, and measurements were carried out at each point. Bar 10 µm
scale migration of receptors on the time scale of this measurement, and, second, it is unlikely that the receptors were extracted from the cell membrane during the experiments. If the receptor proteins were extracted, the maps would not have been as reproducible as observed with respect to their local patterns. This conclusion does not exclude the possibility of small-scale migrations of receptors. Figure 19.9 shows a W-mapping over a large area of a living cell. The position at which FN was strongly adhered to the cell was scattered across the surface without recognizable patterns. This result was easily obtained by increasing the contact area, indicating that the use of increased contact area is useful for the quantitative analysis of cell-adhesive interactions and for studying the membrane at the level of the entire cell. 19.3.2 Distribution of Vitronectin Receptors on a Living Osteoblast Cell As the next example, the distribution of vitronectin (VN) receptors was studied [14]. One of the major VN receptor is integrin αv β5 receptor, which only binds to VN with a high specificity [41,47]. VN belongs to a group of adhesive glycoproteins that play key roles in the adhesion of cells to their surrounding matrix, and its involvement in various kinds of diseases has been reported [48]. For example, the locomotion of keratinocytes which plays an important role in the reepithelialization of cutaneous wounds is promoted by VN, and the promotion is inhibited by antibodies raised against the integrin αv β5 receptor [49], which indicates medical importance of VN, especially its interaction with integrin αv β5 receptor.
160
H. Kim et al.
Fig. 19.10. Immunohistochemical image of the distribution of integrin β5 subunits on MC3T3-E1 cells. Bar 10 µm
Figure 19.10 shows an example of immunofluorescence images of integrin β5 subunits on MC3T3-E1 cells. This integrin subunit only makes a heterodimer with the integrin αv subunit; thus, the image in Fig. 19.10 was almost equivalent with a distribution of integrin αv β5 receptors, which are one of the major VN receptors. There have been several reports on the localization of VN receptors, including integrin αv β5 , around focal contacts in cultured cells [50–52]. In this figure, the integrin β5 subunit seemed to be localized on the cell membrane, in agreement with previous findings. Localization was particularly conspicuous at the tiplike end of cells.
Fig. 19.11. Images of a murine osteoblastic cell taken with a an optical microscope and b a fluorescence microscope, and c mapping of W using an AFM. The square in a corresponds to the area of 32 × 32 µm2 in which the measurement using the AFM was performed
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
161
To study the distribution of binding sites for VN, a specific ligand of integrin αv β5 , on the cell membrane, and to compare the result with that for the integrin β5 subunit, measurements of cell-adhesive strength between VN and the cell were carried out using an AFM. Figure 19.11 shows a W-mapping between VN and the receptors on the cell surface, and a comparison of it with the immunohistochemical findings. Results of AFM measurements indicated that VN receptors, including integrin αv β5 , were localized in a similar manner as obtained by immunohistochemistry. The distribution of high values of W was had a similar pattern to that of the fluorescence image, but was not completely the same. This might be because, first, all of the integrin β5 subunits on the top and bottom sides of the cell were stained in immunohistochemistry, but only those on the dorsal side of the cell surface were probed in the AFM experiment, and, second, there are VN receptors other than integrin αv β5 . Though there are other kinds of VN-binding molecules, a comparison with the distribution of integrin αv β5 has special importance, because integrin αv β5 interacts only with VN, and their biological role might be controlled by VN. From the abovementioned results, it was concluded that VN receptors are localized and some of the peripheries of the cell were covered with receptors. These results indicated that this AFM-based method is suitable for studying the membrane receptors at the level of a single cell, and a comparison of the results of AFM measurements with immunohistochemistry is useful to study distributions of receptors on the cell surface. 19.3.3 Quantification of the Number of Prostaglandin Receptors on a Chinese Hamster Ovary Cell Surface Cell-adhesive interactions have been studied at the single-cell level using the AFM and notable results were obtained. This AFM-based method does not require artificial cellular treatment, such as cell fixation, and it is easy to make a quantitative evaluation of receptor expression, because the work to separate a molecular pair, which is extensively variable, has been referred to as a parameter. In these previous studies, cell-adhesive strength was successfully measured; however, the numbers of receptors could not be quantified. For quantitative evaluation, values of cell-adhesive strengths, which can be measured using a microbead probe, should be normalized with that of single molecular interaction. To quantify the numbers of receptors, many single ligand–receptor interactions should be measured one by one, but it is not so easy and a general method is desired. For such a purpose, it is useful to genetically introduce a green fluorescent protein (GFP) molecule to the receptors [53–55]. The advantage is, firstly, many kinds of receptors, the distributions of which have been observed with a microbead, will be quantified by normalizing the value with that of the single GFP:anti-GFP antibody interaction when the GFP is set as the target of the measurement of the distribution by immobilizing the anti-GFP antibody to the microbead, and, secondly, it is also useful to compare the result of AFMbased receptor mapping with that of fluorescence observation. Though we can also identify the distribution of the receptor by the fluorescence of GFP molecules, there are advantages with the use of the AFM. Firstly, it is not necessary to consider the photobleaching problem; hence, it has the potential to detect a low number of target
162
H. Kim et al.
molecules with high reliability. Secondly, quantitative evaluation is easy in the case of the AFM-based method. A prostaglandin (PG) receptor, EP3, which is known to be a target of many kinds of medicines was investigated [25]. PGs exert a variety of actions in various types of tissues and cells. Milton and Wendlandt [56] were the first to suggest PGE2 as a central mediator of fever. Indeed, E type PGs are powerful fever inducers when injected into the brain, and the level of PGE2 increased in the preoptic area during lipopolysaccharide-induced fever, and indomethacin completely abolished both lipopolysaccharide-induced fever and increased the levels of PGE2 in the preoptic area [56–60]. The actions of PGE2 are mediated through rhodopsin-type cell surface receptors [61]. These receptors are four subtypes of the PGE2 receptor: EP1, EP2, EP3, and EP4 [61]. Ushikubi et al. [62] used mice lacking each subtype of the PGE receptor to address PGE2 -mediated fever generation. They found that only the mice lacking the EP3 receptor failed to show febrile responses. This study thus clearly demonstrated that PGE2 -mediated fever generation by acting on the EP3 receptor. By taking such clinical importance into account, the distribution of EP3 receptors was studied at the single-cell level using an AFM. For the identification, GFP, which was genetically introduced to the extracellular region of EP3 receptor on a chinese hamster ovary (CHO) cell (the cell is denoted as GFP+/CHO cell hereafter), was used as a target of interaction measurement, and an AFM-based measurement of the interaction between the GFP and its antibody, which was immobilized on a microbead tip, was carried out (Fig. 19.12). The force curves obtained were transformed to force–extension curves, the separation work, W, was calculated, and a W-mapping between the anti-GFP antibody and GFP on a GFP+/CHO cell surface was obtained (Fig. 19.13). The result was also compared with the immunohistochemical findings at the same area. The results of both AFM measurement and immunohistochemistry indicated that EP3 receptors were distributed nonuniformly on the GFP+/CHO cell surface in small patches. To compare the AFM finding with that of fluorescence, the former was averaged and the latter was subdivided into 4 × 4 regions (Fig. 19.13d,e) and the correlation of each data point
Fig. 19.12. Force curves obtained between the GFP+/CHO cell and anti-GFP antibody where a microbead was used as a probe. GFP green fluorescent protein, CHO Chinese hamster ovary. Bar = 5 nm
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
163
Fig. 19.13. Distribution of EP3 receptors on the GFP+/CHO cell surface. a Optical microscope image of a living GFP+/CHO cell on which the measurement of W was performed. The area surrounded by a solid line was subdivided into 1-µm2 units, and measurements of W were carried out at each point using a microbead probe. b A map of W between the anti-GFP antibody and the GFP+/CHO cell surface. c Immunohistochemical image obtained using a confocal microscope. d The W map obtained with 16 × 16 data points was averaged to a 4 × 4 map to compare with it the fluorescence image. Values were normalized using the weakest value. e The fluorescence image was subdivided into 4 × 4 data points, and the intensity of each point was calculated. Values were also normalized using the weakest value. f Correlation between the intensity of the confocal image and the result for W, obtained by comparing each data point between d and e
was studied (Fig. 19.13f). There was a good correlation between the intensity of the confocal image and the value of W, which confirmed the reliability of the result of W-mapping. The advantages of the AFM measurement are, firstly, that the AFM can detect small amounts of receptors, which are difficult to visualize by immunohistochemistry (in this case, the fluorescence of GFP was enhanced with the immunohistochemical method because of too weak fluorescence), without any chemical treatment, such as cell fixation, and, secondly, it is very easy to compare the relative numbers of receptors at each position of the cell membrane individually with a high signal-to-noise ratio by comparing the values of W. These results indicated that the AFM-based method is suitable for studying the membrane receptors at the single-cell level, and a comparison of the results of AFM measurements with immunohistochemistry is useful to study the distributions of receptors on the cell surface. To quantify the number of EP3 receptors, a single molecular interaction between GFP and the anti-GFP antibody was measured on the GFP+/CHO cell surface. Because the cell surface is too soft, the cells were easily deformed by applying
164
H. Kim et al.
a loading force and the contact area increased even if a usual AFM tip was used. To overcome this problem, compression-free force spectroscopy [35, 36] was used to make the contact area as small as possible (Fig. 19.14). Figure 19.14b shows typical images of the force curve between the anti-GFP antibody and the GFP+/CHO cell obtained with compression-free force spectroscopy. Measurements were also performed on a normal CHO cell surface, on which GFP was not introduced, in a control experiment (Fig. 19.14c) and no interaction was observed. The values of W obtained using the method were plotted, and discrete populations of W were obtained (Fig. 19.15). The value of W to separate a single GFP molecule from its antibody on a cell surface was then estimated to be about 1.5 × 10−18 J. This result was firstly obtained by applying compression-free force spectroscopy to the measurement of interaction on soft materials, such as a living cell surface. Using this value, we can roughly quantify the W-mapping. For example, there are about 880 EP3 receptor molecules in the area of Fig. 19.13, and if we assume the area of the cell surface to be 5000 µm2 [20], about 1.7 × 104 receptors exist on the cell surface. These results indicate that the AFM-based method is one of the more powerful ways to quantify the number of receptors on a living cell surface.
19.4 Further Application of the AFM to the Study of Single-Cell Biology 19.4.1 Manipulation of Expressed mRNAs in a Living Cell Using an AFM The AFM was used not only for the nanoprobe to measure interaction strength, but also as a mechanical nanomanipulator to pick up biomolecules from a living cell without serious damage. In this section, examples of a method to study expressed mRNAs in a single living cell are introduced [18, 19]. In this method, the AFM tip is inserted into a living cell body, and expressed mRNA is harvested by the AFM tip. Details of the procedure are as follows. Living cell lines were grown in culture dishes and cells were washed with culture medium devoid of serum components and used for AFM experiments. The AFM tip was positioned above a target cell under an AFM combined with an optical microscope. The AFM tip was approached using a step motor and the contact of the AFM tip with the cell surface was detected by the deflection signal of the AFM cantilever. Higher loading forces were applied to insert the AFM tip into the cell membrane. The AFM tip was inserted into the cytoplasm region around the nucleus to collect mRNA. After the AFM tip had been inserted into a single cell, it was placed into a PCR tube and PCR was performed. For the first example, the β-actin mRNA expression of individual living cells was examined using rat fibroblast-like VNOf90 cells and mouse osteoblast-like MC3T3E1 cells (Fig. 19.16a). Although β-actin mRNAs are usually distributed throughout the cytoplasm uniformly, they are localized to the leading edge of the cells when the cells start to migrate [63,64]. Thus, the single cells, which were surrounded by other cells that inhibit the migration of the target cells, were chosen as targets to study
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
165
Fig. 19.14. Observation of single molecular interactions on a cell surface. a Compression-free force spectroscopy. Piezo movement was reversed from approaching the tip to retraction immediately before the start of the upward deflection of the cantilever. b Typical images of the force curve between anti-GFP antibody and the GFP+/CHO cell using compression-free force spectroscopy. c Typical images of the force curve between the anti-GFP antibody and the normal CHO cell in the control experiment. Approach of the AFM tip to the cell surface was confirmed by a small deflection of the cantilever. Bars 2 nm
β-actin mRNA expression. PCR products for rat and mouse β-actin mRNAs were detected as shown in the even-numbered lanes of Fig. 19.16b and c. In the negative control, PCR products were not detected without the insertion of the tip into the cell (odd-numbered lanes in Fig. 19.16b, c). Experiments for the detection of β-actin mRNA and the negative control were performed alternately. In Table 19.1, results of the detection of β-actin mRNA from single VNOf90 cells are presented. The AFM-based mRNA extraction was performed on 102 single living cells. The number of assays against single cells ranged from one to six. The time interval between one assay and the next one against the same cell ranged from 5 to 60 min. When six sets of assays had performed against three single cells, the following results were obtained. In cases of two single cells, PCR products were detected in all six assays. In the case of another cell, PCR products were detected five times out of six assays. Seventy-two positive results were obtained when one assay was tried for 73 single cells. In total, 189 assays were performed on 102 single living cells, and 182 positive results were obtained (96.3%). The other seven negative results may have resulted for the following three reasons: (1) the procedure for inserting the tip into the cell was not successful; (2) the PCR was not successful;
166
H. Kim et al.
Fig.19.15.Determination of W which corresponds to the strength of a single molecular interaction. a W obtained using compression-free force spectroscopy. b Population of the values of W obtained. All data points were plotted on the axis of W
(3) β-actin mRNA did not exist in the cell where the tip was inserted. The first two reasons are considered to be experimental errors, and the last one is considered to result from the probability of the existence of mRNA. Nevertheless, the high probability of detecting β-actin mRNA indicated that it could be used as a positive marker for examining the expression of other genes which would exclude the first two reasons from consideration. A population of mammalian cells grown in culture medium contains about one million mRNAs per cell. Fibroblast cells are reported to contain 2000–3000 copies of β-actin mRNA per cell [65]. This method can be applied to detect other mRNAs whose copies number more than several thousand per cell. In the next trial, it was examined whether this AFM-based method can be used to determine more than one kind of gene at the same time or not. As shown in Fig. 19.17a, VNOf90 cells expressed FN protein as well as β-actin. AFM tips inserted into the single living cells were placed into PCR solution that contained primers for both β-actin and FN. In total, ten single living cells were assayed and the results of five single cells are shown in Fig. 19.17b. PCR products for both β-actin and FN mRNAs were detected in all of the ten different single cells. In order to examine time-dependent gene expression of single living cells, the response of rat VNOf90 cells to serum has been used as a model for studying the changes in c-fos gene expression (Fig. 19.18). The cells were induced to enter a quiescent state with culture medium containing 0% fetal bovine serum (FBS) for
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
167
Table 19.1. Detection of β-actin messenger RNA from single VNOf90 cells Number of assays Number of β-actin Number Total number of β-actin Total number of against single cells positive cases of cells positive cases assays in each case 1 1 2 2 2 3 3 4 4 4 5 5 5 6 6 6
1 0 2 1 0 3 2,1 or 0 4 3 2,1 or 0 5 4 3,2,1 or 0 6 5 4,3,2,1 or 0
72 1 4 1 0 1 0 12 3 0 4 1 0 2 1 0 102
72 0 8 1 0 3 0 48 9 0 20 4 0 12 5 0 182
72 1 8 2 0 3 0 48 12 0 20 5 0 12 6 0 189
Fig. 19.16. Extraction of messenger RNAs (mRNAs) from living cells. a The experimental procedure. A target cell was positioned underneath the AFM tip through the observation of an inverted optical microscope combined with AFM. The AFM tip was then lowered onto the cell and inserted, and held for approximately 45 s to allow the tip to bind the cell ingredient containing mRNA with physical absorption. The tip was then lifted off the cell and placed into a polymerase chain reaction (PCR) tube. b, c β-Actin mRNA expression of five rat VNOf90 cells (b) and two mouse MC3T3-E1 cells (c) was examined as shown in the even-numbered lanes. In the negative control (odd-numbered lanes), the tip contacted the medium without its insertion into the cell
168
H. Kim et al.
Fig. 19.17. Detection of both β-actin and FN mRNAs from single living cells. Immunocytochemistry demonstrated VNOf90 cells expressing both FN (a) and β-actin (b). Bar 100 µm. Both β-actin and FN mRNAs were detected in all five different single cells (c)
Fig. 19.18. Cell population analysis. a Response of VNOf90 cells to serum and cycloheximide (CHX) was used as a model for studying time-dependent gene expression. VNOf90 cells were induced to enter a quiescent state by serum deprivation for 24 h and then stimulated by the addition of culture medium containing 10% fetal bovine serum (FBS) and 10 µg/mL CHX. The addition of serum and CHX induced changes in gene expression, including c-fos mRNA. b Relative mRNA levels of the cell population were measured with quantitative PCR. The ordinate represents the natural logarithm of the relative initial amount of mRNA
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
169
24 h and then stimulated by addition of 10% FBS and 10 µg/mL cycloheximide (CHX) for 2 h. CHX is well known to induce increases in c-fos mRNA expression level to prevent degradation of c-fos mRNA. The solution was then changed to culture medium containing no FBS and 10 µg/mL CHX for 1 h. Finally, the solution was changed to culture medium containing 10 µg/mL CHX (condition A) or no CHX (condition B). Total RNA from confluent cells on 60-mm dishes at different times was extracted and quantification of mRNA was carried out with quantitative PCR. The β-actin mRNA levels were more constant than the c-fos mRNA levels in both conditions A and B (Fig. 19.18b). Although the c-fos mRNA levels decreased rapidly in the absence of CHX from 3 to 5 h following stimulation, they decreased gradually in the presence of CHX from the cell population analysis. The detection of c-fos and β-actin mRNAs from single cells, termed A1, A2, and A3 cells (from condition A), was examined with an AFM-based manipulation method (Fig. 19.19). The change
Fig. 19.19. Single-cell analysis. a Time-dependent mRNA expression of single living cells was examined in both condition A (A1–A3) and condition B (B1–B3). The ordinate represents the natural logarithm of the number of initial mRNA copies, and the abscissa represents time in hours. b Number of β-actin mRNAs detected with single-cell assay. n.d. product was not detected with quantitative PCR
170
H. Kim et al.
of β-actin mRNA levels was also small compared with that of c-fos mRNA levels in single-cell analysis as well as in the cell population analysis. Although the expression profile of c-fos mRNA in the A1 cell was very similar to the result obtained from the cell population analysis, the A2 and A3 cells showed different expression profiles. These results suggest that c-fos mRNA expression levels are not uniform among individual cells. Although c-fos mRNA was not detected in the A3 cell at 3.5 h, β-actin mRNA was detected, indicating that the experimental errors were able to be eliminated. In condition B (B1, B2, and B3 cells), c-fos mRNA expression was not observed in any of the cases at 5 h following stimulation. This result corresponded to the result from the cell population analysis. The expression of c-fos mRNA in the B1 cell was observed only at 3 h. In the B2 cell, c-fos mRNA expression was detected at 3 h and at 3.5 h and not after 4 h. In the B3 cell, c-fos mRNA expression was detected at 3 h and not at 3.5 h, and was then detected again at 4 h. This type of pattern was also observed in the A3 cell. These results indicate the possibility that the fluctuation of c-fos mRNA expression level occurs at the single-cell level. Figure 19.19b shows the histogram of the number of β-actin mRNAs bound to AFM tips. In most cases, their number was equal to or less than 20. In conclusion, a powerful new technique to study gene expression in mammalian single cells was provided. The results of these studies demonstrate that time-dependent gene expression patterns in single cells are not always similar to those in large cell populations. At the single-cell level, gene expression patterns sometimes fluctuate up and down. These phenomena will aid in the understanding of the properties of individual cells. 19.4.2 Manipulation of Membrane Receptors on a Living Cell Surface Using an AFM The extraction of membrane protein molecules has been examined using an AFM for establishment of a new technique to harvest and identify critical molecular markers on the cell surface as shown in Fig. 19.20 [15–17]. In this study, the AFM tip was modified with bifunctional covalent cross-linkers, and membrane proteins were indiscriminately picked up using AFM force measurement mode. In Fig. 19.21, typical force curves, which were obtained on a glass surface (Fig. 19.21a) and on the surface of the live fibroblast cell (Fig. 19.21b) using an unmodified silicon nitride tip, are presented. In the case of Fig. 19.21a, there was no indentation on either the tip or the sample, whereas in the case of Fig. 19.21b, the sample was deformed under the vertical stress inflicted by the tip. The curve in Fig. 19.21c represents a typical force curve obtained on a living cell surface using a tip modified with a bifunctional covalent cross-linker. Although the approach part of the curve is similar to that of the curve in Fig. 19.21b, its return part is characterized by a prolonged downward deflection which was abruptly ended after a sample extension of approximately 115 nm. The extension of the sample was calculated as the difference between the cantilever deflection (ordinate) and the distance of the piezo movement (abscissa) from the sample surface. The curve shows that an adhesive interaction was established between the tip and the sample surface during their contact of approximately 300–400 ms and that, during the retraction period of the curve, the cell surface was lifted by a vertical distance of approximately 115 nm, whereas the cantilever was
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
171
Fig. 19.20. Extraction of receptors. An AFM tip was modified with a cross-linker and brought to the living cell surface. A covalent bond was formed between the cross-linker on the tip and receptors. Receptors were then picked up with the retraction of the AFM tip
Fig. 19.21. Force curves obtained under three different conditions: a against a bare glass surface using an unmodified tip, b against a cell surface using an unmodified tip, and c against a cell surface using a tip modified with a cross-linker. The gray line corresponds to approach and the black line corresponds to retraction for all the force curves
pulled down by approximately 7 nm. The tensile force at the final separation of the tip from the cell surface was calculated as approximately 0.42 nN in this case by knowing the force constant of the cantilever to be approximately 0.06 nN/nm. When unmodified silicon nitride tips were used on live cells, 90% of the force curves (N = 50) showed no positive indication of adhesive interactions, and the remaining 10% exhibited weak interaction forces of less than approximately 50 pN with the cell surface. To verify that the relatively strong separation force observed in Fig. 19.21c corresponded to the extraction of membrane proteins through covalent bond formation, the modified cantilevers used on cell surface to obtain more than 50 force curves were remounted on the AFM after treating them with glycine/glycinamide to deactivate
172
H. Kim et al.
the remaining cross-linkers. Force curves were then recorded for a silicon substrate modified with another bifunctional cross-linker. It was expected that if proteins were on the tip they would form covalent bonds with the cross-linkers on the substrate and force curves representing protein stretching would be obtained. Because the length of protein extension is an important parameter in protein stretching together with the tensile force, it is suitable to present the results of the force curve measurements in terms of force–extension curves. In Fig. 19.22, a typical force–extension curve, which was obtained using a used tip on a cell surface against a modified silicon surface, is presented together with a histogram of the extended length before the final separation of the tip from the substrate. For the force–extension curve in Fig. 19.22a, the force of the final separation was approximately 1.3 nN, proving the establishment of a covalently bonded system between the tip and the substrate. Proof for this comes from a report on the strength of covalent bonds under mechanical stress. Grandbois et al. [66] reported their theoretical prediction that the weakest bond in a similar cross-linking system to that in this study would be the Si–C bond, and concluded from their experimental results that covalent bond rupture occurs at 2.0±0.3 nN (with a distribution ranging mainly between 1.0 and 2.0 nN) at a loading rate of 10 nN/s. In a similar loading rate range, 1.0–2.0 nN of the final rupture force is also most often obtained for carbonic anhydrase using a covalently immobilized tip or substrate under the AFM [67, 68]. The histogram in Fig. 19.22b shows that the tensile material covalently sandwiched between the used tip and the modified silicon substrate was stretched up to 200–250 nm with a mean extension length of approximately 85 nm, strongly suggesting that the extracted macromolecules, most probably proteins, were transferred from the cell surface to the AFM tip during the abovementioned operation. Figure 19.23 shows a histogram of rupture forces obtained on a cell surface using a modified AFM tip. In this histogram, the force values mainly clustered between 0.4 and 0.8 nN, with a mean and deviation of 0.45 ± 0.22 nN. Because a force larger than 1 nN is infrequent and may involve multiple adsorption events, a force range
Fig. 19.22. Confirmation of the extraction of membrane receptors using an AFM. a F–E curve obtained when a modified tip used on the cell surface was tested on a modified silicon surface with a crosslinker, as characterized by a total extension of approximately 85 nm and a final rupture force of 1.3 nN. b The maximal extension length of 44 curves (final rupture force >1.0 nN) is collected in a histogram
19 Application of AFM to the Study of Expressed Molecules in or on a Cell
173
Fig. 19.23. Rupture forces obtained on the cell surface with cross-linkermodified tips
of 0.4–0.6 nN was considered as the most likely force range required to extract or, according to Bell, “uproot” intrinsic membrane proteins, including glycoproteins, from the cell membrane [69]. In conclusion, a method to pick up the membrane receptors is suggested using an AFM tip on which reactive cross-linkers are modified. This technology is especially remarkable for the establishment of a method to identify expressed characteristic marker proteins on a cell surface at the single-molecule level.
References 1. Reya T, Morrison SJ, Clarke MF, Weissman IL (2001) Nature 414:105 2. Pardal R, Clarke MF, Morrison SJ (2003) Nat Rev Cancer 3:895 3. Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, Faber J, Levine JE, Wang J, Hahn WC, Gilliland DG, Golub TR, Armstrong SA (2006) Nature 442:818 4. Schena M, Shalon D, Davis RW, Brown PO (1995) Science 270:467 5. Lockhart DJ, Dong H, Byrne MC, Follettie MT, Gallo MV, Chee MS, Mittmann M, Wang C, Kobayashi M, Horton H, Brown EL (1996) Nat Biotechnol 14:1675 6. Cawley S, Bekiranov S, Ng HH, Kapranov P, Sekinger EA, Kampa D, Piccolboni A, Sementchenko V, Cheng J, Williams AJ, Wheeler R, Wong B, Drenkow J, Yamanaka M, Patel S, Brubaker S, Tammana H, Helt G, Struhl K, Gingeras TR (2004) Cell 116:499 7. Bertone P, Stolc V, Royce TE, Rozowsky JS, Urban AE, Zhu X, Rinn JL, Tongprasit W, Samanta M, Weissman S, Gerstein M, Snyder M (2004) Science 306:2242 8. Cheng J, Kapranov P, Drenkow J, Dike S, Brubaker S, Patel S, Long J, Stern D, Tammana H, Helt G, Sementchenko V, Piccolboni A, Bekiranov S, Bailey DK, Ganesh M, Ghosh S, Bell I, Gerhard DS, Gingeras TR (2005) Science 308:1149 9. Rudkin GT, Stollar BD (1977) Nature 265:472 10. Bauman JG, Wiegant J, Borst P, van Duijn P (1980) Exp Cell Res 128:485 11. Nederlof PM, Robinson D, Abuknesha R, Wiegant J, Hopman AH, Tanke HJ, Raap AK (1989) Cytometry 10:20
174
H. Kim et al.
12. Raap AK, van de Rijke FM, Dirks RW, Sol CJ, Boom R, van der Ploeg M (1991) Exp Cell Res 197:319 13. Kim H, Arakawa H, Osada T, Ikai A (2002) Colloids Surf B 25:33 14. Kim H, Arakawa H, Osada T, Ikai A (2003) Ultramicroscopy 97:359 15. Afrin R, Arakawa H, Osada T, Ikai A (2003) Cell Biochem Biophys 39:101 16. Afrin R, Yamada T, Ikai A (2004) Ultramicroscopy 100:187 17. Afrin R, Ikai A (2006) Biochem Biophys Res Commun 348:238 18. Osada T, Uehara H, Kim H, Ikai A (2003) J Nanobiotechnol 1:2 19. Osada T, Uehara H, Kim H, Ikai A (2004) Adv Clin Chem 38:239 20. Uehara H, Osada T, Ikai A (2004) Ultramicroscopy 100:197 21. Ducker WA, Senden TJ, Pashley RM (1991) Nature 353:239 22. Ducker WA, Senden TJ, Pashley RM (1992) Langmuir 8:1831 23. Lee GU, Kidwell DA, Colton RJ (1994) Langmuir 10:354 24. Takano H, Kenseth JR, Wong SS, O’Brien JC, Porter MD (1999) Chem Rev 99:2845 25. Kim H, Arakawa H, Hatae N, Sugimoto Y, Matsumoto O, Osada T, Ichikawa A, Ikai A (2006) Ultramicroscopy 106:652 26. Smith BL, Schaffer TE, Viani M, Thompson JB, Frederick NA, Kindt J, Belcher A, Stucky GD, Morse DE, Hansma PK (1999) Nature 399:761 27. Kim H, Arakawa H, Osada T, Ikai A (2002) Appl Surf Sci 188:493 28. Kim H, Tsuruta S, Arakawa H, Osada T, Ikai A (2004) Ultramicroscopy 100:203 29. Sneddon IN (1965) Int J Eng Sci 3:47 30. Weihs TP, Nawaz Z, Jarvis SP, Pethica JB (1991) Appl Phys Lett 59:3536 31. Ikai A (1996) Surf Sci Rep 26:263 32. Haga H, Sasaki S, Morimoto M, Kawabata K, Ito E, Abe K, Sambongi T (1998) Jpn J Appl Phys Part 1 37:3860 33. Haga H, Sasaki S, Kawabata K, Ito E, Ushiki T, Sambongi T (2000) Ultramicroscopy 82:253 34. Charras GT, Lehenkari PP, Horton MA (2001) Ultramicroscopy 86:85 35. Sekiguchi H, Arakawa H, Okajima T, Ikai A (2002) Appl Surf Sci 188:489 36. Sekiguchi H, Arakawa H, Taguchi H, Ito T, Kokawa R, Ikai A (2003) Biophys J 85:484 37. Romberger DJ (1997) Int J Biochem Cell Biol 29:939 38. Sekiguchi K, Hakomori S (1980) Proc Natl Acad Sci USA 77:2661 39. Brumfeld V, Werber MM (1993) Arch Biochem Biophys 302:134 40. Hynes RO (1987) Cell 48:549 41. Ruoslahti E, Pierschbacher MD (1987) Science 238:491 42. Hynes RO (1992) Cell 69:11 43. George EL, Georges-Labouesse EN, Patel-King RS, Rayburn H, Hynes RO (1993) Development 119:1079 44. Duband JL, Nuckolls GH, Ishihara A, Hasegawa T, Yamada KM, Thiery JP, Jacobson K (1988) J Cell Biol 107:1385 45. Felsenfeld DP, Choquet D, Sheetz MP (1996) Nature 383:438 46. Choquet D, Felsenfeld DP, Sheetz MP (1997) Cell 88:39 47. Smith JW, Vestal DJ, Irwin SV, Burke TA, Cheresh DA (1990) J Biol Chem 265:11008 48. Preissner KT, Seiffert D (1998) Thromb Res 89:1 49. Kim JP, Zhang K, Chen JD, Kramer RH, Woodley DT (1994) J Biol Chem 269:26926 50. Wayner EA, Orlando RA, Cheresh DA (1991) J Cell Biol 113:919 51. Stuiver I, Ruggeri Z, Smith JW (1996) J Cell Physiol 168:521 52. DePasquale JA (1998) Histochem Cell Biol 110:485 53. Rocheville M, Lange DC, Kumar U, Patel SC, Patel RC, Patel YC (2000) Science 288:154 54. Bouvier M (2001) Nat Rev Neurosci 2:274 55. Zhang J, Campbell RE, Ting AY, Tsien RY (2002) Nat Rev Mol Cell Biol 3:906
19 Application of AFM to the Study of Expressed Molecules in or on a Cell 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69.
175
Milton AS, Wendlandt S (1970) J Physiol 207:76P Stitt JT (1986) Yale J Biol Med 59:137 Sirko S, Bishai I, Coceani F (1989) Am J Physiol 256:R616 Kluger MJ (1991) Physiol Rev 71:93 Sehic E, Szekely M, Ungar AL, Oladehin A, Blatteis CM (1996) Brain Res Bull 39:391 Narumiya S, Sugimoto Y, Ushikubi F (1999) Physiol Rev 79:1193 Ushikubi F, Segi E, Sugimoto Y, Murata T, Matsuoka T, Kobayashi T, Hizaki H, Tuboi K, Katsuyama M, Ichikawa A, Tanaka T, Yoshida N, Narumiya S (1998) Nature 395:281 Lawrence JB, Singer RH (1986) Cell 45:407 Kislauskis EH, Zhu X, Singer RH (1997) J Cell Biol 136:1263 Latham VM Jr, Kislauskis EH, Singer RH, Ross AF (1994) J Cell Biol 126:1211 Grandbois M, Beyer M, Rief M, Clausen-Schaumann H, Gaub HE (1999) Science 283:1727 Wang T, Ikai A (1999) Jpn J Appl Phys Part 1 38:3912 Alam MT, Yamada T, Carlsson U, Ikai A (2002) FEBS Lett 519:35 Bell GI (1978) Science 200:618
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates? Amalisa Relini · Ornella Cavalleri · Claudio Canale · Tiziana Svaldo-Lanero · Ranieri Rolandi · Alessandra Gliozzi
Abstract. Amyloids are insoluble, thread-like protein aggregates which are associated with a number of severe diseases. Different techniques currently employed to characterize the structural features of amyloids, such as optical methods, electron microscopy, X-ray diffraction, nuclear magnetic resonance and atomic force microscopy (AFM), are briefly described in the first part of this chapter. Among these techniques, AFM plays a relevant role, as it is capable of imaging samples in different environments without requiring particular preparation, such as staining or fixation procedures. It allows the different steps of the aggregation process to be followed, from early oligomers to fibrils, providing nanometric details of their morphology. AFM has been used to probe the effects of organic and inorganic surfaces on the aggregation process and the interaction of amyloid with model membranes. Finally we review how the statistical analysis of AFM images and force spectroscopy measurements can give quantitative information on mechanical properties of fibrils. Key words: Atomic force microscopy, Tapping mode atomic force microscopy, Amyloid aggregates, Amyloid fibrils, Surface-induced protein aggregation
Abbreviations Aβ AFM CR DRA EM EPR HypF-N
Amyloid β Atomic force microscopy Congo red Dialysis-related amyloidosis Electron microscopy Electron paramagnetic resonance N-terminal domain of the Escherichia coli hydrogenase maturation factor HypF NMR Nuclear magnetic resonance sBLM Supported lipid bilayer membrane SEM Scanning electron microscopy SMA Immunoglobulin light chain variable domain SSNMR Solid-state nuclear magnetic resonance STEM Scanning transmission electron microscopy STM Scanning tunneling microscopy TEM Transmission electron microscopy
178
A. Relini et al.
20.1 Introduction Amyloid deposition is involved in severe pathological conditions whose number has been gradually increasing in the last few years owing to the recent identification of amyloid deposits associated with known diseases [1]. Amyloids are insoluble thread-like protein aggregates sharing specific tinctorial and structural features. The name “amyloid” was attributed to these aggregates in the nineteenth century owing to the fact that upon staining they displayed the same behavior as starch (amylon in Greek). At present forty different diseases involve extracellular or intracellular amyloid formation [2], although they are properly named “amyloidoses” only in the case of extracellular amyloid deposition; representative examples are listed in Table 20.1. In some cases large amounts of amyloid aggregates are deposited in a variety of organs, such as kidney, liver, heart or spleen; in other cases, amyloid deposition affects specific tissues or systems, such as the muscular-skeletal system in dialysis-related amyloidosis (DRA) or the brain in Alzheimer’s and Parkinson’s diseases. A different protein or peptide is associated with each disease; however, despite the differences in structure and amino acid sequence, they all give rise to amyloid fibrils with shared structural features and properties. Moreover, since 1998 a number of studies have shown that under appropriate solution conditions even proteins unrelated to amyloid disease can undergo amyloid aggregation, resulting in prefibrillar aggregates with cytotoxic activity and eventually in amyloid fibrils with the same characteristics as those displayed by disease-associated amyloids [3]. These facts have led to the suggestion that the ability to undergo amyloid aggregation may be considered a generic property of proteins resulting from the physicochemical properties of the polypeptide main chain, which is the same for all proteins. The amino acid sequence, and hence the presence of different side chains, only plays a role in determining the details of the aggregation process and the polypeptide arrangement in the amyloid structure.
Table 20.1. Representative diseases associated with extracellular or intracellular amyloid formation. For a complete list, see [1, 2] Disease
Protein or peptide involved
Alzheimer’s disease Spongiform encephalopathies Parkinson’s disease Amyotrophic lateral sclerosis Huntington’s disease Primary systemic amyloidosis Hemodialysis-related amyloidosis Lysozyme systemic amyloidosis Type 2 diabetes Injection-localized amyloidosis Cataract
Amyloid β peptide Prion protein (full length or fragments) α-Synuclein Superoxide dismutase Huntingtin (with polyQ expansion) Immunoglobulin light chains or fragments β2 -Microglobulin Lysozyme mutants Amylin Insulin γ-Crystallins
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
179
Remarkably, amyloid aggregation is also an evolutionary conserved mechanism involved in the formation of structures with specific biological functions [4]. In fact, it is employed by bacteria such as Escherichia coli to build extracellular fibrils (curli) which promote cell adhesion to inert surfaces and mediate binding to host proteins [5]; it is found in mammalian melanocytes, where the melanin pigment is polymerized onto amyloid-like fibrous striations formed in the melanosomes by the glycoprotein fragment Pmel17 [6]; it is also exploited in the formation of aerial hyphae by bacteria such as Streptomyces coelicolor [7] or filamentous fungi such as Neurospora crassa [8]. A natural protective amyloid (chorion) is the major component of silkmoth eggshell [9]. In addition, aggregates of yeast prion proteins allow the transmission of phenotypic traits without the involvement of specific genes. For example, the inactivation of the yeast prion protein Ure2p sequestered in the fibrillar prionic form is associated with the URE3 phenotype enabling cells to thrive on nitrogen-poor sources [10]. Amyloid formation is a consequence of the increased aggregation propensity exhibited by proteins under particular conditions, such as misfolding or fragmentation. In these cases, either owing to the shift of the dynamic equilibrium between the protein conformational states towards more unfolded forms, or to the different folding properties induced by the cleavage of the polypeptide chain, groups that are normally buried within the protein are exposed to the aqueous environment, thus increasing the tendency of protein monomers to aggregate into stabler, polymeric forms (Fig. 20.1); in some cases, natively unfolded proteins in the absence of their binding partners undergo amyloid aggregation [2]. Aggregation proceeds according to a nucleation and growth mechanism, often displaying a lag phase. Protein monomers first assemble into globular oligomeric aggregates, which give rise to more complex prefibrillar structures such as rings or thin filaments (protofilaments), eventually evolving into amyloid fibrils (Fig. 20.2).
Fig.20.1.Tapping mode atomic force microscopy (AFM) image (height data, scan size 2.6 µm, Z range 50 nm) of insulin fibrils obtained in vitro after 13 days of incubation at pH 2 and 57 ◦ C (unpublished results)
180
A. Relini et al.
Fig. 20.2. Aggregation process of a disease-unrelated protein (the N-terminal domain of the Escherichia coli hydrogenase maturation factor HypF, HypF-N) in the presence of 30% trifluoroethanol, pH 5.5. Tapping mode images were acquired in liquid after a a few hours, b 3 days and c 5 days and in air after d 3 days, e 8 days and f 11 days. Scan size a 670 nm, b 5.3 µm, c 1.9 µm, d 4.8 µm, e 3.6 µm and f 3.9 µm. Z range a 20 nm, b 45 nm, c 120 nm, d 40 nm, e 80 nm and f 100 nm. (Modified from [68]. Copyright 2004, with permission from Elsevier)
All amyloid fibrils exhibit similar features when viewed with the electron microscope and the atomic force microscope. Typically, they appear as relatively straight, unbranched strands up to several micrometers in length formed by two to six protofilaments twisted together. Different fibril morphologies are observed, ranging from tightly coiled ropelike structures to flat ribbons. Fibril diameter measured by the electron microscope is about 6–12 nm, while protofilaments are around 2 nm in diameter. When measured with the atomic force microscope in air, these sizes are slightly reduced (as discussed below). Amyloid fibrils are characterized by a typical X-ray diffraction pattern, with reflections that are consistent with the so-called cross-β conformation of the polypeptide chain (Fig. 20.3a). This involves β-strands oriented perpendicular to the fibril axis and forming hydrogen bonds that are parallel to the latter; this arrangement gives rise to extended β-sheets that are essentially continuous along the fibril direction (Fig. 20.3b). The reflections observed in the X-ray diffraction pattern correspond, respectively, to the interstrand and intersheet distances. The high degree of order of the arrangement of the polypeptide chain has been proposed to give rise to a specific, regular binding of the organic dyes thioflavin T and Congo red (CR), traditionally used to identify amyloid fibrils; the ordered dye binding results in fluorescence enhancement for thioflavinT and a red shift from 490 to around 540 nm of the maximum absorbance together with birefringence effects for CR [11, 12].
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
181
Fig. 20.3. a X-ray diffraction pattern measured from partially aligned amyloid β peptide (1–42) ˚ and 10 A ˚ characteristic of the cross-β structure. b Model fibrils, displaying reflections at 4.7 A of an amyloid fibril showing the assembly of the polypeptide chains into a cross-β structure. Two parallel β-sheets form a protofilament, while four protofilaments twist around one another forming the fibril. (a From [36], with permission. Copyright 2005 Wiley-Liss. b Modified from [132]. Copyright 2004, with permission from Elsevier)
20.2 Techniques and Methods Used To Study Amyloid Aggregates From a physical point of view an amyloid is a macroscopic object formed by distinct elements hierarchically ordered according to their magnitude scale. In fact an amyloid is a bunch of strands (fibrils), formed by yarns (proto-filaments) either twisted together or placed side by side in a ordered way. Protofilaments, in turn, are made up of tidily arranged macromolecules enriched in the β structure. To study the different elements of such a hierarchical structure, whose details span several orders of magnitude, from angstroms to millimeters, investigators must use techniques that provide information at different magnitude scales. Accordingly, inspection is achieved by optical microscopy, electron microscopy (EM) and atomic force microscopy (AFM). Mean aggregation properties and molecular configuration are studied by light and neutron scattering, circular dichroism, Fourier-transform infrared spectroscopy, X-ray diffraction, electron diffraction, electron paramagnetic resonance (EPR) and solid-state nuclear magnetic resonance (SSNMR) [13, 14]. According to the historical details contained in the interesting reviews by Sipe and Cohen [15] and Kyle [16] it appears that our understanding of amyloid follows the development of microscopic, spectroscopic and diffractometric techniques.
182
A. Relini et al.
Indubitably, the clinical need to recognize amyloid in relation to symptoms and diseases gave rise to the first studies of amyloid morphology. Today, optical fluorescence and polarized microscopy and immunoelectron microscopy are the techniques more commonly used for this purpose. The first takes advantage of fluorescent dyes that selectively bind amyloid, the second takes advantage of antibodies labeled with gold particles. Subsequently, the need to study the amyloid molecular structure to understand the causes of its formation promoted the use of diffractometric (X-ray, electron diffraction, neutron diffraction), spectroscopic (SSNMR and EPR) and ultramicroscopic (AFM and scanning transmission electron microscopy, STEM) techniques. 20.2.1 Optical Methods The connection between the presence in tissue of amyloid deposits in different pathological states was discovered by the work of great pathologists of the nineteenth century who could rely only on their eyes and the optical microscope. The term “amyloid,” used in 1838 by the German botanist Matthias Schleiden for a normal amylaceous constituent of plants [16], was then employed by the German physician Rudolph Wirchow to describe abnormal tissues in the brain. Wirchow, in 1854, described “cerebral corpora amylacea” of abnormal macroscopic appearance that were stained blue by iodine. The subsequent addition of sulfuric acid turned them violet. Since these were the reactions used to characterize cellulose, Wirchow concluded that cellulose was the substance underlying the abnormal tissues. At those times the difference between starch and cellulose was not quite clear: cellulose, starch and dextrin were considered isomers of the same carbohydrate, while amyloid was the substance obtained by treating cellulose with sulfuric acid [17]. Only a few years later (1859) Friedreich and Kekulé [18] found that amyloid had a high content of nitrogen and it was made up of proteins rather than of carbohydrate. On the basis of optical microscopy observation, in the first years of the twentieth century scientists believed that amyloids were amorphous. In 1927 Divry and Florkin [19], by using polarized optical microscopy, found that amyloid is birefringent and therefore it is an ordered material with anisotropic optical properties. The intrinsic birefringence of amyloid is increased by CR. Amyloid aggregates stained with CR appear orange in white polarized light and apple-green if placed between crossed polarizers (Fig. 20.4). This effect, which is called apple-green birefringence, is really caused by the coexistence of light absorption anisotropy (circular dichroism), refractive index anisotropy (linear birefringence) and light dispersion. After Divry and Florkin, congophilia and green birefringence became the criteria adopted to recognize amyloid. Presently, it remains the principal diagnostic criterion even if the chemical and physical bases of CR–amyloid interaction are not well understood and the criterion is not quantitative. CR is the sodium salt of benzidinediazobis(1-naphthylamine-4sulfonic acid). It is a secondary diazo compound soluble in organic solvents such as ethanol, and yields a red colloidal suspension in water. Its UV–vis absorption spectrum shows a characteristic, intense peak around 498 nm in aqueous solution, at low concentration. The aggregation of the dye tends to redshift the absorption
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
183
Fig. 20.4. Congo red stained amyloid plaques associated with A Alzheimer’s disease, B Gerstmann–Straüssler–Scheinker disease and C Down’s syndrome and observed I in linearly polarized white light and II between crossed polarizers showing apple-green birefringence. (From [29], with permission. Copyright 2003 National Academy of Sciences, USA)
spectrum, whereas binding to cellulose fibres or amyloid fibrils produces a blueshift. CR is also fluorescent and its fluorescence intensity changes when it is bound to amyloid. Its interaction with amyloid is not specific and it binds also other materials which have a fibrous structure. Hydrogen bonding [20–24], hydrophobic interactions [21, 25], van der Waals forces [22, 25] and ionic forces [26–28] have been proposed to be the major interactions between CR and amyloid. The use of a polarizing microscope, which can build optical images associated with transmission, birefringence, dichroism and optical orientation, allowed Jin et al. [29] to state that CR molecules are aligned along the amyloid fibers with their transition moment parallel to the fiber longitudinal axis. These authors, on the basis of this finding, were able to demonstrate the orientation of amyloid fibrils in plaques and blood vessels. While in plaques fibrils have a radial orientation with a disordered center, in blood vessels fibrils roughly follow the contour of the blood vessel. 20.2.2 Electron Microscopy Although birefringence suggested the presence of some ordered structure in amyloid, only EM studies revealed its fibrillar structure. Using EM, Cohen and Calkins [29] observed straight, rigid fibrils in sections of fixed tissues containing amyloid bundles. The samples were prepared by negatively staining them and displayed fibril widths ˚ range, and lengths in the 1000–16,000-A ˚ range. Since further in the 60–130-A
184
A. Relini et al.
studies showed that the fibrillar structure was common to all the forms of amyloid investigated, the fibrillar morphology was adopted as the second criterion to identify amyloid. Nowadays, scanning electron microscopy (SEM) and transmission electron microscopy (TEM) are used to image amyloid using both shadow-casting and negativestaining. Fibril staining is performed using uranyl acetate, lead citrate, platinum– palladium and phosphotungstic acid. These different preparations produce similar results for different types of amyloids. In EM images, fibrils from different proteins, either isolated by gentle physical methods or grown in vitro, show similar ˚ wide, morphologies, compatible with a common ultrastructure, fibrils 75–100 A ˚ consisting of filamentous subunits (protofilaments) having a diameter of 25–35 A. Negative-staining produces quite flat images of fibrils characterized by two dark edges, suggesting a possible fibril structure of a ribbon formed by protofilaments running side by side. The understanding of the amyloid structure by EM images took advantage of various methods of image analysis and sample preparation. Serpell et al. [31], using superposition and averaging of images of fibril cross section, reconstructed the substructure of fibrils of different proteins, such as transthyretin, lysozyme (Fig. 20.5), monoclonal immunoglobulin λ light chain, apolipoprotein AI and amyloid A protein. In cryo-EM, the biological sample is inspected at liquid nitrogen temperature. Very fast sample freezing in ethane slush produces vitreous or noncrystalline ice to prevent sample dehydration. In these conditions, sample dimensions and morphology should be better preserved without distortions due to stains or fixatives. Jimenez et al. ˚ resolution, were able to determine [32], using cryo-EM and image processing at 25-A
Fig. 20.5. Electron microscopy analysis of resin-embedded and thin-sectioned ex vivo lysozyme amyloid fibrils. a Cross sections of embedded fibrils; b pattern obtained from the averaging of cross-sectional views of thin-sectioned fibrils; c contour map showing the possible positions of the subunits or protofilaments. (From [36], with permission. Copyright 2005 Wiley-Liss)
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
185
the 3D structure and protofilament packing of amyloid fibrils formed by the SH3 domain of phosphatidylinositol-3′ -kinase, ex vivo lysozyme fibrils [33] and insulin fibrils [34]. Serpell et al. [35, 36] resolved the β-sheet structure as striations across the fibrils in fibrils formed by the Amyloid β (Aβ) (11–25) peptide. The structure of the URE3 prion has been studied by Baxa et al. [37, 38], while the oligomeric intermediates of β-lactoglobulin and the Aβ (13–21) peptide have been studied by Carrotta et al. [39] and Dong et al. [40], respectively. Alternatively, the native condition of the sample in EM can be preserved in “environmental SEM”, sometimes referred to as “low-vacuum SEM”. The low-vacuum environment in the sample chamber (up to 20 Torr) allows users to investigate samples in their natural state (wet, hydrated, uncoated, etc.). Using this technique, Donald and coworkers have studied spherulites formed by fibrils of bovine insulin [41,42] and β-lactoglobulin [43] under conditions in which the solvent has not been dried off. Structural information has also been obtained by STEM. By using this technique, Goldsbury et al. [44] found out that the Aβ (1–40) protofilaments have a mass per unit length of about 19 kDa/nm, and coupling TEM and STEM data, Antzutkin et al. [45] found a similar mass per unit length of about 19 kDa/nm for protofilaments of Aβ (1–42) peptide. 20.2.3 X-ray Diffraction The molecular architecture of amyloid fibrils was first determined by X-ray diffraction at the end of the 1960s [46,47]. For diffraction, fibrils are usually suspended in an aqueous solution and either mechanically or magnetically oriented. The diffraction patterns of amyloid fibrils are very similar for fibrils formed in vitro and for those extracted ex vivo and are independent of the fibril-forming protein [48, 49]. They contain two predominant reflections (Fig. 20.3a): a strong, sharp reflection at about ˚ on the meridian that corresponds to the direction of the fibril axis and a diffuse 4.7 A ˚ on the equator. The first arises from the hydrogen-bonding disreflection at 10–13 A tance between β-strands in a β-sheet, while the second corresponds to the intersheet distance, which can slightly vary with the side chain content of the peptide. The diffraction pattern indicates that the β-strands run perpendicular to the fibril axis. X-ray diffraction is unable to provide more detailed information such as side chain conformation, which is available only from single-crystal X-ray diffraction. More recently, Makin et al. [50] and Nelson et al. [51] were able to produce nanocrystals and microcrystals of peptide fragments that have characteristics of amyloid fibrils and from X-ray data could accurately describe β-strand and side chain structure and interactions. To date, X-ray fibril diffraction has not revealed low-angle reflections that can be used to calculate the size and the number of protofilaments. 20.2.4 Nuclear Magnetic Resonance The other technique that provides information on the molecular structure of amyloid fibrils is nuclear magnetic resonance (NMR). Solution NMR is unable to provide the whole detailed molecular structure of amyloid fibrils since these are too large, but it
186
A. Relini et al.
is advantageously used to study solvent accessibility [52, 53] and the early fibrillogenesis process [54]. More details on amyloid molecular structure are provided by SSNMR, which can be used to determine interatomic distances (both intramolecular and intermolecular), place constraints on backbone and side-chain torsion angles, and identify tertiary and quaternary contacts [55, 56]. 20.2.5 Atomic Force Microscopy The atomic force microscope is now one of the most useful instruments to monitor amyloid fibrils and their precursors (Fig. 20.6) [57]. It is used not only for imaging, but also for manipulating nano-sized samples and measuring molecular interaction forces. After the first papers in the second half of the 1990s [58,59], the use of AFM to study amyloid increased rapidly. The atomic force microscope has a resolution comparable to that of the electron microscope, but yields images on the basis of a very different principle. Therefore, the images produced with the two techniques show different details and different features that together can provide interesting structural information. The capability of AFM to produce a 3D image of the sample surface is particularly useful for studying fibril morphology. The shape of twisted fibrils, observed either by EM or AFM, can yield information about the fibril ultrastructure. Perutz et al. [60] observed that the fibril shape contour depends on the number of protofilaments comprising the fibril, as the way protofilaments overlap depends on their number. Two filaments overlap symmetrically along the center of the fibril, three filaments overlap alternately on either side, four filaments overlap alternately on both sides and in the center, and so on (Fig. 20.7); this results in distinguishable shape contours which can be used to identify the number of constituent protofilaments [61]. Other important advantages of AFM are its capability to image samples in different environments and the fact that the sample does not require particular preparation,
Fig. 20.6. Tapping mode AFM images of amyloid fibrils. a Lysozyme fibrils (amplitude data) obtained in vitro at 57 ◦ C and pH 2 (unpublished results). b, c HypF amyloid fibrils (height data) obtained in vitro at 25 ◦ C and pH 5.5 in the presence of 30% trifluoroethanol. c is an enlargement of the region corresponding to the black square in b. The arrow indicates a thin filament proposed to represent a protofilament. Scan size a 3.2 µm, b 2.8 µm and c 0.58 µm. Z range b, c 40 nm. (b, c Modified from [68]. Copyright 2004, with permission from Elsevier)
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
187
Fig. 20.7. Fibril models made of plastic tubes. From left to right, the number of tubes employed to build the fibril increases from two to six, giving rise to different shape contours of the resulting fibrils. (Reprinted from [61]. Copyright 2004, with permission from Elsevier)
such as staining or fixation procedures. Samples are usually prepared by depositing a drop of the protein solution, either in the monomeric or in the aggregated form, depending on the purpose of the study, on a flat substrate. Samples can then be imaged under a liquid in the hydrated state, or dried and imaged in air. In the presence of salts, washing the sample before drying may be convenient to avoid salt crystallization; however, washing can also remove protein aggregates loosely bound to the substrate. Therefore, it can be useful to compare the results obtained with different deposition protocols. The most commonly used substrates are mica and graphite because of their very low surface roughness and ease of preparation [62–66]. In some cases substrates are covered by a self-assembled organic layer in order to modify the substrate surface properties and therefore control protein adhesion to the surface. In some cases mica has been alkylamine-functionalized or covered by polylysine to modify the surface charge density and therefore tune the electrostatic forces at the interface [63, 67]. In other cases, the hydrophilic mica surface has been changed into a hydrophobic one through functionalization with octadecyltrichlorosilane [63]. Since usually fibrillization is a slow process with respect to the time required to acquire AFM images, it is possible to follow fibril formation by taking images at different times [62, 68]. Since intermolecular specific forces can be measured by AFM (force spectroscopy) the interaction forces between protein molecules undergoing fibrillogenesis have been measured [69] as well as the force necessary to unzip the β-sheets [70]. Moreover, AFM can measure length, width and height of fibrils. The reliability of these measurements depends on the resolution of the atomic force microscope as well as on avoiding systematic errors related to characteristics of the instrument. The atomic force microscope works as a profilometer: a sharp tip connected to a cantilever is scanned in a raster pattern on the sample surface. Usually the radius of curvature of the tip is on the order of 10 nm, but tips with a radius of curvature as small as 2.0 nm are available. The local tip–sample interaction deflects the cantilever. Cantilever deflections are recorded and reported as a function of the in-plane position (x–y coordinates) of the tip. This results in a reconstruction of the topography of the sample surface. The vertical (z) resolution is different from the horizontal (x–y) resolution. The first is limited by the resolution of the system that records the
188
A. Relini et al.
cantilever deflection and can be a fraction of angstrom, while the resolution in the horizontal plane is determined by various factors. In the case of a sample with an atomically flat surface only the very end of the atomic force microscope tip makes contact, and the tip end can be atomically sharp even for a macroscopically blunt tip. Under this condition, AFM is capable of reaching atomic resolution. When the surface topography of the sample presents high (on the order of nanometers), steep features and the sample is hard and nondeformable the horizontal resolution depends on both the tip radius of curvature and the sample surface topography [71,72]. In the case of soft deformable samples, the resolution depends on the size of the tip–sample interaction area. The other factor that affects the reliability of the size measurements by AFM is the presence of systematic errors. In height measurements, the tip–sample interaction plays an important role. The presence of repulsive forces, for instance, of electrostatic origin, increases the measured values since the tip does not touch the sample [73]. Sample deformation, on the other hand, decreases the measured value of the height. In particular, when the sample is dried to facilitate its adhesion to the mica substrate, this procedure may result in a reduction in the size of the imaged objects owing to dehydration and/or deformation. Previous control measurements yielded a correction factor, in the case of samples dried under mild vacuum, of about 2.5 for monomeric globular proteins [68]; in the case of amyloid fibrils, however, this factor may be larger (about 5) [61]. In the horizontal plane the measured object size is affected by the finite size of the tip. The AFM image is distorted, since it is the convolution of the sample shape with a transfer function depending on the tip shape. This effect increases the measured value of the fibril width. Keller [74] showed that the Legendre transform of the true surface is related to the Legendre transforms of the distorted image surface and the tip surface. In principle, knowing the tip surface function, one can calculate the transform of the true surface and obtain the true surface by inverse Legendre transform. Practically this is a cumbersome operation and is rarely done. More often, corrections of the measured widths are done by using simple geometrical models of tips and objects [61]. Tip effects should be properly taken into account to avoid misinterpretation of AFM images. For instance, since the tip size can decrease the AFM lateral resolution, fibrils (or protofilaments) closer than 20–30 nm (which is the typical size of the tip radius of curvature) do not appear separated in the AFM image and they are sometimes interpreted as a single fibril formed by two protofilaments placed side by side. Moreover, since AFM is a local technique, it gives information about local features which may not be representative of the general structure of the sample. To avoid this drawback, images should be collected from different regions of samples prepared in the same conditions.
20.3 Monitoring the Aggregation Process by AFM The AFM technique has been extensively used to monitor the different steps involved in the aggregation process and the detailed features of fibrils and their intermediates. This is a key issue whose final goal is to develop appropriate strategies, based on structural information, to prevent or counteract amyloid formation. When studying
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
189
the time-course of protein aggregation, the sample must be imaged at different times during the process. Such an approach, based on time-lapse imaging, has been named by some authors “time-lapse AFM” [62, 75–77]. By carrying out the aggregation experiments directly in the AFM fluid cell, it is possible to monitor the growth of individual protofibrils and fibrils by imaging the same sample area at different times. This procedure has been applied, in particular, to follow the spontaneous assembly of amylin fibrils in vitro [62], obtaining information on fibril morphology and growth rate. In addition, a peculiar feature of this approach is the capability of collecting data on growth direction, as the same structures are probed as a function of time. For amylin, fibril growth was bidirectional, occurring by elongation at both fibril ends. Another approach consists of withdrawing aliquots of the sample solution at different times during the aggregation reaction. The deposition on the mica substrate is performed just before imaging; this method allows possible effects of the substrate on the growth process to be avoided or at least minimized. This procedure was first employed by Harper et al. [78] to study fibril formation by Aβ protein in the presence of seeds and by Ionescu-Zanetti [79] to monitor the assembly of immunoglobulin light chain. We used the same approach to study the aggregation process of the N-terminal domain of the hydrogenase maturation factor HypF (HypF-N), a protein not related to any disease but capable of undergoing amyloid aggregation in the presence of trifluoroethanol [60,80]. HypF-N aggregation takes place according to a hierarchical path: early globular structures assemble into crescents which subsequently close into large rings; the latter organize into ribbonlike structures, eventually assembling into differently supercoiled mature fibrils (Fig. 20.2). Inspection by scanning tunneling microscopy (STM) shows that the globules display an asymmetric shape and tend to assume a rather well-defined reciprocal orientation, suggesting the presence of specific directional interactions [68]. Disaggregation experiments performed with acylphosphatase [81] suggest that amyloid formation occurs in discrete steps whose reversibility is increasingly difficult as the aggregate size increases. AFM data, often used in combination with EM measurements [82,83], were employed to build models of fibril assembly for different proteins [76,79,83,84]. Fibril formation is proposed to result from the hierarchical assembly of protofilaments; however, different mechanisms of assembly, such as protofilament intertwining or lateral association, can act in parallel, giving rise to fibrils of different sizes, therefore explaining the polymorphism often exhibited by mature fibrils [76,83,85–87]. From the comparative analysis of height and phase-shift images acquired in tapping mode, Jansen et al. [76] could unveil interesting structural details of insulin fibrils. They observed the coexistence of twisted ribbonlike structures, parallel tubular fibrils and segmented fibrils, likely due to different protofilament association mechanisms. Images obtained by zooming in on a bundle of fibrils show fibril structural details which are interpreted as the entrance of a fibrillar canal and would support the model of amyloids as water-filled nanotubes [60]. High-resolution AFM data obtained on glucagon fibrils showed that, within a single supercoiled mature fibril, individual subunits can associate with in-phase alignment, out-of-phase alignment, or twisting [88]. The aggregation process can be strongly affected by solution conditions. It has been shown that increasing concentrations of ethanol induce different morphologies of insulin aggregates, which at fixed aggregation time undergo structural modifica-
190
A. Relini et al.
Fig. 20.8. AFM images (height data) of insulin aggregates obtained after 24 h of incubation at pH 1.9 and 60 ◦ C in the presence of ethanol: a 20 wt %; b 70 wt %. (Modified from [89]. Copyright 2005 American Chemical Society)
tions from long, thin fibrils to annular doughnut structures [89] (Fig. 20.8); changing the cosolvent from ethanol to glycerol or trifluoroethanol, which results in different stability of the native state, gives rise to different aggregate morphologies [90]. In general, pH and ionic strength are very important factors modulating the features of the aggregation process; this indicates that electrostatic interactions play an important role in determining fibril growth and morphology. Stine et al. [91] investigating the conditions which influence the assembly of Aβ, defined solution conditions for the formation of either oligomeric or fibrillar assemblies. Gosal et al. [92] demonstrated that depending on protein concentration, pH and ionic strength two different assembly pathways are involved in β2 -microglobulin aggregation. One pathway leads to the formation of straight amyloid fibrils, while the other one produces semiflexible fibrils which share an antibody epitope with other toxic oligomeric species and are unable to convert into straight amyloid fibrils. Amyloid aggregation has also been monitored in the presence of factors which may accelerate or slow down the process. Specific antibodies [93] have been shown to inhibit aggregation, while charged polymers, such as glycosaminoglycans, enhance amyloid formation [94, 95]. Much information has been gained up to now, by AFM and other techniques, on the structure of fibrils, which are the stable end products of the aggregation process; however, attention has been gradually shifted from mature fibrils to their precursors [96], and is now mainly focussed on the structural features of protofibrils and early globular aggregates [97–99], which have been demonstrated to be directly involved in cytotoxic effects [100–102].
20.4 Effect of Surfaces on the Aggregation Process Protein unfolding and aggregation are likely to be influenced by surfaces and several investigations have been carried out to unveil this aspect. Two main reasons justify this research. The first is that most of the morphological inspections are carried out on
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
191
fibrils deposited on surfaces. Therefore, it is quite important to understand whether the interaction with surfaces introduces artifacts in the system under analysis. The second is that understanding the physicochemical interaction of the proteins with surfaces helps in understanding the factors that affect misfolding and fibrillization. This issue is relevant from the biomedical point of view since biological surfaces such as collagen and cell membranes seem to induce amyloid formation. Being a surface technique, AFM is well suited to study the influence of surfaces on the aggregation process and can be advantageously coupled to spectroscopic techniques to correlate the morphology of the aggregates formed under different conditions with their spectroscopic features. Several AFM studies have investigated the influence of surfaces and their physicochemical properties on the fibrillization process. One of the first studies on the subject was by Sluzky et al. [103], who investigated the kinetics of aggregation of insulin in the presence of hydrophobic Teflon surfaces by UV absorption spectroscopy and quasi-elastic light scattering. They found an increase in the rate of aggregation by increasing the effective Teflon surface area, suggesting the importance of interactions with hydrophobic surfaces for the aggregation process. More recently, Giacomelli and Norde [104] investigated the influence of Teflon surfaces on the structure of the Aβ (1–40) peptide by dynamic light scattering and circular dichroism at different pH values. Upon adsorption on the Teflon surface they found a conversion of the random structure (observed at pH 10) and of the β-sheets (observed at pH 7) into α-helical structures. However, by increasing the peptide surface density, lateral interactions between the adsorbed peptide molecules induced the formation of β-sheet structures. Goldsbury et al. [62] studied amylin fibrillization by AFM measurements in liquid. They compared the morphology of fibrils grown in solution and subsequently deposited on a mica surface with the morphology of fibrils directly assembled on mica. Aggregation in solution produces polymorphic fibrils with the majority of the fibrils being characterized by a (6.8 ± 0.5)-nm height and a 25-nm twisting periodicity, in agreement with EM results. On the other hand, assembly on mica results in thinner fibrils that do not exhibit any twisting periodicity. The predominance, in the latter case, of protofibrils suggests that the formation of mature fibrils is hindered owing to constraints imposed by fibril immobilization on the surface during assembly. The effect of a mica surface on aggregation was also addressed by Zhu et al. [63], who investigated the fibrillogenesis of a recombinant amyloidogenic light chain variable domain, SMA. The authors analyzed by ex situ AFM SMA aggregates formed on mica sheets that had been incubated in the SMA solution. They found that the aggregation kinetics on the surface were significantly accelerated since aggregates were observed in faster times and at lower protein concentrations compared with aggregation in solution. The increase of local monomer concentration and the partial protein unfolding resulting from adsorption on the surface could account for the observed increase in the aggregation rate. As also observed by Goldsbury et al. [62] in the case of amylin, surface binding slows down protofibrillar assembly into higher-order and twisted fibrils. The effect of surface charge in favoring fibrillization in biological environments has been recently demonstrated for β2 -microglobulin, which is involved in dialysis-related amyloidosis (DRA). AFM measurements showed that under physiopathological conditions (37 ◦ C and pH 6.4, as in the presence of inflammation),
192
A. Relini et al.
β2 -microglobulin fibrillogenesis takes place in the presence of fibrillar collagen, while no fibrils are obtained without collagen. Fibrillization also occurs if collagen is replaced with a positively charged poly(l-lysine)-coated mica surface; this result strongly supports the hypothesis that the positively charged regions along the collagen fibers may act as catalysts for fibril formation, and suggests an explanation for the strict specificity of DRA for the tissues of the skeletal system (Fig. 20.9) [67]. A number of studies have been reported that investigated the aggregation of different Aβ peptides in the presence of various surfaces. Kowalewsky and Holtzman [64] followed in situ by tapping mode AFM the aggregation of the Aβ (1–42) peptide on mica, i. e., a hydrophilic surface, and graphite, i. e., a hydrophobic surface. On mica, Aβ formed globular aggregates, about 5.0 nm in height, highly mobile on the surface with a tendency to coalesce and assemble into linear structures resembling the protofibrillar aggregates observed by ex situ AFM imaging of solution aggregated-Aβ peptide. While the affinity of the peptide for mica is likely driven by the charged residues of the extracellular domain, the interaction of the transmembrane domain with the hydrophilic mica surface is unfavored. The formation of droplike pseudomicellar species could therefore arise from the balance among the different contributions to peptide–surface interaction. According to this interpretation, anionic hydrophilic surfaces, such as cell membranes in tumoral cells [105], could induce the formation of globular oligomeric aggregates. Other anionic surfaces (including sodium dodecyl sulfate micelles) and glycosaminoglycans such as heparin have been shown to induce fibrillization of various peptides and proteins [106]. A completely different scenario was observed on hydrophobic graphite surfaces. In this case, after a few minutes of exposure to the Aβ solution, the graphite became covered with elongated parallel structures running along three directions at 120◦ angles to each other. These structures were interpreted in terms of β-sheets with extended peptide chains perpendicular to the long axis of the aggregates stabilized by hydrophobic interactions between the
Fig. 20.9. Surface plot of a tapping mode AFM image (height data) of β2 -microglobulin ex vivo amyloid material obtained from the femoral head of a patient affected by dialysis related amyloidosis and showing amyloid fibrils crowding around a collagen fiber. Modified from [67]
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
193
basal plane of graphite and the hydrophobic residues of the polypeptide chain. Such interactions could be important from the physiological point of view since they could account for the association between the Aβ peptide and lipoprotein particles occurring in vivo. Sheets formed by parallel narrow lines oriented along threefold symmetry directions have also been observed in another AFM/STM investigation of Aβ (1–42) aggregation on graphite [65]. A different aggregation pathway was reported by Losic et al. [66], who investigated by ex situ AFM the aggregation of Aβ (1–40) on graphite. The authors observed the deposition of globular aggregates preferentially aligned along surface step edges on graphite surfaces that had been incubated in Aβ (1–40) peptide solution for times ranging from 2.0 min to 5.0 days. These aggregates, initially separated, slowly associated to form linear fibrils running along the steps. Similar results were obtained in control experiments performed on Aβ (1–42). In contrast, preaggregated Aβ (1–40) oligomers were found to preferentially adsorb on the graphite basal plane. Ha and Park explored the use of N-hydroxysuccinimide ester activated solid surfaces to immobilize amyloid seeds of Aβ (1–42) [107] and insulin [108]. From ex situ AFM analysis of fibrils grown on surfaces exposing different kinds of seeds, preaggregated oligomers were found to be the most efficient, among fresh monomers, preaggregated oligomers and fibrils, in accelerating amyloid growth. These observations suggest that the generation of Aβ (1–42) oligomers and their deposition might be an important early step in inducing further growth of neuritic plaques. The influence of lipid surfaces on aggregation is another important issue to address owing to its physiological relevance. Sharp et al. [109] carried out a comparative study of insulin denaturation and fibrillization in the presence of cationic and anionic lipid surfaces and in bulk solution. By coupling Fourier transform infrared attenuated total reflection and AFM, the authors showed that charged lipid surfaces increase the protein unfolding kinetics but reduce the content of β-sheet structures within the adsorbed protein, thus resulting in smaller fibrils and larger amorphous aggregates.
20.5 Interaction with Model Membranes The mechanisms by which amyloid peptides or oligomeric aggregates of amyloid proteins cause cytotoxicity and the appearance of clinical symptoms of diseases are not completely understood. There is, however, a general agreement on the hypothesis that protein/peptide aggregation causes membrane perturbation and disruption of the ionic balance across the membrane. In particular, the influx of calcium ions into the cells seems to be among the earliest biochemical modifications underlying cytotoxicity and tissue damage [3]. Membrane permeation could be the result of the insertion in the cell membrane of porelike structures, formed by the assembly of oligomeric peptides or proteins. An alternative mechanism could be the loss of structural organization of the lipid bilayer spreading over the entire membrane surface, causing lesions that alter the barrier properties of the membrane. To unravel the molecular mechanisms by which early oligomers interact with cells, several studies have been performed on model
194
A. Relini et al.
membrane systems (monolayers, supported bilayers, liposomes) [110]. In particular, AFM imaging of supported lipid bilayer membranes (sBLM) in the presence of amyloid peptides or proteins is one of the most powerful approaches to tackle this problem. A supported lipid bilayer can be produced by depositing a drop of vesicle suspension on freshly cleaved mica. Owing to the interaction with the support, vesicles spontaneously coalesce forming a bilayer [111]. Alternatively, a sBLM can be produced by depositing on a mica surface a bilayer film using the Langmuir– Blodgett technique [112]. Several AFM studies on amyloid protein/peptide interaction with sBLM have recently appeared [113–117]. In a recent paper, Quist et al. [116] reported that amyloid peptides and proteins are able to form porelike structures, when they are interacting with sBLM. In this study not only AFM but also other complementary techniques such as circular dichroism, spectrometry, gel electrophoresis and electrical recordings were employed. The investigation was extended to different amyloidogenic peptides, including Aβ (1–40), α-synuclein, amylin, A Bri, A Dan and serum amyloid A. AFM images of these molecules inserted in sBLM (Fig. 20.10) showed a heterogeneous population of multimeric peptide complexes with an outer diameter in the 8.0–12-nm range often displaying a central channel-like concavity (1.0–2.0 nm in diameter). From these images, complemented by the information obtained with the above mentioned techniques, the authors concluded that membrane permeation was induced by ion flow through the channels formed by the peptide.
Fig. 20.10. AFM images of porelike structures formed by a amyloid β, b amylin and c serum amyloid A. Image sizes are a, b 25 nm and c 20 nm. (Modified from [116]. Copyright 2005 National Academy of Sciences, USA)
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
195
The experiments performed by Green et al. [112] support a completely different viewpoint. These authors used time-lapse AFM to monitor sample changes subsequent to protein addition and their possible relation to defects originally present in the sample. Lipid bilayers, deposited on a mica surface using the Langmuir– Blodgett technique, were imaged in the presence of amylin and Aβ peptide. No porelike structures were seen by AFM inspection; instead, halo regions appeared around the original lesions of the bilayer and expanded with time. Therefore, the authors concluded that lipid depletion via micellization, rather than pore formation, is the main cause of permeability disruption. When lipid loss occurs and micelles leaving the bilayer are formed, transient defects are produced that permeabilize the membrane. Therefore, these authors critically discussed the hypothesis of channel formation by amyloid peptides. They propose that either the annular protofibrillar Aβ peptide or α-synuclein just adsorb on the lipid surface without spanning the bilayer, or that amyloid aggregates adsorb around lipid bilayer defects, giving rise to porelike structures. To date no common consensus has been reached; such disagreement is a hallmark of the complexity of the problem. In our own laboratory we analyzed the interaction of the model protein HypF-N and its prefibrillar aggregates with phospholipid monolayers at the air–water interface, vesicles and sBLM [68, 117]. We showed that HypF-N prefibrillar aggregates interact with lipid monolayers and permeabilize liposomes. Owing to these results, we aimed to image possible changes that might occur specifically at the surface
Fig. 20.11. Tapping mode AFM images (height data) of supported membranes made of a phosphatidylserine (PS) and b PS after incubation with HypF aggregated for 48 h in the presence of 30% trifluoroethanol. Sections performed along the lines indicated in the images are also reported
196
A. Relini et al.
of supported lipid bilayers upon protein interaction. Figure 20.11a shows the typical morphology and the section analysis of a phosphatidylserine sBLM, while Fig 11b shows the morphology and the section analysis of the sBLM upon interaction with HypF-N aggregates. The surface morphology changes in the presence of the aggregates, with respect to that imaged in their absence. Moreover, the section analysis clearly shows the appearance of steps whose height is well below the thickness of the bilayer itself. Figure 20.12 displays the quantitative analysis of the membrane surface measured both in the absence of the protein and when the lipid surface is exposed to HypF-N incubated at different aggregation times. Interestingly, no change in step height is observed when the bilayer is exposed to mature fibrils. It is likely that in the presence of early oligomers a thinner phase is formed and the reduced step height corresponds to the difference in height of two phases: a pure lipid phase and a phase hosting the protein. Figure 20.13 shows two possible models of membrane thinning. This patchwork structure, involving possible lipid loss, might be responsible for nonspecific membrane leakage. Lipid loss following insertion of amylin monomers/oligomers in the bilayer has been previously reported [118].
Fig. 20.12. Histogram of PS membrane step height measured from surface profiles of AFM height images without HypF a and after incubation with HypF aggregated for b 48 h, c 72 h and d 13 days
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
197
Fig. 20.13. Models of membrane thinning resulting from the interaction with the protein
In conclusion, this body of results demonstrates the importance of AFM imaging to understand the action of amyloid proteins on membranes and, at the same time, reveals that many issues are still to be solved.
20.6 Physical Properties of Fibrils Obtained by AFM A quantitative description of the fibril mechanical properties can be obtained from a statistical analysis of AFM images. During deposition on a surface, the conformation of a long filament (either a polymer such as DNA or an amyloid fibril) is affected by the transition from the 3D structure in solution to the 2D structure on the surface. Depending on the filament–surface interaction, one can envisage two different scenarios: either filaments can freely equilibrate on the surface before being captured in a particular configuration or filaments are trapped on the surface and adhere to it without equilibrating. In the former case, the filaments deposited on the surface represent an ensemble of the lowest-energy conformations in two dimensions and the analysis of 2D images can provide meaningful information on the filament structure. From polymer theory, in the case of conformational equilibration in two dimensions, the mean square of the end-to-end distance of the filament R2 depends on the filament contour length L according to the relation [98]
L 2P − 2P 2 1− e , (20.1) R 2D = 4PL 1 − L
where P is the fibril persistence length, i. e., the decay length over which the fibril keeps memory of its original orientation. On the other hand, when molecules adhere to the surface without equilibrating, their conformation resembles a projection on the surface of their previous 3D structure in solution. Of course, since the filament contour length has to be conserved during adsorption, the conformation of the molecule on the surface will somehow deviate from a true mathematical projection. In this case the mean square of the end-to-end distance of the filament R2 depends on the filament contour length L according to the relation [119]
2 L P 4 1 − e− P . (20.2) R proj = PL 1 − 3 L
Therefore, by plotting R2 as a function of the contour length L it is possible to understand whether molecules have been trapped to the surface or have equilibrated on
198
A. Relini et al.
it. From the fit of the above reported equations to the experimental data, it is possible to evaluate the fibril persistence length, which gives a measure of the fibril flexibility. This approach has been exploited to investigate the mechanical properties of different kinds of “biological filaments.” First applied to the study of DNA molecules (Fig. 20.14) [119–122], such analysis has subsequently been applied to vimentin intermediate filaments [123] and to amyloid fibrils [124]. By this approach, Hatters et al. [124] analyzed 2D images (in that case obtained by EM) of apolipoprotein C-II fibrils. From the statistical image analysis the authors estimated a fibril persistence length of 36 nm using the 2D equilibration model. Since the persistence length is related to the Young’s modulus Y and to the area moment of inertia I through the relation YI = kB TP [125], from knowledge of P it is possible to evaluate the fibril Young’s modulus and therefore obtain information on the fibril mechanical properties. Thanks to the good vertical resolution of AFM, further information on the fibril structure can be obtained through a careful analysis of AFM height maps. In fibrils exhibiting a twisted structure, the height along the fibril backbone displays maxima and minima separated by a fixed twist angle. Knowles et al. [126] performed an analysis of the torsional correlation along the fibril axis on fibrils obtained from three different proteins, insulin, an 84-residue SH3 domain of PI3 kinase and an 11residue fragment (105–115) from transthyretin. They found an exponential decay of the torsional correlation along the fibrils with correlation lengths of (3.7±0.1) µm for insulin, (6.3±0.15) µm for SH3 and (12.6±0.9) µm for transthyretin. These distances over which the fibrils maintain their torsional correlation are remarkably large for structures a few nanometers in diameter and indicate a high level of molecular organization within the fibril core despite the presence of portions of unstructured protein outside the core [127]. The analysis by Knowles et al. suggests that two mechanisms can contribute to the loss of angular order along the fibril: thermal fluctuations and structural defects. According to their model the torsional correlation
Fig. 20.14. Mean-square end-toend distance R2 , as a function of the DNA contour length L. The symbols represent the R2 values of DNA fragments in different buffer conditions. The line represents R2 evaluated from (20.1) with a DNA persistence length of 53 nm. The inset is an expansion of the first part of the curve. (Reproduced from [119]. Copyright 1996, with permission from Elsevier)
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
199
length is given by: ξ=
Φ02 ζ
2 , + 1/(βCT )
(20.3)
where Φ0 is the angular change associated with each defect, ζ is the 1D defect density, β = 1/(kB T), the inverse temperature, and CT is the torsional rigidity. Indeed, the torsional correlation length, calculated under the assumption of zero defect density and by using an independent estimate of the torsional rigidity obtained by measuring the shear modulus from force spectroscopy [128], is in good agreement with the experimentally determined correlation length, suggesting that thermal fluctuations are the major contribution to the loss of angular order. The authors also examined the interactions between single protofilaments responsible for stabilizing the helical fibril structure. Protofilaments, which are characterized by a finite bending rigidity CB , have to curve in order to organize themselves into a fibril. The associated
l 2 elastic energy cost can be calculated as ΔE el = CB k2 (s) ds, where k is the filament 0
r curvature k = r 2 +(h/π) 2 , where 2h is the helix pitch and r is the helix radius. In order to obtain stable fibrils, the elastic energy penalty associated with protofilament bending must be compensated by an energy gain resulting from interprotofilament interactions. From the evaluation of CB obtained from the analysis of AFM images of single insulin protofilaments, the authors calculated the elastic energy cost associated with protofilament bending. They determined a lower bound of 310 kB T /µm for the energy associated with the attractive interaction between two insulin protofilaments; this value is 2 orders of magnitude larger than the interaction between single sickle hemoglobin fibers [129]. A different approach for investigating the elastic properties of fibrils consists of using the atomic force microscope tip to stretch individual fibrils. The reversible mechanical unzipping of different Aβ peptides, Aβ (1–40), Aβ (1–42), Aβ (25–35) and the Lys28-acetylated Aβ (25–35), has been studied through the acquisition of force versus distance curves [70,130,131]. From the analysis of force curves obtained during stretch–relaxation cycles, the authors found two characteristic behaviors: fully reversible, nonlinear elasticity and force plateaus with constant force steps during stretch. Force plateaus were not observed in the presence of globular aggregates, but only in the presence of fibrillar structures. From the statistical analysis of the force steps, the authors identified unit force steps characteristic of each Aβ peptide with values ranging from about 20 to 40 pN. These force values were interpreted as arising from the unzipping of a single β-sheet from the fibril. A recent and very interesting mechanical characterization of amyloid fibrils was performed by Smith et al. [128]. The authors investigated the nanoscale properties of insulin fibrils by combining imaging and force spectroscopy. Fibrils were suspended onto nanoscale grooves fabricated in a silicon substrate by using a focused ion beam (Fig. 20.15A). Two-filament fibrils were considered for the analysis. The surface was imaged until a fibril suspended over a groove was identified. Force–distance curves were then acquired in different locations by moving the AFM probe along the suspended fibril. As reported in Fig. 20.15B, four distinct regimes characterize the curves. At large tip–sample distances there is no interaction between the tip and the
200
A. Relini et al.
Fig. 20.15. AFM image of two-filament insulin amyloid fibrils deposited onto a patterned gold substrate. a A suspended fibril can be observed. b Force versus piezo displacement curves obtained on the suspended fibril. Four regions can be identified: 1 noncontact, 2 linear elastic, 3 nonlinear, 4 atomic force microscope tip pressing on the surface. c AFM image obtained after the acquisition of the force curves, confirming fibril breakage. (Modified from [128], with permission. Copyright 2006 National Academy of Sciences, USA)
sample and the cantilever is not deflected. By decreasing the separation between the tip and the sample, the tip eventually comes into contact with the fibril and a linear elastic region is observed. In this region the deflection of the fibril under the load is the horizontal distance between the coupled cantilever–fibril deflection curve and the cantilever deflection originating from the point of contact. Increasing the load, one encounters a nonlinear regime until the fibril breaks and the tip makes contact with the surface of the groove beneath. From the analysis of the nonlinear elastic region and from the point of fracture, the authors could estimate the fibril strength to be (0.6 ± 0.4) GPa. To evaluate the fibril stiffness the authors analyzed the linear region of the curves where the effective spring constant is the serial cantilever–fibril spring constant. The effective elastic constant is found to decrease along the suspended fibril when moving from the edge to the center of the grooves. By modeling the fibril as a continuous material strongly adsorbed to the substrate and undergoing a transverse load at the atomic force microscope tip position, the authors estimated a Young’s modulus of (3.3±0.4) GPa, a shear modulus of (0.28 ± 0.2) GPa and a bending rigidity of (9.1 ± 1) × 10−26 Nm2 for insulin fibrils. The fibril mechanical properties obtained from force spectroscopy analysis were found to nicely agree with the results obtained from a statistical analysis of AFM images of insulin fibrils deposited on mica. Remarkably, twofilament insulin fibrils are found to exhibit mechanical properties similar to those of stiff materials such as spider silk or even steel.
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
201
Fig. 20.16. Average length of insulin fibrils incubated in a monomer solution as a function of the incubation time determined from AFM images (points), expected by solving numerically the differential equations corresponding to a nucleation-dependent growth mechanism in the absence of breakage events (dashed line) and including breakage events (shaded region calculated by varying the breakage rate between the previously determined upper and lower limits, see text). (Reproduced from [128], with permission. Copyright 2006 National Academy of Sciences, USA)
Another interesting issue addressed in this paper concerns the study, by AFM and light scattering measurements, of fibril growth kinetics. By studying the growth of preformed fibrils added to a solution of insulin monomers, the authors could show that fibril breakage plays an important role in the fibril growth process. They found out a fibril breakage constant of (1.7±1.3)×10−8 /s. As shown in Fig. 20.16, the increase of the average fibril length as a function of time calculated in the absence of breakage (dashed line) is larger than the values experimentally determined from the analysis of AFM images (points). The latter values are in good agreement with the expected values obtained by solving numerically the differential equations corresponding to a nucleation-dependent growth mechanism including breakage events (shaded region calculated by varying the breakage rate between the previously determined upper and lower limits). Since fibril growth occurs only at the ends of the fibrils, the increase in the number of free ends resulting from the breakage of existing fibrils can play a key role in the process of fibril growth through monomer incorporation and could be of fundamental importance in the understanding of fibrillogenesis. Furthermore, the increase in concentration of free ends could have a protective effect by clearing monomers and preventing their assemblies into toxic oligomers. Acknowledgements. The authors are deeply indebted to Nadia Marano (St. Lawrence University) and Massimo Stefani (Florence University) for critical reading of the manuscript.
References 1. 2. 3. 4. 5.
Stefani M (2004) Biochim Biophys Acta 1379:5 Chiti F, Dobson CM (2006) Annu Rev Biochem 75:333 Stefani M, Dobson CM (2003) J Mol Med 81:678 Kelly JW, Balch WE (2003) J Cell Biol 161:461 Chapman MR, Robinson LS, Pinkner JS, Roth R, Heuser J, Hammar M, Normark S, Hultgren SJ (2002) Science 295:851
202
A. Relini et al.
6. Berson JF, Theos AC, Harper DC, Tenza D, Raposo G, Marks MS (2003) J Cell Biol 161:521 7. Claessen D, Rink R, de Jong W, Siebring J, de Vreugd P, Hidde Boersma FG, Dijkhuizen L, Wösten HAB (2003) Gene Dev 17:1714 8. Kwan AHY, Winefield RD, Sunde M, Matthews JM, Haverkamp RG, Templeton MD, Mackay JP (2006) Proc Natl Acad Sci USA 103:3621 9. Iconomidou VA, Chryssikos GD, Gionis V, Vriend G, Hoenger A, Hamodrakas SJ (2001) FEBS Lett 499:268 10. Chien P, Weissman JS, DePace AH (2004) Annu Rev Biochem 73:617 11. Dobson CM (2006) Protein Pept Lett 13:219 12. Krebs MRH, Bromley EHC, Donald AM (2005) J Struct Biol 149:30 13. Antzutkin ON (2004) Magn Reson Chem 42:231 14. Makin OS, Serpell LC (2005) FEBS J 272:5950 15. Sipe JD, Cohen AS (2000) J Struct Biol 130:88 16. Kyle RA (2001) Br J Haematol 114:529 17. Puchtler H, Sweat F (1965) J Histochem Cytochem 14:123 18. Friedreich N, Kekulé A (1859) Virchows Arch Path Anat Klin Med 16:50 19. Divry P, Florkin M (1927) C R Soc Biol 97:1808 20. Puchtler H, Sweat F, Levine M (1962) J Histochem Cytochem 10:355 21. Glenner GG, Eanes ED, Bladen HA, Linke RP, Termine JD (1974) J Histochem Cytochem 22:1141 22. Mera SL Davies JD (1983) J Pathol 141:547 23. Turnell WG, Finch JT (1992) J Mol Biol 227:1205 24. Pigorsh E, Elhaddaoui A, Turrel S (1994) Spectrochim Acta A 50:2145 25. Horobin RW (1980) J Microsc (Oxford) 119:345 26. Klunk WE, Pettegrew JW, Abraham DJ (1989) J Histochem Cytochem 37:1273 27. Puchtler H, Waldrop FS, Meloan SN (1985) Appl Pathol 3:5 28. Katenkamp D, Stiller D (1972) Histochemie 29:37 29. Jin L-W, Claborn KA, Kurimoto M, Geday MA, Maezawa I, Sohraby F, Estrada M, Kaminsky W, Kahr B (2003) Proc Natl Acad Sci USA 100:15294 30. Cohen AS, Calkins E (1959) Nature 183:1202 31. Serpell L, Sunde M, Benson M, Tennent G, Pepys M, Fraser P (2000) J Mol Biol 300:1033 32. Jimenez JL, Guijarro JI, Orlova E, Zurdo J, Dobson CM, Sunde M, Saibil HR (1999) EMBO J 18:815 33. Jimenez JL, Tennent G, Pepys M, Saibil HR (2001) J Mol Biol 311:241 34. Jimenez JL, Nettleton EJ, Bouchard M, Robinson CV, Dobson CM, Saibil HR (2002) Proc Natl Acad Sci USA 99:9196 35. Serpell LC, Smith JM (2000) J Mol Biol 299:225 36. Stromer T, Serpell LC (2005) Microsc Res Techn 67:210 37. Baxa U, Taylor KL, Wall JS, Simon MN, Cheng N, Wickner RB, Steven AC (2003) J Biol Chem 278:43717 38. Baxa U, Cheng N, Winkler DC, Chiu TK, Davies DR, Sharma D, Inouye H, Kirschner DA, Wickner RB, Steven AC (2005) J Struct Biol 150:170 39. Carrotta R, Bauer R, Waninge R, Rischel C (2001) Protein Sci 10:1312 40. Dong J, Apkarian RP, Lynn DG (2005) Bioorg Med Chem 13:5213 41. Krebs MR, Bromley EH, Rogers SS, Donald AM (2005) Biophys J 88:2013 42. Krebs MR, Macphee CE, Miller AF, Dunlop IE, Dobson CM, Donald AM (2004) Proc Natl Acad Sci USA 101:14420 43. Bromley EH, Krebs MR, Donald AM (2005) Faraday Discuss 128:13 44. Goldsbury CS, Wirtz S, Muller SA, Sunderji S, Wicki P, Aebi U, Frey P (2000) J Struct Biol 130:217
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates? 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82.
203
Antzutkin ON, Leapman RD, Balbach JJ, Tycko R (2002) Biochemistry 41:15436 Eanes ED, Glenner GG (1968) J Histochem Cytochem 16:673 Bonar L, Cohen AS, Skinner MM (1969) Proc Soc Exp Biol Med 131:1373 Sunde M, Serpell LC, Bartlam M, Fraser PE, Pepys MB, Blake CC (1997) J Mol Biol 273:729 Sunde M, Blake C (1997) Adv Protein Chem 50:123 Makin OS, Atkins E, Sikorski P, Johansson J, Serpell LC (2005) Proc Natl Acad Sci USA 102:315 Nelson R, Sawaya MR, Balbirnie M, Madsen AØ, Riekel C, Grothe R, Eisenberg D (2005) Nature 435:773 Olofsson A, Sauer-Eriksson AE, Ohman A (2006) J Biol Chem 281:477 Olofsson A, Ippel JH, Wijmenga SS, Lundgren E, Ohman A (2004) J Biol Chem 279:5699 Mandal PK, Pettegrew JW, McKeag DW, Mandal R (2006) Neurochem Res 31:883 Lynn DG, Meredith SC (2000) J Struct Biol 130:153-73 Tycko R (2006) Q Rev Biophys 39:1 Gosal WS, Myers SL, Radford SE, Thomson NH (2006) Protein Pept Lett 13:261 Stine WB, Snyder SW, Ladror US, Wade WS, Miller MF, Perun TJ, Holzman TF, Krafft GA (1996) J Protein Chem 15:193 Kayed R, Bernhagen J, Greenfield N, Sweimeh K, Brunner H, Voelter W, Kapurniotu A (1999) J Mol Biol 287:781 Perutz MF, Finch JT, Berriman J, Lesk A (2002) Proc Natl Acad Sci USA 99:5591 Relini A, Rolandi R, Bolognesi M, Aboudan M, Merlini G, Bellotti V, Gliozzi A (2004) Biochim Biophys Acta 1690:33 Goldsbury C, Kistler J, Aebi U, Arvinte T, Cooper GJ (1999) J Mol Biol 285:33 Zhu M, Souillac PO, Ionescu-Zanetti C, Carter SA, Fink AL (2002) J Biol Chem 277:50914 Kowalewsky T, Holtzman DM (1999) Proc Natl Acad Sci USA 96:3688 Wang Z, Zhou C, Wang C, Wan L, Fang X, Bai C (2003) Ultramicroscopy 97:73 Losic D, Martin LL, Aguilar M-I, Small D (2006) Biopolymer 84:519 Relini A, Canale C, De Stefano S, Rolandi R, Giorgetti S, Stoppini M, Rossi A, Fogolari F, Corazza A, Esposito G, Gliozzi A, Bellotti V (2006) J Biol Chem 281:16521 Relini A, Canale C, Torrassa S, Rolandi R, Gliozzi A, Rosano C, Bolognesi M, Plakoutsi G, Bucciantini M, Chiti F, Stefani M (2004) J Mol Biol 338:943 McAllister C, Karymov MA, Kawano Y, Lushnikov AY, Mikheikin A, Uversky VN, Lyubchenko YL (2005) J Mol Biol 354:1028 Karsai Á, Mártonfalvi Z, Nagy A, Grama L, Penke B, Kellermayer MSZ (2006) J Struct Biol 155:316 Stedman MS (1988) J Microsc 152:611 Griffith JE, Grigg DA (1993) J Appl Phys 74:R83 Muller DJ, Engel A (1997) Biophys J 73:1633 Keller D (1991) Surf Sci 253:353 Goldsbury C, Green J (2005) Mol Biol 299:103 Jansen R, Dzwolak W, Winter R (2005) Biophys J 88:1344 Podestà A, Tiana G, Milani P, Manno M (2006) Biophys J 90:589 Harper JD, Lieber CM, Lansbury PT (1997) Chem Biol 4:951 Ionescu-Zanetti C, Khurana R, Gillespie JR, Petrick JS, Trabachino LC, Minert LJ, Carter SA (1999) Proc Natl Acad Sci USA 96:13175 Marcon G, Plakoutsi G, Canale C, Relini A, Taddei N, Dobson CM, Ramponi G, Chiti F (2005) J Mol Biol 347:323 Calamai M, Canale C, Relini A, Stefani M, Chiti F, Dobson CM (2005) J Mol Biol 346:603 Cardoso I, Goldsbury CS, Müller SA, Olivieri V, Wirtz S, Damas AM, Aebi U, Saraiva MJ (2002) J Mol Biol 317:683
204
A. Relini et al.
83. Goldsbury C, Frey P, Olivieri V, Aebi U, Müller SA (2005) J Mol Biol 352:282 84. Arimon M, Diez-Perez I, Kogan MJ, Durany N, Giralt E, Sanz F, Fernandez-Bousquets X (2005) FASEB J 19:1344 85. Kad NM, Thomson NH, Smith DP, Smith DA, Radford SE (2001) J Mol Biol 313:559 86. Kad NM, Myers SL, Smith DP, Smith DA, Radford SE, Thomson NH (2003) J Mol Biol 330:785 87. Anderson M, Bocharova OV, Makarava N, Breydo L, Salnikov VV, Baskakov IV (2006) J Mol Biol 358:580 88. De Jong KL, Incledon B, Yip CM, DeFelippis MR (2006) Biophys J 91:1905 89. Dzwolak W, Grudzielanek S, Smirnovas V, Ravindra R, Nicolini C, Jansen R, Loksztejn A, Porowski S, Winter R (2005) Biochemistry 44:8948 90. Grudzielanek S, Jansen R, Winter R (2005) J Mol Biol 351:879 91. Stine WB, Dahlgren KN, Krafft GA, LaDu MJ (2002) J Biol Chem 278:11612 92. Gosal WS, Morten I, Hewitt EW, Smith DA, Thomson NH, Radford SE (2005) J Mol Biol 351:850 93. Legleiter J, Czilli DL, Gitter B, DeMattos RB, Holtzman DM, Kowalewski T (2004) J Mol Biol 335:997 94. Yamin G, Munishkina LA, Karymov MA, Lyubchenko YL, Uversky VN, Fink AL (2005) Biochemistry 44:9096 95. Suk JY, Zhang F, Balch WE, Linhardt RJ, Kelly JW (2006) Biochemistry 45:2234 96. Huang TH, Yang DS, Plaskos NP, Go S, Yip CM, Fraser PE, Chakrabartty A (2000) J Mol Biol 297:73 97. Catharino S, Buchner J, Walter S (2005) Biol Chem 386:633 98. DeMarco ML, Silveira J, Caughey B, Daggett V (2006) Biochemistry 45:15573 99. Morozova-Roche LA, Zamotin V, Malisauskas M, Ohman A, Chertkova R, Lavrikova MA, Kostanyan IA, Dolgikh DA, Kirpichnikov MP (2004) Biochemistry 43:9610 100. Dahlgren KN, Manelli AM, Stine WB, Baker LK, Krafft GA, LaDu MJ (2002) J Biol Chem 277:32046 101. Zamotin V, Gharibyan A, Gibanova NV, Lavrikova MA, Dolgikh DA, Kirpichnikov MP, Kostanyan IA, Morozova-Roche LA (2006) FEBS Lett 580:2451 102. Kayed R, Sokolov Y, Edmonds B, McIntire TM, Milton SC, Hall JE, Glabe CG (2004) J Biol Chem 279:46363 103. Sluzky V, Tamada JA, Klibanov AM, Langer R (1991) Proc Natl Acad Sci USA 88:9377 104. Giacomelli CE, Norde W (2003) Biomacromolecules 4:1719 105. Zhao H, Jutila A, NurminenT, Wickström SA, Keski-Oja J, Kinnunen PKJ (2005) Biochemistry 44:2857 106. Necula M, Chirita CN, Kuret J (2002) J Biol Chem 278:46674 107. Ha C, Park CB (2006) Langmuir 22:6977 108. Ha C, Park CB (2005) Biotechnol Bioeng 90:848 109. Sharp JS, Forrest JA, Jones RAL (2002) Biochemistry 41:15810 110. Gorbenko GP, Kinnunen PKJ (2006) Chem Phys Lipids 141:72 111. Jass J, Tjärnhage T, Puu G (2000) Biophys J 79:3153 112. Green JD, Kreplak L, Goldsbury C, Blatter XL, Stolz M, Cooper GS, Seelig A, Kistler J, Aebi U (2004) J Mol Biol 342:877 113. Lin H, Bhatia R, Lal R (2001) FASEB J 15:243 114. Yip CM, McLaurin JA (2001) Biophys J 80:1359 115. Yip CM, Darabie AA, McLaurin JA (2002) J Mol Biol 318:97 116. Quist A, Doudevski I, Lin H, Azimova R, Ng D, Frangione B, Kagan B, Ghiso J, Lal R (2005) Proc Natl Acad Sci USA 102:10427 117. Canale C, Torrassa S, Rispoli P, Relini A, Rolandi R, Bucciantini M, Stefani M, Gliozzi A (2006) Biophys J 91:4575
20 What Can Atomic Force Microscopy Say About Amyloid Aggregates?
205
118. Engel MFM, Yigittop H, Elgersma RC, Rijkers DTS, Liskamp RMJ, de Kruijff B, Höppener JWM, Killian JA (2006) J Mol Biol 356:783 119. Rivetti C, Guthold M, Bustamante C (1996) J Mol Biol 264:919 120. Zuccheri G, Scipioni A, Cavaliere V, Gargiulo G, De Santis P, Samorì B (2002) Proc Natl Acad Sci USA 98:3074 121. Valle F, Favre M, De Los Rios P, Rosa A, Dietler G (2005) Phys Rev Lett 95:158105 122. Anselmi C, DeSantis P, Scipioni A (2005) Biophys Chem 113:209 123. Mücke N, Kreplak L, Kirmse R, Wedig T, Herrmann H, Aebi U, Langowski J (2004) J Mol Biol 335:1241 124. Hatters DM, MacRaild CA, Daniels R, Gosal WS, Thomson NH, Jones JH, Davis JJ, MacPhee CE, Dobson CM, Howlett GJ (2003) Biophys J 85:3979 125. Landau LD, Lifshitz EM (1980) Statistical physics, part 1, 3rd edn. Pergamon, Oxford 126. Knowles TPJ, Smith JF, Craig A, Dobson CM, Welland ME (2006) Phys Rev Lett 96:238301 127. Hoshino M, Katou H, Hagihara Y, Hasegawa K, Naiki H, Goto Y (2002) Nat Struct Biol 9:332 128. Smith JF, Knowles TPJ, Dobson CM, MacPhee CE, Welland ME (2006) Proc Natl Acad Sci USA 103:15806 129. Jones CW, Wang JC, Ferrone FA, Briehl RW, Turner MS (2005) Biophys J 88:2433 130. Kellermayer MSZ, Grama L, Karsai Á, Nagy A, Kahn A, Datzi ZL, Penke B (2005) J Biol Chem 280:8464 131. Karsai Á, Nagy A, Kengyel A, Mártonfalvi Z, Grama L, Penke B, Kellermayer MSZ (2005) J Chem Inf Model 45:1641 132. Dobson CM (2004) Semin Cell Dev Biol 15:3
21 Atomic Force Microscopy: Interaction Forces Measured in Phospholipid Monolayers, Bilayers, and Cell Membranes Zoya Leonenko · David Cramb · Matthias Amrein · Eric Finot
Abstract. Atomic force microscopy (AFM) is a powerful technique which is commonly used to image surfaces at the nanoscale and single-molecule level, as well as to investigate physical properties of the sample surface using a technique known as force spectroscopy. In this chapter, we review our recent research where we used AFM to investigate physical properties of phospholipid monolayers, bilayers, and cell membranes. We describe the experimental procedures for AFM imaging, force measurements, and theoretical models to analyze force spectroscopy data. The data obtained allowed correlations between AFM topography and local adhesion and mechanoelastic properties of supported lipid bilayers in water, supported pulmonary surfactant films in air, and the plasma membrane of epithelial type II cells. Finally, AFM was applied to help elucidate the effect of anesthetics and cholesterol present in the lipid films. Key words: Phospholipid, Monolayer, Bilayer, Pulmonary surfactant, Bovine lipid extract surfactant, Epithelial type II cell, Atomic force microscopy, Adhesion, Cholesterol, Halothane
21.1 Introduction Lipid membranes are integral components of the cellular organization in all living organisms [1]. As the major constituents of the biological membranes that form the outer boundary of cells and organelles they are involved in a vast number of biological processes. Their functions are to compartmentalize the cells, sort, regulate, and mediate biomolecular interactions. These complex functions are possible owing to the diverse physical-chemical properties of these membranes, which include two-dimensional fluidity, material elasticity, chemical diversity, and rich phase behavior [2]. Understanding the physical and chemical properties of lipid membranes [3, 4] and how they control biological processes is critical to our understanding of the molecular mechanisms of many diseases and finding a cure for them. The variety in structure and physical properties of cell membranes governs their biomolecular assembly and plays an important role in many fundamental biological processes [5]. For example, small, condensed rafts of spingolipids, cholesterol, and raft-associated protein play a crucial role in signaling, intracellular trafficking of lipids, platelet activation, membrane fusion, and protein binding [6, 7]. Lipid rafts present in a cell membrane serve as entry and exit sites for microbial pathogens, such as influenza virus, measles virus, and HIV. Lipid–protein mixtures not only form membranes. Pulmonary surfactant [8, 9], which is a specific lipid–protein complex,
208
Z. Leonenko et al.
forms a molecular film that covers the lung’s interface with the air [10]. The major functions of pulmonary surfactant are to reduce the surface tension of the airspace– liquid interface, provide stability to the alveolar structure, and reduce the work of breathing [11]. Understanding the lateral distribution and interactions among the lipids and proteins in pulmonary surfactant monolayers is extremely important in order to develop methods of preventing alveolar collapse in cases of surfactant deficiency such as in premature infants or dysfunctional surfactant such as in adult respiratory distress syndrome (ARDS) [12]. Supported phospholipid bilayers (SPBs) [13] and monolayers are accepted and convenient model systems which mimic many fundamental properties of biological membranes and pulmonary surfactant. SPBs are composed of phospholipids adsorbed to a planar solid support. They also are of importance for biotechnology applications, as they form membrane-mimetic materials useful for creation of new classes of biosensors, diagnostic tools, and high-throughput characterization platforms for rapid and early detection of interactions between cells and their environment [14, 15]. Such lipid films can serve as templates into which functional biomolecules can be incorporated, creating a powerful platform for the development of nanosensor and nanodevices. For biomedical and nanotechnology applications it is important to control processes involving the interaction of lipid membranes and monolayers with DNA and proteins, as well as polymers and inorganic nanoparticles. Adhesion of particles to the supported lipid films is important in biotechnology and medicine applications, for example, to study toxicity of nanoscale air pollutants and also for the development of effective pulmonary drug delivery systems [16]. Atomic force microscopy (AFM) has proven to be a valuable tool for imaging “soft” biological samples such as lipid films [17–21] especially in intermittent contact mode [22–24]. With AFM, one can not only image membranes in their native hydrated state. One can also measure the forces of interaction between the AFM probe and the sample, obtaining information about the physical properties of biological systems [25–29]. Measuring the physical and chemical properties of lipid membranes and monolayers involves measuring a multitude forces of weak noncovalent bonds (e. g., electrostatic, van der Waals, and/or hydrogen bonds) or hydrophobic interactions between geometrically complementary surfaces. To understand a response of a biological system to its environment one needs to measure the physical and chemical properties of the macromolecular assembly, such as membranes and monolayers, and how these properties are affected by mechanical stress, thermal change, and/or molecular adsorption. It concerns the mechanical properties such as two-dimensional fluidity and anisotropic elasticity of the membrane as well as thermodynamic properties underlying its rich phase behavior. Those bulk properties can also be investigated by a thermodynamic approach such as calorimetry, densitometry, contact angle measurement, or analysis of adsorption kinetics using surface plasmon resonance or ellipsometry. On a molecular level, molecular dynamics simulation has been employed, providing an insight into the organization of lipid structures and the stability. AFM is an ideal tool to measure directly the molecular force from 0.1 pN up to micronewtons between two surfaces or even single molecules. On approach of the AFM probe tip to the sample, the force–distance curve can be used to characterize surface properties, such as van der Waals and electrostatic forces, solvation,
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
209
hydration, and compression-related steric forces [30]. The retraction force curves often show a hysteresis referred to as an adhesion “pull-off” event, which can be used to estimate the adhesion forces. Many experimental force data are now available in the literature, and theoretical models have been developed for the analysis of forces acting between two solid surfaces [31–36]. In contrast, only a few publications have addressed force measurements of soft thin layers [37] such as thiol self-assembled monolayers [38–41] and phospholipid bilayers [42–44]. Here we review our recent research on interaction forces measured with AFM on SPBs in water and supported phospolipid monolayers in air, as models for biological membrane and pulmonary lung surfactant films, and on epithelial type II cells which cover the lung’s interface with the air.
21.2 Phase Transitions of Lipid Bilayers in Water The complex structural dynamics of phospholipid bilayers are governed by temperature-dependent parameters such as average interfacial area per lipid, bilayer thickness, and disorder of the hydrophobic tails. Although SPBs differ from free-standing membranes such as those found in vesicles or liposomes, they serve as good models for biological membranes. The SPB can exist in several lamellar phases: gel phase, liquid-crystalline or fluid phase, subgel phase, and ripple phase [45]. Often, phase behavior is dominated by the main Lβ –Lα (gel–fluid) transition. A large number of stable, metastable, and transient lamellar gel structures can be adopted by different lipids. For saturated phosphatidylcholines, such as 1,2-dipalmitoyl phosphatidylcholine (DPPC), there are four recognized lamellar phases, namely, a liquidcrystalline phase (Lα ) or phases with ordered hydrocarbon chain arrangements, a ripple phase (Pβ ), a gel phase (Lβ ), and “subgel” or “crystal” phase (Lc ) [46]. Phase-related bilayer properties have been examined using a variety of experimental techniques [47–50]. The phase transition from the gel phase to the liquidcrystalline phase is shown schematically in Fig. 21.1. Physical properties of lipids such as density and thickness have been studied by differential scanning calorimetry, pressure calorimetry or dilatometry, and densitometry and other methods [51, 52]. Temperature-dependent AFM studies allow for visualizing these phase transitions and provide valuable information about changes in structural and physical properties of bilayers during phase transitions. Temperature-dependent AFM studies involving imaging bilayers have been reported [53–57]. In addition, we have examined the force measurements as a function of temperature [58]. In this section we review our results on force measurements in SPBs during phase transitions associated with melting or anesthetic incorporation. Dioleoyl phosphatidylcholine (DOPC; lyophilized or in chloroform solution) (Sigma, Oakville, ON, USA) and DPPC (Avanti Polar Lipids, Alabaster, ME, USA) were used without further purification. Phosphate buffer (pH 6.8) and distilled, nanopure water were used in the preparation of all vesicles. Freshly cleaved ASTMV-2 quality, scratchfree ruby mica (Asheville-Schoonmaker Mica) was used as the solid support. Supported planar bilayers were prepared for AFM imaging by the method of vesicle fusion [59, 60]. All vesicles were prepared as previously reported [59]. Aliquots of
210
Z. Leonenko et al.
Fig. 21.1. Domain formation in a phospholipid bilayer supported on mica. The fluid phase domains correspond to a lower thickness than the gel-phase domains
the vesicle solution were deposited on freshly cleaved pure mica. After a controlled period of time the mica was gently rinsed with nanopure water. Surfaces were imaged with an atomic force microscope (Pico SPM, Molecular Imaging) equipped with an AFMS-165 scanner. The nominal spring constants of Au–Cr-coated MAClevers used were approximately 0.6 N/m. The tip radius of curvature was typically 25 nm. The scan rate was 20 µm/s. All imaging and force measurements were performed in a liquid cell under nanopure water. For temperature-dependence experiments the temperature was varied from 22 to 65 ◦ C, within 0.1 ◦ C accuracy with a 1 ◦ C/min ramp. 21.2.1 Morphology Change During Lamellar Phase Transition 21.2.1.1 Temperature-Dependent Phase Transition A DPPC phospholipid bilayer normally exists in the gel phase at room temperature. When heated above the melting transition temperature it transforms into the fluid phase. Domains of lower thickness are readily formed during this process. Such domains are shown schematically in Fig. 21.1 and are easily observed by AFM imaging (Fig. 21.2). A DPPC supported bilayer was imaged while it was heated from room temperature to 65 ◦ C and also while it was cooled back to room temperature. The topography images (Fig. 21.2) show the typical formation of lipid domains during the phase transition. In Fig. 21.2, panel A, a DPPC bilayer is in the gel phase, Lβ , at 22 ◦ C. When sample was heated, several phase transitions were observed as a function of increasing temperature: a broad main transition was observed at 42–52 ◦ C, and another transition at 53–60 ◦ C within the fluid phase. The first transition, at 42–52 ◦ C, is attributed to the main Lβ –Lα transition, and the second transition, between 53–60 ◦ C, to the formation of a fluid disordered phase, perhaps with interdigitated or partly interdigitated lipid chains. Slowly cooling the system back showed that the two higher-temperature transitions at 42–52 and 53–60 ◦ C are reversible and that the bilayer can be restored to its initial thickness [61]. A DOPC bilayer, which exists in a fluid phase at room temperature, when heated shows thickness reduction as well as an increase in surface coverage [62], but no domain formation. Previous studies have established the Lβ –Lα phase transition temperature for DPPC at 41 ◦ C [63]. The presence of two broad transitions that we observed on supported bilayers can be attributed to the effect of the mica support and to separating the phase transitions
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
211
Fig. 21.2. Atomic force microscopy (AFM) topography images (scan size 1.5 µm by 1.5 µm) showing a phase transition in a 1,2-dipalmitoyl phosphatidylcholine (DPPC) bilayer. A DPPC bilayer at room temperature, 22 ◦ C. B DPPC bilayer with excess of halothane (60 µl) [57]. C 50 ◦ C—melting transition [61]. The scale bars show the deviation in height
in the bilayer’s upper and lower leaflets. A similar effect was reported for supported bilayer transitions observed using differential scanning calorimetry by Yang and Appleyard [64]. 21.2.1.2 Phase Transition Due to Incorporation of Halothane Understanding the changes in physical and chemical properties of biological membranes owing to the incorporation of anesthetics is of great interest for understanding the mechanism of anesthetic action. Currently, it is understood that anesthetic directly affects membrane proteins; however, because they are lipophilic, they partition into the membrane and may interact with membrane proteins via the lipid bilayer. AFM imaging of a DPPC bilayer (Fig. 21.2, panel B) shows domain formation in the presence of general anesthetic halothane similar to what we observed during
212
Z. Leonenko et al.
heating. We have shown that changes in bilayer thickness induced by halothane are similar to the first melting transition [58]. For a DPPC SPB containing halothane, only one transition at 52–60 ◦ C was evident during heating and cooling. The presence of halothane changes the structural properties of the bilayer in such a way that the other transitions observed for the pure bilayer are not possible. We assume that in the presence of halothane, the water interface layer is disturbed by halothane molecules, thus eliminating the low-temperature transition. In spite of the visual similarity between an anesthetic-induced domain formation (Fig. 21.2, panel B) and the heat-induced gel–liquid phase transition (Fig. 21.2, panel C), observed by AFM imaging, the mechanism of anesthetic action is likely to be different from the effect of membrane melting. We assume that the physical properties of the domains produced in lipid bilayers by halothane and by temperature are also different. To address these physical properties we used the atomic force microscope also as a force apparatus. 21.2.2 Change in Forces During Phase Transition 21.2.2.1 Force Measurements In AFM force measurements, the probe is moved towards the sample and cantilever deflection is measured as a function of the extension of the piezoelectric tube. During approach to the surface, the attractive and repulsive forces measured are characterized by van der Waals and electrostatic interactions as well as solvation, hydration, and compression-related steric forces [26, 36, 52, 65]. The probe is then retracted from the surface. The retraction force curves often show a hysteresis referred to as an adhesion pull-off event, which can be used to determine adhesion forces between the probe and the sample. In our experiments, we imaged the supported bilayer first in tapping MAC mode. After the image was complete and had been saved we switched the regime into the contact mode and then using the same cantilever, we measured forces of interaction by positioning the atomic force microscope tip on different locations using the image obtained in tapping mode. It is a well-known fact that the force interactions depend on the velocity of the surface approach [26], especially for soft, viscoelastic materials. We varied the frequency sweep over about four decades, from 0.01 to 10 Hz. The Z-range was 50 nm. Note that varying the scan frequency affects simultaneously the velocity at which the tip is withdrawn from the bilayer and the contact time between the tip and the sample. At a fixed applied load, increasing the frequency sweep rate increases the tip velocity, but decreases the contact time. Data were collected over a time period of 2 h at ten locations on the bilayer and a total of 100 force curves were analyzed for each sample. Force curves were saved as plots and then converted into text files for analysis. After the raw data had been collected they were converted into the “force versus separation plots”, averaged, and analyzed. Here we review our results on force measurements in DOPC and DPPC SPBs during heating above TM from 22 to 65 ◦ C. We also compare them with results obtained on bilayers saturated with halothane at 20 ◦ C.
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
213
21.2.2.2 AFM Force Analysis Force measurements using the atomic force microscope (described in Sect. 21.2.2.1) give us raw data of cantilever deflection δV versus piezo movement z piezo . Raw data were transformed numerically using the MatLab environment. Raw data (Z c versus Z p ) were then converted into force F versus surface–tip separation D using Hooke’s law: F = kZ c , where k is the spring constant of the cantilever, and the geometric relationship D = Z c − Z p for incremental changes. Statistical data were extracted from a large set of measurements. The bin size of the histogram was determined statistically by dividing the range of measurements by the square root of the number of measurements. The average of the adhesion force and the standard deviation were obtained from the analysis of the whole set of force curves, which were represented in histogram form, P(x), and then adjusted using Gaussian and Poisson distributions. 21.2.2.3 Adhesion Forces To extract information about adhesion forces we analyzed the retrace part of the force curve. Raw data of the cantilever deflection corresponding to the retraction of the tip from the surface, after contact, are shown in Fig. 21.3. Force data were collected as a function of temperature for a DPPC SPB. When a pure DPPC bilayer was heated, the temperature increase resulted in a gradual change of the force profile. The force required to detach the tip from the bilayer increased from 3 nN at 36 ◦ C to 10 nN at 50 ◦ C. In addition, the jump out of the contact was not abrupt, compared with that for a “solid-like” surface (i. e., mica). The distance the tip travels before it separates from the surface increased with increasing temperature. As van der Waals forces are known to vary only slightly with temperature, the change in the profile must be due to tip–bilayer interactions that depend on the bilayer phase. Incorporation of halothane into DPPC also leads
Fig. 21.3. Cantilever deflection dZ c as a function of the piezotube elongation when the tip is retracted from the sample. DPPC bilayers deposited on mica were measured under three conditions: a the DPPC bilayer at 22 ◦ C; b the DPPC bilayer at 50 ◦ C; c the DPPC bilayer with halothane incorporated at 22 ◦ C
214
Z. Leonenko et al.
to a significant change in the force profile. As a bilayer adsorbed on mica is known to screen the van der Waals forces [66], the huge adhesion force observed cannot originate solely from the van der Waals forces. This adhesion behavior can be compared with that observed in the phase transition on DPPC from the gel to the liquid crystal observed on increasing the temperature (Fig. 21.3, plots b and c). Halothane induces then a transition of the bilayer into the fluid phase in a similar manner as melting, and induces the increase in adhesion, although the adhesion with excess halothane was greater than the adhesion observed on pure DPPC at elevated temperature. To analyze the force histogram (Fig. 21.4), we employed Poisson statistics. It has been demonstrated that the force of a unit interaction between an atomic force microscope tip and a surface can be determined from a statistical analysis of a series of detachment force measurements. For a statistical analysis based on adhesion force originating from a discrete number n of individual interactions or bonds, Fs was used. The total force distribution follows Poisson statistics, where both the adhesion force Fadh and the variance σ originate from a number of individual tip–surface bonds, n: Fadh = n Fs
and σ 2 = n Fs2 .
(21.1)
The force of one bond, Fs , is therefore given by the square of the variance of the force divided by the mean adhesion force Fadh . The value n is the ratio between Fadh and Fs . This analysis has the advantage that knowledge of the mean radius of curvature is not required [67] and gives information on the nature of the bilayer. This analysis allows one to correlate microscale changes in the bilayer physical properties with
Fig.21.4.Distribution of adhesion force Fadh , measured in phospholipid bilayers in water: a DPPC Fs = 0.3 nN, n = 3; b DPPC Fs = 0.5 nN, n = 6 [58]; c DPPC + halothane Fs = 0.5 nN, n = 14 [58]
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
215
Table 21.1. Experimental adhesion force Fadh and its standard error σ on a 1,2-dipalmitoyl phosphatidylcholine (DPPC) bilayer at room and elevated temperature and DPPC with halothane, measured at room temperature; calculation of the number of bonds n and the mean force of a single bond Fs
Fadh (nN) σ (nN) Fs (nN) n
DPPCa 22 ◦ C
DPPCb 22 ◦ C
DOPCa 22 ◦ C
DPPCa 60 ◦ C
DPPC + halothaneb 22 ◦ C
1 0.6 0.3 3
1.5 0.8 0.4 4
3 1.2 0.5 6
5 2 0.8 6
7 2 0.5 14
DOPC dioleoyl phosphatidylcholine a From [61] b From [58]
the molecular structure and the mobility of individual lipid molecules. The lipids in biological membranes have a high lateral mobility since they can easily exchange their positions. The fluidity of the membrane can be then thought of as the lateral motion of the constituents of the membranes. This fluidity should depend on the chain length and the composition of the membranes. The mobility of tails can be seen as the number of molecules n in contact with the AFM probe. The larger n is, the higher the mobility. The change in mobility increases the number of bonds n and therefore Fadh and σ. The statistical distribution of the adhesion forces (Table 21.1) clearly demonstrates that adhesion forces observed for a pure DPPC bilayer increase with heating and increase even more with halothane incorporation. A fluid-phase DPPC bilayer at high temperature is typically ascribed to a higher number of tip–sample bonds [61]. The force of a single bond is increased in a fluid phase. In addition, in a fluid phase, n increases on decreasing the speed of the tip motion v. The number of bonds, n, involved is higher in the fluid phase. It has been shown previously that the force required to remove a single lipid from a bilayer is of the order 30–140 pN [68,69], depending on the lipid and on the technique employed. The lowest values that we measure are at least twice this range, suggesting that we do not observe single lipid binding events. Instead, our unit “bonds” must represent patches containing more than three lipids. The detachment force between the patch and the tip must be smaller than the extraction force of the entire patch otherwise we would see evidence of lipids coating the tip. This is consistent with our previous work where continuous imaging of bilayers displayed no resolution degradation or changes in force curves, which would result from uncontrollable lipid coating the atomic force microscope tip. The effect of halothane is similar to the effect of melting. A DPPC bilayer with halothane corresponds to higher adhesion force, higher standard error, and a larger number of single binding events between the atomic force microscope tip and the sample. This number, n = 14, is more than 2 times larger than for a fluid phase produced by melting. The single force Fs was not altered significantly by the presence of halothane. The presence of halothane in the bilayer increases,
216
Z. Leonenko et al.
therefore, the fluidity of the DPPC molecules in the bilayer. The number of binding events, n, could also be higher owing to the contribution of halothane molecules themselves, interacting with the atomic force microscope tip. The unit force Fs for fluid-phase DPPC is 0.8 compared with 0.5 for the DOPC bilayer (Table 21.1); the umber of unit bonds is 6 for both fluid-phase DPPC and DOPC bilayers. This correlation permits evaluation of a bilayer fluidity and is in a good agreement with imaging data. Adhesion forces depend, therefore, on the physical properties of the bilayer and the force analysis gives valuable information on the nature of the bilayer. 21.2.2.4 Repulsive Forces Repulsive forces were extracted by analyzing the approach part of experimental force curves. Figure 21.5 shows plots of the repulsive forces measured while the atomic force microscope tip was approaching the sample surface of pure mica for a DPPC bilayer at room temperature, a DPPC bilayer at 50 ◦ C [61], and a DPPC bilayer with an excess of halothane [58]. For pure DPPC, three exponential regimes clearly exist, one at the long range from 4 to 7 nm, another at short range between 0.3 and 4 nm, and the third, steep one in close contact with mica, which we will not consider. The long-range interactions can be explained by electrostatics. Electrostatic repulsion likely originates from an effective charge density at the bilayer–water interface, where zwitterionic headgroups interact with water molecules. The approximate expression of the electrostatic force (also called double-layer force) is given by [70] F = 4πσRψT exp(−K D D) ,
(21.2)
where R is the tip radius, σ is the surface charge, ψT is the tip potential, and K D is the inverse of the Debye length. K D can be estimated to be 0.25/nm.
Fig. 21.5. Logarithmic representation of the force acting on the DPPC layer at 22 ◦ C (squares), on the DPPC layer at 50 ◦ C (circles), and on the DPPC layer with halothane at 22 ◦ C (triangles)
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
217
The long-range force does not seem to be affected by temperature, as the two curves in the log plot are almost parallel (Fig. 21.5). In the presence of halothane, (Fig. 21.5), the long-range electrostatic forces appear to be reduced, in the range above 4 nm. Therefore, halothane clearly changes the surface charge density of the DPPC bilayer by replacing water molecules and reducing water shielding of the lipid bilayer, thereby decreasing the magnitude of the electrostatic forces. The second, short-range regime, between 0.3 and 4 nm, is attributed to the deformation of the bilayer. Most studies on the bilayer report a jump corresponding to the piercing of the bilayer for forces ranging from approximately 3 to 25 nN [71–73]. This jump into contact is observed when the force gradient exceeds the spring constant of the cantilever k, which is the case when k < 1 N/m. Although the force sensitivity can be altered, a high cantilever spring constant (k = 30 N/m) was chosen for monitoring all the data during the penetration. The force gradient attributed to the layer indentation was measured to be around 10 N/m and the force needed to puncture the lipid bilayer is in good agreement with that found in other studies. The repulsive force in the short-range regime at the contact of the bilayer cannot be easily treated using common elastic mechanical models such as Hertz’s theory, or any other models based on continuum mechanics. The heads of the phospholipids are responsible for a crystalline elasticity, i. e., a change of the head-to-head distance (high elasticity but low rupture). The lipid chain behaves like rubber (low elasticity), and this behavior is related to the change of entropy (increase of order when one presses the layer). It results in the property that the cell membranes can be easily deformed in the plane but do not resist the expansion (distance between the headgroup cannot vary more than 2%). With our AFM experiments we are examining very small, discrete areas where nanoindentation induces not only macroscopic compression of the bilayer, but also nanoscale structural changes in the bilayer. Therefore, a molecular approach, considering steric forces [42], is preferable. The exponential increase of the repulsive force likely originates from tail steric repulsion induced by the atomic force microscope tip compressing the bilayer. The tails in the bilayer will resist compression, because compression will increase the local concentration of the alkyl segments, increasing the free energy of the system. The steric forces of compression act on the bilayer until the loading force is close to the yield force, when the tip penetrates through the bilayer. The breakthrough force decreases as the temperature increases, and a DPPC bilayer at elevated temperatures shows similar instability in the force plot as a fluid-phase DOPC bilayer and a fluidphase 1,2-dioleoyl-3-trimethylammonium propane (DOTAP) bilayer [44]. The steric force per unit area that the bilayer exerts on the tip can be modeled using the formalism developed in the work of de Gennes [74]. This model was further developed by Cappella and Dietler [26], where interaction between two polymer surfaces was considered. We modified this formula, considering the interaction of the AFM probe and the bilayer, which gives the formula for the steric repulsion force: Fste = 100kTRLγ 3/2 exp(−2πD/L) ,
(21.3)
where k is the Boltzmann constant, R is the tip radius of curvature, T is the temperature, γ is the tail density (inverse of the area per lipid), D is the distance between
218
Z. Leonenko et al.
the atomic force microscope tip and the mica surface, and L is the uncompressed bilayer thickness. Similar to the equation for the two interacting polymer surfaces, this formula is valid in the range of distances where 0.2 < D/L < 0.9. Since the tip radius is not well defined, it is difficult to estimate the tail density from the fit, but the transition between the two exponential regimes can be a good way to measure the thickness of the bilayer L. The steric forces can, however, be used to determine the thickness of the layer adsorbed at the surface in good agreement with the image cross-section analysis. With the temperature increase, the steric force departs from a pure exponential law owing to the increase of the lateral mobility of the phospholipid tails. The transition between the electrostatic regime and the steric regime where the tip is at the top of the bilayer can be an effective way to estimate the bilayer thickness. At 22 ◦ C, a transition occurs at D = 4 nm, corresponding to a thickness L = D/0.9 = 4.5 nm, which agrees with what is known from the literature (4.7 nm). With increasing temperature, the magnitude and the range of the steric force decrease contrary to the change in the electrostatic force. The steric force in a fluid phase, at elevated temperatures, is reduced by 50% compared with that a gel phase, as in Fig. 21.5. This can be explained by the reduction in the tail density with temperature. The tail density is reduced from ˚ 2 at room temperature and 1/50 to 1/67.1 (considering an area per lipid of 50 A 2 ◦ ˚ 67.1 A at 65 C [75–77]), which makes 25% of difference and therefore a change of 40% in the force. In a fluid regime, i. e., at T > 50 ◦ C, the shape of the steric part of the force–distance curve starts to depart from an exponential dependence and is similar to the linear abrupt jump observed for a fluid-phase DOTAP bilayer at room temperature. It appears that (21.3) is only valid for a gel-phase DPPC bilayer, where lateral motion of the lipids is lower. With halothane incorporation (Fig. 21.5, triangles) the force is mostly governed by the bilayer deformation, described by steric forces. The induced phase transition of the DPPC has already been characterized by a decrease in the “effective thickness”. The major difference with the temperature-induced phase transition comes from the fact that the incorporation of halothane changes the electrostatic interaction, which was not the case during the melting transition. Our force measurements confirmed that assumption and clearly showed the differences in the physical properties of the bilayer in thin domains, produced by the melting transition and partitioning of halothane. When the adhesion forces are compared for DPPC and DOPC bilayers, they are similar for the fluid-phase DPPC bilayer at elevated temperatures and the fluidphase DOPC bilayer at room temperature. The unit force Fs for fluid-phase the DPPC bilayer is 0.8 compared with 0.5 for the DOPC bilayer (Table 21.1); the number of unit bonds is 6 for both fluid-phase DPPC and DOPC bilayers. This correlation permits evaluation of a bilayer fluidity and is in a good agreement with imaging data. Therefore, adhesion forces depend on the physical properties of the bilayer, and force analysis gives important information on the nature of the bilayer. In the repulsive part of the force plot using a steric force model also provides an important insight into the properties of the bilayer and ability of lipids to resist compression. Steric forces reduce with the temperature increase for the DPPC bilayer and correlate well with low steric forces for the fluid-phase DOPC bilayer at room temperature. In general, a fluid-phase bilayer can be characterized by the reduced
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
219
repulsive forces, compared with the gel-phase bilayer. The major contribution to this change is provided by changes in steric forces. Lipid compression is much easier in the fluid phase owing to the higher disorder of tails and mobility of lipids, and lower tail density. Individual bond forces can be estimated for lipid bilayers and changes in binding number can be used as a distinctive feature of the increase of bilayer fluidity. A force analysis revealed that the interaction between the tip and the surface was governed by both the electrostatic forces, for longer range, and the steric forces, at shorter range. Electrostatic repulsion likely originates from an effective charge density at the bilayer–water interface, where zwitterionic headgroups interact with water molecules. The steric forces can be used to determine the thickness of the layer adsorbed at the surface in good agreement with the image cross-section analysis. With the temperature increase, the steric force departs from a pure exponential law owing to the increase of the lateral mobility of the phospholipid tails. The transition between the electrostatic regime and the steric regime where the tip is at the top of the bilayer can be an effective way to estimate the bilayer thickness. The temperature-induced phase transition of DPPC has been characterized by a decrease in the, “effective thickness” and we also observed similar behavior for bilayers containing halothane [58]. The major difference comes from the fact that the incorporation of halothane changes the electrostatic interaction, which was not the case during the melting transition. Halothane is a small polar molecule and is assumed to increase the polarity of the bilayer at the interface, and has been shown to replace water molecules bound to lipids at the interface. Our force measurements confirmed that assumption and clearly showed the differences in the physical properties of the bilayer in thin domains, produced by the melting transition and partitioning of halothane.
21.3 Force Measurements on Pulmonary Surfactant Monolayers in Air Pulmonary surfactant is a specific mixture of phospholipids and surfactant-specific proteins. It forms a molecular film at the interface of the hydrated lung epithelium with the air and thereby reduces the surface tension of the interface to near zero. This is required for normal respiration and structural stability of the lung. In adult respiratory distress syndrome (ARDS), surfactant function fails. As a result, the lung is less compliant, the gas exchange area is reduced, and blood oxygenation is strongly decreased. The surfactant film is also the first barrier of airborne particles in polluted air. When inhaled, ultrafine particles with an aerodynamic diameter of 100 nm or less produce a major health threat. They cross the surfactant film, cross the lung epithelium, and enter the bloodstream. We describe force spectroscopy experiments aimed at a better understanding of surfactant failure in ARDS. Force spectroscopy was used to study the interactions between airborne particles and the surfactant. Particles (modeled by an atomic force microscope tip) interact with the surfactant film before they get into close contact because of the electrical surface potential of the film. This interaction depends on surface potential, which is locally strongly variable, depending on the film composition.
220
Z. Leonenko et al.
Fig. 21.6. Supported lipid monolayer and multilayers on mica, transferred from the liquid–air interface by Langmuir–Blodgett deposition; these multilayers are shown on the AFM topography image in Fig. 21.7a
We have shown that the ability of surfactant to reduce surface tension is associated with a distinct molecular architecture of the surfactant film. The functional pulmonary surfactant forms a pattern of molecular monolayer areas and areas of lamellar stacks, cross-linked to the monolayer at the air–aqueous interface (Fig. 21.7a). We recently found that cholesterol in excess of a physiological proportion abolishes surfactant function, and surfactant fails to lower the surface tension upon compression [78]. This failure of function is associated with a change in the molecular architecture [79]. The formation of multilayer stacks, which is a characteristic feature of functional surfactant, does not occur in the presence of excess cholesterol. We performed atomic force measurements on supported pulmonary surfactant films in air to address the effect of cholesterol on the physical properties of lung surfactant films. We used bovine lipid extract surfactant (BLES). BLES is a hydrophobic extract of bovine lung lavage that differs from natural surfactant in the lack of surfactant-specific proteins SP-A and SP–D and cholesterol. Phosphatidylcholines represent 80% of its mass with half of the phosphatidylcholines being the disaturated DPPC. Between 5 and 10 mass % is the negatively charged phosphatidylglycerol, and two hydrophobic surfactant-associated proteins (SP-B, SP-C). BLES in nonbuffered normal saline (pH 5–6) with a phospholipid concentration of 27 mg/ml was a kind gift from the manufacturer (BLES Biochemical, London, ON, Canada). Cholesterol was purchased from Sigma Chemicals. A solution of 1:1:1 methanol, chloroform, and BLES by volume was first vortexed and then spun at 100g for 5 min. The methanol–water phase was discarded and the BLES in chloroform was retained and either 5 or 20% of cholesterol (by mass) with respect to phospholipids in chloroform was added. Each solution was then dried under N2 and resuspended with Goerke’s buffer [140 mM NaCl, 10 mM N-(2hydroxyethyl)piperazine-N ′ -ethanesulfonic acid and 2.5 mM CaCl2 ; pH 6.9] to obtain an aqueous suspension of BLES and cholesterol at a concentration of 27 mg/ml phospholipids. BLES solutions were spread at the air–liquid interface. Supported planar monolayers on mica were prepared using Langmuir–Blodgett technique and transferred on mica support when films were compressed to a surface tension of 25 mN/m. Supported films were imaged in air using a NanoWizard atomic force microscope from JPK Instruments, Germany. All measurements were performed at 25 ◦ C, over five different velocities, collecting each time ten force curves. Silicon cantilevers from Micromash, Spain, were used with cantilever spring constants of 0.6 and 0.7 N/m as determined by a thermofluctuation method using JPK SPM software. Forces were measured on several different structural areas of the film, by positioning the atomic force microscope tip after the image had been collected.
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
221
21.3.1 Adhesion Measurements: Monolayer Stiffness and Function Atomic force measurements helped us to elucidate how the loss of mechanical stability is related to the local mechanical properties of the film. Films containing 5 and 20% w/w cholesterol were compared. The presence of 20% cholesterol in BLES resulted in a decrease of the observed adhesive interaction, and an increase in the rigidity of the film. The force measurements were performed both on the monolayer area and on a first bilayer adjacent to the monolayer. The typical height of the bilayer area is 4.5 nm, and that of the monolayer is 2.5 nm. Typical areas where forces were measured are shown in Fig. 21.7a. For BLES with 20% cholesterol, only a monolayer was present and was used for force measurements (Fig. 21.7a, point 1). BLES with 5% cholesterol at the bilayer area has an adhesion of 60 nN, which is higher than that of the monolayer. BLES with 20% cholesterol has a lower adhesion than the BLES monolayer with 5% cholesterol. It is interesting that the “single force” Fs was slightly lower in the presence of 20% cholesterol, 3 nN compared with 4 nN with 5% cholesterol, but the number of bonds decreased significantly from 15 and 10 to 7 (Table 21.1 [79]). Therefore, the presence of 20% cholesterol decreases the fluidity of the monolayer. A force histogram (Fig. 21.8) was used to estimate the single force Fs and the number of bonds n. It is interesting that Fs was slightly lower in the presence of 20% cholesterol, 3 nN compared with 4 nN with 5% cholesterol, but n decreased significantly from 15 and 10 to 7 (Table 21.2). Therefore, the presence of 20%cholesterol decreases the fluidity of the monolayer.
Fig. 21.7. A AFM topography image of supported bovine lipid extract surfactant (BLES) film, scan area 15 µm by 15 µm. B Adhesion force map image 64 by 64, collected on the same area. The adhesion force map was collected while forces of interaction at each of 64 by 64 points were measured. The adhesion peak of the retrace plot was used to generate the adhesion force map image. 1 area of a typical monolayer; 2 area of a typical bilayer, adjacent to a monolayer
222
Z. Leonenko et al.
Table 21.2. Experimental adhesion force Fadh and its standard error σ on a bovine lipid extract surfactant (BLES) bilayer with with 5% cholesterol, a BLES monolayer with 5% cholesterol, and a BLES monolayer with 20% cholesterol, measured at room temperature; calculation of the number of bonds n and the mean force of a single bond Fs . Rate 200 nm/s, load 20 nN. Adhesion force on pure mica does not exceed 1.5 nN
Fadh (nN) σ (nN) Fs (nN) n
BLES 20% cholesterol monolayer
BLES 5% cholesterol monolayer
BLES 5% cholesterol bilayer
20 8 3 7
40 12 4 10
60 15 4 15
Fig. 21.8. Distribution of adhesion force, Fadh , in air: a BLES monolayer + 20% cholesterol; b BLES monolayer + 5% cholesterol; c BLES bilayer + 5% cholesterol
The increase in the adhesion force for a bilayer as compared with a monolayer may be understood by an increase in the contact area between the AFM probe and the sample at a given load. BLES with 20% cholesterol shows the lowest adhesion force, compared with monolayer and multilayer areas of BLES with 5% cholesterol. This indicates that adhesion properties of the film were considerably altered by the incorporation of 20% cholesterol, decreasing adhesive interaction, and increasing the rigidity of the film. The lipid molecules are not mobile enough in the presence of 20% cholesterol and cannot be easily rearranged around the tip to increase the contact area and adhesion force. The increased rigidity of the film and decreased adhesion that we observed may play an important role in preventing the monolayer–bilayer conversion. 21.3.2 Repulsive Forces: The Interaction of Charged Airborne Particles with Surfactant The repulsive forces (Fig. 21.9) observed in air between the lipids film and the AFM probe were analyzed using two theoretical models for electrostatic interaction. The atomic force microscope tip was modeled once as a point charge and once as
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
223
Fig. 21.9. Electrostatic force measured in air on BLES film with 0% cholesterol (top) and with 20% cholesterol (bottom). Experimental data are shown by dots and theoretical fits with the point-charge model are drawn in solid lines
a charged semisphere. Here q is a charge of the atomic force microscope tip, k is a constant, ε is the dielectric permittivity of the medium, and r is the separation distance between the tip and the medium. The supported BLES surfactant monolayer is composed of lipid molecules, which are organized and oriented in such a way that the dipole moment is directed along the surface normal. Therefore, the patch of the monolayer can be treated as two charged planes, each having charge Q + and Q − and surface charge density σ+ and σ− , with the separation distance δo , which corresponds to the thickness of the monolayer and is equal to 2.5 nm. The surface charge density of the sample, σ, is related to the surface potential V , which we have determined experimentally using Kelvin probe force microscopy [80, 81] and used here as a parameter. In the case of a point charge, the total electrostatic force acting on q is given by Vq F= 2δ0 A0
where
β
α
−√ 1 + α2 1 + β2
,
(21.4)
D =α, R0 D + δ0 =β. R0 For the semisphere model, the force acting on the semisphere tip with a radius Rso will be Vq 1 + β12 − 1 + α12 − 1 + (β1 − γ)2 + 1 + (α1 − γ)2 , Fs = 2δ0 A0 (21.5)
224
Z. Leonenko et al.
where D + RS0 = α1 , R0 D + RS0 + δ0 = β1 , R0 RS0 =γ. R0 Kelvin probe microscopy was used [81] to determine the surface potential of the lipid layer in air. Cholesterol induces a nonuniform distribution of the electric surface potential even at small physiological concentrations, showing nanoscale domains in the Kelvin image, while the AFM topography image remains uniform. Such domains vary in surface potential from −0.6 V, measured for 0% cholesterol, to −0.4 V, at 20% cholesterol. Both models we used fit the experimental data well, showing the presence of electrostatic interactions in the net force measured between the tip and the sample. To better fit the experimental data we varied the charge on the atomic force microscope tip and the size of the charged disks, R0 , which experimentally corresponds to the size of domains already observed with both AFM and Kelvin probe force microscopy imaging. We modeled the AFM probe also with a charged sphere; in this model the effective charge depends on the radius of the tip. Samples of BLES with 20% cholesterol exhibit stronger electrostatic repulsion at longer distances, for all models used. Interestingly, the electrostatic force was found to be more sensitive to the size of these domains than to the changes in the surface potential observed owing to the effect of cholesterol. The experimental data and the theoretical fit correlate well and both revealed that the size of the domains is approximately 5 times larger for the film containing 20% cholesterol than that for the film with no cholesterol present. This confirms the important role of cholesterol—it produced small domains which differ by electrostatic potential, initially invisible with AFM, which causes the increase in the total electrostatic force. These findings correlate with the results of molecular dynamics simulations [82], where cholesterol was shown to order individual lipids adjacent to cholesterol molecules, but abolished the long–range cooperativity between lipids, characteristic for the gel phase with true latticelike spatial order, and altered the lipid electrostatic potential [83]. In the presence of cholesterol [84] it was shown that the membrane electrostatic potential has a much larger variance depending on the distance along the bilayer normal, compared with that for a pure DPPC bilayer.
21.4 Interaction Forces Measured on Lung Epithelial Cells in Buffer Elevated levels of ultrafine particles in air pollution are associated with increased morbidity and mortality [85]. Unlike larger particles, ultrafine particles reach the peripheral lung and are able to cross the various barriers of the lung, including
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
225
the lipid–protein layer of pulmonary surfactant at the air–alveolar interface and the cell membranes of the epithelial and endothelial cells underneath and reach the bloodstream [86–89]. The toxicity of inhaled nanoparticles entering the body through the lung is thought to be initially defined by the electrostatic and adhesive interaction of the particles with the wall of the lung, and lung epithelial cells. Because of their substantial uptake, ultrafine particles are also of great interest for delivery of drugs to the peripheral lung and the body. Despite the importance of the early penetration steps in the biological effects of inhaled particles, the nature of the interactions with the various barriers of the lung remain poorly understood. A detailed study, therefore, promises substantial progress in the understanding of the health threat posed by fine and ultrafine air pollutants and the effectiveness of drug delivery by aerosols. The adhesive interaction includes electrostatic, van der Waals, steric, and hydrophobic forces as well as the line tension between particle, cell, and aqueous medium and depends on the elastic and plastic properties of the cell. The free energy due to these interactions will be minimized upon particle wetting by the lining layer. The lower the surface free energy of the particle, the less it will be wetted by the cell. A thermodynamic model using the “wettability criterion” was indeed successful in predicting passive particle uptake by cells [90]. Another thermodynamic analysis combined with a molecular dynamics simulation found negative line tension values for nanometer-sized particles, but positive values for those an order of magnitude larger [91]. The engulfment of particles in the nanometer range by unbalanced capillary forces was first suggested by Shanahan [92]. We demonstrated that the initial thermodynamic aspects and the time course of the uptake of nanoparticles by lung epithelial cells can be studied using AFM [93] in the force measurement mode. We investigated the first steps of the interaction of nanoparticles with lung epithelial cells using the atomic force microscope as a force apparatus. The apex of theatomic force microscope tip can be used as a model of the nanoparticles, thereby enabling the monitoring of the interaction forces between the nanoparticles and the cell over time. The adhesion force and the work of adhesion have been investigated and correlated with the mechanical properties of the cell. 21.4.1 Cell Culture/Force Measurement Setup Lung epithelial cells, type II, were isolated from male, pathogen-free SpragueDawley rats weighing 150–250 g. The cells were cultured at 6 × 108 cells per milliliter in Dulbecco’s modified Eagle’s medium supplemented with 1% lglutamine, containing 10% heat-inactivated fetal bovine serum, and 300 µl of a penicillin–gentamycin solution in γ-irradiated culture dishes. The cells were grown on light microscope glass cover slips. The cells were kept in an incubator under 5% CO2 and at37 ◦ C until they created a monolayer with approximately 80% of confluence (4–6 days). At this stage, the phenotype of the lung epithelial cells had changed from type II to type I in that the cells were spread out flat on the interface and ceased to produce surfactant. Lung epithelial cells were first imaged using a light microscope (Zeiss Axiovert 200). For force spectroscopy, a medium-covered cover slip with the cells was mounted in the liquid cell of the atomic force microscope. The setup used allows for
226
Z. Leonenko et al.
the observation of the sample and the atomic force microscope cantilever across the cover slip from below by an inverted light microscope at the same time as force spectroscopy (or AFM imaging) is performed. The tip of the atomic force microscope was thus placed over a suitable cell and cell location was under light-microscope control (Fig. 21.10a). Figure 21.10b shows a typical force scan where the cantilever deflection was monitored as a function of the sample position. At first, the probe approached the cell with the lever undeflected (horizontal line of the force trace), indicating no interaction. The tip then made contact with the sample and became deformed. Forces of interaction between the probe and the cell were then measured in medium at 35 ◦ C, by approach of the tip to the cell interface until a preset load was reached (trace of force curve). The force was acquired as a function of the tip–sample separation h. After the tip has come into contact with the cell, h becomes the penetration depth for the tip into the cell. Separate force curves were acquired for a preset load of 100, 200, and 300 pN. The tip was now kept in contact with the cell for a predetermined time (delay time), while the preset load was kept constant under feedback control (i. e., when the load decreased, the tip was moved forward). We varied the delay time from 0 to 1800 s. Thereafter, the tip was retracted and the adhesion observed. Cantilevers (Micromash), with a spring constant of 57 mN/m, and pyramidal silicon nitride tips were used (tip height 2.9 nm, tip radius less than 20 nm (typical 10 nm), tip angle (face to face) 25−45◦ (top to around 300 nm down). Cells were kept in the medium at all times. The spring constant for each cantilever was measured before and after the experiment using the procedures implemented by the manufacturer of the atomic force microscope
Fig. 21.10. Optical microscopy image of epithelial type II cells and atomic force microscope cantilever. For force spectroscopy, an atomic force microscope tip was placed over the central region of a cell. Trace and retrace force curves were measured on a cell. Force versus tip–sample separation is shown in b. The approach part (trace) reveals how the cell deforms upon contact by the tip. The retrace or withdraw part of the curve reveals a large adhesion peak which is often split in multiple peaks. The sketches explain the physical situation for each part of the curve. The hatched area denotes the work of adhesion (adhesion energy) between the tip and the cell. The atomic force microscope tip was allowed to stay in contact for various times (delay time) in position $ 3
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
227
(thermal noise method). The sensitivity and the spring constant were found to be not different before and after the experiment. Forces of interaction between the probe and the cell were then measured in the medium at 35 ◦ C. 21.4.2 Mechanical Properties The approach part of the force curve was used to determine the penetration depth of the tip into the cell as well as the mechanical properties of the sample. Reviews on this topic are given in [94, 95].The measurement of the penetration depth will depend on the accurate determination of the contact point between the tip and the cell. A cantilever spring constant lower than or similar to that of the cell is preferable. In this case, we can assume that the contact point corresponds in the trace force curve to the first upward deflection. After the tip has come into the contact with the cell, D becomes the penetration depth h of the tip into the cell. To evaluate the mechanical properties of the sample from the approach force profile, the Sneddon model was employed [26, 96]. Note that the use of macroscopic physical law should be reserved for homogeneous bulk materials. Young’s modulus should be viewed as an effective parameter to analyze the trend of the in-plane organization of supramolecular structures. The effective parameter can differ from one experiment to another. A more sophisticated approach was to model the mechanical response of the cell membrane explicitly taking into account the mechanics of the cell membrane and that of the underlying cytoskeleton [97]. As the penetration depth h can reach 1 µm, the tip shape must be considered as a function of h. The effective Young’s modulus, E ′ , was estimated point by point for h considering the three regions determined by the geometry of the tip [98]. For h between 0 and 10 nm, the AFM probe was considered as a semisphere with a radius R of 10 nm : F 3 E sphere = √ 3/2 . 4 Rh
(21.6)
For further penetration, the tip profile was viewed as a cone whose half angle varies linearly from α1 = 25◦ at h = 10 nm to α2 = 45◦ when h reaches 300 nm: √ 2 F . (21.7) E cone = tan α h 2 For an indentation deeper than 300 nm, α is considered constant to α2 . Figure 21.11 shows the change in the Young’s modulus when increasing the penetration depth h. At first, the tip experienced a layer of increasing stiffness to a penetration depth of about 100 nm, at which point the resistance of the cell to penetration reached a maximum. Thereafter the cell became more compliant again. The stiff layer may represent the outer layer of the cytoskeleton anchored to the cell membrane. The Young’s modulus of the membrane can be thought of as the surface tension of the membrane (0.5 N/m) divided by the thickness of the membrane (around 10 nm), i. e., a pressure of around 50 MPa. The lipid head is responsible
228
Z. Leonenko et al.
Fig. 21.11. When analyzing the approach curve, one can determine the Young’s modulus of the sample surface. Young’s modulus, which gives a measure of cell elasticity in our case, was found to depend on indentation depth, showing two distinctive parts of the plot: when the peripheral part of the cell was penetrated, up to about 100 nm into the cell, the cell became increasingly stiffer. The stiffness then dropped sharply upon further penetration until it leveled out after penetration to about 200 nm deep
Fig. 21.12. Top: Indentation depth as a function of time. The load was kept constant at 200 pN. At time zero, the indentation was about 500 nm. Over time, the tip first penetrated rapidly and then slowly from 500 to 1300 nm. This is evidence of a time-dependent response of the cell to the tip. Bottom: Adhesion energy as a function of time. The tip was kept in contact with the cell at a constant load of 200 pN. The longer the tip stays in contact with the cell, the larger the adhesion that is observed. Statistical analysis shows that considerable changes occur during the first 100 s, after which the mean value of the adhesion energy no longer changes
for crystalline elasticity, i. e., the change in the head-to-head distance can change (high elasticity but low rupture). The lipid chain behaves with rubber elasticity (low elasticity), which is related to the change of entropy (increase of order upon pressure). It results in the properties that the cell membranes can be easily deformed in the plane but do not resist the expansion (increase between the headgroup cannot vary more than 2%).
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
229
Further measurements of the Young’s modulus of cells have been performed with human platelet cells (E = 1−50 kPa) [99], cardiac cells (E = 100 kPa), skeletal muscle cells (E = 25 kPa), and endothelial cells (1–7 kPa) [100]. The analysis of the forced indentation of the epithelial cells by the tip reveals the fate of a nanoparticle coming into contact with the lung’s epithelium and then being actively loaded onto a cell. This situation will indeed occur in the lung. The lung’s epithelium towards the air is covered by a thin aqueous layer and a molecular film of pulmonary surfactant at the air–water interface. Transmission electron microscopy of lung thin sections as well as scanning electron microscopy of the lung have indeed shown that particles trapped between the surfactant layer and the epithelial cells strongly deform the cells [101, 102]. The penetration depth was also investigated as a function of the contact time between the tip and the cell at a given load. Figure 21.12 shows that the tip first penetrated the cell down to 500 nm in less than 100 s. Equilibrium was reached for approximately 1-µm depth (around half of the cell thickness). We assume that the tip is either actively taken up or the cell rearranges its plasma membrane and the cytoskeleton elements to accommodate the tip. 21.4.2.1 Adhesion Work After the initial indentation, the tip was kept loaded onto the cell with 1 nN for preset times ranging from 0 to 900 s under feedback control. Thereafter, the tip was retracted and the tip adhesion was acquired. The retrace curve was used to calculate the adhesion energy Γ between the tip and the cell: Γ =
h2 h1
Fretrace (h i )(h i+1 − h i ) .
(21.8)
Fretrace (h i ) is the adhesion force at h i . The point of the initial contact between the tip and the cell surface, h 1 , was determined from the approach part of the force curve; h 2 corresponds to the last jump out of contact. Upon the penetration into the cell, the adhesion energy also increases. This indicates that the tip becomes engulfed by the membrane and the area of interaction increases (Fig. 21.12). Our results suggest that the process of particle uptake by lung epithelial cells occurs over approximately 100 s. This appears to be the time necessary for the nanoparticle to increase its surface area in contact with the cell and, hence, its adhesive interaction. Other authors have also shown [103] that the increase in adhesion between a polystyrene sphere to mica with increasing load or contact time is due to the plastic deformation. Load dependence and contact time dependence also indicate plastic and viscoelastic deformation [104]. A linear dependence of adhesion force on the reduced radius Reff = R1 R2 /(R1 + R2 ), where R1 and R2 are the radii of the two particles, was found by Heim et al. [105]. Particles of 50–100 nm in diameter are not phagocytosed [106, 107]. They enter cells in the absence of clathrin and caveolin, associated with active uptake. Even
230
Z. Leonenko et al.
Fig. 21.13. Correlation between the indentation depth (Fig. 21.12, top) and the adhesion energy (Fig. 21.12, bottom) for all the measurements (around 100 points). The deeper the indentation, the higher is the adhesion energy
when an actin-based mechanism has been ruled out [107], nanoparticles inside a cell are not surrounded by membrane and, hence, are not taken up through vesicle formation. The AFM investigation suggests the adhesive interaction in particle uptake can now be measured by our approach and used as a significant parameter test. Particles with different surface properties and different geometries can be mimicked by modified atomic force microscope tips. The adhesive interaction of a particle with the plasma membrane in the lung depends also on the particle history. Airborne particles, after crossing a surfactant layer at the air–water interface of a cell culture dish, have been shown to be more readily taken up than particles added to the media [108]. They may have become coated with a film of pulmonary surfactant and, as a consequence, have penetrated the epithelial cells differently. Whether this is related to a change in adhesive interaction can be tested by coating atomic force microscope tips with surfactant.
21.5 Conclusions Measurements of forces with the atomic force microscope can determine the local changes in the mechanical and electrical properties induced by the absorption of proteins within lipid monolayers or bilayers. Such studies have offered new perspectives in the identification and characterization of the pulmonary surfactant films in air, model membranes in liquid media, as well as cell membranes in media. In air, the adhesion forces of particles are known to be 10 times higher than in aqueous media. For a given applied load, the differences in the magnitude of the force do not depend on the contact area, which is similar in air or in a liquid. In fact, the van der Waals forces are screened under a liquid, thereby reducing the total adhesion forces. However, the adhesion forces can be modulated by varying the
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes
231
contact area between the particle and the layer. We have shown that the temperature or the anesthetic incorporation can largely modify the fluidity of the membrane. Using the AFM probe as a nanoparticle model is an idea to evaluate the ability of the particle to interact with the cell. Although many kinds of force curve cannot yet be interpreted, for instance, the role of the hydrophobic attraction between tails, our work is a starting point to understand how the particle can penetrate through the lipid layer. For example, the nonnegligible role of the electrical forces in air has been proved and the electrical potential of the layer is of importance in the adsorption process involving the electrical barrier. The mechanical properties of the cell, and in general the plastic deformation, have been investigated. The role of the membrane as a mechanical barrier is clearly observed by measuring the Young’s modulus as a function of the penetration depth. Those properties depend on many parameters: the particle speed, the particle size, the applied load. Also, the duration of the contact is crucial in the incorporation ability of the particles. The complexity of the phenomenon cannot of course be satisfactorily and fully described, but some empirical trend can be locally and statistically defined. Acknowledgements. The authors gratefully acknowledge financial support from the National Sciences and Engineering Research Council of Canada, the Canadian Institute for Health Research, the University of Calgary, and the University of Burgundy and also thank C. Le Grimellec and M. Moskovits for critical reading of the manuscript and helpful discussion.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Cullis PR, Dekruijf B (1979) Biochim Biophys Acta 559:399–420 Seelig J, Seelig A (1980) Q Rev Biophys 13:19–61 Brown DA, London E (1998) J Membr Biol 164:103–114 Nagle JF, Tristram-Nagle S (2000) Biochim Biophys Acta Rev Biomembr 1469:159–195 Bloom M, Evans E, Mouritsen OG (1991) Q Rev Biophys 24:293–397 Brown DA, London E (2000) J Biol Chem 275:17221–17224 Brown DA, London E (1998) Annu Rev Cell Devel Biol 14:111–136 Vangolde LMG, Batenburg JJ, Robertson B (1988) Physiol Rev 68:374–455 Lewis JF, Jobe AH (1993) Am Rev Respir Dis 147:218–233 Possmayer F (1988) Am Rev Respir Dis 138:990–998 Vedhuizen R, Nag K, Orgeig S, Possmayer F (1998) Biochim Biophys Acta Mol Basis Dis 140:890–108 Rooney SA, Young SL, Mendelson CR (1994) FASEB J 8:957–967 Tamm LK, McConnell HM (1985) Biophys J 47:105–113 Tien HT, Ottova AL (1998) Electrochim Acta 43:3587–3610 Heyse S, Ernst OP, Dienes Z (1998) Biochemistry 37:507–522 Muller RH, Mader K, Gohla S (2000) Eur J Pharma Biopharma 50:161–177 Leonenko ZV, Carnini A, Cramb DT (2000) Biochim Biophys Acta 1509:134–140 Zasadzinski J, Helm CA, Longo ML, Weisenhorn AL, Gould SAC, Hansma PK (1991) Biophys J 59:755–760 Hui SW (1995) Mol Membr Biol 12:45–50 Singh S, Keller DJ (1991) Biophys J 60:1401–1410 Beckmann M, Nollert P, Kolb HA (1998) J Membr Biol 161:227–233
232
Z. Leonenko et al.
22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.
Han W, Lindsay SM, Jing T (1996) Appl Phys Lett 69:4111–4113 Han W, Lindsay SM (1998) Appl Phys Lett 72:1656–1658 Fasolka MJ, Mayes AM (2001) Annu Rev Mater Res 31:323–355 Lee GU, Kidwell DA, Colton RJ (1994) Langmuir 10:354–357 Cappella B, Dietler G (1999) Surf Sci Rep 34:1–104 Senden TJ, Drummond CJ, Ke’kicheff P (1994) Langmuir 10:358–362 Cleveland JP, Schäffer TE, Hansma PK (1995) Phys Rev B 52:8692–8695 Israelachvili J (2005) Q Rev Biophys 38:331–337 Livadaru L, Kovalenko A (2006) Nano Lett 6:78 Biggs S (1995) Langmuir 11:156–162 Dufrene YF, Boland T, Schneider JW, Barger WR, Lee GU (1998) Faraday Discuss 111:79 Finot E, Lesniewska E, Mutin J-C, Goudonnet J-P (1999) J Chem Phys 111:6590–6598 Finot E, Lesniewska E, Mutin J-C, Goudonnet J-P (2000) Langmuir 16:4237–4244 Finot E, Lesniewska E, Mutin J-C, Goudonnet J-P (2000) Appl Surf Science 161:316–322 Butt HJ, Cappella B, Kappl M (2005) Surf Sci Rep 59:1–152 Finot E, S Roualdes, Kirchner M, Rouessac V, Berjoan R, Durand J, Goudonnet J-P, Cot L (2002) Appl Surf Sci 187:326–338 Mollica V, Relini A, Rolandi R, Bolognesi M, Gliozzi A (2000) Eur Phys J E 3:315 Van EW der Vegte, Hadziioannou G (1997) Langmuir 13:4357–4368 Dimitriadis EK, Horkay F, Maresca J, Kachar B, Chadwick RS (2002) Biophys J 82:2798– 2810 Legay G, Finot E, Meunier-Prest R, Cherkaoui-Malki M, Latruffe N, Dereux A (2005) Biosens Bioelectron 21:627–636 Franz V, Loi S, Muller H, Bamberg E, Butt H-J (2002) Colloids Surf B 23:191–200 Butt HJ, Franz V (2002) Phys Rev E 66:1601 Loi S, Sun G, Franz V, Butt HJ (2002) Phys Rev E 66:1602 Tenchov B, Koynova R, Rapp G (2001) Biophys J 80:1873–1890 Meyer HW, Semmler K, Rettig W, Pohle W, Ulrich AS, Grage S, Selle C, Quinn PJ (2000) Chem Phys Lipids 105:149–166 Bizzotto D, Nelson A (1998) Langmuir 14:6269–6273 Tamm LK (1985) Biol Chem Hoppe-Seyler 366:859–859 Kalb E, Frey S, Tamm LK (1992) Biochim Biophys Acta 1103:307–316 Nollert P, Kiefer H, Jahnig F (1995) Biophys J 69:1447–1455 Exerowa D (2002) Adv Colloid Interface Sci 96:75–100 Ebel H, Grabitz P, Heimburg T (2001) J Phys Chem B 105:7353–7360 Giocondi MC, Vie V, Lesniewska E, Milhiet PE, Zinke-Allmang M, Le C Grimellec (2001) Langmuir 17:1653–1659 Leonenko Z, Cramb D (2002) Nano Lett 2:305–309 Leonenko ZV, Merkle DD, Lees-Miller SP (2002) Langmuir 18:4873–4884 Tokumasu F, Jin AJ, Dvorak JA (2002) J Electron Microsc 51:1–9 Leonenko Z, Cramb D (2004) Can J Chem 82:1–11 Leonenko Z, Finot E, Cramb D (2006) Biochim Biophys Acta Biomembr 1758:487–492 Leonenko ZV, Carnini A, Cramb DT (2000) Biochim Biophys Acta 1509:131–147 Brian AA, McConnell HM (1984) Proc Natl Acad Sci USA 81:6159–6163 Leonenko ZV, Finot E, Ma H, Dahms T, Cramb DT (2004) Biophys J 86:3783–3793 Shamrakov LG, Leonenko Z, Finot E, Cramb DT (2006) Aust J Chem 59:386–389 Biltonen RL, Lichtenberg D (1993) Chem Phys Lipids 64:129–142 Yang J, Appleyard J (2000) J of Phys Chem B 104:8097–8100 Binnig G, Quate CF, Gerber C (1986) Phys Rev Lett 56:930–933 Israelachvili J (1992) Intermolecular and surface forces. Academic, New York Williams JM, Han T, Beebe TP (1996) Langmuir 12:1291–1295
38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67.
21 Atomic Force Microscopy: Interaction Forces in Liquid and Cell Membranes 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105.
233
Marrink SJ, Berger O, Tieleman P (1998) Biophys J 74:931–943 Evans K, Evans MF, Laker MF (1995) Ann Clin Biochem 32:261–280 Butt HJ (1991) Biophys J 60:1438–1444 Butt HJ, Franz V (2002) Phys Rev E 66:1601 Carnini A, Phillips HA, Shamrakov LG, Cramb DT (2004) Can J Chem 82:1139–1149 Garcia-Manyes S, Oncins G, Sanz F (2005) Biophys J 89:4261–4274 de Gennes PG (1987) Adv Colloid Interface Sci 27:189–209 Petrache HI, Dodd SW, Brown MF (2000) Biophys J 79:3172–3192 Nagle JF, Tristram-Nagle S (2000) Biochim Biophys Acta Rev Biomembr 1469:159–195 Janiak MJ, Small DM, Shipley GG (1979) J Lipid Res 20:183–199 Gunasekara L, Schurch S, Schoel WM (2005) Biochimi Biophys Acta Mol Cell Biol Lipids 1737:27–35 Leonenko Z, Finot E, Amrein M (2007) Ultramicroscopy 107:948–953 Leonenko Z, Rodenstein M, Dohner J, Eng LM, Amrein M (2006) Langmuir 22:10135– 10139 Leonenko Z, Gill S, Baoukina S, Monticelli L, Doehner J, Gunasekara L, Felderer F, Rodenstein M, Eng LM, Amrein M (2007) Biophys J 93:674–683 Chiu SW, Jakobsson E, Scott HL (2001) J Chem Phys 114:5435–5443 Smondyrev AM, Berkowitz ML (1999) J Chem Phys 111:9864–9870 Smondyrev AM, Berkowitz ML (1999) J Chem Phys 110:3981–3985 Atkinson RW, Anderson HR, Strachan DP, Bland JM, Bremner SA, de Leon AP (1999) Eur Respir J 13:257–265 Oberdorster G, Sharp Z, Atudorei V, Elder A, Gelein R, Lunts A, Kreyling W, Cox C (2002) J Toxicol Environ Health 65:1531–1543 Geiser M, Rothen-Rutishauser B, Kapp N, Schurch S, Kreyling W, Schulz H, Semmler M, Hof VI, Heyder J, Gehr P (2005) Environ Health Perspect 113:1555–1560 Nemmar A, Hoet PHM, Vanquickenborne B, Dinsdale D, Thomeer M, Hoylaerts MF, Vanbilloen H, Mortelmans L, Nemery B (2002) Circulation 105:411–414 Takenaka S, Karg E, Roth C, Schulz H, Ziesenis A, Heinzmann U, Schramel P, Heyder J (2001) Environ Health Perspect 109:547–551 Chen H, Langer R, Edwards DA (1997) J Colloid Interface Sci 190:118–133 Bresme F, Quirke N (1999) J Chem Phys 110:3536–3547 Shanahan MER (1990) J Phys D Appl Phys 23:321–327 Le C Grimellec, Lesniewska E, Giocondi MC et al (1998) Biophys J 75:695–703 Costa KD (2003) Dis Markers 19:139–154 Radmacher M (2002) Methods Cell Biol 68:67–90 Sneddon JN (1965) Int J Eng Sci 3:47 Mc MElfresh, Baesu E, Balhorn R, Belak J, Allen MJ, Rudd RE (2002) Proc Natl Acad Sci USA 99:6493–6497 Stark RW, Drobek T, Weth M, Frickef J, Heckel WM (1998) Ultramicroscopy 75:161 Radmacher M, Fritz M, Kacher CM, Cleveland JP, Hansma PK (1996) Biophys J 70:556– 567 Mathur AB, Collinsworth AM, Reichert WM, Kraus WE, Truskey GA (2001) J Biomech 34:1545–1553 Bachofen H, Schurch S, Possmayer F (1994) J Appl Physiol 76:1983–1992 Gehr P, Geiser M, Im V Hof, Schurch S, Waber U, Baumann M (1993) Microsc Res Tech 26:423–436 Biggs S, Spinks G (1998) J Adhes Sci Technol 12:461–478 Reitsma M, Craig VSJ, Biggs S (2000) J Adhes 74:125–142 Heim LO, Blum J, Preuss M, Butt H-J (1999) Phys Rev Lett 83:3328–3331
234
Z. Leonenko et al.
106. Geiser M, Rothen-Rutishauser B, Kapp N, Schurch S, Kreyling W, Schulz H, Semmler M, Hof VI, Heyder J, Gehr P (2005) Environ Health Perspect 113:1555–1560 107. Rothen-Rutishauser BM, Schurch S, Haenni B, Kapp N, Gehr P (2006) Environ Sci Technol 40:4353–4359 108. Blank F, Rothen-Rutishauser BM, Schurch S, Gehr P (2006) J Aerosol Med 19 (in press)
22 Self-Assembled Monolayers on Aluminum and Copper Oxide Surfaces: Surface and Interface Characteristics, Nanotribological Properties, and Chemical Stability E. Hoque · J. A. DeRose · B. Bhushan · H. J. Mathieu
Abstract. The formation of self-assembled monolayer (SAM) films onto aluminum and copper oxide surfaces by reaction with 1H,1H,2H,2H-perfluorodecylphosphonic acid (PFDP), octadecylphosphonic acid (ODP), decylphosphonic acid (DP), octylphosphonic acid (OP), and 1H,1H,2H,2H-perfluorodecyldimethylchlorosilane (PFMS) is discussed in this chapter. The properties and chemical stability of the films have been investigated using complementary surface analysis techniques: X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), friction force microscopy (FFM), a derivative of AFM, contact angle measurements, and Fourier transform infrared reflection/absorption spectroscopy (FT-IRRAS). XPS data confirm the presence of alkylphosphonate (perfluorinated and nonperfluorinated) and perfluorosiloxy molecules in the PFDP/Al, ODP/Al, DP/Al, OP/Al, and PFMS/Al SAMs formed on aluminum oxide surfaces. The sessile drop static contact angles of deionized water on PFDP/Al and PFMS/Al are typically more than 130◦ and on ODP/Al, DP/Al, and OP/Al typically more than 125◦ , indicating that Al surfaces reacted with alkylphosphonic acids and alkylsilanes are very hydrophobic. The surface roughness for PFDP/Al, ODP/Al, DP/Al, OP/Al, PFMS/Al, and unmodified Al is approximately 35 nm, as determined by AFM. The critical surface tension for PFDP/Al has been determined to be approximately 11 mJm−2 (mN m−1 ) by the Zisman plot method compared with 16, 20, 21, and 25 mJm−2 for PFMS/Al, ODP/Al, DP/Al, and OP/Al, respectively. PFDP/Al gives the lowest adhesion and friction force, while unmodified Al gives the highest. The adhesion and friction forces for ODP/Al and DP/Al SAMs are in-between those of PFDP/Al and Al. The influence of relative humidity, temperature, and sliding velocity on the friction and adhesion behavior has also been studied. Failure mechanisms of SAMs have been investigated by wear tests. The chemical stability of ODP/Al, PFDP/Al, DP/Al, OP/Al, and PFMS/Al SAMs has been tested by exposure to warm nitric acid (pH 1.8, 30 min, 60– 95 ◦ C). The XPS data and stability against harsh chemical conditions indicate that a type of bond forms between a phosphonic acid or silane molecule and the oxidized Al surface. Stability tests using warm nitric acid (pH 1.8, 30 min, 60–95 ◦ C) show ODP/Al SAMs to be most stable, followed by PFDP/Al, DP/Al, PFMS/Al, and OP/Al. Hydrophobic, low adhesion, and robust Al surfaces have useful applications for microelectromechanical/nanoelectromechanical systems (MEMS/NEMS), such as the digital micromirror device. These studies are expected to aid in the design and selection of proper lubricants and antistiction coatings for MEMS/NEMS. The PFMS SAM on Cu is found to be extremely hydrophobic, typically having sessile drop static contact angles of more than 130◦ for deionized water and a critical surface tension of 14 mJ m−2 . FFM shows a significant reduction in the adhesive force and friction coefficient of PFMS-modified Cu (PFMS/Cu) compared with unmodified Cu. Treatment by exposure to harsh conditions shows that a PFMS/Cu SAM can withstand boiling nitric acid (pH 1.8), boiling water, and warm sodium hydroxide (pH 12, 60 ◦ C) solutions for at least 30 min. Furthermore, no SAM degradation is observed when PFMS/Cu is exposed to warm nitric acid solution for up to 70 min at 60 ◦ C or 50 min at 80 ◦ C. XPS and FT-IRRAS data reveal a coordination of the PFMS Si atom with a cuprate (CuO) molecule present on the oxidized Cu substrate. The data give good evidence that the stability of the SAM film on the PFMS-modified oxidized Cu surface is largely due to
236
E. Hoque et al.
the formation of a siloxy–copper (–Si–O-Cu–) bond via a condensation reaction between silanol (–Si–OH) and copper hydroxide (CuOH). Extremely hydrophobic (low surface energy) and stable PFMS/Cu SAMs films could be useful for surface passivation, corrosion inhibition and/or as antiwetting/low-adhesion promoters in microelectronical/nanoelectromechanical devices or on heat-exchange surfaces (dropwise condensation). Key words: Self-assembled monolayer, Aluminum, Copper, Microtribology/nanotribology, Chemical stability, Phosphonic acid, Silane, X-ray photoelectron spectroscopy, Atomic force microscopy, Friction force microscopy, Hydrophobic, Surface energy, Robust, Adhesion, Friction, Wear, Siloxy–copper, Condensation reaction, Dropwise condensation
22.1 Introduction A self-assembled monolayer (SAM) is the spontaneous assembly of a single layer of molecules on an appropriate solid surface via chemisorption from solution or the gas phase. The molecule consists of a head group that chemisorbs onto a substrate surface, a tail group that makes up the outer surface of the SAM film, and a spacer (backbone) chain that connects the head and tail groups (Fig. 22.1). The degree of close packing in SAM films depends on many factors, such as molecular structure, intermolecular (electrostatic and dipole–dipole) interactions, and the affinity of the head group for the substrate surface. SAMs provide a convenient, flexible, and simple method for engineering the interfacial properties of metal, metal oxide, and semiconductor surfaces. SAMs especially provide a wide range of properties suitable for many technological applications which offer a number of advantages over other films: (1) chemisorption of the molecules, resulting in a strong adhesion to the underlying substrate surface; (2) nearly any substrate can be coated; (3) the thickness of the film can be varied with nanometer precision by changing the molecular chain length; (4) the chemical composition of the SAM can be controlled, allowing a wide range of groups to be incorporated in both the alkyl chain and the
Fig. 22.1. A schematic for chemisorbed self-assembled momolayer (SAM) molecules on a substrate surface
22 Self-Assembled Monolayers on Al and Cu
237
tail group; (5) the films are generally crystalline; and (6) the wetting (hydrophilic versus hydrophobic) properties can be fine-tuned by altering the tail group (from low-energy trifluoromethyl (–CF3 ) groups to high-energy hydroxyl (–OH) or carboxyl (–COOH) groups) [1,2]. SAMs are versatile and have potential for application in many areas [3]. Some of the potential applications of SAMs include (1) protective coatings against corrosion [4–7], (2) lubrication [8–12], (3) reduction or enhancement of adhesion [13], (4) chemical binding, (5) biological sensors, (6) microcontact printing, (7) nonlinear optics, and (8) control of electronic properties at organometallic interfaces [3]. The development of electronic devices where the active constituents are organic molecules (rather than traditional solid-state semiconductors based on Si or GaAs) has attracted considerable attention over the last 10 years [14, 15]. Theoretical studies suggest that organic molecules might be able to mimic the function of common circuit elements utilized in microelectronics—conductors, rectifiers, transistors, logic gates [16, 17]. The physical properties of SAMs that remain relatively unexplored are magnetic/electro-optical responses, quantum effects, and electrowetting [18]. Owing to ever-increasing miniaturization of electronics and advances in microelectromechanical systems/nanoelectromechanical systems (MEMS/NEMS), SAMs on Al surfaces are an area of immense interest for a wide range of applications [19– 21]. Typical SAMs, such as alkanethiols reacted with Au and Cu [22, 23], alkylsilanes reacted with silica [23], n-alkanoic acids reacted with oxidized Al [24, 25], and alkylphosphonic acids reacted with tantalum oxide [26, 27], have been studied extensively. However, not much information is found in the literature on phosphonic acid based SAMs on Al [28–31]. Phosphonic acids are one important class of selfassembling molecules owing to their ability to react with a range of metal oxides [28]. The friction properties of alkanethiols [32] and alkylsilanes [33] have been studied as a function of the alkyl chain length, with the conclusion that shorter-chain SAMs exhibit higher friction coefficients than longer-chain ones owing to a more disordered film structure. Recently, Al-based substrates have gained increasing attention, particularly after the inception of the digital micromirror device [34,35]. The efficiency, power output, and steady-state operation of MEMS/NEMS devices can be critically influenced by adhesion, friction, and wear [36–41]. The necessity for an ultrathin lubricant film, such as a SAM, to minimize adhesion, friction, and wear between surfaces in contact for MEMS/NEMS is clear. The poor durability of most SAMs under ambient or more extreme conditions makes them impractical for most applications. The introduction of fluorocarbon groups in the alkyl chain of the silane or phosphonic acid molecule causes the SAM to be significantly more stable compared with one composed of only hydrocarbon (nonperfluorinated) groups [22, 42]. Recently, the properties, i. e., chemical composition, stability, hydrophobicity, surface energy, root mean square (rms) roughness, adhesion, and friction, of perfluoroalkylsiloxy and perfluoroalkylphosphonate and non-perfluoroalkylphosphonate SAMs on Al have been studied and compared using surface sensitive techniques, such as X-ray photoelectron spectroscopy (XPS), contact angle measurement (CAM), atomic force microscopy (AFM), and friction force microscopy (FFM) [10, 11, 19, 42]. SAM formation via reaction of alkylsilanes and alkylphosphonic acids with Al surfaces relies on hydroxylation of the oxide (alumina) layer. Generally, chemisorption of alkylphosphonic acids occurs by proton dissociation to
238
E. Hoque et al.
form an alkylphosphonate species. Silanols (–Si–OH) (hydrolyzed silanes) and phosphonic acids undergo a condensation reaction with surface-bound aluminohydroxyl (−Al−OH) species to form siloxyaluminum and aluminophosphonate compounds plus H2 O as a by-product: R3 Si−OH + −Al−OH → R3 Si−O−Al− + H2 O and R−PO(OH)2 + −Al−OH → R−(OH)OP−O−Al− + H2 O [42]. SAMs formed by reaction of alkylsilanes with silicon oxides and alkylthiols with non-oxidized Cu and Au have been investigated extensively [5, 43–45]. The poor durability of these films under ambient or more extreme conditions is a major concern for the implementation of SAMs in everyday applications. Formation of SAMs on Cu substrates has attracted interest for both fundamental and applied research, e. g., corrosion inhibition [6,7,46–48] of Cu used in microelectronic device fabrication and heat-exchange surfaces which exploit dropwise condensation [49– 52]. Cu is destined to replace Al alloy as the principle very large scale integration or ultra-large-scale integration interconnection material due to its lower electrical and thermal resistivity and greater resistance to electromigration. Despite these advantages, one of the major obstacles for using it as an interconnection material is its susceptibility to corrosion and oxidation. Although alkylthiol SAMs and thin films on Cu have shown the ability to inhibit corrosion [6, 22, 46], the durability of alkylthiol/Cu SAM films is insufficient for most practical applications. Heat transfer is most efficient via condensation of a fluid medium onto a heat-exchange surface. Dropwise condensation has been known to produce heat -transfer coefficients around 10 times that of filmwise condensation and, for this reason considerable attention has been paid over the past few decades toward the development of suitable dropwisecondensation promoters. However, the main problem has been the durability of these promoters [49, 53]. Very recently, thorough studies of SAMs formed on oxidized Cu substrates via reaction with 1H,1H,2H,2H-perfluorodecyldimethylchlorosilane (PFMS) have appeared. The morphology, chemical composition, wetting properties, adhesive force, coefficient of friction, and stability of the PFMS/Cu SAM have been analyzed using complementary techniques: XPS, AFM, FFM, FT-IRRAS, and CAM [21, 54]. This chapter is based upon the published works of Hoque and coworkers [10, 11, 19, 21, 42, 54].
22.2 Substrate Preparation 22.2.1 Aluminum Two types of Al substrates have been used for alkylphosphonate and alkylsilane SAM formation: (1) Al deposited onto Si pieces and (2) pieces of commercially available Al sheet (mirror quality). Si pieces of approximately 10 mm × 15 mm × 0.5 mm
22 Self-Assembled Monolayers on Al and Cu
239
were cut from a commercially available Si wafer and ultrasonically cleaned in ethyl acetate, 2-propanol, and deionized (DI) water for 10 min each to minimize adsorbed contamination. In-between each cleaning step, samples are blown dry using Ar gas. Then, 99.98% pure Al was DC sputter coated (Edwards Coating System) in a high-vacuum chamber onto the Si pieces. The sample holder was rotated, passing sequentially above the Al target to produce a uniform Al film approximately 1 µm thick. Al pieces of approximately 10 mm × 12 mm × 0.5 mm were cut from a 300 mm × 210 mm × 0.5 mm specially produced, mirrorlike, polished Al plate (Alcan, Singen, Germany). The protective polymer layer was first peeled off the Al samples and then they were degreased ultultrasonically using the procedure employed for Si substrates. The final cleaning step (and simultaneously surface oxidation) was done using a radiofrequency oxygen plasma for 10 min with a power of 50 W and an O2 partial pressure of 0.3 Torr (Plasmaline 415, barrel-type asher, Tegal, CA, USA). The same cleaning method was applied to all Al substrates modified with phosphonic acids [11, 19, 20]. 22.2.2 Copper Circular pure Cu (99.98%) discs of 3-mm thickness and 12-mm diameter were used for PFMS/Cu SAMs. The Cu discs were mechanically polished with silicon carbide (SiC) paper of grades 180–4000 in series followed by cloth pad micropolishing using diamond paste and a lubricant. To make substrates with a variety of surface roughness values, the polishing was stopped at various steps (grades of SiC paper). After all the polishing steps had been performed, the Cu substrate had a highly reflective “mirror” finish. Then, all the substrates were cleaned in a similar way to that used for Si substrates. Finally, they were oxidized by immersion in a 10% H2 O2 solution for 5–10 min at room temperature [21, 22, 54].
22.3 Phosphonic Acid and Silane Based SAMs on Al and Cu 22.3.1 SAM Preparation 22.3.1.1 Materials All precursors and solvents used for making alkylphosphonate and alkylsilane SAMs were of standard commercial grade. They were used as received without any further purification. All substrates used were cleaned with 99.5% pure ethyl acetate (CH3 COOCH2 CH3 ) and 99.7% pure 2-propanol [(CH3 )2 CH–OH], both purchased from Merck (Darmstadt, Germany). The alkylsilane and alkylphosphonic acid molecules utilized for Al and Cu modification were PFMS with ten carbon atoms [CF3 (CF2 )7 (CH2 )2 (CH3 )2 SiCl; 97%; ABCR, Karlsruhe, Germany; CAS 74612-30-9]; n-octadecylphosphonic acid (ODP), with 18 carbon atoms
240
E. Hoque et al.
[H3 C(CH2 )17 PO(OH)2 ; 100%; Alpha Aesar, Karlsruhe, Germany; CAS 4724-474]; 1H,1H,2H,2H-perfluorodecylphosphonic acid (PFDP), with ten carbon atoms [CF3 (CF2 )7 (CH2 )2 PO(OH)2 ; 95%; Unimatec, Ibaraki-Ken, Japan; CAS 80220-639]; n-decylphosphonic acid (DP), with ten carbon atoms [H3 C(CH2 )9 PO(OH)2 ; 98%; ABCR, Karlsruhe, Germany; CAS 5137-70-2]; and n-octylphosphonic acid (OP), with eight carbon atoms [H3 C(CH2 )7 PO(OH)2 ; 99%; ABCR, Karlsruhe, Germany; CAS 4724-48-5]. The solvents used for making the alkylsilane and alkylphosphonic acid solutions were n-hexane [CH3 (CH2 )4 CH3 ; 98.5%] and ethanol (CH3 CH2 OH; 99.8%), both purchased from Merck (Darmstadt, Germany), respectively [10, 11, 19, 21, 42, 54]. 22.3.1.2 SAM Formation SAMs were formed on Al and Cu by liquid-phase deposition. A PFMS solution of 0.1 wt % was prepared using n-hexane as the solvent. Phosphonic acid solutions of 0.1 wt % were prepared by dissolving the phosphonic acid powder in ethanol (phosphonic acid solution must be agitated for 15 min in order to completely dissolve the powder). Oxidized Al and Cu substrates were immersed in the PFMS solution for 50 min at room temperature under ambient conditions. A reaction time of 50 min is sufficient to completely cover the aluminum and copper oxide surfaces with alkylsiloxy molecules [20–22, 54–56]. Oxidized Al substrates were dipped into phosphonic acid solution for a reaction time of 24 h at room temperature under ambient conditions. After each reaction, the samples were rinsed several times using fresh solvent in a beaker with hand agitation for approximately 30 s to remove physisorbed molecules from the substrate surface. Afterwards, the samples were blown-dry carefully with Ar gas, giving the final PFMS/Al, PFMS/Cu, ODP/Al, DP/Al, OP/Al, and PFDP/Al SAMs. Then, the PFMS/Al and PFMS/Cu samples were heated at 150 ◦ C under vacuum (base pressure of 3.3 × 10−2 mbar) for 1 h [10, 11, 19, 21, 42, 54]. 22.3.2 Surface and Interface Characterization XPS analysis was performed using a Kratos AXIS ULTRA spectrometer equipped with a concentric hemispherical analyzer in the standard configuration. All the spectra were acquired using a monochromatic Al Kαmono X-ray radiation source operated at 15.0 kV and 150 W in ultrahigh vacuum (base pressure below 10 × 10−9 mbar). The core level single peaks were measured using a pass energy of 20 eV and an acquisition time of 180 s. Samples were grounded (earthed) to prevent charging and electron charge compensation was also applied. Under these conditions, the energy resolution is calibrated by the full width at half maximum measured for the Ag 3d5/2 and Au 4 f 7/2 peaks, which is found to be 0.7 eV for both. The angle-resolved XPS (AR-XPS) measurements were done at different takeoff angles (detection angle), namely, 90◦ , 75◦ , 60◦ , 45◦ , 30◦ , and 20◦ , by tilting the sample holder with respect to the average plane of the substrate surface. The spectra were referenced to the
22 Self-Assembled Monolayers on Al and Cu
241
alkyl C 1s photoelectron peak at 284.8 eV. Quantitative analysis of elements on the surface was done using peak areas of the XPS multiplex after linear-type background subtraction. Then, deconvolution of high-resolution XPS peaks was done by mixed Gaussian–Lorentzian fit using standard CASA XPS processing software (version 2.2.95). The estimated relative error for all XPS data used for elemental quantification was ±2% [11, 19, 21, 42, 54]. CAM was performed using a GBX Digidrop apparatus. Sessile drop static CAMs were made with ultrapure H2 O (Fluka) at 40–60% relative humidity and 23–25 ◦ C to determine the wetting properties of nonreacted PFMS and phosphonic acid reacted Al and Cu, as well as the treated SAMs. The reported CAM values are the average of at least five measurements. Drops with typical volumes of less than 5 µl were used for static measurements. A drop of water was released from the Teflon tip of a syringe onto the sample surface for the static measurement. The absolute error for all contact angles determined by the instrument software is ±3◦ . Fourier transform IR reflection/absorption spectroscopy (FT-IRRAS) was carried out with a PerkinElmer SpectrumTM optical bench using a Spectratech grazing angle FT-80 system equipped with a wire grid polarizer. In some cases, FT-IRRAS spectra were obtained without the wire grid polarizer. Most measurements were done with polarized light at grazing incidence (80◦ angle of incidence with respect to the surface normal) and a 90◦ polarization angle (electric field vector, E, direction with respect to the sample surface; p-polarized). All spectra were recorded by collecting 256 scans over the 4000 to 400-cm−1 spectral region with a resolution of 4 cm−1 . The critical surface tension, γcrit , a measure of interfacial surface energy, was determined from the Zisman plot of the sessile drop static contact angle versus the alkane surface tension of the homologous series of n-alkane liquids: hexadecane (C16 H34 ; 27.5 mNm−1 surface tension), dodecane (C12 H24 ; 25.6 mNm−1 surface tension), undecane (C11 H24 ; 24.7 mNm−1 surface tension), and decane (C10 H22 ; 23.7 mNm−1 surface tension). The alkane contact angles were measured in the same manner as for the water CAM data at 23–25 ◦ C under atmospheric conditions. The critical surface tension is determined from the abscissa intercept of the best-fit plot of the cosine of the static contact angles (θ) versus the known surface tensions (measured at 20 ◦ C) for the n-alkane liquids (γcrit found at cos θ = 1) [22]. 22.3.2.1 Substrates Aluminum Figure 22.2 shows representative Al 2 p high-resolution XPS spectra (deconvoluted) taken on oxidized Al. The peaks deconvoluted at the binding energies of 71.2 and 71.6 eV correspond to metallic Al 2 p3/2 and Al 2 p1/2 , respectively. The fitted peaks with binding energies of 72.1 and 74.0 eV can be attributed to aluminum suboxide (AlOx ) and aluminum oxide (Al2 O3 ), respectively [19]. The chemical surface composition of clean and unmodified Al was determined by XPS to be 37, 49, and 14% for Al, O, and C, respectively. The presence of oxide on the Al surface is required for both the phosphonic acid and the silane reactions. The unmodified Al substrate was found to be hydrophilic with a water contact angle of less than 10◦ .
242
E. Hoque et al.
Fig. 22.2. Al 2 p core level high-resolution X-ray photoelectron spectroscopy (XPS) spectra of unmodified, oxidized Al. The full widths at half maximum (FWHM) of the deconvoluted peaks are between 0.4 and 1.5 eV. From [19]
Copper High-resolution core level XPS Cu 2 p3/2 and C 1s spectra are illustrated in Fig. 22.3. XPS is capable of resolving neutral Cu [Cu(0)], Cu with a valence of +1 [Cu(I)], and Cu with a valence of +2 [Cu(II)] through the analysis of the core level Cu 2 p peaks and the corresponding “shake-up” satellite peaks, as well as the X-ray excited Cu Auger electron L2,3 M4,5 M4,5 line. These so-called “shake-up” satellite peaks arise on the high binding energy side owing to the simultaneous excitation of outer-shell 3d electrons during the ejection of core level 2 p electrons. For the clean polished Cu surface, the intensity of the satellite peak feature is very low, as shown in Fig. 22.3a, relative to the main photoemission peak. However, for the polished and oxidized samples (Fig. 22.3b), the increase in intensity of the “shakeup” satellite structure feature (S1 and S2 ) extending from around 939 to 946 eV indicates an increasing proportion of cupric [Cu(II)] oxide, i. e., CuO, at the surface. So it is noteworthy that for the XPS spectra of nonoxidized Cu sample surfaces (Fig. 22.3a), the lack of the “shake-up” satellite peak indicates the existence of Cu mainly in a metallic Cu(0) or cuprous [Cu(I)], i. e., Cu2 O, form. The peak of Cu 2 p3/2 at 933.9 eV [57] as shown in Fig. 22.3b can unambiguously be attributed to CuO on the basis of the presence of the typical characteristic “shake-up” peak. The peak at 569.40 eV resulting from a Cu L2,3 M4,5 M4,5 Auger line (data shown later) corresponds to a Cu(II) species of CuO. It suggests that the oxidation process converts Cu/Cu2 O to CuO and/or Cu(OH)2 . The Cu 2 p3/2 component observed at 932.9 eV [21, 22, 58–60] is not readily assigned, because Cu(0) and Cu2 O have, within the resolution of conventional XPS measurements, identical binding energies (±0.1 eV) [46]. The deconvoluted third peak, at 935.1 eV, indicates the presence of Cu(OH)2 which is formed at some point between oxidation and silanization of the Cu. On the basis of the XPS data, it is clear that the polished and oxidized Cu surface contains Cu2 O and/or metallic Cu, CuO, and Cu(OH)2 with corresponding energies [21,22,54,58,60] of 932.9, 934.0, and 935.1 eV, respectively. The chemical surface composition of clean, oxidized, and unmodified Cu was determined from
22 Self-Assembled Monolayers on Al and Cu
Fig. 22.3. Deconvoluted highresolution core level Cu 2 p3/2 XPS spectra obtained for a clean, polished Cu, b clean, polished, and oxidized Cu, and c C 1s high-resolution deconvoluted XPS spectra taken on clean, polished, and oxidized Cu. The FWHM of the deconvoluted peaks are between 1.0 and 2.6 eV. (From [22])
243
244
E. Hoque et al.
XPS multiplexes and was found to be 36, 45, and 19% for Cu, O, and C, respectively. The presence of oxide on the Cu surface is needed for the silane reaction; therefore, in order to form SAM films with silane, the Cu surface must be oxidized. The characteristic oxidized carbon species of alcohol or ether type (C–OH/C–O–C) and ester, ketone, or carboxyl type (CO–O–C/C–CO–C/CO–O–H) adsorbed from the air are observed on clean polished and oxidized Cu surfaces (Fig. 22.3c) owing to atmospheric exposure between oxidation and XPS analysis [6, 22, 54]. Similar oxygenated carbon species were also observed in the deconvoluted C 1s spectra taken on unmodified Al [11]. The unmodified Cu substrate was found to be hydrophilic with a water contact angle of less than 15◦ for the range of surface roughness values considered. 22.3.2.2 Alkylphosphonic Acid and Alkylsilane Based SAMs on Al A schematic for SAM formation after reaction of OP or PFMS with oxidized Al is shown in Fig. 22.4. The most probable bonding scheme for OP with Al is monodentate as shown in Fig. 22.4a. The lattice spacing of Al may permit bidentate (Fig. 22.4b) or tridentate bonding (Fig. 22.4c) as well [61]. The most probable bonding scheme (monodentate) after reaction of PFMS with Al is shown in Fig. 22.4d. XPS Analysis PFDP SAM on Al. Figure 22.5 shows high-resolution deconvoluted XPS C 1s spectra performed for PFDP powder deposited onto an In foil (Fig. 22.5a) and a PFDP/Al SAM (Fig. 22.5b). The peak component in Fig. 22.5a at the binding energy of about 284.8 eV can be attributed to C–H/C–C groups which arise from the methylene (–CH2 –) groups of the PFDP molecules and partly from adventitious carbon contamination. The characteristic oxidized carbon species of alcohol or ether type (C–OH/C-O–C) and ester, ketone, or carboxyl type (CO–O–C/C–CO–C/CO–OH) adsorbed from the ambient air are observed [6, 19, 22]. The peaks (Fig. 22.5a) at high binding energies of 290.1 and 292.3 eV can be attributed unambiguously to the difluoromethylene (–CF2 –) functional backbone and trifluoromethyl (–CF3 ) tail groups of the PFDP molecule [62]. The ratio of the peak areas for the CF2 and CF3 functional groups is 6.8±0.3, which agrees well with the expected value of 7. The same functional groups are found in the deconvoluted C 1s XPS spectra (Fig. 22.5 b) of PFDP/Al. The ratio of the peak areas for the –CF2 – and –CF3 functional groups is 6.2±0.3, which is close to 7 as well. The binding energies of all the deconvoluted functional groups are listed in Table 22.1. These results, plus the fact that phosphorous was also detected on PFDP/Al (data shown latter), confirm the presence of a chemically homogeneous perfluorophosphonate film on the Al surface. XPS data revealed that airborne carbonaceous contaminants (Fig. 22.3c) are present on the unmodified Al. It is also noteworthy that oxygenated carbon contamination species are also found in the deconvoluted C 1s XPS spectra for PFDP/Al (Fig. 22.5b) as well. The organic contamination species
22 Self-Assembled Monolayers on Al and Cu
245
Fig. 22.4. A proposed bonding scheme for the reaction of n-octylphosphonic acid (OP) or 1H,1H,2H,2H-perfluorodecyldimethylsilane (PFMS) with aluminum oxide. The OP bond with Al in the a monodentate, b bidentate, or c tridentate forms. d A proposed bond (monodentate) after reaction of PFMS with oxidized Al
that remain adsorbed on the surface, despite the reaction with PFDP, are in very low quantities compared with perfluorophosphonates (quantities determined from the XPS and CAM data). The complete displacement of adsorbed contamination would provide a higher surface density of adsorption sites, resulting in a more closely packed array of molecules and, consequently, higher water contact angles [19]. Interface analysis reveals information about the binding of molecules with the substrate. Figure 22.6 shows representative deconvoluted spectra for O 1s taken on unmodified Al, a PFDP/Al SAM, and PFDP powder pressed onto an indium (In) foil (PFDP/In) and for Al 2 p taken on a PFDP/Al SAM. The O 1s and Al 2 p spectra for a PFDP/Al SAM were taken with a takeoff angle of 20◦ . The O 1s spectrum for Al is resolved into three components as shown in Fig. 22.6a. The components
246
E. Hoque et al.
Fig. 22.5. Deconvoluted high-resolution core level C 1s XPS spectra carried out on a 1H,1H,2H,2H-perfluorodecylphosphonic acid (PFDP) in bulk and b PFDP/Al. The FWHM of the deconvoluted peaks are between 1.0 and 2.3 eV. (From [42]) Table 22.1. The corresponding binding energy values for functional groups from deconvoluted peaks in the C 1s high-resolution X-ray photoelectron spectroscopy (XPS) spectra measured for 1H,1H,2H,2H-perfluorodecylphosphonic acid (PFDP) powder on In, PFDP/Al, and 1H,1H,2H,2H-perfluorodecyldimethylchlorosilane (PFMS)/Al. (From [42]) Substrates
CF3
CF2
CO–O–C/ C–CO–C/ C–CO–H
C–O
C–H/C–C
PFDP/In PFDP/Al PFMS/Al
292.3 292.6 292.4
290.1 290.4 290.2
287.8 287.8 287.6
285.7 285.7 285.4
284.8 284.8 284.8
22 Self-Assembled Monolayers on Al and Cu
247
at the binding energies of 530.6 eV (I), 531.6 eV (II) , and 532.6 eV (III) can be attributed to aluminum oxide, AlOH, and –C–O– (adsorbed), respectively [42]. The O 1s spectrum for a PFDP SAM is resolved as illustrated in Fig. 22.6b. Components I, II, and III are found at binding energies of 530.5, 531.3, and 532.8 eV, respectively. A significant reduction of adsorbed organic contamination is observed for PFDP/Al. The O 1s spectrum for PFDP/In is fitted with two components as shown in Fig. 22.6c which are assigned to P=O and P–OH at binding energies of 530.8 and 532.2 eV, respectively. After the reaction of PFDP with Al, the two peaks for P=O and P-OH observed on PFDP/In disappeared, presumably due to the binding of PFDP with the Al surface, probably resulting in an Al–O–P bond with a binding energy of 531.3 eV (Fig. 22.6b). Figure 22.6d shows the resolved Al 2 p spectra for a PFDP/Al SAM.
Fig. 22.6. Resolved core level high-resolution XPS spectra: a O 1s for clean, oxidized (unmodified) Al, b O 1s for a PFDP/Al SAM [resolved peaks are aluminum oxides (I), –AlOH (II), and –CO– (III)], c O 1s for bulk PFDP (PFDP/In), and d Al 2 p for a PFDP/Al SAM. The FWHM of the fitted peaks are between 0.5 and 1.5 eV for Al 2 p and between 1.6 and 1.7 eV for O 1s. (From [42])
248
E. Hoque et al.
Fig. 22.6. (continued)
The components at binding energies of 71.2 and 71.7 eV correspond to metallic Al 2 p3/2 and Al 2 p1/2 peaks, while the peak at a binding energy of 72.2 eV can be assigned to aluminum suboxide (AlOx ). The peak with a binding energy of 73.9 eV is attributed to aluminum oxide. The additional peak at a binding energy of 74.6 eV could be due to the presence of aluminophosphonate groups (Al–O–P) [42]. Nonperfluorinated Alkylphosphonic Acid SAMs on Al. C 1s high-resolution XPS spectra obtained for hydrocarbon phosphonate SAMs on ODP/Al, DP/Al, and OP/Al are presented in Fig. 22.7. The C 1s core level high-resolution XPS spectra measured for ODP/Al, DP/Al, and OP/Al are fitted as illustrated in Fig. 22.7a–c, respectively. Peaks at a binding energy of 284.8 eV (Fig. 22.7) can be attributed unambiguously to C–C/C–H functional groups which arise mostly from the –CH2 and –CH3 functional backbone (spacer) and tail groups of the phosphonate molecules and partly from carbon contamination due to ambient exposure [11, 62].
22 Self-Assembled Monolayers on Al and Cu
249
Fig. 22.7. Deconvoluted high-resolution core level C 1s XPS spectra obtained for a n-octadecylphosphonic acid (ODP/Al), b n-decylphosphonic acid (DP/Al), and c noctylphosphonic acid (OP/Al). The FWHM of the deconvoluted peaks are between 1.0 and 1.2 eV. (From [11])
250
E. Hoque et al.
Table 22.2 demonstrates the binding energies of all the resolved functional groups as illustrated in Fig. 22.7. It has been proposed that airborne carbonaceous contaminants can block the adsorption sites of aluminum oxide and hydroxide [63]. It is noteworthy that the XPS data (Fig. 22.7c) reveal the presence of oxidized carbon species of alcohol or ether type only for the case of OP/Al, whereas no characteristic oxidized carbon species are detected for ODP/Al and DP/Al (Fig. 22.7a, b). However, the organic contamination species that remain adsorbed on OP/Al are in very low quantities compared with unmodified Al, which is inferred from the XPS and CAM data (discussed later). The alkyl chain length has a strong influence on the molecular packing during self-assembly—the longer the chain length, the better the orientation of the molecules on the surfaces. The longer chains are better able to self-assemble owing to an increase in van der Waals attractive forces with greater chain length, because the strength of the van der Waals interactions per adsorbate is proportional to the number of methylene units [23]. XPS data for ODP/Al and DP/Al indicate the almost complete displacement of adsorbed atmospheric organic contamination. The DP and ODP molecules allow for a higher density of adsorption on Al, resulting in a more densely packed array of molecules and, consequently, higher contact angles and lower critical surface tension when compared with OP/Al. Thus, the XPS analysis confirms the presence of a chemically uniform film for ODP/Al and DP/Al, while the continued presence of residual contamination on OP/Al induces the formation of a film with less densely packed molecules [11]. The chemical surface composition (atomic percent) and elemental ratios (C to P) quantified from the high-resolution core level XPS multiplex performed on ODP/Al, DP/Al, and OP/Al are summarized in Table 22.3. Phosphorous was detected in the spectra, which is a key indication of the presence of a phosphonate molecule on the Al surface. Table 22.3 presents the variation of Table 22.2. The corresponding binding energy values for functional groups typically found due to the presence of phosphate or surface contamination. Each peak representing the group is deconvoluted from C 1s high-resolution XPS spectra taken on n-octadecylphosphonic acid (ODP)/Al, n-decylphosphonic acid (DP)/Al, and n-octylphosphonic acid (OP)/Al. (From [11]) Carbon species
Binding energy (eV)
C–C/C–H C–O
ODP/Al
DP/Al
OP/Al
284.8 –
284.8 –
284.8 285.7
Table 22.3. Atomic percent concentration (±2%) for the elements O, C, P, and Al and the C-to-P ratio quantified from the XPS multiplex of O 1s, C 1s, P 2s, and Al 2 p measured at 90◦ takeoff angle on untreated ODP/Al, DP/Al, and OP/Al. (From [11]) Substrates
O
C
P
Al
C/P
ODP/Al DP/Al OP/Al
32.0 39.9 43.1
38.4 24.8 21.6
1.95 2.15 2.05
27.6 33.1 33.2
19.7 11.6 10.5
22 Self-Assembled Monolayers on Al and Cu
251
atomic concentration with alkyl chain length. The elemental ratios of C to P for ODP/Al, DP/Al, and OP/Al are 19.7, 11.6, and 10.5, which approach the expected values of 18, 10, and 8, respectively, but the large deviation in the case of OP/Al is due to the presence of adventitious hydrocarbon contamination. For ODP/Al and DP/Al, the smaller deviation from the theoretical ratio stems probably from the upright position of the molecules, thus resulting in a higher atomic percentage of carbon compared with that of phosphorous owing to the higher exit thickness for the energetically lower phosphorous photoelectrons. Thus, from the XPS data it can be concluded that a chemically uniform ODP, DP, and OP film forms on Al [11]. Angle resolved XPS (AR-XPS) measurements are useful for obtaining information at different film depths. The depth of the film from which the photoelectrons escape decreases with an increase in takeoff angle. Therefore, AR-XPS measurements were performed on ODP/Al, DP/Al, and OP/Al to determine the variation of the phosphonate film elemental composition and thickness, t. The variation in chemical surface composition of ODP/Al is shown in Fig. 22.8. This AR-XPS data plot shows that the signal of all the elements, except carbon, decreases concomitantly with an increase of the takeoff angle. The quantity of carbon increases and that of phosphorus decreases with an increase in takeoff angle as expected for an alkylphosphonate film. The phosphonate head group is expected to bind with a hydroxylated site on the oxidized Al surface. If the molecules follow the above hypothesis, C-to-P ratios should increase with an increase in takeoff angle. The AR-XPS data confirm the C-to-P ratio increase with takeoff angle. The ARXPS data obtained for DP/Al and OP/Al also show a very similar variation in the chemical surface composition with respect to the takeoff angle. Thus, AR-XPS data give evidence that alkylphosphonic acid reacted with Al forms a monolayer assembly with the phosphonate head group bonded to the oxidized Al surface. The effective film thicknesses for ODP/Al, DP/Al, and OP/Al determined by AR-XPS were 2.1, 1.6, and 1.6 nm, respectively, which agree with ellipsometric data reported in the literature [11, 28].
Fig. 22.8. Angle-resolved XPS on ODP/Al: atomic percent concentration (quantified from the multiplex of Al 2 p, O 1s, C 1s, and P 2s) and C-to-P ratio versus takeoff angle; measurements were performed at takeoff angles of 90◦ , 75◦ , 60◦ , 45◦ , 30◦ , and 20◦ . (From [11])
252
E. Hoque et al.
PFMS SAM on Al. Figure 22.9 shows a deconvoluted C 1s high-resolution XPS spectra performed for PFMS/Al. The resolved peak at the binding energy of 283.9 eV is assigned to SiCHx , which is present in the PFMS molecule. The peak component at the binding energy of about 284.8 eV can be attributed to C–H/C–C groups which arise partly from the –CH2 – groups of the PFMS molecule and partly from adventitious carbon contamination. The peaks at the high binding energies of 290.1 and 292.3 eV can be unambiguously attributed to the –CF2 – and –CF3 functional backbone and tail groups of the PFMS molecule [62]. The ratio of the peak areas for the –CF2 – and –CF3 groups is 6.6 ± 0.3, which agrees well with the expected value of 7. Characteristic adsorbed oxygenated carbon species observed for PFDP/Al (Fig. 22.5b) are also seen for PFMS/Al. The binding energies of all the deconvoluted functional groups are listed in Table 22.1. These results indicate the presence of a chemically homogeneous perfluorosiloxy film on the Al surface [42]. The adsorbed atmospheric, organic contamination species on PFMS/Al are in very low quantities compared with the fluorocarbon species (inferred from the XPS data). The complete displacement of adsorbed contamination would provide a higher surface density of SAMs, resulting in closer-packed molecules and, consequently, higher water hydrophobicity. The high-resolution Al 2 p spectrum carried out on PFMS/Al is deconvoluted as shown in Fig. 22.9b. The peaks at binding energies of 71.4 and 71.9 eV correspond to metallic Al 2 p3/2 and Al 2 p1/2 , respectively. The fitted peaks with binding energies of 72.3 and 73.6 eV can be attributed to aluminum suboxide and aluminum oxide, respectively. The peak at 74.3 eV is likely due to a Si–O–Al bond which could form between hydrolyzed silane (silanol) and aluminum hydroxide [19, 42]. This extra peak is observed for γ -glycidoxypropyltrimethylsilane films on Al [64] and bis[1,2(triethoxysilyl)]ethane films on Al [65]. This peak is not seen in the deconvoluted Al 2 p spectrum taken on unmodified Al. The Si 2 p peak is shown in Fig. 22.9c with a binding energy of 102.2 eV. For Si 2 p, binding energies of around 102–102.5 eV have been found for siloxane molecules, such as polydimethylsiloxane [64, 66, 67], which has a polymer network based on Si–O–Si linkages [42]. Critical Surface Tension (Surface Energy) Analysis Figure 22.10a shows the best fit for the Zisman plot of the sessile drop static contact angle of four n-alkane liquids (hexadecane, dodecane, undecane, and decane) on PFDP/Al versus the alkane surface tension [68]. As all the alkane liquids used completely wet the unmodified Al surface, a Zisman plot was not possible. This information plus the fact that H2 O formed a contact angle of less than 10◦ indicates that the unmodified Al critical surface tension is between 28 and 72 mNm−1 . The critical surface tension determined from the best fit for PFDP/Al is 11.4 ± 1.5 mNm−1 [19]. With use of the Zisman plot method, the critical surface tensions for PFMS/Al, ODP/Al, and DP/Al were measured to be 16.5±1.5, 19.9 ± 3, and 21.4 ± 3 mNm−1 [19, 42]. Although the critical surface tensions for ODP/Al and DP/Al are comparable, a statistically significant difference (smaller) in critical surface tension is observed for PFDP/Al. If one compares the critical surface tension value determined for PFDP/Al with that typically quoted for polytetrafluoroethylene (PTFE; ‘Teflon’) [23], then one concludes that the significantly lower PFDP/Al critical sur-
22 Self-Assembled Monolayers on Al and Cu
253
Fig. 22.9. Deconvoluted core level high-resolution a C 1s, b Al 2 p, and c Si 2 p XPS spectra carried out on PFMS/Al. The FWHM of the deconvoluted peaks are between 1.0 and 2.3 eV for C 1s and are 0.5 eV for Al 2 p1/2 and Al 2 p3/2 , 1.5 eV for aluminum oxide and AlOSi, and 1.7 eV for –SiO–. (From [42])
254
E. Hoque et al.
Fig. 22.10. a Zisman plot for PFDP/Al (root mean square, rms, roughness 31.8 ± 0.5 nm). The critical surface tension, which correlates with the surface energy, is extrapolated from the plot cos θ = 1, where θ is the contact angle. b Histogram of the water contact angles (±3◦ ) for unmodified Al, PFDP/Al, PFMS/Al, ODP/Al, DP/Al, and OP/Al, respectively. (From [11, 42])
face tension is due to the fact that a large fraction of the perfluoroalkylphosphonate molecules have the perfluorinated chain oriented normal to the surface, with the CF3 groups sticking directly outwards (a PTFE surface would be dominated by CF2 groups) [42]. CAM Analysis Water CAMs, shown in Fig. 22.10b, were measured on unmodified Al, PFDP/Al, PFMS/Al, ODP/Al, DP/Al, and OP/Al to determine the hydrophobicity of the sample surfaces (hydrophilic oxidized unmodified Al versus hydrophobic –CF3 -terminated and –CH3 -terminated alkane chains). CAMs are also sensitive to the orientation of the molecules in the SAM. In well-ordered monolayers, the terminal methyl group of the
22 Self-Assembled Monolayers on Al and Cu
255
alkyl chain is oriented outwards, reducing access of the water drop to the substrate. An extreme improvement in hydrophobicity was observed for perfluorophosphonic acid and silane modified Al when compared with oxidized, unmodified Al: H2 O contact angles of 137◦ for PFDP/Al and 133◦ for PFMS/Al and less than 10◦ for unmodified Al. However, ODP, DP, and OP exhibited similar contact angles (more than 125◦ ), but, again, much higher than that of unmodified Al. The data confirm that PFDP/Al has the lowest critical surface tension as the water contact angle generally increases with a decrease in surface energy. PFDP/Al has a higher hydrophobicity than ODP/Al, DP/Al, or OP/Al [19, 42]. 22.3.2.3 PFMS-Based SAM on Cu Figure 22.11 shows a proposed schematic representation of the 1H,1H,2H,2Hperfluorodecyldimethylsiloxy–copper bond for the PFMS/Cu SAM surface. The same type of bond has been proposed to explain the strong adhesion of Cu monolayers on silica (SiO2 ) films [69]. In addition, it has been observed as a by-product of copper silicon oxide (Cu–SiOx ) synthesis [70].
Fig. 22.11. Proposed scheme for the 1H,1H,2H,2Hperfluorodecyldimethylsiloxy–copper bond leading to SAM formation after PFMS modification of oxidized Cu. (From [21])
XPS and FT-IRRAS Analysis of PFMS/Cu Figure 22.12a shows a C 1s core level high-resolution XPS spectrum taken on PFMS/Cu. The peak component at the binding energy of 283.8 eV is attributed to SiCHx which arises from the one –CH2 – and two –CH3 groups coordinated with Si in the alkyl chain of the PFMS molecule. The binding energy at 284.8 eV is assigned to CHx which arises from the one non-Si-coordinated –CH2 – present in the PFMS molecule and a small amount from adventitious carbon contamination. The characteristic oxidized carbon species of alcohol or ether type (C–O–) and
256
E. Hoque et al.
Fig. 22.12. a C 1s core level high-resolution deconvoluted XPS spectra obtained on PFMS/Cu and b Fourier transform IR reflection/absorption spectroscopy (FT-IRRAS) nonpolarized spectra taken for bulk PFMS liquid as a reference (upper spectrum) and a PFMS/Cu SAM (lower spectrum). (From [21])
ester, ketone, or carboxyl type (CO–O–C/C–CO–C/CO–O–H) adsorbed from the air are observed owing to atmospheric exposure between oxidation and XPS analysis. Similar oxygenated carbon species are observed in the deconvoluted C 1s spectra taken on polished only and polished, oxidized Cu (Fig. 22.3c). The high binding energy peaks of 291.1 and 293.3 eV can be unambiguously attributed to the –CF2 – and –CF3 groups, respectively, present in the perfluoroalkyl chain of PFMS. The ratio of the peak areas for –CF2 – and –CF3 is 6.6±0.3, which is close to the expected value of 7. Figure 22.12b shows FT-IRRAS nonpolarized spectra for a thick liquid film of PFMS deposited onto an unrelated substrate (as a reference) and a PFMS/Cu SAM. For the thick PFMS liquid film, absorption peaks for C–F stretching modes
22 Self-Assembled Monolayers on Al and Cu
257
are observed at approximately 1236, 1201, and 1145 cm−1 , with relative intensities (absorptions) of 65, 100, and 80, respectively. For PFMS/Cu, the absorption peaks for C–F stretching modes are found at approximately 1248, 1219, and 1154 cm−1 , with relative intensities of 100, 82, and 52, respectively. Thus, the FT-IRRAS data agree with the XPS data concerning the presence of fluorocarbon species on PFMS/Cu [21]. For PFMS/Cu, the Cu 2 p3/2 , Si 2 p, O 1s, and X-ray excited Cu Auger lines have been examined carefully. Figure 22.13 a (upper part) shows the deconvoluted Cu 2 p3/2 spectrum of nonmodified polished, oxidized Cu. The Cu 2 p3/2 component observed at 932.9 eV is not readily assigned, because Cu2 O is shifted from Cu(0) by only 0.1 eV and these two species are, thus, not distinguishable. The most distinguished feature in the spectrum is the pronounced satellite peak in the 2 p region for Cu(II); these peaks are separated from the 2 p3/2 peaks of Cu(0) and Cu(I) by around 10 eV. The peak due to CuO assigned at 934 eV is shifted to higher energy by 1.1 eV from the Cu(0)/Cu2 O peak. The deconvoluted third peak, at 935.1 eV, can readily be attributed to Cu(OH)2 , which is again shifted 1.1 eV from the CuO peak on the high-energy side. Cu(OH)2 is formed on the surface at some point between oxidization and silanization of Cu. Although CuO and Cu(OH)2 are readily identifiable, they overlap with the Cu(0) and Cu2 O peaks, resulting in an overall wider peak. Figure 22.13a (lower part) shows a representative deconvoluted Cu 2 p3/2 spectrum for PFMS/Cu. In this spectrum, only one main peak, at 932.9 eV, is identified unambiguously which is at the energy position of Cu(I)/Cu(0). The loss of the satellite peak denotes a change from Cu(II) to Cu(I). The chemical state identification can also be done by using the most intense component of the X-ray excited Auger L2,3 M4,5 M4,5 transition. Figure 22.13b (upper part) shows X-ray excited Auger spectra for unmodified oxidized Cu. The broad peak at 569.4 eV corresponds to Cu(II). Figure 22.13b (lower part) shows the X-ray excited Auger spectrum for PFMS/Cu. After reaction with PFMS, two distinct peaks appear at 568.2 and 570.4 eV corresponding to metallic Cu(0) and Cu(I), respectively, which confirms again the presence of Cu(I) for PFMS/Cu [21]. Figure 22.13c shows the high-resolution XPS Si 2 p spectrum. The only observed peak is at a binding energy of 102.6 eV. For Si, binding energies of around 102–102.5 eV have been found for siloxane molecules, such as polydimethylsiloxane a polymer which has a repeating chemical structure with subunits of [– O–Si(CH3 )2 –O–Si(CH3 )2 ]n where n is the number of subunits. Once the PFMS molecule has been hydrolyzed, a silanol (–Si–OH) forms and HCl is produced as a by-product. At this point in the reaction, it is quite possible that PFMS can dimerize to form bis(perfluorodecyldimethyl)disiloxane [F3 C(CF2 )7 (CH2 )2 (CH3 )2 Si–O– Si(CH3 )2 (CH2 )2 (CF2 )7 CF3 ], which can physisorb onto the oxidized Cu surface. However, in general, a film composed of molecules physisorbed onto an oxide surface would not withstand the harsh treatments (boiling nitric acid, boiling water, warm sodium hydroxide) that PFMS/Cu was subjected to for this study. A previous study showed that the stability of alkylsiloxy films on oxidized Cu correlates with perfluorination of the alkylsilane molecule (thus, nonperfluorinated alkylsiloxy films formed on oxidized Cu tend to be less stable against harsh treatment than perfluorinated ones) [22]. Also, reaction of PFMS with Au substrates leads to a film composed of physisorbed bis(perfluorodecyldimethyl)siloxane molecules on the Au surface (PFMS/Au). On the basis of CAM data (not reported here), PFMS/Au degrades rapidly after exposure to warm nitric acid (60 ◦ C, pH 1.8, 10 min). The stability of
258
E. Hoque et al.
Fig. 22.13. a Deconvoluted high resolution Cu 2 p3/2 spectra for unmodified Cu (upper part) and PFMS/Cu (lower part), b photoelectron Auger spectra on unmodified Cu (upper part) and PFMS/Cu (lower part), c Si 2 p peak for PFMS/Cu, and d O 1s spectra for unmodified Cu (upper part) and PFMS/Cu (lower part). (From [21])
22 Self-Assembled Monolayers on Al and Cu
259
the PFMS/Cu film against harsh treatments can be explained by the hypothesis that silanol is capable of reacting with copper hydroxide (CuOH) on the substrate surface (as is the case for free silanol molecules which react with surface-bound silanols on silicon oxide), then the similarity between the siloxane bond (–Si–O–Si–) and the siloxy–copper bond (–Cu–O–Si–) proposed above becomes obvious. Thus, this peak in the Si spectrum at 102.6 eV is a direct indication that such a bond has formed for PFMS/Cu [21]. Figure 22.13d (upper part) shows a high-resolution XPS O 1s spectrum taken on unmodified oxidized Cu. The careful inspection of the O 1s XPS spectra is very important to verify the chemical changes on Cu after PFMS modification observed by deconvolution of Cu 2 p3/2 , Auger, and Si 2 p spectra. The O 1s lines (Fig. 22.13d, upper part) at 529.8, 530.8, and 531.7 eV are attributed to CuO, Cu2 O, and Cu(OH)2 , respectively. The line at 533.2 eV is due to an oxidized carbon species. Figure 22.13d (lower part) shows the XPS O 1s spectrum for PFMS/Cu. The lack of the CuO peak indicates that Cu(II) has disappeared after modification with PFMS. The new line at 532.2 eV is assigned as SiO. The disappearance of the CuO and Cu(OH)2 peaks for PFMS/Cu supports further the conclusion that a siloxy–copper bond has formed [21]. FT-IRRAS is able to probe molecular conformation changes in SAM films [28]. In the case of metal surfaces, IR radiation with the electric field direction polarized perpendicular to the plane of incidence (0◦ polarization or s-polarized) does not result in observable dipole transitions, so only p-polarized (electric field direction polarized parallel to the plane of incidence) IR radiation can be used. The vibrational modes of the perfluoroalkyl chain molecules composing the SAM film are listed in Table 22.4. Table 22.4. IR vibrational modes (±4 cm−1 ) for PFMS/Cu self-assembled monolayer (SAM). d − asymmetric stretch mode, d + symmetric stretch mode, r rocking mode, δ bending mode, and wag wagging mode. (From [21]) Mode
Wavenumber (cm−1 )
νa (CH2 ), d− νs (CH2 ), d+ νa (CF2 ); δ(CH2 ) bending ν(CF2 ) progression, axial CF2 stretch ν(CF2 ) progression, axial CF2 stretch νa (CF2 ), and r(CF2 ) νa (CF2 ) + νa (CF3 ); δ(CCC), ν(CC) νa (CF2 ) + νa (CF3 ) νs (CF2 ); δ(CF2 ) ν(CC) stretch ν(CC) stretch νa (Si–O) νs (Si–O) νz (CF2 ) progression δ(Si–O) r(CF2 ) wag(CF2 ) wag(CF2 ) wag(CF2 ), r(CF2 )
2924 2853 1444 1375 1332 1245 1217 1205 1151 1136 1117 1040 1002 873 822 710 665 623
260
E. Hoque et al.
Figure 22.14a shows a FT-IRRAS spectrum taken from 3000 to 2800 cm−1 at a polarization angle of 90◦ ( p-polarized) for a PFMS/Cu sample after exposure to ambient conditions for more than 240 days. The symmetric (d+ ) and asymmetric (d− ) C-H stretching modes for the –CH2 – groups present in the spacer perfluoroalkyl chain are assigned at 2853 and 2924 cm−1 , respectively, which are in agreement with data reported in the literature, 2854 (d+ ) and 2925 cm−1 (d− ), for perfluorinated phosphonate SAMs on Al [28]. Although the head group of the phosphonate SAM discussed in [28] differs from that of the PFMS/Cu SAM discussed here, the spacer chains and tail groups (perfluorinated n-alkyl chains) are homologues.
Fig. 22.14. a FT-IRRAS spectrum taken of a PFMS/Cu sample after exposure to ambient conditions for more than 240 days at a polarization angle of 90◦ (direction of E with respect to the surface) from 2000 to 600 cm−1 showing C–F stretching mode peaks (dashed lines at 1245 and 1154 cm−1 ) and b FT-IRRAS spectrum taken of the same PFMS/Cu sample at a polarization angle of 90◦ from 3000 to 2800 cm−1 showing C–H stretching mode peaks (dashed lines at 2853 and 2924 cm−1 ). (From [21])
22 Self-Assembled Monolayers on Al and Cu
261
The IR absorption spectra of homologous molecules can be directly compared [21]. Figure 22.14b shows a FT-IRRAS spectrum taken from 2000 to 600 cm−1 at a 90◦ polarization angle for the same PFMS/Cu sample as in Fig. 22.14a. The asymmetric and symmetric C–F stretching mode vibrations for –CF2 – spacers of SAMs are attributed at 1245 and 1151 cm−1 , having relative intensities of 68 and 100, respectively, while the d− stretching mode vibrations of C–F for the –CF3 tail and –CF2 – groups are assigned at 1205 cm−1 with a relative intensity of 75. An interesting feature is observed in the high-frequency region of the spectra taken of PFMS/Cu. The absorption intensity of the O–H stretch assigned at 3350 cm−1 is reduced dramatically for PFMS/Cu when compared with unmodified oxidized Cu, which is in good agreement with the XPS data. It can be argued that this decline in the O–H mode occurs because –Cu(OH)2 is reduced after reaction with PFMS, resulting in a siloxy–copper bond. In addition, a Si–O stretching mode is seen at 1040 cm−1 which is in good agreement with the literature [21]. CAM Analysis for PFMS/Cu CAM data were obtained to determine the wetting properties of unmodified Cu and PFMS/Cu of different surface roughness. Water drop static contact angle versus rms roughness measured on PFMS/Cu for five different surface roughness values is shown in Fig. 22.15a. The rms roughness values for the five samples are 34 ± 2, 115 ± 8, 193 ± 12, 284 ± 12, and 450 ± 8 nm. The corresponding CAM values are 132, 145, 152, 155, and 157◦ , respectively. The CAM value for unmodified Cu for all surface roughness values considered is less than 15◦ . Contact angles are dependent on the surface roughness and energy. CAM data clearly show the dramatic difference in hydrophobicity between unmodified Cu (contact angles of less than 15◦ ) and PFMS/Cu (contact angles of more than 130◦ ). Figure 22.15b illustrates the variation of the measured critical surface tension, γcrit , with rms surface roughness for PFMS/Cu. Therefore, to more precisely quantify the critical surface tension using the Zisman method, γcrit was measured as a function of surface roughness and a value was extrapolated from the plot for a perfectly smooth surface. Three Cu substrates with different surface roughness values were reacted with PFMS. The critical surface tensions were found to be 14 ± 1.5, 13 ± 1.5, and 11 ± 1.5 mJm−2 for PFMS/Cu samples with rms surface roughness of 35 ± 2, 115 ± 8, and 284 ± 33 nm, respectively. If the plot is extrapolated plot to 0-nm rms roughness for a perfectly smooth surface, the critical surface tension (which is proportional to the surface energy) of PFMS/Cu would be approximately 14.5 ± 1.5 mJm−2 [54]. 22.3.3 Nanotribological Properties FFM Analsyis The adhesion and friction tests were conducted using a commercial AFM system (Dimension 3000, Nanoscope IIIa controller, DI, Santa Barbara, USA). Squarepyramidal Si3 N4 tips with a nominal 30-50-nm tip radius on gold back-coated triangular Si3 N4 cantilevers with a typical spring constant of 0.58 Nm−1 were used.
262
E. Hoque et al.
Fig. 22.15. a Water contact angle versus rms surface roughness. The measurements were performed on five PFMS/Cu samples with different surface roughness values. The contact angle on unmodified Cu for all surface roughness values is less than 15◦ . b Critical surface tension versus surface roughness. Data were obtained for three PFMS/Cu samples with different surface roughness values. (From [54])
Adhesion can be calculated either using force calibration plots or from the negative intercepts on the horizontal axes of the friction force versus normal load plots [10]. Both methods generally yield similar results. The force calibration plot technique was used in this study. The coefficient of friction was obtained from the slope of the friction force versus normal load plots. Normal loads typically ranged from 5 to 100 nN. Friction force measurements were performed at a scan rate of 1 Hz along the fast scan axis and over a scan size of 2 µm. The fast scan axis was perpendicular to the longitudinal direction of the cantilever. The influence of relative humidity on the adhesion was studied in an environmentally controlled chamber. Relative humidity was controlled by introducing a mixture of dry and moist air
22 Self-Assembled Monolayers on Al and Cu
263
inside the chamber. The temperature was maintained at 22 ± 1 ◦ C. The temperature range studied was 20–110 ◦ C. The relative humidity was maintained at 50 ± 5% during these measurements. The effect of sliding velocity on friction force was monitored under ambient conditions using a high-velocity piezo stage designed for achieving high relative sliding velocities on a commercial AFM setup [10]. The traveling distance of the sample, i. e., the scan size, was set at 25 µm, while the scan frequency was varied between 0.1 (5 µm s−1 ) and 100 Hz (5000 µm s−1 ). Wear tests were conducted using a single-crystal natural diamond tip and a Si tip. The diamond tip, ground to the shape of a three-sided pyramid, had an apex angle of 80◦ and a tip radius of about 50 nm. The tip was bonded with epoxy to a gold-coated 304 stainless steel rectangular cantilever beam (20 µm thick, 0.2 mm wide, and approximately 2 mm long). Wear tests were performed on a 1 µm × 1 µm scan area at the desired normal load and at a scan rate of 1 Hz. After each wear test, a 3 µm × 3 µm area was imaged and the average wear depth was calculated. Alkylphosphonate SAMs on Al (PFDP/Al, ODP/Al, DP/Al, OP/Al) Adhesive force arises from the presence of a thin liquid film, such as a mobile lubricant or adsorbed water layer, which causes menisci to build up around asperities, on surfaces in contact or near contact, as a result of surface energy effects. The intrinsic attractive force arising from meniscus contributions depends on the surface tension of the film and the corresponding contact angle and may result in high friction and wear. Figure 22.16a gives the adhesive force, friction force, and coefficient of friction measured at controlled temperature and humidity using AFM and Fig. 22.16b shows the friction force versus normal load plots for various SAMs measured under the same conditions. The unmodified Al substrate shows much higher adhesive force than when a SAM film is present. Between DP and ODP, DP showed higher adhesive force, which suggests that molecules with longer chains have lower adhesion. The SAMs with longer chains produce a flexible monolayer. When the atomic force microscope tip penetrates this flexible layer, the SAM molecules stick to it very easily; thus, for these SAMs, the adhesive force is lower. PFDP/Al, which has the highest water contact angle, showed the lowest adhesion. This tendency of adhesion is very consistent with the change in surface energy. AFM adhesive force linearly correlates with the work of adhesion, or, in this case, is equivalent to the surface energy [10, 71]. The coefficient of friction is higher for the unmodified Al substrate than when a SAM is present. The coefficient of friction of a SAM with a fluorocarbon backbone chain is load-dependent; it is lower in the range 0–30 nN than in the range 30–70 nN. The coefficient of friction for a fluorocarbon SAM in the high-load range is higher than that for a SAM with a hydrocarbon backbone chain. This difference might be attributed to the higher stiffness of the fluorocarbon backbone. For a fluorocarbon chain, it is harder to rotate the backbone structure owing to the much larger size of F atoms compared with H atoms [72]. The larger size of the trifluoromethyl groups suggests a more densely packed arrangement at the outermost portion of the layer. This implies that a fluorocarbon backbone is more rigid than a hydrocarbon one. The C–C bonds of alkyl chains can freely rotate around. A molecular spring or brush model has been used to explain why less compliant SAMs show higher friction [73, 74]. The SAMs with a higher spring constant or
264
E. Hoque et al.
stiffer backbone may need more energy to be elastically deformed during sliding; therefore, friction is higher for these SAMs. In terms of a chain-length effect, it has been reported that the coefficient of friction for SAM films decreases with the carbon backbone chain length (n) up to 12 C atoms (n ∼12) [75]. A chain length effect for the coefficient of friction was obvious for DP and ODP [10, 11]. The influence of relative humidity on the adhesive force and the coefficient of friction was studied, as shown in Fig. 22.17a. In the case of unmodified Al, adhesive force increases with relative humidity owing to formation of menisci. The DP/Al and
Fig. 22.16. a Adhesive force, friction force, and coefficient of friction and b friction force versus normal load plots for unmodified Al and various SAMs on Al. The coefficient of friction values for each SAM are indicated in parentheses. All measurements were made at 22 ◦ C and 50% relative humidity (RH) using atomic force microscopy. (From [10, 11])
22 Self-Assembled Monolayers on Al and Cu
265
Fig. 22.16. (continued)
ODP/Al SAMs show almost no change in adhesive force with humidity. The highly hydrophobic nature of these monolayers means that the formation of water menisci is negligible at all humidity values studied. A hydrophobic layer with a fluorocarbon backbone chain shows the same tendency. Figure 22.17a also shows the relationship between the relative humidity and the coefficient of friction. It is observed that the coefficient of friction first increases slightly and then decreases at higher relative humidity. This effect is attributed to the ability of an adsorbed water layer to act as a lubricant. The influence of temperature on adhesive force is shown in Fig. 22.17b. The adhesive force increases for unmodified Al up to 50 ◦ C and then decreases to a stable value for higher temperatures. This tendency for unmodified Al could be attributed to the desorption of water molecules from the surface. Once most of the water molecules are desorbed, the adhesive force becomes constant. The adhesive force for SAMs on Al showed a similar tendency, but the underlying cause may not be the same as for unmodified Al. Adhesive force for hydrophobic SAMs showed a weak dependency on humidity; however, small amounts of water could be seen to affect the surface hydrophobicity. The increase in adhesive force could be caused by the melting of the SAM film. A typical melting point for a linear carbon chain molecule, such as hexadecanol [CH3 (CH2 )15 OH], is 50 ◦ C [76]. With an increase in temperature, the SAM film softens, thereby increasing the real area of contact and, as a consequence, the adhesive force. Once the temperature is higher than the melting point, the mechanism changes from boundary lubrication, i. e., a solid SAM film, to liquid lubrication, i. e., a melted SAM film [10, 74]. Figure 22.18 shows the effect of sliding velocity on friction force. Friction force increases slowly at lower sliding velocities for unmodified Al and SAMs on Al. The friction force increases rapidly at higher sliding velocities for all samples except for PFDP/Al. The increase in friction force at higher velocities is the result of impacts
266
E. Hoque et al.
Fig. 22.17. a Dependence of the adhesive force and coefficient of friction upon RH at 22 ◦ C and b temperature dependence of the adhesive force at 50% RH for unmodified Al and various SAMs on Al. (From [10])
22 Self-Assembled Monolayers on Al and Cu
267
Fig. 22.18. Dependence of the friction force upon the sliding velocity for unmodified Al and various SAMs on Al measured at 22 ◦ C, 50% RH, and 70-nN normal load. (From [10])
with asperities and the corresponding high frictional energy dissipation at the sliding interface. The “molecular spring” model [71] has been extended to explain the velocity-dependent increase in friction force for compliant SAM molecules [77]. The SAM molecules reorient under the tip normal load. As the tip proceeds along the scan direction, the molecules are “relieved” of the tip load and reorient to their initial position. At high velocities, the SAM molecules do not have sufficient time to reorient after the atomic force microscope tip had slid past and are believed to remain oriented in a particular direction [10]. Figure 22.19a shows the relationship between the decrease in surface height and normal load for unmodified Al and various SAMs on Al subjected to microwear tests. Figure 22.19b summarizes the critical normal load beyond which point the surface drastically decreases (evidenced by a sudden change in the slope). ODP/Al shows higher critical load than DP/Al owing to its longer chain. The SAMs with fluorocarbon backbone chains showed a better wear resistance than the hydrocarbon backbone chains and unmodified Al. The SAMs PFDP/Al, DP/Al, and ODP/Al, and unmodified Al showed critical loads of 20, 10, 20, and 0.35 µN, respectively. A mechanism for the failure of compliant SAMs during wear tests has been proposed [71]. It is believed that the SAMs fail mostly owing to shearing of the molecule at the head group, i. e., pulling the molecules off the substrate [10]. PFMS SAM on Cu (PFMS/Cu) Figure 22.20 illustrates the adhesive force, coefficient of friction, and friction force versus normal load measured for unmodified Cu and PFMS/Cu. A significant reduction in adhesive force is observed for PFMS/Cu when compared with unmodified Cu, as shown in Fig. 22.20a. The coefficient of friction of unmodified Cu is almost double that of PFMS/Cu (Fig. 22.20b). The friction force versus normal load plot for unmodified Cu and PFMS/Cu is presented in Fig. 22.20c. Reduction of adhesion and friction for PFMS/Cu is consistent with its low critical surface tension (data discussed in Sect. 22.3.2.3).
268
E. Hoque et al.
Fig. 22.19. a Decrease in surface height as a function of normal load after one scan cycle and b comparison of critical loads needed for failure during wear tests for unmodified Al and various SAMs on Al measured at 22 ◦ C and 50% RH. (From [10])
22.3.4 Chemical Stability To understand the strength of interaction between the substrate (aluminum or copper oxide) and perfluoroalkylsiloxy molecules or perfluorinated and nonperfluorinated alkylphosphonate molecules, different types of harsh chemical treatment (stability tests) have been performed. Concentrated nitric acid (HNO3 ) was mixed with DI water to make a dilute nitric acid solution of pH 1.5–1.8. PFMS/Al, PFDP/Al, ODP/Al, DP/Al, OP/Al, and PFMS/Cu samples were exposed to the nitric acid solution for 30–70 min at 60–100 ◦ C. PFMS/Cu film samples were exposed to DI water for 30–70 min at 100 ◦ C. Sodium hydroxide (NaOH) pellets were mixed with DI water to make a NaOH solution of pH 12.0. PFMS/Cu samples were exposed to the NaOH solution for 30–70 min at 60–80 ◦ C. All the surface roughness values were determined from AFM data over an area of 40 × 40 µm2 [11, 19, 21, 42, 54].
22 Self-Assembled Monolayers on Al and Cu
269
Fig. 22.20. Tribology data: a adhesive force, b coefficient of friction, and c friction force versus normal load for unmodified Cu and PFMS/Cu. (From [54])
270
E. Hoque et al.
Alkylphosphonate and Alkylsiloxy SAMs on Al ODP SAM (ODP/Al). XPS, CAM, and surface roughness data for an ODP SAM before and after exposure to warm nitric acid (pH 1.8, 30 min, 60–95 ◦ C) are shown in Table 22.5. The surface elemental composition (O, C, P, and Al) and C-to-P ratio were determined from XPS high-resolution O 1s, C 1s, P 2s, and Al 2 p spectra measured for an ODP/Al SAM. The ODP/Al SAM was exposed to nitric acid solution for 30 minutes at 60, 80, and 95 ◦ C. After exposure to nitric acid at 60–95 ◦ C, the elemental composition and the C-to-P ratio remain similar to those seen before. The C-to-P ratio is close to the expected value of 18. The CAM values for untreated and 60, 80, and 95 ◦ C treated ODP/Al SAMs are 130◦ , 127◦ , 124◦ , and 123◦ , respectively, with corresponding rms surface roughness values of 31±1, 36±2, 38±2, and 36±2 nm. Therefore, no significant change in chemical composition, hydrophobicity, or surface roughness is seen after exposure to warm nitric acid solution, indicating a high degree of robustness for the ODP/Al SAM [42]. Table22.5.Atomic percent concentration (±2%) for O, C, P, and Al and the C-to-P ratio quantified from the XPS multiplex of O 1s, C 1s, P 2s, and Al 2 p measured at 90◦ takeoff angle on an ODP/Al SAM before and after exposure to HNO3 (pH 1.8, 30 min) at 60, 80, and 95 ◦ C; the corresponding contact angle measurement (CAM) (±3◦ ) and root mean square (rms) surface roughness values are also shown. (From [42]) Substrates ODP/Al ODP/Al (60 ◦ C) ODP/Al (80 ◦ C) ODP/Al (95 ◦ C)
O
C
P
Al
C/P
CAM (°)
rms (nm)
23.2 23.0 23.1 25.0
51.3 55.8 53.4 53.1
2.4 2.7 2.5 2.5
23.1 18.5 21.0 19.4
21.3 20.7 21.3 21.2
130 127 124 123
31 ± 1 36 ± 2 38 ± 2 36 ± 2
PFDP SAM (PFDP/Al). Table 22.6 shows XPS, CAM, and surface roughness data for a PFDP/Al SAM before and after exposure to warm nitric acid (pH 1.8, 30 min, 60– 95 ◦ C). The rms surface roughness values of the PFDP/Al SAM measured before treatment and after treatment at 60, 80, and 95 ◦ C are 34 ± 1, 33 ± 1, 38 ± 1, Table 22.6. Atomic percent concentration (±2%) for F, O, C, P, and Al and the F-to-P ratio quantified from the XPS multiplex of F 1s, O 1s, C 1s, P 2s, and Al 2 p measured at 90◦ takeoff angle on a PFDP/Al SAM before and after exposure to HNO3 (pH 1.8, 30 min) at 60, 80, and 95 ◦ C; the corresponding CAM (±3◦ ) and rms surface roughness values are also shown. (From [42]) Substrates PFDP/Al PFDP/Al (60 ◦ C) PFDP/Al (80 ◦ C) PFDP/Al (95 ◦ C)
F
O
C
P
Al
F/P
CAM (°)
rms (nm)
39.1 35.5 32.2 28.6
21.9 23.2 26.0 26.9
17.0 19.5 19.9 22.4
2.0 1.8 1.7 1.6
20.0 20.0 20.2 20.5
18.5 19.7 18.9 17.9
137 139 140 123
34 ± 1 33 ± 1 38 ± 1 41 ± 2
22 Self-Assembled Monolayers on Al and Cu
271
and 41 ± 1 nm, respectively and the corresponding contact angles are 137◦ , 139◦ , 140◦ , and 123◦ . At 95 ◦ C, F and P decrease by about 20%, while O increases approximately by 20%, maintaining the F-to-P ratio almost constant, close to the expected value of 17. Although there is no significant change in surface roughness, the CAM values decrease by approximately 10%. As the temperature increases, hydrolysis of any bond between the phosphonate head group and aluminum oxide surface occurs more rapidly, resulting in a faster decomposition of the PFDP/Al film. The noticeable change in XPS and CAM data after exposure to warm nitric acid at 95 ◦ C for 30 min probably indicates some degradation of the PFDP/Al SAM, but no significant change, and thus no degradation, is seen at lower temperatures [42]. DP SAM (DP/Al). XPS, CAM, and surface roughness data for the DP/Al SAM before and after exposure to warm nitric acid solution (pH 1.8, 30 min, 60–80 ◦ C) are shown in Table 22.7. The CAM values measured for the untreated and the 60 and 80 ◦ C treated nitric acid DP/Al SAM are 129◦ , 75◦ , and 40◦ , respectively, with corresponding rms surface roughness values of 36 ± 3, 41 ± 6, and 51 ± 2 nm. Although no significant change in surface roughness is noticed after exposure to warm nitric acid at 60 ◦ C, C and P decrease by approximately 30 and 70%, respectively, while O increases by around 30%. At 80 ◦ C, C and P are found to decrease by approximately 55 and 95%, respectively, while O is found to increase by around 45%. The contact angle decreases by almost 70%, while rms surface roughness increases by almost 45%. The C-to-P ratio becomes much higher than the expected value of 10 after both treatments, unlike in the case of the ODP/Al SAM, where it remains almost constant. This is probably due to degradation of the SAM film followed by an increase in contamination. The data indicate the start of degradation of the DP/Al SAM after exposure to nitric acid at 60 ◦ C and dramatic degradation at 80 ◦ C with little of the DP/Al SAM left on the Al surface. Then, it appears that the nitric acid begins to etch the Al surface, giving rise to greater surface roughness [42]. Table22.7.Atomic percent concentration (±2%) for O, C, P, and Al and the C-to-P ratio quantified from the XPS multiplex of O 1s, C 1s, P 2s, and Al 2 p measured at 90◦ takeoff angle on a DP/Al SAM before and after exposure to HNO3 (pH 1.8, 30 min) at 60 and 80 ◦ C; the corresponding CAM (±3◦ ) and rms surface roughness values are also shown. (From [42]) Substrates DP/Al DP/Al (60 ◦ C) DP/Al (80 ◦ C)
O
C
P
Al
C/P
CAM (°)
34.2 45.2 50.1
30.9 23.0 13.1
2.5 0.7 0.1
32.4 33.2 36.7
12.3 32.9 131
129 75 40
rms (nm) 36 ± 3 41 ± 6 51 ± 2
PFMS SAM (PFMS/Al). Table 22.8 shows XPS, CAM, and surface roughness data for a PFMS/Al SAM before and after exposure to warm nitric acid (pH 1.8, 30 min, 60 ◦ C). The rms surface roughness values measured for the PFMS/Al SAM before and after treatment are 30±2 and 67±4 nm and the corresponding contact angles are 133 and 135◦ , respectively. After the treatment, F and Si decrease by approximately
272
E. Hoque et al.
Table 22.8. Atomic percent concentration (±2%) for F, O, C, Si, and Al and the F-to-Si ratio quantified from the XPS multiplex of F 1s, O 1s, C 1s, Si 2s, and Al 2 p measured at 90◦ takeoff angle on a PFMS/Al SAM before and after exposure to HNO3 (pH 1.8, 30 min) at 60 ◦ C; the corresponding CAM (±3◦ ) and rms surface roughness values are also shown. (From [42]) Substrates PFMS/Al PFMS/Al (60 ◦ C)
F
O
C
Si
Al
F/Si
CAM (°)
51.0 35.0
9.1 20.5
31.6 29.6
2.8 2.0
5.5 12.9
18.3 17.5
133 135
rms (nm) 30 ± 2 67 ± 4
30% and C decreases by around 8%, while O and Al almost double, maintaining the F-to-P ratio very close to the expected value of 17. The surface roughness increases by more than double. Nitric acid at 60 ◦ C degrades the PFMS/Al SAM and, as a result, the aluminum oxide is etched, giving rise to greater surface roughness, which may also explain the slight increase in the CAM values. After exposure to nitric acid at 80 ◦ C for 30 min, XPS and CAM data indicate that the PFMS/Al SAM is completely eradicated from the Al surface. Again, this allows the nitric acid to etch the aluminum oxide surface and finally the Al film begins to separate from the Si substrate [42]. OP SAM (OP/Al). XPS, CAM, and surface roughness data for the OP/Al SAM before and after exposure to warm nitric acid (pH 1.8, 30 min, 60 ◦ C) are listed in Table 22.9. The contact angles measured on the OP/Al SAM are 125◦ and 25◦ before and after treatment, with a corresponding rms surface roughness of 30 ± 1 and 41 ± 3 nm, respectively. After treatment in nitric acid at 60 ◦ C, C and P decrease by approximately 50 and 95%, respectively, while O and Al increase by around 50 and 20%, respectively. The C-to-P ratio becomes much higher than the expected value of 8 after treatment, unlike in the case of the ODP/Al SAM, where it remains almost constant. As for the DP/Al case, this is probably due to degradation of the SAM film followed by an increase in contamination. The CAM values decrease by approximately 80% and the rms surface roughness increases by almost 35%. As OP tends to form a less densely packed SAM on Al than the longer alkyl chain phosphonates [11], this may explain why the OP/Al SAM is eliminated almost completely after exposure to nitric acid at 60 ◦ C. At 80 ◦ C, the OP/Al SAM is eradicated even faster and the exposed Al film is attacked by the nitric acid and eventually detaches from the Si substrate [42].
Table22.9.Atomic percent concentration (±2%) for O, C, P, and Al and the C-to-P ratio quantified from the XPS multiplex of O 1s, C 1s, P 2s, and Al 2 p measured at 90◦ takeoff angle on an OP/Al SAM before and after exposure to HNO3 (pH 1.8, 30 min) at 60 ◦ C; the corresponding CAM (±3◦ ) and rms surface roughness values are also shown. (From [42]) Substrates OP/Al OP/Al (60 ◦ C)
O
C
P
Al
C/P
CAM (°)
rms (nm)
40.0 48.5
24.7 12.3
2.4 0.1
32.9 39.0
10.3 123
125 25
30 ± 1 41 ± 3
22 Self-Assembled Monolayers on Al and Cu
273
Comparison between SAMs on Al. Comparing the resistance to warm nitric acid of all the SAMs leads to the conclusion that ODP/Al and PFDP/Al SAMs are the most stable, followed by DP/Al, PFMS/Al, and OP/Al SAMs. Recently, the authors have observed that SAMs formed by reaction of perfluorodecanoic acid [CF3 (CF2 )8 COOH] with Al [78, 79] are less stable than ODP/Al and PFDP/Al SAMs, but are more stable than DP/Al, PFMS/Al, or OP/Al SAMs (unpublished data). Another interesting observation is that among SAMs of equal carbon chain length (ten carbons), the PFDP/Al SAM is the most stable, followed by DP/Al and PFMS/Al SAMs, thus indicating that phosphonates may make a more stable bond with Al. Between PFDP/Al and DP/Al SAMs, the PFDP/Al SAM is much more stable. The fluorocarbon chains adopt a helical conformation and, consequently, possess larger cross sections than those of simple hydrocarbon chains. Furthermore, fluorocarbon chains are more rigid than hydrocarbon chains and they are more resistant to oxidation and corrosion. All these factors explain the enhanced stability of perfluorinated alkyl chain SAMs against very harsh conditions, despite the greatly reduced van der Waals interaction between perfluorinated chains. Probably the water molecules, hydronium ions (H3 O+ ), or hydroxyl ions (OH− ) in an acid solution are better obstructed from reaching the SAM head groups by a perfluorinated alkyl chain film and, therefore, are less able to degrade it. The data also show that nonperfluorinated alkyl phosphonic acids with 18 carbon atoms form robust phosphonate SAMs on Al. By the same logic, perfluorinated alkylsilanes with 18 carbon atoms also should form robust SAMs on Al which are even more stable than those with ten carbon atoms [42]. PFMS SAM on Cu (PFMS/Cu) Chemical State After Treatment. Figure 22.21 shows deconvoluted C 1s core level high-resolution XPS spectra obtained for PFMS/Cu treated in boiling nitric acid (pH 1.8, 30 min), PFMS/Cu treated in boiling water (30 min), and PFMS/Cu treated in warm NaOH (pH 12, 60 ◦ C, 30 min). All the functional groups found in the deconvoluted C 1s spectra for PFMS/Cu (Fig. 22.12a) are also observed in the fitted C 1s spectra for treated PFMS/Cu (Fig. 22.21). The ratio of the peak areas for –CF2 – and –CF3 always remains very close to the expected value of 7. The total carbon contamination of oxygenated and nonoxygenated species for NaOH-treated PFMS/Cu is found to be approximately 14%, whereas for acid-treated and watertreated PFMS/Cu, it is around 22 and 29%, respectively. The increase in surface contamination after treatment may partially explain the slight decrease in contact angle. Thus, the XPS data show PFMS/Cu to have significant resistance to boiling nitric acid, boiling water, and warm NaOH (60 ◦ C) for 30 min [54]. Nitric Acid Treatment. XPS, CAM, and surface roughness data for PFMS/Cu before and after exposure to nitric acid (pH 1.8, 80–100 ◦ C, 30–70 min) are listed in Tables 22.10 and 22.11. The surface elemental composition (Cu, F, O, C, and Si) and the F-to-Si ratio were determined from high-resolution XPS Cu 2 p, F 1s, O 1s, C 1s, and Si 2 p spectra obtained for PFMS/Cu. For nontreated and treated PFMS/Cu, the CAM values obtained are 132, 133, 137, and 145◦ with a corresponding rms surface roughness
274
E. Hoque et al.
Fig. 22.21. Deconvoluted C 1s core level high-resolution XPS spectra obtained for a HNO3 treated PFMS/Cu (pH 1.8, 100 ◦ C, 30 min), b boiling water treated PFMS/Cu (30 min), and c warm NaOH treated PFMS/Cu (pH 12 , 60 ◦ C, 30 min). (From [54])
22 Self-Assembled Monolayers on Al and Cu
275
Table 22.10. Atomic percent concentration (±2%) for Cu, F, O, C, and Si and the F-to-Si ratio quantified from the XPS multiplex of Cu 2 p, F 1s, O 1s, C 1s, and Si 2 p measured at 90◦ takeoff angle on PFMS/Cu before and after exposure to HNO3 (pH 1.8, 80 ◦ C) for 30, 50, and 70 min; also shown are the corresponding CAM (±3◦ ) and rms surface roughness values. (From [54]) Substrates
Cu
F
O
C
Si
F/Si
CAM (°)
rms (nm)
PFMS/Cu PFMS/Cu (30 min) PFMS/Cu (50 min) PFMS/Cu (70 min)
5.9 8.0 7.7 12.0
48.9 45.4 47.2 41.1
11.5 14.3 13.2 15.7
30.9 29.7 29.2 28.9
2.8 2.6 2.7 2.3
17.5 17.5 17.5 17.9
132 133 137 145
34 ± 2 38 ± 2 46 ± 2 107 ± 5
Table 22.11. Atomic percent concentration (±2%) for Cu, F, O, C, and Si and the F-to-Si ratio quantified from the XPS multiplex of Cu 2 p, F 1s, O 1s, C 1s, and Si 2 p measured at 90◦ takeoff angle on PFMS/Cu before and after exposure to boiling HNO3 (pH 1.8, 100 ◦ C) for 30, 50, and 70 min; also shown are the corresponding CAM (±3◦ ) and rms surface roughness values. (From [54]) Substrates
Cu
F
O
C
Si
F/Si
CAM (°)
rms (nm)
PFMS/Cu PFMS/Cu (30 min) PFMS/Cu (50 min) PFMS/Cu (70 min)
5.9 8.1 16.0 18.3
48.9 46.3 34.9 29.4
11.5 14.5 19.1 21.7
30.9 28.5 28.0 28.9
2.8 2.6 2.0 1.7
17.5 17.7 17.5 17.3
132 133 144 134
34 ±2 39 ±2 115 ±4 202 ±10
of 34–107 nm (Table 22.10). The F-to-Si ratio remains very close to the expected value of 17 in all cases. For the 30- and 50-min treatments, no significant changes in XPS spectra, CAM values, or surface roughness values are noticed. However, after exposure at 80 ◦ C for 70 min, F decreases by almost 15% and Cu almost doubles. Interestingly, although the change in chemical composition is obvious, the contact angle increases by approximately 10% and the rms surface roughness by almost 3 times. It appears that the SAM film was completely eradicated in some areas of the PFMS/Cu surface and the warm nitric acid solution was able to etch directly the Cu, giving rise to a greater surface roughness, which probably explains the increase in contact angle via the lotus effect. Thus, a PFMS/Cu SAM can resist treatment in nitric acid at 80 ◦ C for up to 50 min. The data for the PFMS/Cu SAM exposed to boiling nitric acid (100 ◦ C) for 30–70 min are shown in Table 22.11. The XPS and CAM data are found to be similar before and after the exposure to boiling nitric acid for 30 min, while SAM degradation is observed after the treatments for 50 or 70 min [54]. In all cases, the F-to-Si ratio remains very close to the expected value of 17. For nontreated and treated PFMS/Cu, the CAM data were measured to be between 132◦ and 144◦ with a corresponding rms surface roughness of 34–202 nm (Table 22.11). After 50-min treatment, F, Si, and C decrease by approximately 25, 25, and 15%, respectively, while O increases by 50% and Cu increases about 2.5 times. Surface roughness increases by about 3 times and contact angles by about 10%. As already mentioned, a dramatic increase in surface roughness, resulting in an increase in contact angle, can be an indication of SAM degradation. After 70-min exposure, O increases by almost 85% and Cu increases by about 3 times, while F
276
E. Hoque et al.
Fig. 22.22. Water contact angle and rms surface roughness versus harsh treatment (HNO3 acid at 60, 80, and 100 ◦ C for 50 and 70 min) of PFMS/Cu (polished Cu, initial surface roughness less than 40 nm). The dashed lines indicate the surface roughness and the contact angle for untreated PFMS/Cu. (From [54])
and Si drop by approximately 40% and C drops by around 15%. Although surface roughness increased by almost 6 times, no significant change in contact angle was noticed, probably due to the off-setting effects already mentioned. The PFMS/Cu SAM can survive exposure to boiling nitric acid for up to 30 min without any noticeable degradation. The change in contact angle and surface roughness versus nitric acid treatment temperature for both 50 and 70 min is shown in Fig. 22.22. No change is noticed in either parameter for PFMS/Cu subjected to nitric acid at 60 ◦ C, leading to the conclusion that the PFMS/Cu SAM film remains intact until higher temperatures [54]. Boiling Water Treatment. Table 22.12 lists XPS, CAM, and surface roughness data for PFMS/Cu before and after exposure to boiling water (30–70 min). The water CAM values are between 132◦ and 106◦ for nontreated and treated PFMS/Cu with a corresponding rms surface roughness of 20–34 nm. O increases by approximately 40% while F decreases by 10% after the exposure to boiling water for all times. In all cases, the F-to-Si ratio is found to be very close to the expected value of 17. Surface roughness (rms) is found to remain about the same up to 70 min, although it is approximately 30% lower than that found for nontreated PFMS/Cu. A 20% reduction in contact angle is noticed for the 70-min treatment. Thus, the PFMS/Cu Table 22.12. Atomic percent concentration (±2%) for Cu, F, O, C, and Si and the F-to-Si ratio quantified from the XPS multiplex of Cu 2 p, F 1s, O 1s, C 1s, and Si 2 p measured at 90◦ takeoff angle on a PFMS SAM before and after exposure to boiling H2 O for 30, 50, and 70 min; also shown are the corresponding CAM (±3◦ ) and rms surface roughness values. (From [54]) Substrates
Cu
F
O
C
Si
F/Si
CAM (°)
PFMS/Cu
5.9
48.9
11.5
30.9
2.8
17.5
132
PFMS/Cu (30 min) PFMS/Cu (50 min) PFMS/Cu (70 min)
3.6 4.1 3.1
44.2 43.0 43.1
17.6 17.1 17.2
32.1 33.3 34.2
2.5 2.4 2.4
17.7 17.9 18.0
120 114 106
rms (nm) 34 ± 2
22 ± 2 20 ± 2 23 ± 2
22 Self-Assembled Monolayers on Al and Cu
277
SAM film is only slightly degraded by exposure to boiling H2 O for up to 30 min. Further degradation of the PFMS/Cu SAM film is seen after exposure to boiling H2 O for times longer than 30 min [54]. Sodium Hydroxide Treatment. The water CAM and surface roughness data measured on PFMS/Cu before and after exposure to warm NaOH (pH 12, 60–80 ◦ C, 30–70 min) are shown in Table 22.13. The water CAM values are between 118◦ and 87◦ with a corresponding rms surface roughness of 27–29 nm. After exposure to NaOH at 80 ◦ C for 30, 50, and 70 min, water drop contact angles are found to be 100◦ , 32◦ , and 35◦ , respectively, with a corresponding rms surface roughness of 24–29 nm. CAM data show that PFMS/Cu is only slightly degraded by exposure to NaOH (pH 12, 60 ◦ C) for up to 30 min. However, serious degradation of the PFMS/Cu SAM film is seen after exposure to NaOH for longer times or at higher temperatures [21, 54]. Table 22.13. CAM (±3◦ ) values for PFMS/Cu before and after exposure to NaOH (pH 12, 60–80 ◦ C, 30–70 min) and the corresponding rms surface roughness. (From [54]) 60 ◦ C
80 ◦ C
Substrate
CAM (°)
rms (nm)
CAM (°)
rms (nm)
PFMS/Cu FMS/Cu (30 min) PFMS/Cu (50 min) PFMS/Cu (70 min)
132 118 100 87
34 ± 2 29 ± 3 28 ± 3 27 ± 3
132 100 32 35
34 ± 2 29 ± 3 24 ± 1 25 ± 1
22.4 Summary In this chapter, the interfacial properties and chemical stability of SAM films formed on oxidized aluminum or copper substrates by reaction with PFMS, PFDP, OP, DP, and ODP have been discussed. The properties of the SAM films have been studied using complementary techniques, such as XPS, AFM, FFM (a derivative of AFM), CAM, and FT-IRRAS. The chemical stability of PFMS, PFDP, ODP, DP, and OP SAMs formed on Al has been tested. XPS and CAM data confirm the presence of nonperfluorinated and perfluorinated phosphonate and perfluorosiloxy SAMs on oxidized Al. CAM data showed that ODP/Al, DP/Al, and OP/Al are highly hydrophobic, resulting in contact angles of more than 125◦ and critical surface tensions of approximately 20, 20, and 25 mJm−2 , respectively. PFDP/Al and PFMS/Al were shown to be extremely hydrophobic, having contact angles of more than 130◦ and critical surface tensions of approximately 11 and 16 mJm−2 , respectively. The results of a comparative study for the nanoscale tribological properties of PFMS/Al, PFDP/Al, ODP/Al, DP/Al, and OP/Al were also reported. Each SAM surface, PFMS/Al, PFDP/Al, DP/Al, and ODP/Al, shows much lower adhesion and friction than the unmodified Al. PFDP/Al,
278
E. Hoque et al.
which has a fluorocarbon tail and backbone chain resulting in the lowest critical surface tension, shows the lowest adhesion, while DP/Al and ODP/Al, having a hydrocarbon tail and backbone chain yielding higher critical surface tensions, exhibit comparable adhesion which is higher than that of PFDP/Al. The friction force for SAMs with a fluorocarbon backbone chain is higher than that for SAMs with a hydrocarbon backbone chain, in spite of having a lower critical surface tension. The influence of environment (relative humidity and temperature), sliding velocity, and wear was studied and the corresponding friction mechanisms were discussed. The SAMs show a negligible dependence of adhesive force on the relative humidity. However, there is some decrease in the coefficient of friction. This effect could be attributed to lubrication by the adsorbed water layer. The adhesive force for SAMs with a hydrocarbon backbone chain shows a temperature dependence. The adhesive force for SAMs with a fluorocarbon backbone chain, on the other hand, is temperature-independent. The friction force for the hydrocarbon SAMs on Al shows little dependence on the sliding velocity. Fluorocarbon SAMs show larger friction force as well as a larger sliding velocity dependence. The wear behavior of the SAMs is mostly determined by the molecule–substrate bond strength. The long hydrocarbon chain molecules of ODP/Al and the fluorocarbon backbone chains of PFDP/Al have higher critical loads than DP/Al with its short hydrocarbon chain molecules and unmodified Al. From a tribological point of view, it is found that SAMs with compliant and long hydrocarbon backbone chains and those with fluorocarbon backbone chains show highly desirable behavior. XPS and CAM data give an indication that PFMS and PFDP chemisorb onto the oxidized Al, which accounts in part for the good stability of the SAMs. These stability tests provide evidence that ODP/Al produces a more densely packed phosphonate SAM on Al, while DP/Al and OP/Al form a less densely packed SAM. ODP/Al is the most stable SAM, followed by PFDP/Al, DP/Al, PFMS/Al, and OP/Al. The robustness of SAMs on Al correlates with the alkyl chain length and perfluorination. Hydrophobic, low-adhesion, and robust aluminum surfaces have useful applications for MEMS/NEMS, such as the digital micromirror device. The results of these studies are expected to be useful in the choice of proper lubricants and antistiction coatings for MEMS/NEMS devices [10, 11, 19, 42]. XPS and CAM data confirm the presence of perfluorinated alkyl chains on the PFMS/Cu surface. CAM data illustrate the dramatic improvement in hydrophobicity of Cu after the reaction with PFMS yielding contact angles of more than 130◦ and a surface energy of 14 mJm−2 . On the other hand, contact angles are found to be less than 15◦ for unmodified Cu. PFMS/Cu also has lower adhesion and friction force than that of unmodified Cu. Thus, the good agreement between CAM and FFM data indicates that the reaction of PFMS with Cu produces a densely packed perfluorinated alkyl chain SAM. The stability of PFMS/Cu was tested by subjecting it to harsh conditions. The PFMS/Cu SAM film can survive boiling water, boiling nitric acid solution, and warm sodium hydroxide solution for up to 30 min with little or no degradation, as determined by XPS and CAM data. The robustness of the PFMS/Cu SAM film is attributed, at least in part, to the formation of a siloxy– copper (–Si–O–Cu–) bond after reaction of PFMS with oxidized Cu. This conclusion is based upon XPS and FT-IRRAS data indicating the presence of copper hydroxide [Cu(OH)2 ] and the absence of siloxy (–SiO–) groups on oxidized Cu before reaction
22 Self-Assembled Monolayers on Al and Cu
279
with PFMS, the presence of siloxy groups and the absence of copper hydroxide on PFMS/Cu, and the resistance of PFMS/Cu to harsh treatments generally indicating chemisorption. The extremely hydrophobic and stable SAMs on oxidized Cu could have useful applications in corrosion inhibition for microelectronics/nanoelectronics and/or heat-exchange surfaces exploiting dropwise condensation [21, 54].
References 1. Ulman A (1996) Chem Rev 96:1533–1554 2. Ulman A, Evans SD, Shnidman Y, Sharma R, Eilers JE, Chang JC (1991) J Am Chem Soc 113:1499–1506 3. Schreiber F (2000) Prog Surf Sci 65:151–256 4. Scherer J, Vogt MR, Magnussen OM, Behm RJ (1997) Langmuir 13:7045–7051 5. Jennings GK, Laibinis PE (1996) Colloids Surf A 116:105–114 6. Tan YS, Srinivasan MP, Pehkonen SO, Chooi SYM (2004) J Vac Sci Technol A 22:1917– 1925 7. Tremont R, De Jesús-Cardona H, Garcia-Orozco J, Castro RJ, Cabrera CR (2000) J Appl Electrochem 30:737–743 8. Xiao X, Hu J, Charych DH, Salmeron M (1996) Langmuir 12:235–237 9. Blackman GS, Mate CM, Philpott MR (1990) Phys Rev Lett 65:2270–2273 10. Bhushan B, Cichomski M, Hoque E, DeRose JA, Hoffmann P, Mathieu HJ (2006) Microsyst Technol 12:588–596 11. Hoque E, DeRose JA, Kulik G, Hoffmann P, Mathieu HJ, Bhushan B (2006) J Phys Chem B 110:10855–10861 12. Bhushan B (ed) (2005) Nanotribology and nanomechanics—an introduction. Springer, Berlin 13. Kim S, Choi GY, Ulman A, Fleischer C (1997) Langmuir 13:6850–6856 14. Salomon A, Cahen D, Lindsay S, Tomfohr J, Engelkes VB, Frisbie CD (2003) Adv Matter 15:1881–1890 15. Halik M, Klauk H, Zschieschang U, Schmid G, Dehm C, Schütz M, Maisch S, Effenberger F, Brunnbauer M, Stellacci F (2004) Nature 431:963–966 16. Rawlett AM, Hopson TJ, Amlani I, Zhang R, Tresek J, Nagahara LA, Tsui RK, Goronkin H (2003) Nanotechnology 14:377–384 17. Troisi A, Ratner MA (2004) Nano Lett 4:591–595 18. Love JC, Estroff LA, Kriebel JK, Nuzzo RG, Whitesides GM (2005) Chem Rev 105:1103– 1169 19. Hoque E, DeRose JA, Hoffmann P, Mathieu HJ, Bhushan B, Cichomski M (2006) J Chem Phys 124:174710–174715 20. Hoque E, DeRose JA, Hoffmann P, Mathieu HJ (2006) J Surf Anal 13:178–184 21. Hoque E, DeRose JA, Houriet R, Hoffmann P, Mathieu HJ (2007) Chem Mater 19:798–804 22. Hoque E, DeRose JA, Hoffmann P, Mathieu HJ (2006) Surf Interface Anal 38:62–68 23. Ulman A (1991) An introduction to ultrathin organic films: from Langmuir-Blodgett to self-assembly. Academic, San Diego 24. Allara DL, Nuzzo RG (1985) Langmuir 1:45–52 25. Allara DL, Nuzzo RG (1985) Langmuir 1:52–66 26. Brovelli D, Häner G, Ruiz L, Hofer R, Kraus G, Waldner A, Schlösser J, Oroszlan P, Ehrat M, Spencer ND (1999) Langmuir 15:4324–4327 27. Textor M, Ruiz L, Hofer R, Rossi A, Feldman K, Hähner G, Spencer ND (2000) Langmuir 16:3257–3271 28. Pellerite MJ, Dunbar TD, Boardman LD, Wood EJ J Phys Chem B 16:3257–3271
280
E. Hoque et al.
29. Liakos IL, Newman RC, McAlpine E, Alexander MR (2004) Surf Interface Anal 36:347– 354 30. Kelley TW, Boardman LD, Dunbar TD, Muyres DV, Pellerite MJ, Smith TP (2003) J Phys Chem B 107:5877–5881 31. Adolphi B, Jähne E, Busch G, Cai X (2004) Anal Bioanal Chem 379:646–652 32. Frisbie CD, Rozsnyai LF, Noy A, Wrighton MS, Lieber CM (1994) Science 265:2071– 2074 33. Kasai T, Bhushan B, Kulik G, Barbieri L, Hoffmann P (2005) J Vac Sci Technol B 23:995–1003 34. Wei G, Bhushan B, Jacobs SJ (2004) Ultramicroscopy 100:375–389 35. Hornbeck LJ (2001) MRS Bull 26:325–327 36. Tambe NS, Bhushan B (2005) Nanotechnology 16:1549–1558 37. Bhushan B (1998) Tribology issues and opportunities in MEMS. Kluwer, Dordrecht 38. Bhushan B (2003) J Vac Sci Technol B 21:2262–2296 39. Bhushan B, Liu H (2004) Nanotechnology 15:1785–1791 40. Bhushan B (2007) Springer handbook of nanotechnology, 2nd edn. Springer, Heidelberg 41. Bhushan B (1999) Handbook of micro/nano tribology, 2nd edn. CRC, Boca Raton 42. Hoque E, DeRose JA, Hoffmann P, Bhushan B, Mathieu HJ (2007) J Phys Chem C 111:3956–3962 43. Pellerite MJ, Wood EJ, Jones VW (2002) J Phys Chem B 106:4746–4754 44. Laiho T, Leiro JA, Heinonen MH, Mattila SS, Lukkari J (2005) J Electron Spectrosc Relat Phenom 142:105–112 45. Sung MM, Kim Y (2001) Bull Korean Chem Soc 22:748–752 46. Whelan CM, Kinsella M, Ho HM, Maex K (2004) J Electrochem Soc 151:B33–B38 47. Skolnik AM, Hughes WC, Augustine BH (2000) Chem Educator 5:8–13 48. Laibinis PE, Whitesides GM (1992) J Am Chem Soc 114:9022–9028 49. Rose JW (2002) Proc Inst Mech Eng Part A J Power Energy 215:115–128 50. Leach RN, Stevens F, Langford SC, Dickinson JT (2006) Langmuir 22:8864–8872 51. Izumi M, Kumagai S, Shimada R, Yamakawa N (2004) Exp Therm Fluid Sci 28:243–248 52. Ganzevles FLA, van der Geld CWM (2002) Int J Heat Mass Transfer 45:3233–3243 53. Das AK, Kitty HP, Marto PJ, Andeen GB, Kumar K (2000) J Heat Transfer Trans ASME 122:278–286 54. Hoque E, DeRose JA, Hoffmann P, Bhushan B, Mathieu HJ (2007) J Chem Phys 126:114706–114713 55. Somlo B, Gupta V (2001) Mech Mater 33:471–480 56. Hansal WEG, Hansal S, Pölzler M, Kornherr A, Zifferer G, Nauer GE (2006) Surf Coat Technol 200:3056–3063 57. Du T, Luo Y, Desai V (2004) Microelectron Eng 71:90–97 58. Briggs D, Seah MP (1990) Practical surface analysis by Auger and X-ray photoelectron spectroscopy, vol 1, 2nd edn. Wiley, New York 59. Du T, Tamboli D, Desai V, Seal S (2004) J Electrochem Soc 151:G230–G235 60. Hernandez J, Wrschka P, Oehrleinc GS (2001) J Electrochem Soc 148:G389–G397 61. Ramsier RD, Henriksen PN, Gent AN (1988) Surf Sci 203:72–88 62. Moulder JF, Stickle WF, Sobol PE, Bomben KD (1992) Handbook of X-ray photoelectron spectroscopy. PerkinElmer, MN 63. Allara DL, Parikh AN, Judge E (1994) J Chem Phys 100:1761–1764 64. Leung YL, Zhou MY, Wong PC, Mitchell KAR (1992) Appl Surf Sci 59:23–29 65. Franquet A, Biesemans M, Terryn H, Willem R, Vereecken J (2006) Surf Interface Anal 38:172–175 66. Wagner CD, Passoja DE, Hillery HF, Kinisky TG, Six HA, Taylor WT (1982) J Vac Sci Technol 21:933–944
22 Self-Assembled Monolayers on Al and Cu
281
67. Schnyder B, Lippert T, Kötz R, Wokaun A, Graubner V-M, Nuyken O (2003) Surf Sci 532–535:1067–1071 68. Korösi G, Kováts ES (1981) J Chem Eng Data 26:323–332 69. Xu X, He J-W, Goodman DW (1993) Surf Sci 284:103–108 70. da Cruz RS, de Silva JM, Arnold U, Sercheli MS, Schuchardt U (2002) J Braz Chem Soc 13:170–176 71. Bhushan B, Liu H (2001) Phys Rev B 63:245412-1–245412-11 72. Chidsey CED, Loiacono DN (1990) Langmuir 6:682–691 73. Clark ES (1999) Polymer 40:4659–4665 74. Liu H, Bhushan B (2002) Ultramicroscopy 91:185–202 75. McDermott M, Green J, Porter M (1997) Langmuir 13:2504–2510 76. Lide DR (ed) (2004) Handbook of chemistry and physics, 85th edn. CRC, Boaca Raton 77. Tambe NS, Bhushan B (2005) Nanotechnology 16:2309–2324 78. Wallace RM, Chen PJ, Henck SA, Webb DA (1995) J Vac Sci Technol A 13:1345–1350 79. Chen PJ, Wallace RM, Henck SA (1998) J Vac Sci Technol A 16:700–706
23 High Sliding Velocity Nanotribological Investigations of Materials for Nanotechnology Applications Nikhil S. Tambe · Bharat Bhushan
Abstract. The advent of microstructures/nanostructures and the subsequent miniaturization of moving components for various nanotechnology applications, such as microelectromechanical/nanoelectromechanical systems, have given paramount importance to the tribology and mechanics on the nanoscale. Most of these microdevices/nanodevices and components operate at very high sliding velocities (of the order of tens of millimeters per second to a few meters per second). Research conducted on various materials, coatings and lubricants has revealed a strong velocity dependence of friction and adhesion on the nanoscale. However, these investigations have been rendered inadequate owing to the inherent limitations on the highest sliding velocities achievable with commercial atomic force microscopes (AFM) that allow the study of nanoscale phenomena. The development of a new AFM-based technique has enabled high sliding velocities that are of engineering importance to be reached. By incorporating high speed piezo stages in a commercial AFM it is possible to conduct fundamental studies at sliding velocities of scientific as well as engineering importance. The utility of the technique is demonstrated through the mapping of nanoscale friction and wear as a function of sliding velocities. These maps provide significant understanding of the dominant friction mechanisms as well as the conditions at which they transition and would potentially enable selection of materials, coatings and lubricants during the design of nanotechnology applications. Key words: High sliding velocity, Nanotribology, Nanofriction mapping, Nanowear mapping
23.1 Bridging Science and Engineering for Nanotribological Investigations Nanotechnology, defined literally as any technology performed on a nanoscale that has applications in the real world (Feynman 1960), has spurred the development of innovative microsystems/nanosystems with the discovery of novel materials, processes and phenomena on the microscale/nanoscale and has led to the rapid advancement of microlectromechanical/nanoelectromechanical systems (MEMS/NEMS) and their various biological and biomedical applications, BioMEMS. Recent years have seen a multitude of new emerging applications in this field. Commercial applications such as the microfluidic devices that can manipulate tiny amounts of fluids, “labon-chip” sensors used for drug delivery, accelerometers used for automobile air bag deployment and digital micromirror devices used in high-definition TVs and video projectors in homes and theaters are just the tip of the iceberg. In fact these MEMS/NEMS are now believed to be the next logical step in the “silicon revolution.” Visionaries and leading scientists and researchers, presenting at the National Nanotechnology Initiatives workshop on nanotechnology in space exploration held
284
N.S. Tambe · B. Bhushan
in Palo Alto, CA, USA in August 2004, have slated the emerging field of nanotechnology to be the next disruptive technology that will have major impact in the next one to three decades. It is estimated that the annual global impact of products where nanotechnology will play a key role will exceed US $1 trillion by 2015 and would require about two million nanotechnology workers (Roco 2003). 23.1.1 Microtribology/Nanotribology Despite the increasing popularity and technological advances in MEMS/NEMS applications the severe tribological (friction and wear) problems tend to undermine their performance and reliability. In fact, several studies have shown that the tribology and mechanics of these devices are the limiting factors to the imminent broad-based impact of nanotechnology on our everyday lives (Komvopoulos 2003; Maboudian and Howe 1997; Bhushan 1998, 2003, 2004). Miniaturization and the subsequent development of MEMS/NEMS require better tribological performance of the system components and a fundamental understanding of basic phenomena underlying friction, wear and lubrication on the microscale and the nanoscale (Bhushan 1997, 1998, 1999a,b, 2001a,b, 2004). The components used in microstructures/nanostructures are very light (on the order of a few micrograms) and operate under very light loads (on the order of a few micrograms to a few milligrams). Going from the macroscale to the microscale the surface area to volume ratio increases considerably and becomes a cause of serious concern from the tribological point of view. On microscales, surface forces such as friction, adhesion, meniscus forces, viscous drag and surface tension that are proportional to area significantly increase and limit the life and reliability of MEMS/NEMS. As a result, friction and wear (on the nanoscale) of lightly loaded microstructures/nanostructures are highly dependent on the surface interactions. The emergence of the new field of microtribology/nanotribology, which pertains to the experimental and theoretical investigations of interfacial processes occurring during adhesion, friction, wear and thin-film lubrication of sliding surfaces on scales ranging from the atomic and molecular scale to the microscale, and its associated techniques (Bhushan 1999a) have provided a viable means of addressing the tribological issues on microscales/nanoscales. Microtribological/nanotribological investigations can be performed using the surface force apparatus (SFA) and the atomic force microscope (AFM) and have already provided valuable insights into the behavior of materials on the nanoscale (Bhushan et al. 1995; Bhushan 1999a). A sharp tip of the AFM developed by Binnig et al. (1986), sliding on a surface simulates microscale/nanoscale contacts, thus allowing high-resolution measurements of surface interactions. 23.1.2 Historical Perspective for Velocity Dependence of Friction Controlling friction and wear is important in all machine components requiring relative motion. Two basic laws of dry (or conventional) friction are generally obeyed
23 Nanotribology at High Velocity
285
over a wide range of applications. These laws are often referred to as Amontons’s laws, after the French physicist Guillaume Amontons (1699) who rediscovered them in 1699; after Leonardo Da Vinci first described them some 200 years earlier. Amontons’s first law states that the friction force is proportional to the normal load. The second law states that friction force (or coefficient of friction) is independent of the apparent area of contact between the contacting bodies. To these two laws, a third law is sometimes added which is often attributed to Coulomb (1785). It states that the kinetic friction force (or coefficient of friction) is independent of the sliding velocity once motion starts. Coulomb also made a clear distinction between static and kinetic friction. These laws have over the years been found to not hold in many cases (Bowden and Tabor 1950, 1964; Moore 1972; Singer and Pollock 1992; Bhushan 1999b), particularly as dimensions of the components and loads used have continued to decrease, tribological and mechanical properties on the microscale to nanoscale have become very important. Various studies on the microscale/nanoscale have indicated a strong normal load and sliding velocity dependence of friction force (Koinkar and Bhushan 1996; Bhushan and Kulkarni 1996; Bouhacina et al. 1997; Baumberger et al. 1999; Riedo et al. 2003; Liu and Bhushan 2002, 2003a,b; Gnecco et al. 2004; Tambe and Bhushan 2004, 2005a–i). Friction on the atomic scale was first reported by Mate et al. (1987). Investigators have subsequently shown that stick–slip on the atomic scale is the result of the energy barrier that needs to be overcome to jump over atomic corrugations on the sample surface and this stick–slip has been theoretically modeled with classical mechanical models (Tomlinson 1929; Tomanek et al. 1991). Further, the potential barrier required to make jumps during stick–slip follows the logarithm of the sliding velocity (Bouhacina et al. 1997; Hoshi et al. 2000; Riedo et al. 2003; Gnecco et al. 2004). Riedo et al. (2002) investigated the kinetics of capillary condensation in nanoscale sliding friction. A logarithmic dependence of friction force on the scan velocity was reported and it was found that while for partially hydrophilic and nanoscale rough surfaces friction decreased logarithmically, for partially hydrophobic surfaces nanoscale friction increased with velocity. The velocity dependence studied was explained with the help of a model based on the kinetics of capillary condensation.
23.2 Need for Speed: Extending the AFM Capabilities for High Sliding Velocity Studies Extensive as the research efforts have been to characterize and understand the velocity dependence of friction (Fig. 23.1), inherent instrument limitations on the highest sliding velocities achievable with the commercial AFM (less than 100 µm/s) (Tambe and Bhushan 2005a) have stymied research pursuits geared towards obtaining a fundamental understanding of failures resulting from high relative sliding velocities found in many real-world applications (tens of millimeters per second to a few meters per second). Many MEMS/NEMS devices (some examples are shown in Fig. 23.2) are designed to operate at very high rotational speeds, ranging from well above 100,000 rpm as in the case of micromotors (Bhushan 2004) to well above
286
N.S. Tambe · B. Bhushan
Fig. 23.1. Brief summary of research efforts to study the velocity dependence of friction from atomic scale to microscale/nanoscale over a range of sliding velocities. (AFM atomic force microscope) Fig. 23.2. Examples of nanotechnology devices that operate at high sliding velocities: electrostatic micromotor (Tai et al. 1989), microturbine (Spearing and Chen 1997), microgear drive (http://www.sandia.gov) and planetary gear system (Lehr et al. 1996)
1,000,000 rpm for microgas turbines (Frechette et al. 2005). Given the size of these devices (the micromotor has a rotor diameter of 120 µm), the relative sliding velocities that they operate at can reach several meters per second. With the lack of sufficient experimental evidence, theoretical formulations have remained limited (Riedo et al. 2002; Tambe 2005). A study to investigate the velocity dependence of friction, adhesion and wear for sliding velocity ranges of engineering importance would be crucial to the future design and development of microstructures/nanostructures and devices for various nanotechnology applications.
23 Nanotribology at High Velocity
287
23.2.1 Modifications to the Commercial AFM Setup There are a number of commercial AFMs that have been available on the market since 1989. Depending on the design of the AFM, i. e., whether the cantilever tip is scanned or the sample is scanned, AFMs can be broadly classified as largesample and small-sample AFMs, respectively (Bhushan 1999a). Since the tip is held stationary in the small-sample AFM, it provides better resolution and there is less noise contamination of the measured data, particularly for small scan sizes. However, a small-sample AFM, as the name suggests, inherently requires the sample studied to be of small size and weight; otherwise the piezo scanner operation is hindered. A large-sample AFM provides the capability of studying larger samples and also provides the added flexibility needed for mounting samples on custom-built stages for performing a wide array of studies. To achieve higher sliding velocities between the AFM cantilever tip and the sample surface, the primary approach has been to incorporate a customized stage, capable of providing high sliding velocity, into the commercial AFM setup. To date various techniques have been developed that modify the basic AFM setup in order to enhance its measurement capabilities and broaden the scope of tribological studies into higher sliding velocity regimes. One approach has been to mount samples on a shear wave transducer and then drive the transducer at very high frequencies (in the megahertz range) (Yamanaka and Tomita 1995; Scherer et al. 1999; Marti and Krotil 2001; Reinstadtler et al. 2003). In this manner, velocities on the order of a few millimeters per second can be achieved. In techniques employed by researchers using a shear wave transducer, the modulation amplitude, however, remains very small, in the nanometer range. Moreover friction force measurements as made by these researchers do not directly provide a fundamental understanding of nanoscale friction since they use a combination of oscillators, one to achieve high velocity and another for scanning. This does not provide a good estimate for nanoscale friction resulting from actual relative sliding between two components as oscillation of the tip (or the sample) while scanning changes the adhesive force at the tip–sample interface and influences stick–slip behavior. The main challenges while modifying the AFM setup to achieve higher velocities are maintaining a constant velocity profile during the entire scan duration and at the same time scanning on larger surface areas to get a better understanding of nanotribological properties. An alternative approach for accomplishing such a modification is using either a motorized stage or a piezo stage with large amplitude (approximately 100 µm) and relatively low resonance frequency (a few kilohertz). The most important parameters that need to be taken into consideration before implementing such modifications are the maximum travel (or scan size), the resolution, the accuracy and repeatability (hysteresis in forward and reverse scans), a constant velocity requirement over the entire scanning length, and the maximum sample size and weight allowable (this is particularly important for piezo stages). Motorized stages can prove to be relatively inferior to piezo stages owing to vibration-related issues, particularly in the direction perpendicular to the scanning axis (along the normal load axis of the cantilever tip). Such a stage can result in the cantilever tip being subjected to a constant change in normal load and the “chatter” can contaminate measurements. Piezo stages provide better performance in this respect. However, the disadvantage in using piezo stages
288
N.S. Tambe · B. Bhushan
is that they can be limited by the maximum travel, the scanning frequency and the sample size. To achieve high velocities (in the tens of millimeters per second range and beyond) using a piezo stage implies that both the scanning frequency and the scan size should be relatively high. In most instances, however, there is a trade-off between the maximum travel and the frequency of the piezo stage. Piezo stages can introduce mechanical noise as a result of “ringing” which occurs because of poor piezo tracking. Open loop operation (i. e., no active feedback) can result in the piezo losing its tracking accuracy at high frequencies and thus the requirement of having a constant velocity over the entire scan length is not met. Instead, therefore, a closed loop operational mode can be employed; however, this is generally at the expense of the maximum travel distance (scan size) achievable. We discuss two approaches as developed by Tambe and Bhushan (2005a) and Tao and Bhushan (2006, 2007). They incorporate custom-calibrated piezo stages into a commercial AFM setup. 23.2.1.1 Large Amplitude Piezo Stage Tambe and Bhushan (2005a) modified a commercial AFM setup (D3100, Nanoscope IIIa controller, Digital Instruments, Santa Barbara, CA, USA) by incorporating a custom-calibrated piezo stage (P621.1CL, HERA Nanopositioner with capacitive feedback, Polytec PI, Karlsruhe, Germany). This piezo stage has an unloaded resonance frequency of 800 Hz and a maximum scan length of 100 µm. The piezo stage drive units are based on P-885 PICMA® low-voltage multilayer piezo ceramics, having an extremely stiff friction-free flexure system with excellent guiding accuracy (small tilt angles over the entire travel range). This design dispenses ultrahigh power electronics that can cause conventional piezo stacks to heat excessively when operated at high frequencies (D. Rego, personal communication 2003). Table 23.1 lists specifications of the piezo stage. In the modified setup, the single-axis piezo stage is oriented such that the scanning axis is perpendicular to the long axis of the AFM cantilever (this corresponds to the 90◦ scan angle mode of the commercial AFM). Scanning is achieved by providing a triangular voltage pulse to the piezo amplifier. Figure 23.3a shows a schematic of the experimental setup and a cross-sectional schematic of the piezo stage. The piezo crystal drives the flexure and the sample mounted on top of it at the desired frequency, corresponding to the input voltage pulse. The displacement is monitored using an integrated capacitive feedback sensor, located diametrically opposite to the piezo crystal as shown in Fig. 23.3a. The capacitive sensor has a stationary Table 23.1. Specifications of the P621.1CL piezo stage Closed loop travel Closed loop linearity Stiffness Maximum load Lateral force limit
100 µm 0.01% 0.35 N/µm 10 N 10 N
Tilt (perpendicular to scan axis) Electrical capacitance Unloaded resonance frequency Resonance frequency at 20g Operating temperature range
3 µrad 1.5 µF (±20%) 800 Hz (±20%) 520 Hz (±20%) −40 to 120 ◦ C
23 Nanotribology at High Velocity
289
Fig. 23.3. a Modifications made in the commercial AFM setup using the single-axis piezo stage and a cross-sectional view showing construction details of the piezo stage. The integrated capacitive sensors are used as feedback sensors to drive the piezo. The piezo stage is mounted on the standard motorized AFM base and is operated using independent amplifier and controller units driven by a frequency generator (not shown)
290
N.S. Tambe · B. Bhushan
Fig. 23.3. b Operational flowchart showing the various components of stage operation. (From Tambe and Bhushan 2005a)
target element mounted on the stage block and a moving probe element mounted on the flexure. The capacitance change, corresponding to the stage displacement, gives an indication of the amount of displacement. The stage can be operated in both open loop and closed loop operational modes. In the closed loop mode, the capacitive sensor signals are used as feedback by the piezo controller to provide better guiding and tracking accuracy. The closed loop position control of piezoelectricdriven stages using capacitive feedback sensors provides linearity of motion better than 0.01% with nanometer resolution and a stable drift free motion (Anonymous 2003).1 Figure 23.3b shows an operational flowchart for the piezo stage. For scanning purposes, a triangular voltage pulse with 0–10-V maximum peak-to-peak voltage is input to the piezo amplifier using a function generator. The triangular pulse gives constant displacement with respect to time and thereby ensures a constant scanning velocity, which is necessary during friction measurements. The input voltage pulse is fed into the piezo amplifier and controller circuit, where it was amplified to 0–100-V range and used for driving the piezo crystal. The piezo displacement is monitored by the capacitive sensor and in closed loop operational mode is fed back to the controller. The piezo stage displacements as measured by the capacitive sensors at two high frequencies in open and closed loop modes of operation are shown in Fig. 23.4. The open loop mode of operation (i. e., when no feedback was employed) had the advantage that piezo tracking was much better during the reversals in the direction of the stage motion while scanning, as seen in Fig. 23.4. However, in the absence of active tracking, the piezo displacements stray away considerably from the input triangular profile. This is particularly evident at 250-Hz scan frequency when piezo 1
The piezo stage drive units, based on P-885 PICMA® low-voltage multilayer piezo ceramics, have an extremely stiff friction-free flexure system with excellent guiding accuracy (typically less than 5 µrad pitch/yaw over the entire travel range). The closed loop position control of piezoelectric-driven stages using capacitive feedback sensors provides linearity of motion better than 0.01% with nanometer resolution and a stable drift-free and hysteresis-free motion.
23 Nanotribology at High Velocity
291
Fig. 23.4. Piezo stage displacements (in volts) as measured using the capacitive sensor for open loop and closed loop operation at 100and 250-Hz scanning frequencies; and piezo hysteresis under open and closed loop operational modes at 100-Hz scan frequency. (From Tambe and Bhushan 2005a)
“ringing”, high-frequency noise, becomes a significant factor. The capacitive sensor can be used for providing feedback to the piezo controller and this improves the piezo tracking considerably. In this case, the stage reversals are not linear and so stage velocity is not constant towards the ends when the stage reverses direction. However, the tracking is significantly improved as a result of the active feedback loop and this ensures that the velocity remains constant over most of the stage travel. Nonetheless the piezo does show hysteretic behavior for a 100-Hz scan frequency (Fig. 23.4). Hysteresis (area enclosed by the curve) is higher in the closed loop operational mode, but this mode also gives a linear response over a large portion of the total scan length. The linear response implies a constant velocity and hence is more useful for friction force measurements. The commercial AFM setup modified with this customcalibrated nanopositioning stage is capable of achieving high sliding velocities up to 10 mm/s. The output of the capacitive sensor was calibrated to get the actual scan size in microns. A capacitive sensor output voltage of 1 V corresponds to a scan size of 10 µm. This calibration data were obtained from the stage manufacturer (Physik Instrumente, Karlsruhe, Germany). Figure 23.5 shows the calibration chart for the piezo stage over a large range of scanning frequencies from 1 to 500 Hz and for scan sizes of 0–25 µm. The chart was obtained by running the stage in closed loop operational mode and monitoring the capacitive sensor signals for different
292
N.S. Tambe · B. Bhushan
Fig. 23.5. Calibration chart for the piezo stage giving scan sizes over a range of operating voltages (0–10 V) and frequencies (1–500 Hz) for the closed loop operational mode. Solid lines and dashed lines correspond to constant input voltage and constant velocity, respectively. (From Tambe and Bhushan 2005a)
input voltage pulses. Constant input voltage lines (solid lines) indicate that the stage performance is that of a critically damped system. The constant velocity lines (dashed lines) indicate the range of sliding velocities achievable with the modified AFM setup. The maximum input voltage to drive the piezo is 10 V and the shaded area on the calibration chart indicates a region that was beyond the operational limit of the piezo stage. It should be noted that the stage has a maximum scan length of 100 µm. On the basis of the calibration chart, the input voltage pulse can be selected in order to obtain a specific scan size at a given scanning frequency or for determining the scanning frequency necessary at a given scan size to achieve the required scanning velocity. 23.2.1.2 Ultrasonic Linear Piezo Drive Tao and Bhushan (2006, 2007) have developed another piezo-based technique for achieving sliding velocities up to 200 mm/s. The system includes a customcalibrated piezo stage or a piezo ultrasonic linear drive (M663.465, Physik Instrumente), a stage controller (C865, Physik Instrumente), and a custom-designed software application for operation control (Fig. 23.6a,b). The control application sends motion parameters to the controller, and the controller calculates the exact position of each step and sends the position signal to the stage. The piezo stage has a maximum scan length of 19 mm and a maximum velocity of 400 mm/s. Table 23.2 lists specifications of the stage. The key part of the stage, shown in Fig. 23.6a, is a rectangular monolithic piezoceramic plate (the stator) with
23 Nanotribology at High Velocity
293
Fig. 23.6. a An ultra-high-velocity piezo stage, b stage and control system, c a digital proportional integral derivative servo filter (courtesy of Physik Instrumente), and d block diagram of the data collection and processing system used for frictional force measurement. (From Tao and Bhushan 2006) Table 23.2. Parameters of the M-663.465 piezomotor stage Travel range Design resolution Max. velocity Maximum acceleration Maximum normal load capacity
19 mm 0.1 µm 400 mm/s 100,000 mm/s2 0.5 kg
two excitation electrodes which are resonated using a 12 V power supply. Depending on the desired direction of motion, the left or the right electrode is excited to produce high-frequency eigenmode oscillations up to 200 kHz. Simultaneous eigenmodes result in quasi-elliptical motion. An alumina friction tip (pusher) attached to the plate pushes a slider with a glued friction bar which rests on a set of bearings. Through its contact with the friction bar, the piezoceramic plate provides micro impulses and drives the slider forward or backward. While the longitudinal oscillation component provides the energy as driving force, the transverse component serves to change the pressure of the friction tip against the friction bar. The transverse oscillation
294
N.S. Tambe · B. Bhushan
energy determines the maximum frictional force and hence the holding and driving force of the stage. The stage transfers motion through friction and can produce accelerations up to 20g and velocities up to 400 mm/s with a repeatability of about 50 nm. The stage is equipped with an optical position reference photosensor. It is located approximately in the middle of the travel range and is used to reference the absolute position of the stage within 1-µm repeatability. Before operation, an initialization process is performed to locate the home position (reference) point by the photosensor. During motion, the increments of the linear scale from a home position point are converted to determine the position using a linear optical encoder. The stage controller includes a motion processor, which contains a trajectory generator. The trajectory generator receives profile parameters from the host computer and calculates a new desired position at each time step based on parameters and the current position of the stage obtained using the position sensor (Fig. 23.6b). The output of the trajectory generator is combined with the actual encoder position to calculate a 32 bit position error passed through a proportional integral derivative (PID) type digital servo filter (Fig. 23.6c). The resultant value from the filter is then sent to an external amplifier using a pulse width modulation signal generator and then to the stage. The servo filter used is a PID algorithm, with velocity and acceleration feedforward terms and an output scale factor. An integration limit provides an upper bound for the accumulated error. An optional bias value can be added to the filter calculation to produce the final motor output command. A limiting value for the filter output provides additional constraints. The PID digital servo filter is expressed as (Anonymous 2002) ⎡ Outputn = ⎣ K p E n + K d E k − E (k−1) +
n j=0
⎤
E j × K i /256 + K vff (CmdVel/4) + K aff (CmdAccel × 8)⎦
× K out /65536 + bias ,
(23.1)
where E n and E k are the accumulated error terms, K i is the integral gain, K d is the derivative gain, K p is the proportional gain, K aff is the acceleration feed forward, K vff is the velocity feed forward, bias is the DC motor offset, and K out is the scale factor for the output command To achieve a desired motion profile, the parameters in (23.1) have to be adjusted. For smooth motion and minimum settling time, the system automatically switches between two servo-loop parameter sets (each consisting of proportional, integral, and derivative terms, velocity feed forward and acceleration feed forward) during motion. The switching is done as a function of the distance between the current position and the target position. As long as the distance exceeds a certain value (default value 10 µm), the first parameter set is used. When the stage reaches a position within the value, the second PID set immediately becomes active.
23 Nanotribology at High Velocity
295
Application software was custom-developed using Microsoft Visual Basic 6.0 to calibrate the system and manipulate the stage. A software interface (a set of functions) is provided by the manufacturer of the stage for Visual Basic user application development. Utilizing these functions, the application reads the status and the parameters from the stage, changes the system parameters (such as PID parameters), receives motion parameters from user input, and sends commands to the stage controller through the serial port of the host computer. The purpose of stage calibration is to find an optimal set of parameters for the desired motion profile and to prevent vibration during motion. A critical requirement for high-velocity tests is constant velocity during motion. For the stage calibration, the application could read the exact motion profile from the stage. A block diagram of the data collection and processing system used for the friction force data measurement is shown in Fig. 23.6d. In the modified setup, the single-axis piezo stage is oriented such that the scanning axis is perpendicular to the long axis of the AFM cantilever (this corresponds to the 90◦ scan angle mode of the commercial AFM). For a scan size as large as 1000 µm, the tilt of sample is a serious problem. The translation limit for the Z piezo of a DI 3100 AFM is about 6.5 µm. For a 1-mm scan length, an inclined angle of 0.34◦ could result in a vertical translation of 6 µm. In order to level the sample over 1-mm
Fig. 23.7. a The tilt stage for sample leveling on the AFM. b Scan lines before and after leveling on 100 µm. (From Tao and Bhushan 2006)
296
N.S. Tambe · B. Bhushan
length, a tilt stage was designed (Fig. 23.7a). A 76.2-mm-diameter micrometer head (Scherr Tumico, St James, MN, USA) with a resolution of 2.54 µm was mounted on a plate. The distance between the supporting points of the plate is 100 mm. The piezo stage is fixed on the top surface of the tilt stage. For a scan size of 1000 µm on the sample, the height difference is 25.4 nm for a 2.54-µm height change on the 100-mm scale. In order to level the sample, the AFM was used to scan a small
Fig. 23.8. Comparison of displacement and velocity profiles of the ultra-high-velocity piezo stage operated using default parameters and optimal parameters at a 200 mm/s on 1000 µm and b 400 mm/s on 5000 µm
23 Nanotribology at High Velocity
297
size (10 µm) and the scanned lines were observed. If the line was inclined, then the micrometer head was carefully adjusted until the height difference was close to zero. Then a larger scan size was used (20 µm), and the sample was leveled again by adjusting the micrometer head. Increasing the scan length up to 100 µm (the limit of the DI 3100 scan size) resulted in the sample being leveled. An example of a surface scan line before and after leveling on a 100-µm scan size is shown in Fig. 23.7b. After leveling for 100 µm, the scan length of the AFM was set to zero and the velocity stage was operated. Thus, the scan length was increased from 100 to 1000 µm and the micrometer head was adjusted accordingly so that the center line of the Z piezo was within 6 µm. The PID parameters were adjusted to obtain constant velocity during the scan length. At the high-velocity regime, constant velocity with ±5% error could only be obtained on part of the scan length (Fig. 23.8). After calibration, the system was operated at the optimal parameters. As shown in Fig. 23.8, there is a significant difference between the velocity profiles from optimal parameters and default parameters. It should be noted that for different velocities and sliding distances, the parameter values might vary. For a given velocity, in order to obtain a constant velocity on the motion curve over at least 50% of the total distance, the moving distance has to be larger than a certain value. For example, for 200 mm/s, the minimum moving distance is 1 mm. When the velocity increases to a maximum of 400 mm/s, the distance also increases. The required minimum sliding distance as
Fig. 23.9. Minimum scan length required to achieve constant velocity as a function of velocity Table 23.3. Parameters of the C-865.161 piezomotor controller Velocity Minimum scan Acceleration/deceleration (mm/s) length (mm) (mm/s2 ) 1 10 20 40 100 200 300 400
0.025 0.025 0.1 0.2 0.5 1 3 5
8000 20000 50000 60000 70000 70000 70000 80000
P1
70 35 15 10 15 15
I1
160 60 50 20 30 30
D1
P2
I2
D2
120 120 120 120 160 100
180 10 100 80 50 50 50 50
500 500 1000 1000 1000 1000 1000 1000
300 30 100 100 100 100 100 100
298
N.S. Tambe · B. Bhushan
a function of velocity is presented in Fig. 23.9 and is listed in Table 23.3. Also listed in Table 23.3 are acceleration/deceleration values and PID control parameters. The acceleration/deceleration values are used by the system for starting/stopping the motion. When the stage starts moving, it accelerates at the acceleration value provided until it reaches the programmed velocity. It continues in motion at the programmed velocity, and then decelerates using the deceleration value until it stops at the specified position. The ratio of the constant velocity in the whole motion profile is dominated by the acceleration/deceleration values. At lower target velocities, lower values of acceleration/deceleration were used to prevent overshooting of the stage. As shown in (23.1), the PID parameters are used by the system to calculate the next step position based on the commanded position and the real position from sensor feedback. For smooth moving, two sets of PID parameters are used by the system. The two sets of PID parameters switch at the programmed distance of 50 m. Therefore, for moving distance lower than 50 m, the primary PID set is meaningless, and the values are left blank in the table. 23.2.2 Friction Investigations on the Microscale/Nanoscale at High Sliding Velocities During the experiments, the AFM cantilever is held stationary by maintaining a scan size of zero. The AFM controller feedback functioned in the conventional manner and maintained a constant normal load by adjusting the vertical deflection of the cantilever. As the mounted sample is scanned below the AFM tip, the torsional deflections of the tip are recorded by the photodiode detector. The raw signals from the optical detection system are directly routed to a high-speed data acquisition A/D
Fig. 23.10. Comparison of friction data measured for a scan size of 2 µm using the commercial and the modified AFM setup. (From Tambe and Bhushan 2005a)
23 Nanotribology at High Velocity
299
Fig. 23.11. a Comparison of friction force measurements conducted at different relative sliding velocities for a single-crystal silicon sample with native oxide using the commercial and the modified AFM setup. b Comparison of friction force data obtained using the commercial AFM and the high-velocity piezo stage based AFM setups. (a From Tambe and Bhushan 2005a; b from Tao and Bhushan 2006)
300
N.S. Tambe · B. Bhushan
board [National Instruments, NI PCI-6040E (PCI-MIO-16E-4), 12-bit, 16 analog input multifunction DAQ]. The photodiode signals are a direct indication of the friction force encountered by the tip during sliding motion. The data acquisition system is synchronized to collect friction signals, with the input voltage pulse acting as the trigger. Raw friction data is processed using high sampling rates, up to 25 kilosamples/s/channel. Figure 23.10 shows typical trace minus retrace (TMR) signals collected using the commercial AFM setup and the AFM setup modified with the large-amplitude piezo stage. TMR signals are the friction signals measured during the “trace” and “retrace” scans in an AFM. In the case of the commercial AFM setup, noise sets in as a result of piezo “ringing” at high frequencies. Piezo ringing became evident beyond 30 Hz and it influenced the measurements more at higher scan frequencies. This noise often results in an underestimation of the friction force values. The friction data obtained using the modified AFM setup show no piezo ringing even for high frequencies. The capacitive feedback loop maintains a constant velocity profile over the entire scan length. This ensures that the friction data obtained using the modified setup are the correct estimate of the friction resulting from the relative sliding between the tip and the sample even at high velocities. To eliminate any further discrepancies in the friction measurements that could result from any uncharacteristic behavior of the piezo stage during reversals, only the central portion of the TMR signals are considered while calculating the average values of friction. Figure 23.11a shows the actual friction force values calculated over a range of velocities using the commercial and the modified AFM setup as developed by Tambe and Bhushan (2005a). The values plotted are the average value of the TMR friction data over one scan cycle and each point on the graph was obtained by averaging data from at least ten such TMR cycles. Friction data obtained using a commercial AFM at high velocities (more than 100 µm/s) are not always reliable for reasons stated above and hence are not plotted. It is seen that the data obtained using the two modified setups follow the same trend and the friction values are similar to those obtained with measurements made using the commercial AFM, indicating that the data are highly reproducible. Figure 23.11b shows friction force values obtained using a commercial AFM, the large-amplitude piezo stage and the ultrasonic linear piezo drive for the same scan size and normal load. The data obtained by each method are identical for overlapping ranges of sliding velocities. This demonstrates the successful integration of the stage into the commercial AFM and its utility for conducting fundamental studies. It should be noted that to achieve very high sliding velocities (up to 200 mm/s), Tao and Bhushan (2006, 2007) used larger scan sizes. However, the scan size independence of nanoscale friction force above certain scan size values is known from previous studies (Koinkar and Bhushan 1996; Tambe and Bhushan 2005a) and has been corroborated by analytical models developed by Bhushan and Nosonovsky (2003).
23.3 Microscale/Nanoscale Friction and Wear Studies at High Sliding Velocities The development of techniques for achieving high sliding velocities has enabled nanotribological investigations at scales and operating conditions that are pertinent
23 Nanotribology at High Velocity
301
to many nanotechnology applications. It is critical to evaluate MEMS/NEMS component materials and coatings for failures resulting from high static friction and wear. Tambe (2005) conducted investigations on various materials, coatings and lubricants at high sliding velocities using the newly developed techniques and proposed a comprehensive friction model to explain the various dominant friction mechanisms. In this section we discuss techniques for mapping nanoscale friction and wear as a function of sliding velocity and normal load. These maps provide invaluable insights into the various friction regimes on the nanoscale as well as help identify, classify and even demarcate various dominant mechanisms responsible at specific sliding velocities. Another tool we discuss, one that will potentially help researchers further, is a predictive tool for selection of materials for specific nanotechnology applications. The tool is a map of friction and adhesion as a function of the elastic modulus of the material. Such mapping of friction behavior on the nanoscale as a function of material properties is bound to enable a much fundamental yet broader understanding of the material property dependence of friction. 23.3.1 Nanoscale Friction Mapping: Understanding Normal Load and Velocity Dependence of Friction Force Tambe and Bhushan (2005i) have studied nanoscale friction as a function of normal load and sliding velocity and used a novel mapping technique that helps identify, demarcate and classify different friction mechanisms. Figure 23.12 shows the nanoscale friction maps obtained to study the friction force dependence on normal load and sliding velocity for single-crystal silicon with a native oxide layer [Si(100)], highly oriented pyrolytic graphite (HOPG), diamond-like carbon (DLC), aluminum and polymers poly(dimethylsiloxane) (PDMS) and poly(methyl methacrylate) (PMMA). Experiments were conducted with the modified AFM setup as described in Sect. 23.2.1.1 (Tambe and Bhushan 2005a). Measurements were performed using square-pyramidal Si3 N4 tips with nominal radii of 30–50 nm, mounted on gold-coated triangular Si3 N4 cantilevers having a nominal spring stiffness of 0.32 and 0.58 N/m. Experiments were performed over a range of velocities from 5 µm/s to 1 mm/s and in a controlled environment of 20 ± 2 ◦ C and 50 ± 5% relative humidity. A scan size of 2 µm was used for friction measurements. The contour maps were generated from friction force data collected at different normal loads and sliding velocities. The contours represent constant friction force lines and are marked by the value of the friction force in nanonewtons. The contours for each material are characteristic of the friction behavior exhibited by that particular material even though there are some features that can be classified as universal irrespective of the material and are discussed next. On the basis of the specific arrangement of contour lines, the dominant friction mechanisms can be identified. At the bottom half of Fig. 23.12, the most commonly observed features in the contours (or the characteristic contours) as found from our study are summarized and the significance of each is stated. Horizontal contour lines indicate the velocity-independent nature of friction force. This behavior is found at high velocities for HOPG, at moderately high velocities for aluminum and at low velocities for PMMA. Studies on the the dependence
302
N.S. Tambe · B. Bhushan
Fig. 23.12. Contour maps showing friction force dependence on normal load and sliding velocity and summary of characteristic contour patterns and their significance. HOPG highly oriented pyrolytic graphite, DLC diamond-like carbon, PMMA poly(methyl methacrylate), PDMS poly(dimethylsiloxane). (From Tambe and Bhushan, 2005i)
of nanoscale friction force on velocity (Riedo et al. 2003) indicate that friction force becomes constant relative to sliding velocity when the atomic scale stick–slip occurring at low sliding velocities loses its dominance. A constant friction force with respect to sliding velocity would appear as a horizontal contour on a friction map and is the reason, for example, for the horizontal contours seen for HOPG at high velocities. Vertical contour lines indicate a normal load independence of friction force. In all the samples studied this behavior is not seen, although the steep contour lines for Si(100), HOPG and aluminum indicate that there is a very small normal load dependence on friction at low velocities. For all practical instances it would be impossible to find a material that shows normal load independence of friction force. Some other characteristic contours are those with slanting lines with either a positive or a negative slope. These friction contours arise from the microscale stick– slip related contributions or from the formation of meniscus bridges by preferential condensation of liquid films at the sliding interface, particularly for hydrophilic interfaces. Researchers have shown that friction force increases with velocity owing to atomic scale stick–slip (Riedo et al. 2003; Gnecco et al. 2004). On a friction map, this increase would be seen as slanted contours with a positive slope, i. e., an increase in friction force as one moves from left to right on the map. In this study,
23 Nanotribology at High Velocity
303
HOPG, DLC and aluminum showed this behavior. Stick–slip can also originate as a result of other mechanisms and result in a decrease in friction force with an increase in sliding velocity (Ruths et al. 2004). This behavior would result in slanted contours with a negative slope on the friction map such as seen for PMMA at high sliding velocities and for PDMS. The friction force arising from meniscus contributions, found for hydrophilic surfaces such as Si(100), results in a drop in friction force with an increase in sliding velocity. A minimum threshold equilibration time is necessary for the formation of stable meniscus bridges at contacting and nearcontacting asperities for a sliding interface (Bouquet et al. 1998; Tambe and Bhushan 2005b). With increasing velocity, fewer meniscus bridges build up at the interface and thus the overall contribution to friction force drops with an increase in velocity. This behavior would manifest itself in the form of slanted contour lines with negative slope on the friction map. Contour maps can also consist of concentric contour lines. DLC appears to show this kind of behavior. A concentric contour map implies a peak friction force corresponding to a critical normal load and sliding velocity. Beyond that point any further increase in either the normal load or the sliding velocity would result in a decrease in the friction force. This kind of behavior typically would imply a phase transformation by formation of a low-friction phase at the interface or localized melting at the contact zone. Phase transformation is known to occur for DLC, resulting in a low-friction graphite-like layer by an sp3 to sp2 phase transition (Grill 1997; Tambe and Bhushan 2005e). Localized melting would arise from very high frictional energy dissipation and is expected particularly in the case of polymer materials. In this study, PMMA appeared to show concentric lines at moderately high velocities; however, for the given range of normal load and sliding velocity, the experimental evidence is not sufficient to support this hypothesis. Another characteristic contour is the one where contour lines change direction suddenly. This implies a sudden change in the dominant friction mechanism. Si(100) and PMMA showed such behavior. In the case of silicon, researchers have reported the formation of a Si(OH)4 layer at the sliding interface at high sliding velocities (Mizuhara and Hsu 1992). It is believed, on the basis of the contour maps, that this effect is initiated at a particular sliding velocity and that this is the reason for the sudden change in friction force. For Si(100), contribution of the meniscus to friction force is also a dominant mechanism; however, both these mechanisms are known to coexist in tandem and result in a decrease in friction force with velocity (Liu and Bhushan 2003b). 23.3.2 Nanoscale Wear Mapping: Wear Studies at High Sliding Velocities For wear studies, previously an AFM-based “continuous microscratch” technique has been developed by Sundararajan and Bhushan (2001). That technique gave a direct dependence of the scratch depth on the applied normal load, thereby giving the critical normal load for failure for materials and coatings. This technique has been used by Liu and Bhushan (2002) to study the wear behavior of various self-assembled monolayer lubricants as well as substrates. The microscratch technique, however, does not provide a dependence of normal load, sliding velocity and the number
304
N.S. Tambe · B. Bhushan
of sliding cycles simultaneously that is needed to generate wear maps. Wear maps give a direct representation of various operating conditions and can help identify and demarcate the corresponding wear mechanisms. On the macroscale, various researchers have used an approach where they run a multitude of experiments and then plot wear maps based on the failure data and the rate of removal of material during wear (Bowden and Tabor 1950, 1964; Singer and Pollock 1992; Bhushan 1999b). Lim and Ashby (1987) and Lim et al. (1987) constructed wear maps using empirical data as well as theoretical analysis. They demonstrated the utility of the wear-mechanism mapping method as a way of classifying and ordering wear data and of showing the relationships between competing wear mechanisms. A novel AFM-based technique was developed by Tambe and Bhushan (2005h) to generate wear maps on the nanoscale by varying the sliding velocity, the number of sliding cycles and the normal load. For the nanowear mapping experiments, Si tips (coated with Ti/Pt) with nominal cantilever stiffness of 3.5 N/m and nominal tip radii of 40 nm were used. For generating nanoscale wear maps, the sample was oscillated using the piezo stage and simultaneously the AFM tip was dragged perpendicular to the direction of motion of the sample (Fig. 23.13). The sample oscillation frequency (scan speed), the AFM tip velocity and the normal load were controlled to achieve appropriate relative sliding velocities, normal loads and a specific number of sliding cycles. The relative sliding velocity was varied by changing the scan speed, while the number of sliding cycles was varied by varying the rate of movement of the AFM tip (i. e., the velocity of the AFM tip while sliding perpendicular to the direction of motion of the sample). This varying rate of movement results in the AFM tip residing for different time intervals on the sample surface and thereby the number of sliding cycles obtained at each location on the sample surface is different. For all wear mapping experiments, a scan size (i. e., the magnitude of sample oscillation) of 5 µm was chosen. After the wear tests, the sample surface was imaged using the same AFM tip but at relatively lower normal loads. Wear maps were first obtained by varying only one parameter out of the three at a time: normal load, sliding velocity and number of sliding cycles. To study the effect of increasing normal load and increasing number of sliding cycles at constant sliding velocity, the AFM tip was programmed to make controlled movements at desired normal loads
Fig. 23.13. Novel technique for wear mapping, achieved by controlled movement of the AFM tip while the sample is sliding perpendicular to the direction of movement of the AFM tip. (From Tambe and Bhushan 2005h)
23 Nanotribology at High Velocity
Fig. 23.14. Wear maps showing effect of normal load and sliding velocity, as well as the effect of the number of sliding cycles for DLC. (From Tambe and Bhushan 2005h)
305
306
N.S. Tambe · B. Bhushan
and velocities. This controlled motion of the AFM tip was achieved using custom software code written in NanoScriptTM (Anonymous 1999). Next, wear maps were generated to study the effect of normal load and sliding velocity simultaneously. To achieve varying sliding velocities across the scan area, the input voltage pulse for the piezo stage was slightly modified. In normal operation a triangular voltage pulse is provided to the piezo stage to achieve scanning operation and to obtain a constant sliding velocity. For the wear mapping experiments, a parabolic voltage pulse was used as the input to drive the piezo. This resulted in a steady increase in sliding velocity across the scan area. The synchronized and controlled movement of the AFM tip and the sample is a novel approach for obtaining nanoscale wear maps and helps generate a visual representation of the sample surface wear as a function of sliding velocity, applied normal load and number of sliding cycles (Tambe and Bhushan, 2005h). Using this approach, one can generate a wear map in one single experiment, as against the rather cumbersome approach where researchers conduct multiple experiments at different normal loads, sliding velocities and for a different
Fig. 23.15. Nanowear map (AFM image and schematic) illustrating the influence of sliding velocity and normal load on the wear of DLC resulting from phase transformation. The curved area shows debris lining and is indicative of the minimum frictional energy needed for phase transformation. (From Tambe and Bhushan 2005h)
23 Nanotribology at High Velocity 307
Fig. 23.16. Utility and impact of nanotribological investigations conducted at high sliding velocities on future design of microelectromechanical systems(MEMS)/nanoelectromechanical systems and other nanotechnology applications. (From Tambe 2005)
308
N.S. Tambe · B. Bhushan
number of sliding cycles and then generate a contour map for the sample wear based on individual data points obtained from each experiment. The wear maps in Fig. 23.14 indicate that wear debris particles are generated only for certain combinations of sliding velocities, normal loads and number of sliding cycles. In these wear maps only one operating parameter was varied at a time. To obtain true wear maps that can allow investigation of different wear mechanisms simultaneously, it is necessary to vary both the normal load and the sliding velocity and investigate the resulting wear. Here we present a nanowear map for a DLC sample that was obtained by simultaneously varying the normal load and the sliding velocity over the entire scan area to investigate wear resulting from phase transformation of DLC. The wear map generated for a normal load range of 0–1000 nN and a sliding velocity range of 0–2.5 mm/s in this fashion is shown in Fig. 23.15. Wear debris was seen to form only for particular sliding velocities and normal loads, i. e., beyond certain threshold frictional energy dissipation; hence, the wear area was curved, indicating that for low velocities and low normal loads there is no phase transformation. For clarity, the wear mark corners are indicated by white dots in the AFM image and the various zones of interest over the entire wear mark are schematically illustrated in Fig. 23.15. Analogous to nanowear mapping, nanofriction maps can also be generated using the same technique by monitoring the evolution of friction force over a period of time. Nanoscale friction maps were discussed in the previous section. In that case, friction force was plotted as a function of the operating parameters: sliding velocity and normal load. By introducing time dependence into the friction maps, one can use the resulting nanofriction maps in conjunction with the nanowear maps to provide valuable information regarding the operating parameter dependence of nanoscale friction and wear. Tambe and Bhushan (2005e) have demonstrated the effectiveness and utility of these techniques when used in tandem while studying the phase transformation related reduction in friction and wear of DLC.
23.4 Closure Nanotribological investigations at high sliding velocities provide not only a fundamental understanding for behavior of materials as individual entities but also give a bigger picture that is necessary for developing newer, better and longer-lasting solutions for the next-generation nanodevices that are more reliable and efficient. Figure 23.16 schematically summarizes how nanotribological investigations at high sliding velocities can impact the design and selection of materials, coatings and lubricants for nanotechnology applications. Fundamental studies can be conducted by emulating the nanoscale contacts using an AFM. New techniques developed for high-speed scanning allow for investigating friction behavior at sliding velocities that are pertinent to the operation of many real-world microdevices/nanodevices and components such as the microgear drive (Fig. 23.16). By developing a fundamental understanding of the failure mechanisms involved, various materials, coatings and lubricants can be evaluated for incorporation into future design and development of nanotechnology applications.
23 Nanotribology at High Velocity
309
References 1. Amontons G (1699) Mem Acad R A 257–282 2. Anonymous (1999) Nanoscope® command reference manual, version 4.42, appendix D: lithography 3. Anonymous (2002) Pilot motion processor user’s guide. Performance Motion Devices, Lincoln 4. Anonymous (2003) User manuals PZ-106E and PZ-62E. http://www.polytecpi.com 5. Baumberger T, Berthoud P, Caroli C (1999) Phys Rev B 60:3928–3939 6. Bennewitz R, Meyer E, Bammerlin M, Gyalog T, Gnecco E (2001) In: Bhushan B (ed) Fundamentals of tribology and bridging the gap between the macro- and micro/nanoscales. Kluwer, Dordrecht, pp 53–66 7. Bhushan B (1997) Micro/nanotribology and its applications. NATO ASI series E: applied sciences, vol 330. Kluwer, Dordrecht 8. Bhushan B (1998) Tribology issues and opportunities in MEMS. Kluwer, Dordrecht 9. Bhushan B (1999a) Handbook of micro/nanotribology, 2nd edn. CRC, Boca Raton 10. Bhushan B (1999b) Principles and applications of tribology. Wiley, New York 11. Bhushan B (ed) (2001a) Modern tribology handbook, vol 2: materials, coatings and industrial applications. CRC, Boca Raton 12. Bhushan B (ed) (2001b) Fundamentals of tribology and bridging the gap between the macroand micro/nanoscales. Kluwer, Dordrecht 13. Bhushan B (2003) J Vac Sci Technol B 21:2262–2296 14. Bhushan B (ed) (2004) Springer handbook of nanotechnology. Springer, Heidelberg 15. Bhushan B, Kulkarni AV (1996) Thin Solid Films 278:49–56 16. Bhushan B, Israelachvili JN, Landman U (1995) Nature 374:607–616 17. Binnig G, Quate CF, Greber C (1986) Phys Rev Lett 56:930–933 18. Bouhacina T, Aime JP, Gauthier S, Michel D (1997) Phys Rev B 56:7694–7703 19. Bouquet L, Charlaix E, Ciliberto S, Crassous J (1998) Nature 396:735–737 20. Bowden FP, Tabor D (1950) The friction and lubrication of solids, part I. Clarendon, Oxford 21. Bowden FP, Tabor D (1964) The friction and lubrication of solids, part II. Clarendon, Oxford 22. Carpick RW, Salmeron M (1997) Chem Rev 97:1163–1944 23. Coulomb CA (1785) Mem Math Phys X Paris 161–342 24. Feynman RP (1960) Eng Sci 23:22–36 25. Frechette LG, Jacobson SA, Breuer KS, Ehrich FF, Ghodssi R, Khanna R, Wong CW, Zhang X, Schmidt MA, Epstein A (2005) J Microelectromech Syst 14:141–152 26. Gnecco E, Bennewitz R, Gyalog T, Loppacher C, Bammerlin M, Meyer E, Guntherodt H-J (2000) Phys Rev Lett 84:1172–1175 27. Gnecco E, Bennewitz R, Pfeiffer O, Socoliuc A, Meyer E (2004) In: Bhushan B (ed) Springer handbook of nanotechnology. Springer, Heidelberg, pp 631–660 28. Grill A (1997) Surf Coat Technol 94–95, 507–513 29. Hoshi Y, Kawagishi T, Kawakatsu H (2000) Jpn J Appl Phys Part 1 39:3804–3807 30. Johnson KL, Woodhouse J (1998) Tribol Lett 5:155–160 31. Koinkar V, Bhushan B (1996) J Vac Sci Technol A 14:2378–2391 32. Komvopoulos K (2003) J Adhes Sci Technol 17:477–517 33. Lehr H, Abel S, Doppler J, Ehrfeld W, Hagemann B, Kamper KP, Michel F, Schulz C, Thurigeen C (1996) Proc SPIE 2906:202–210 34. Lim SC, Ashby MF (1987) Acta Metall 35:1–24 35. Lim SC, Ashby MF, Brunton JH (1987) Acta Metall 35:1343–1348 36. Liu H, Bhushan B (2002) Ultramicroscopy 91:185–202 37. Liu H, Bhushan B (2003a) Ultramicroscopy 97:321–340 38. Liu H, Bhushan B (2003b) J Vac Sci Technol A 24:1528–1538
310
N.S. Tambe · B. Bhushan
39. Maboudian R, Howe RT (1997) J Vac Sci Technol B 15:1–20 40. Marti O, Krotil H-U (2001) In: Bhushan B (ed) Fundamentals of tribology and bridging the gap between the macro- and micro/nanoscales. Kluwer, Dordrecht, pp 121–135 41. Mate CM, McClelland GM, Erlandsson R, Chiang S (1987) Phys Rev Lett 59:1942–1945 42. Mizuhara K, Hsu SM (1992) In: Dowson D et al (eds) Wear particles, Elsevier, Amsterdam, pp 323–328 43. Moore DF (1972) The friction and lubrication of elastomers, 1st edn, Pergamon, New York 44. Reinstadtler M, Rabe U, Scherer V, Hartmann U, Goldade A, Bhushan B, Arnold W (2003) Appl Phys Lett 82:2604–2606 45. Riedo E, Levy F, Brune H (2002) Phys Rev Lett 88:185505-1–185505-4 46. Riedo E, Gnecco E, Bennewitz R, Meyer E, Brune H (2003) Phys Rev Lett 91:084502-1– 084502-4 47. Roco MC (2003) MRS Bull 28:416–418 48. Ruths M, Berman AD, Israelachvili JN (2004) In: Bhushan (ed) Springer handbook of nanotechnology. Springer, Heidelberg, pp 585–591 49. Scherer V, Bhushan B, Arnold W (1999) Surf Interface Anal 27:578–586 50. Singer IL, Pollock HM (1992) Fundamentals of friction: macroscopic and microscopic processes. NATO series E: applied sciences, vol 220. Kluwer, Boston 51. Spearing SM, Chen KS (1997) Ceram Eng Sci Proc 18:11–18 52. Sundararajan S, Bhushan B (2001) J Mater Res 16:437–445 53. Tai YC, Fan LS, Muller RS (1989) Proc IEEE Micro Electro Mech Syst 1–6 54. Tambe NS (2005) Nanotribological investigations of materials, coatings and lubricants for nanotechnology applications at high sliding velocities. PhD dissertation, The Ohio State University. http://www.ohiolink.edu/etd/send-pdf.cgi?osu1109949835 55. Tambe NS, Bhushan B (2004) Nanotechnology 15:1561–1570 56. Tambe NS, Bhushan B (2005a) J Phys D Appl Phys 38:764–773 57. Tambe NS, Bhushan B (2005b) Nanotechnology 16:2309–2324 58. Tambe NS, Bhushan B (2005c) J Vac Sci Technol A 23:830–835 59. Tambe NS, Bhushan B (2005d) Ultramicroscopy 105:238–247 60. Tambe NS, Bhushan B (2005e) Scr Mater 52:751–755 61. Tambe NS, Bhushan B (2005f) Appl Phys Lett 86:061906:1-3 62. Tambe NS, Bhushan B (2005g) Nanotechnology 16:1549–1558 63. Tambe NS, Bhushan B (2005h) Tribol Lett 20:83–90 64. Tambe NS, Bhushan B (2005i) Appl Phys Lett 86:193102 65. Tao Z, Bhushan B (2006) Rev Sci Instrum 77:103705 66. Tao Z, Bhushan B (2007) J Vac Sci Technol A 25:1267–1274 67. Tománek D, Zhong W, Thomas H (1991) Europhys Lett 15:887–892 68. Tomlinson GA (1929) Philos Mag Ser 7(7):905–939 69. Yamanaka K, Tomita E (1995) Jpn J Appl Phys 34:2879–2882 70. Zwörner O, Hölscher H, Schwarz UD, Wiesendanger R (1998) Appl Phys A 66:S263–S267
24 Measurement of the Mechanical Properties of One-Dimensional Polymer Nanostructures by AFM Sung-Kyoung Kim · Haiwon Lee
Abstract. The chapter reviews recent studies on atomic force microscopy (AFM) based measurements of mechanical properties of 1D nanostructures by analyzing their technical aspects and their reliability for the nanostructures with a particular focus on polymer nanofibers prepared by the electrospinning method, a method typically used to produce polymer nanostructures. First we introduce various AFM-based techniques, such as the three-point bending test, electrostatic resonant-contact AFM, nanoindentation, and the tensile test, for measuring the mechanical properties of 1D polymer nanostructures. Recent studies on the mechanical properties of electrospun polymer nanofibers are detailed with a focus on the reinforcement of composite nanofibers and the size effect of electrospun nanofibers. Lastly, for enhancing the test reliability, we state matters that demand special attention in AFM-based measurements of mechanical properties of 1D polymer nanostructures. Key words: Atomic force microscope, Electrospinning, 1D nanostructures, Mechanical properties, Size effect
24.1 Introduction Over the last three decades, nanometer-scale structures have received significant attention owing to their exceptional properties compared with those of bulk materials. Since the discovery of carbon nanotubes (CNTs), there has been increased interest in one-dimensional (1D) nanostructures because of their unique electrical, mechanical, and optical properties, and practical applicability [1]. Besides CNTs, these 1D nanostructures include nanofibers, nanowires, nanobelts, nanoribbons, and nanorods, which are generally prepared by spinning or thermal evaporation. The mechanical properties of the structures are considered very important for the formation of nanofibrous scaffolds with structural integrity in tissue engineering or biomedical engineering, integration into microsystems, and the determination of material applications. The behavior of a material under stress is expressed in terms of mechanical properties such as tensile strength, elongation (or strain), Young’s modulus (or elastic modulus or tensile modulus), and toughness. The tensile strength and elongation correspond to the stress and strain, respectively, at the breaking point of a sample. Young’s modulus and toughness are the slope and lower area of the stress–strain curve, respectively. Young’s modulus describes the rigidity of a material and thus presents high values for metals but low values for elastomers. The toughness refers to the brittleness of the material. The mechanical properties have been found to change
312
S.-K. Kim · H. Lee
as the dimensions are reduced from microscale to nanoscale. The size dependence of the mechanical properties of 1D nanostructures has been demonstrated for various materials such as CNTs [2], polypyrrole nanotubes [3], MoS2 nanotubes [4], gold [5], silicon [6], silver [7], and SiO2 [8] nanowires, ZnO nanobelts [9], SiC nanorods [10], and chromium [11] and silicon [12] nanocantilevers. Their mechanical properties showed different size-dependencies. For example, the elastic modulus of polypyrrole nanotubes dramatically increases with decreasing nanotube thickness. In contrast, the effective Young’s modulus of thin chromium cantilevers decreases as the cantilevers become thinner. On the other hand, the average modulus of gold nanowires, with diameters ranging from 40 to 250 nm, is independent of diameter. For polycrystalline materials, these size effects have been attributed to an increase in the area of grain boundaries as the grains are downsized in nanostructures. That is, dislocations at the crystal planes sheared between individual grains are blocked effectively so that the materials become harder. For the opposite case, this effect is explained by sliding motions at grain boundaries [13]. Polymeric nanofibers, however, have not been investigated as thoroughly owing to the difficulty in preparing and integrating them into macro systems. Recently, a number of reports have been published regarding the preparation of polymeric nanofibers via the electrospinning technique, which has its roots in the current requirement for developing novel and versatile materials [14]. Although numerous methods, such as in situ scanning and transmission electron microscopy testing, are now available to investigate the mechanics of 1D nanostructures, it would be still difficult to measure those of polymeric nanofibers owing to difficulties in manipulating and isolating the structures at the nanoscale. The recent development of atomic force microscopy (AFM) enabled the measurement of extremely small force or deformation for the nanofibers. Here we review recent studies on AFM-based measurements for mechanical properties of 1D nanostructures analyzing their technical aspects and their reliability for the nanostructures with a particular focus on polymer nanofibers prepared by the electrospinning method, a method typically used to produce polymer nanostructures.
24.2 AFM-Based Techniques for Measuring the Mechanical Properties of 1D Polymer Nanostructures Since its development in 1986 by Binnig et al. [15], the atomic force microscope has demonstrated widespread applicability as well as extremely high resolution. Owing to its unique scanning and operating mode, AFM has been applied for analyzing surface properties, even on an insulator, measuring extremely small forces and manipulating and fabricating nanostructures. In the atomic force microscope, a tip at the end of a cantilever experiences a repulsive or attractive force when in contact with a sample surface as a function of tip-to-sample separation. A detector measures the displacement of the cantilever beam caused by the forces acting on the tip. The most common AFM system has incorporated an optical position-sensitive diode as a detection method, to which the laser beam reflected from the back of the cantilever is aligned and is thus displaced on the diode when the cantilever is
24 AFM Measurement of the Mechanical Properties
313
subjected to displacements. This allows the measurement of tiny deformations of a material in the subangstrom range and thus of extremely small forces acting on the tip in the piconewton range. Various promising approaches to the exploration of the mechanical properties of 1D polymer nanostructures have demonstrated that AFM is quite useful in elucidating the mechanics of the structures in the nanometerscale regime. These AFM-based techniques include the three-point bending test, electrostatic resonant-contact AFM, nanoindentation, and the tensile test. The measurement techniques that use the three-point bending test and nanoindentation are based on force versus distance (F/D) curves which are used to measure the vertical force that the atomic force microscope tip applies to the surface. Local variations in the form of the F/D curve indicate variations in the local elastic properties. Figure 24.1 shows a typical F/D curve and the sequential movement of the AFM tip following an F/D curve. At the right side of the curve, the scanner is fully retracted and the cantilever stays undeflected because the tip is not touching the sample (1 in Fig. 24.1). As the scanner extends, the cantilever remains undeflected until it comes close enough to the sample surface for the tip to undergo an attractive van der Waals force. The tip snaps into the surface (2 in Fig. 24.1). Similarly, the cantilever suddenly bends towards the surface. The cantilever deflects gradually as the scanner continues to extend (3 in Fig. 24.1). After applying the maximum force,
Fig. 24.1. a Representative force versus distance curve and b the movement of an atomic force microscopy (AFM) tip at the z position corresponding to each position numbered in a
314
S.-K. Kim · H. Lee
the cantilever deflection retraces the same curve as the scanner is withdrawn from the position at the maximum force (4 and 5 in Fig. 24.1). The slope of the F/D curve shifts depending on the surface conditions. Hard samples give a higher slope of the F/D curve and higher adhesion holds the tip in contact with the surface, bending the cantilever strongly towards the surface as the scanner pulls away from the surface. Similarly, the slope of the F/D curve becomes lower for soft samples. Three-point bending is bending a component by placing the specimen on two supports and then applying a load on it between the supported ends (Fig. 24.2). The elastic modulus (E) of the suspended 1D nanostructures under a vertical load is calculated using the three-point bending theory with the following relationship [16] which is derived by applying the boundary conditions to a mathematical description for deformation of a clamped-clamped beam under a center load: E = FL 3 /(192δI) ,
(24.1)
where F is the force applied by the atomic force microscope tip, L is the suspended length of the nanofiber, δ is the deflection of the nanofiber at its midspan, and I is the second moment of the area of the nanofiber (I = πd 4 /64, where d is the diameter of the nanofiber). When a small deflection is applied by the atomic force microscope tip at the midpoint of a suspended 1D nanostructure, the deflection of the suspended nanostructure is due to tensile and compressive deformation and to shear deformation. Because shear deformation is assumed to be negligible in calculating the theoretical elastic modulus of nanofibers, it is necessary to form slender nanostructures to satisfy the condition (L/d) > 16 [16]. In (24.1), δ is obtained from the force–displacement curves and can be obtained using a literature method [17]. Figure 24.3 shows that the nanofiber underwent elastic deformation whilst under a load from an atomic force microscope tip, because the loading and unloading cycles of the F/D curves coincide with each other. To confirm the possibility of indentation by the atomic force microscope tip, the nanofibers have to be reimaged using the atomic force microscope after the original force– displacement curves of the nanofiber have been measured using the atomic force microscope tip. Because the AFM images can be distorted if a very small fragment of the nanofiber attaches itself to the atomic force microscope tip, the accuracy of the F/D curves should be confirmed by comparing sequential AFM images. However, the bending test provides a simple setup and convenient process that allows high reproducibility and therefore reliability.
Fig. 24.2. A nanofiber deflected by an AFM tip
24 AFM Measurement of the Mechanical Properties
315
Fig. 24.3. Force–displacement curves obtained on the supporting Si structure as a reference (dashed line) and at the midpoint of a suspended nanofiber (solid line). The deflection of the nanofiber δ signifies the difference between the nanofiber and the reference curve at the maximum force applied. (From [18])
Similarly, AFM-based nanoindentation techniques are based on the displacement of a cantilever beam, i.e., a force–distance curve, to determine hardness, modulus, and other mechanical properties of the materials. Here, hardness means average pressure and is calculated as the applied load is divided by the projected contact area of the indentation [19]. The tip, as an indenter, generally has an axisymmetric or pyramidal geometry with a small radius of curvature at the apex because it is blunted with continued operation. For an axisymmetric indenting tip, the elastic modulus (E) of 1D nanostructures placed on a flat substrate can be obtained with a known shape of the tip and the contact area from the following relationships [20]: 2 1 1 − νc2 1 − νs2 dF 1/2 = E A , = + , (24.2) r dΔh π1/2 Er Ec Es where F is the loading force, h is the deformation, A is the contact area, E r is the reduced modulus, E is the Young modulus, ν is the Poisson ratio, and c and s refer to the cantilever and sample, respectively. The deformation Δh is obtained from the force versus displacement curves and the contact area A is usually assumed to be the curvature radius of the tip apex. Ko et al. [21] electrospun polyacrylonitrile (PAN)/CNT composite nanofibers to evaluate the reinforcement. A nonlinear load– deformation relationship became distinct for the PAN/CNT composite nanofibers, the effect of which is attributed to the finite deformation of the fiber under a large load (Fig. 24.4). The elastic modulus of the nanofiber is calculated using the linear portion of the load–deformation curve under low forces (less than 100 nN). The dynamic characteristics of the cantilever beam allowed for the determination of tip–sample interaction forces such as stiffness and adhesion. This dynamic technique is referred to as either mechanical resonant-contact AFM (or atomic force acoustic microscopy) [22,23] or electrostatic resonant-contact AFM [24], according to the excitation mode of the cantilever vibration. To measure the contact stiffness of the surface, resonant-contact AFM monitors the displacement of a cantilever vibrated by modulators such as an ultrasonic transducer, a piezoceramic, or a function
316
S.-K. Kim · H. Lee
Fig. 24.4. Load indentation curves obtained by AFM nanoindentation testing for polyacrylonitrile (PAN)/single-walled carbon nanotube (CNT) nanofibrils. (From [21])
generator. In fact, this technique utilizes a noise spectrum when the cantilever is in contact with a sample. That is, the resonance frequencies of the cantilever depend on the contact stiffness and thus shift relative to those of the free cantilever, which is due to tip–sample interaction forces. Cuenot et al. [24] have made polypyrrole nanotubes suspended over the pores of a poly(ethylene terephthalate) membrane to measure their elastic modulus [24]. To excite the vibration of the cantilever in contact with the nanotubes, an external electric field was applied between the sample holder and the atomic force microscope head using a function generator. The resonance spectrum obtained (Fig. 24.5) shows three peaks; F1 and F2 are due to flexural vibrations of the cantilever, and T1 is due to torsional vibrations. They analyzed the relation between two flexural vibration frequencies of the cantilever by measuring the resonance frequencies on several nanotubes, and the “tip–surface” contact was consecutively modeled. The nanotube stiffness and elastic modulus were determined using the theory of cantilever vibration [25, 26]. To measure the mechanical properties of 1D nanostructures, researchers have used a nanoscale three-point bending test rather than a tensile test. This is due to the difficulties in elongating 1D nanostructures by attaching both ends of the nanostructure to an atomic force microscope tip. For the bending test, the vertical deflection of an atomic force microscope cantilever is examined to obtain the elastic property of the material. In spite of the difficulties in manipulating 1D nanofibers, a tensile test of a single nanofiber was recently conducted using a piezoresistive atomic force microscope cantilever tip [27]. Deflection of the cantilever tip results in a linear change in the resistance, which is converted to load readings (Fig. 24.6). A plot of stress versus strain was successfully drawn for a 700-nm diameter poly(ethylene oxide) (PEO) fiber. In this study, one end of the nanofiber was attached to a movable optical microscope stage and the other end of the nanofiber to a piezoresistive atomic force microscope cantilever tip. The nanofiber was stretched by moving the microscope
24 AFM Measurement of the Mechanical Properties
317
Fig. 24.5. Resonance spectrum obtained from a cantilever– polypyrrole nanotube contact system. The peaks of F1 and F2 correspond to flexural vibrations of the cantilever, and T1 corresponds to torsional vibrations. The inset shows the resonance peak of the free cantilever. (From [24])
Fig. 24.6. The tensile test of the nanofiber using a piezoresistive AFM tip. PEO poly(ethylene oxide). (From [27])
stage and the force was measured via the deflection of the cantilever. The elastic modulus of the PEO nanofiber was found to be about 45 MPa. For electrospun nylon-6,6 nanofibers, Zussman et al. [28] performed a tensile test. Through a nanomanipulation process using a microscope, one end of an individual electrospun nanofiber was attached to an atomic force microscope cantilever and the other to a stainless steel wire. Figure 24.7a shows a series of images of the attached nanofiber undergoing a tensile test. The deflection of the cantilever, δ, was observed using the microscope equipped with a CCD camera. Figure 24.7b presents the stress–strain behavior of nylon-6,6 electrospun nanofibers and a microfiber. During electrospinning, the sharp edge of a grounded collector disk was used to collect and align the electrospun nanofibers. They concentrated on a point that the mechanical behavior of the electrospun nanofibers could be varied by controlling a take-up wheel velocity, expecting the alignment of the macromolecules in the fiber matrix. Indeed, the nanofiber collected at a take-up velocity of 20 m/s revealed an elastic modulus approximately 2 times greater than one collected at a take-up velocity of 5 m/s (Fig. 24.7b). However, the measured mechanical properties had broad distributions that arose from experimental inaccuracies.
318
S.-K. Kim · H. Lee
Fig. 24.7. a A series of images taken during a tensile test for a nylon-6,6 electrospun nanofiber. δ specifies the deflection of the cantilever. b Stress–strain curve of the nylon-6,6 electrospun nanofibers and a microfiber prepared by melt extrusion processing. (From [28])
24.3 Mechanical Properties of Electrospun Polymer Nanofibers Among a large number of synthetic or fabrication methods for generating 1D nanostructures, electrospinning provides a straightforward and inexpensive method of producing polymeric nanofibers with diameters in the range of tens of nanometers [14, 29]. The formation of polymeric nanofibers via electrospinning is based on the elongation of a viscoelastic jet emitted from microfluidic sources under an electric field. Figure 24.8a shows a schematic of the basic setup for electrospinning that is performed with processing variables such as acceleration voltage, solution concentration, viscosity, surface tension, and flow rate. Figure 24.8b shows polymer nanofibers with random orientation, collected on a substrate via electrospinning.
24 AFM Measurement of the Mechanical Properties
319
The polymer nanofibers produced by the electrospinning method may have a rich variety of potential applications, such as filter media, drug delivery, sensor devices, tissue engineering scaffolds, artificial tissues, and reinforced composites [14]. The nanofibers undergo severe structural deformations during electrospinning [30]. Thus, the mechanical properties of the electrospun nanofibers have to be investigated prior to their integration into future applications. The long and flexible electrospun polymer nanofibers have to be oriented to measure the mechanical properties. Kameoka et al. [31, 32] developed a scanned electrospinning system for depositing oriented nanofibers. In this system, a triangular polymeric tip or an arrow-shaped silicon tip was used as a source, and the counter electrode containing a substrate chip was attached to a motor. Controlling the rotating velocity allowed the orientation of nanofibers with a standard deviation of 2% in diameter; the nanofibers were then deposited along the rotating direction on the substrate. Recently, Li and Xia [14] employed a collector split into two strips to uniaxially align the electrospun nanofibers. The strips consisted of metals or highly doped silicon and were separated by a gap ranging from hundreds of micrometers to several centimeters. During electrospinning in this system, the electric field was split into two parts toward the strips in the vicinity of the collector, which contributes to aligning as-spun fibers. In addition, the alignment was attributed to the other electrostatic force formed between the charged fiber and the two grounded electrodes. The AFM image in Fig. 24.9a shows a nanofiber that is linearly oriented on a trench structure of a silicon substrate which is placed in the middle of the two split strips. The inset in Fig. 24.9a shows a magnified AFM image of the fiber. To make a precise measurement, the diameter of the nanofibers deposited on the trench structure is measured from the AFM height profiles, as shown in Fig. 24.9b, top and bottom. The heights of the nanofibers obtained from Fig. 24.9b, top and bottom are 81.6 and 82.7 nm, respectively, and the average heights measured at selected points along the nanofiber can be used to calculate the elastic modulus. In this case, the step height of the trench structure is 102 nm. The measured height profiles also make it possible to confirm whether the nanofibers were deflected because of their own weight before obtaining the force–displacement curves from theatomic force microscope.
Fig. 24.8. a A basic setup for electrospinning and b scanning electron microscope image of the electrospun polymer nanofibers randomly collected on a silicon substrate
320
S.-K. Kim · H. Lee
Fig. 24.9. a An AFM contact mode image of the oriented nanofiber on a Si trench structure by using the split electrodes in an electrospinning process. The inset shows an AFM image taken at high magnification. b The height profiles of a nanofiber taken along the direction of the corresponding line A and line B
24 AFM Measurement of the Mechanical Properties
321
Fig. 24.10. Transmission electron microscope images of a a poly(vinyl alcolhol) (PVA) nanofiber and b a ferritin-incorporated PVA nanofiber
The mechanical properties of polymer nanofibers have been improved using reinforcements, which were embedded in nanofibers by mixing them in an electrospinning solution. Mack et al. [33] prepared PAN composite nanofibers containing graphite nanoplatelets as reinforcements using an electrospinning method. They expected that the addition of graphite nanoplatelets, which were synthesized using graphite-intercalation chemistry, would enhance the mechanical strength of a composite owing to their high surface area and large aspect ratio. The elastic modulus of electrospun composite nanofibers with diameters ranging from 5 to 500 nm was measured by a nanoindentation technique, resulting in an elastic modulus 2 times higher than that of a pristine PAN fiber for 4 wt % graphite nanoplatelet/PAN composite nanofibers. Shin et al. [18] utilized the split collector to investigate the reinforcement of a polymeric composite nanofiber containing ferritin nanoparticles. A three-point bending test resulted in the enhancement of the elastic modulus of poly(vinyl alcohol) (PVA)/ferritin composite nanofibers (6.3–8.1 GPa), which was 2 times higher than that of the pure PVA nanofibers (3.0–3.6 GPa). They suggested that the enhancement was attributed to hydrogen bonds formed between the peptide chains of the ferritin shell and the PVA matrix. The transmission electron microscope images in Fig. 24.10 show the electrospun PVA nanofibers with diameters of about 100 nm. Figure 24.10a shows a very thin and uniform PVA nanofiber, but on the other hand Fig. 24.10b shows that the ferritins are well dispersed in a PVA nanofiber, without any of the aggregation that can sometimes occur during electrospinning. Similarly, CNTs have been incorporated into polymer nanofibers with the expectation that the thermal and electrical conductivity and mechanical properties of nanofibers would be improved. Zhou et al. [34] employed the electrospinning method to prepare PEO and PVA nanofibers containing multiwalled CNTs (MWNTs). It was found that the morphology of the CNT/polymer nanofibers was affected by the embedded CNTs. Therefore, they focused on evaluating the elastic deformation of MWNTs incorporated in nanofibers by considering the tortuosity factor and calculating the scale, rather than investigating the reinforcement. According to their results, the elastic
322
S.-K. Kim · H. Lee
Fig. 24.11. Variation of the measured elastic modulus for poly(2-acrylamido-2-methyl-1propanesulfonic acid) nanofibers as a function of the diameter [36]
deformation of MWNTs that occurs in the solidified composite nanofibers remains to a relatively large degree in the polymers with relatively high elastic moduli. Ko et al. used load indentation curves to measure the elastic modulus of PAN nanofibers filled with single-walled CNTs (SWNTs). They found that the elastic modulus of the composite nanofibers increased with an increase in the SWNT content of the polymer and was improved by up to 120% for nanofibers containing 4 wt % SWNTs. Although the mechanical properties of CNT-embedded composite nanofibers were successfully characterized by AFM-based techniques, the morphologies of composite nanofibers and the alignment of CNTs in nanofibers had to be controlled more precisely to obtain high reliability. The elastic modulus of PEO was measured by Tan et al. [27]. The tensile test of the nanofiber using a piezoresistive atomic force microscope tip demonstrated that the electrospun PEO nanofiber with a relatively bigger diameter of 700 nm had a modulus of 45 MPa, slightly lower than that of roll-cast PEO films (59.5 MPa). More recently, they characterized electrospun poly(l-lactic acid) (PLLA) nanofibers with diameters ranging from 140 to 270 nm using both a tensile test and a threepoint bending test [35]. The elastic modulus of the PLLA nanofibers obtained from the tensile test using a commercial tensile tester was 4.8 ± 1.9 GPa, which was reconfirmed by a three-point bending test (4.7 ± 1.2 GPa). The size dependence of the mechanical properties of 1D nanostructures has been largely demonstrated for various bulk materials but only to a small degree for polymer nanofibers. Recently, electroactive poly(2-acrylamido-2-methyl-1-propanesulfonic acid) (PAMPS) nanofibers with diameters ranging from 50 to 110 nm were electrospun and subjected to a three-point bending test [36]. The elastic modulus of the PAMPS nanofibers revealed size-dependent characteristics (Fig. 24.11). As the diameter of the nanofiber decreased, the elastic modulus of the nanofiber increased exponentially at a diameter of about 70 nm. It was suggested that the transformation of the mechanical properties was due to the orientation of polymer molecules inside nanofibers arising from the strong strain forces in polymer jets. On this subject, there are previous reports showing that polymer chains of the resultant polymer nanofibers are oriented along the fiber axis owing to the high strain rate of polymer jets having
24 AFM Measurement of the Mechanical Properties
323
a magnitude on the order of 104 /s [30, 37–39]. If experimental parameters are identical in the electrospinning process, one may assume that the change in the diameter of a polymer nanofiber is dependent on the strain rate of the ejected polymer jets. There are more detailed arguments regarding the reason for the observed changes in the properties for polymer nanofibers that are based on the effects of either surface tension [40] or molecular orientation [41]. Cuenot et al. [40] used electrostatic resonant-contact AFM to measure the elastic modulus of Ag and Pb nanowires and conductive polypyrrole nanotubes with diameters ranging from 30 to 250 nm. Analyzing the measured stiffness of the structures, they found an abrupt increase in the elastic modulus of the 1D nanostructures when the diameters were smaller than the critical diameter (70 nm for Ag, 100 nm for Pb and polypyrrole [24]). For the polypyrrole nanotubes with a diameter around 35 nm, the measured elastic modulus was about 120 GPa. With the expectation that the elastic modulus would change owing to structural modifications, they measured the elastic modulus of polypyrrole samples synthesized at two different temperatures (−10 and 20 ◦ C). Taking into account the result that there was no effect of synthesis temperature on the elastic modulus, they suggested that there are surface tension effects on the nanostructure stiffness according to the theory of small deflections of clamped beams. Their calculation fitted the experimental data well and supported the conclusion that the observed increase of the elastic modulus with decreasing diameter is essentially due to surface tension effects. More recently, Arinstein et al. [41] dealt with these phenomena by arguing against the existence of surface tension effects. They advocated the concept of the elastic modulus, i.e., the slope of the stress–strain curve. According to their statement, the stress caused by the surface tension forces does not depend on the elongation, while the stress caused by the elastic force is proportional to the fiber elongation (i.e., strain). They proposed that the stretching of polymer molecules (i.e., their orientational ordering) has a major effect on the increase of the elastic moduli of the fibers. Taking into account their observation by X-ray analysis that there had been no dramatic changes in nanofiber structures, they focused on the size of the oriented regions by estimating the orientation correlation length within the amorphous polymer portion using scaling concepts in polymer physics. The estimated cross section of the ordered region was 300 nm, which was about half the critical diameter (abruptly increasing point) for electrospun nylon-6,6 nanofibers. At the moment, we need to consider the difference in their experimental methods. To deduce the elastic modulus of 1D nanostructures, Cuenot et al. [24] measured the bending displacement of the suspended nanostructures, but Arinstein et al. measured the tensile deformation of nanofibers by stretching them with both ends attached, respectively, to an atomic force microscope tip and a stainless steel wire [28]. Because polymers are relatively flexible, the explanation from the latter looks more plausible for the polymer nanofibers prepared by electrospinning. However, surface tension forces would have to be considered in the estimation of the mechanical properties for nanoscale objects.
24.4 Test of Reliability of AFM-Based Measurements Owing to their small size, 1D nanostructures with diameters smaller than several hundred nanometers are difficult to manipulate and measure precisely. Often, one will
324
S.-K. Kim · H. Lee
introduce some assumptions to evaluate the mechanical properties of 1D nanostructures. These frequently cause an unacceptable error when evaluating the mechanical properties using the data obtained. During electrospinning, polymer nanofibers with a noncircular cross section can be produced owing to a strong impingement with a collector surface. Shin et al. [36] demonstrated the ellipsoidal nature of a polymer nanofiber in contact with the collector surface. Figure 24.12 shows the ellipsoidal cross section of the nanofiber electrospun on a silicon substrate. Liu et al. [42] also reported that the electrospun SU8 polymer nanofiber containing quantum dots was tightly bound to the substrate after electrospinning, presenting a semicircular cross section of the nanofiber. This cross-sectional deformation is attributed to insufficient solvent evaporation such that the solidification of the bound fiber occurs on a collecting substrate. The measurement of the elastic modulus using the three-point bending test requires accurate dimensions of the cross section of the nanofibers because the elastic modulus is related to the diameter raised to the fourth power [43]. Researchers usually choose the height as the diameter of a fiber from AFM images rather than the width in order to avoid tip artifacts resulting from its conical shape [44]. In the case of nanofibers deposited on a trench structure, the nanofiber dimensions in the suspended region may be different from those on the supporting region taking into account the cross-sectional deformation. Therefore, the fiber dimensions need to be carefully determined with a correction for the limitations of the atomic force microscope. For the polymer nanofibers with an ellipsoidal cross section, one can use a reduced diameter by defining the height as the short diameter and the width as the long diameter. The second moment of inertia of the section can be replaced as follows: πa3 b πd 4 Ic = Ie , (24.3) 64 64 where Ic is the second moment of the circular area, Ie is the second moment of the ellipsoidal area, d is the reduced diameter of the nanofiber, a is the short diameter (height) of the nanofiber, and b is the long diameter (width) of the nanofiber. There is a linear correlation between the reduced diameter and the measured height, for which the correlation coefficient is 0.96 (Fig. 24.13).
Fig. 24.12. Scanning electron micrograph of a cross section of the deformed nanofiber impinged on a Si wafer [36]
24 AFM Measurement of the Mechanical Properties
325
Fig. 24.13. The relationship between the reduced diameter and the height of the measured nanofibers
Measuring the mechanical properties of 1D nanostructures requires the precise positioning of the atomic force microscope tip regardless of direct or indirect force measurement. A linear F/D curve indicates the agreement of a retracting and an extending cycle (R–E cycle) (Fig. 24.14a). However, a nonlinear F/D curve (Fig. 24.14b) can be obtained owing to the unstable movement of the tip during the R–E cycle, the mislocation of the tip, and the use of a tip with an inappropriate spring constant. The accuracy of force measurements depends in part on the sensitivity of the atomic force microscope, but also in part on the accuracy of the knowledge of the spring constant. The value of the spring constant is very sensitive to geometric variations of the cantilever, which should be verified by a resonant-contact AFM test. Besides the case of nanoindentation, the spring constant of the cantilever used in the mechanical characterization, in fact, can range from 0.15 to 0.5 N/m. The use of these relatively weak spring constants is due to the simultaneous consideration of both force resolutions and tip stability. The higher the force resolution, the greater are the displacements needed to recognize that force. In the linear region of the F/D curves, the slope is related to the elastic modulus of the system. When the cantilever
Fig. 24.14. a Linear and b nonlinear force–distance curves obtained on a suspended nanofiber by applying a normal force of 15 nN
326
S.-K. Kim · H. Lee
is much softer than the sample surface, the slope of the curve mostly reflects the spring constant of the cantilever. Moreover, the deflection of the cantilever can be converted to the equivalent force applied to the cantilever by the relation F = kΔZ (where k is the spring constant and ΔZ is the deflection of a cantilever). The vertical displacement of the piezo is the sum of the deflections of the suspended nanofiber and the cantilever. Thus, to obtain an accurate deflection of a suspended nanofiber, the piezo would have to be calibrated carefully. Figure 24.15 presents atomic force microscope tips with different shapes. When the curvature of the atomic force microscope tip is sharp, indentation may occur for soft materials. To prevent this, apex radii for tip curvatures larger than 30 nm have been used in AFM-based mechanical testing, with the exception of nanoindentation. However, it would be necessary to confirm nanofiber morphology after atomic force microscope testing in order to eliminate any possible errors due to indentation. From this point of view, the atomic force microscope tip with a flat end (Fig. 24.15e) would be preferable to measuring the mechanical properties of 1D polymer nanostructures. Moreover, for suspending the 1D nanostructures, well-defined structures such as an atomic force microscope calibration grating will provide more reliable testing.
Fig. 24.15. Scanning electron microscope images of AFM probes with different tip shapes: a octagonal cone tip, b square-pyramidal tip, c improved cone tip, d pyramidal tip, and e critical dimension reentrant tip (a flared tip that enables three-dimensional scanning of features)
24 AFM Measurement of the Mechanical Properties
327
Acknowledgements. The authors wish to acknowledge the excellent contributions of S.J. Kim and M.K. Shin of the Center for Bio-Artificial Muscle, Department of Biomedical Engineering, Hanyang University, Korea. The research was funded by the National Program for Tera-Level Nanodevices of the Ministry of Science and Technology as one of the 21 Century Frontier Programs.
References 1. Wang ZL (2003) Nanowires and nanobelts—materials, properties and devices, vol II. Kluwer, Norwell 2. Poncharal P, Wang ZL, Ugarte D, de Heer WA (1999) Science 283:1513–1516 3. Cuenot S, Demoustier-Champagne S, Nysten B (2000) Phys Rev Lett 85:1690–1693 4. Kis A, Mihailovic D, Remskar M, Mrzel A, Jesih A, Piwonski I, Kulik AJ, Benoit W, Forro L (2003) Adv Mater 15:733–736 5. Wu B, Heidelberg A, Boland JJ (2005) Nat Mater 4:525–529 6. Heidelberg A, Ngo LT, Wu B, Phillips MA, Sharma S, Kamins TI, Sader JE, Boland JJ (2006) Nano Lett 6:1101–1106 7. Jing GY, Duan HL, Sun XM, Zhang ZS, Xu J, Li YD, Wang JX, Yu DP (2006) Phys Rev B 73:235409 8. Ni H, Li X, Gao H (2006) Appl Phys Lett 88:043108 9. Ni H, Li X (2006) Nanotechnology 17:3591–3597 10. Wong EW, Sheehan PE, Lieber CM (1997) Science 277:1971–1975 11. Nilsson SG, Borrise X, Montelius L (2004) Appl Phys Lett 85:3555–3557 12. Li X, Ono T, Wang Y, Esashi M (2003) Appl Phys Lett 83:3081–3083 13. Schiotz J, Di Tolla FD, Jacobsen KW (1998) Nature 391:561–563 14. Li D, Xia Y (2004) Adv Mater 16:1151–1170 15. Binnig G, Quate CF, Gerber C (1986) Phys Rev Lett 56:930–933 16. Timoshenko SP, Gere JM (1972) Mechanics of materials. Van Nostrand, New York 17. Tombler TW, Zhou C, Alexseyev L, Kong J, Dai H, Liu L, Jayanthi CS, Tang M, Wu SY (2000) Nature 405:769–772 18. Shin MK, Kim SK, Lee H, Kim SI, Kim SJ (2006) Appl Phys Lett 88:193901 19. Doerner MF, Nix WD (1986) J Mater Res 1:601–609 20. Kracke B, Damaschke B (2000) Appl Phys Lett 77:361–363 21. Ko F, Gogotsi Y, Ali A, Naguib N, Ye H, Yang G, Li C, Willis P (2003) Adv Mater 15:1161–1165 22. O’Connor SD, Gamble RC, Eby RK, Baldeschwieler JD (1996) Rev Sci Instrum 67:393–396 23. Rabe U, Kester E, Arnold W (1999) Surf Interface Anal 27:386–391 24. Cuenot S, Fretigny C, Demoustier-Champagne S, Nysten B (2003) J Appl Phys 93:5650– 5655 25. Rabe U, Janser K, Arnold W (1996) Rev Sci Instrum 67:3281–3293 26. Landau L, Lifshitz E (1967) Theory of elasticity. Mir, Moscow 27. Tan EPS, Goh CN, Sow CH, Lim CT (2005) Appl Phys Lett 86:073115 28. Zussman E, Burman M, Yarin AL, Khalfin R, Cohen Y (2006) J Polym Sci Part B Polym Phys 44:1482–1489 29. Reneker DH, Chun I (1996) Nanotechnology 7:216–223 30. Doshi J, Reneker DH (1995) J Electrostat 35:151–160 31. Kameoka J, Orth R, Yang Y, Czaplewski D, Mathers R, Coates GW, Craighead HG (2003) Nanotechnology 14:1124–1129 32. Kameoka J, Verbridge SS, Liu H, Czaplewski DA, Craighead HG (2004) Nano Lett 4:2105– 2108
328
S.-K. Kim · H. Lee
33. Mack JJ, Viculis LM, Ali A, Luoh R, Yang G, Thomas Hahn H, Ko FK, Kaner RB (2005) Adv Mater 17:77–80 34. Zhou W, Wu Y, Wei F, Luo G, Qian W (2005) Polymer 46:12689-12695 35. Tan EPS, Lim CT (2006) Nanotechnology 17:2649–2654 36. Shin MK, Kim SK, Lee H, Kim SI, Spinks GM, Kim SJ (2006) Appl Phys Lett 89:231929 37. Chen Z, Foster MD, Zhou W, Fong H, Reneker DH, Resendes R, Manners I (2001) Macromolecules 34:6156–6158 38. Stephens JS, Chase DB, Rabolt JF (2004) Macromolecules 37:877–881 39. Jaeger R, Schönherr H, Vansco GJ (1996) Macromolecules 29:7634–7636 40. Cuenot S, Fretigny C, Demoustier-Champagne S, Nysten B (2004) Phys Rev B 69:165410 41. Arinstein A, Burman M, Gendelman O, Zussman E (2007) Nat Nanotechnol 2:59–62 42. Liu H, Edel JB, Bellan LM, Craighead HG (2006) Small 2:495–499 43. Roy R, Craig JR (2000) Mechanics of materials, 2nd edn. Wiley, New York 44. Plaschke M, Schafer T, Bundschuh T, Manh TN, Knopp R, Geckeis H, Kim JI (2001) Anal Chem 73:4338–4347
25 Evaluating Tribological Properties of Materials for Total Joint Replacements Using Scanning Probe Microscopy Sriram Sundararajan · Kanaga Karuppiah Kanaga Subramanian
Abstract. Scanning probe methods have been utilized to reveal useful information regarding the tribological phenomena of total joint replacement materials. In particular, studies on the effect of synovial fluid constituents on the tribological response of polymeric materials and the effect of residual stresses on the fretting response of metallic materials are described. The studies show that an adsorbed layer of proteins causes an increase in friction behavior of hydrophobic ultrahigh molecular weight polyethylene (UHMWPE) owing to denaturation upon adsorption. Changing the crystallinity of UHMWPE can affect the friction behavior and have ramifications for mechanisms of protein adsorption. Studies on fretting behavior of metallic (CoCrMo) implant materials showed that significant plastic deformation of the metal may not be a prerequisite for oxide fracture. In addition, results suggest that nearly neutral boundary lubricants may provide significant protection against the onset of fretting. Finally, the chapter concludes with an outlook on how scanning probe methods can be used to provide greater insight into the design of biomaterials. Key words: Total joint replacements, Ultrahigh molecular weight polyethylene, Polyethylene, Atomic force microscopy, Friction, Wear, Proteins
25.1 Introduction This chapter describes examples of scanning probe applications in the field of bioengineering and biomaterials, specifically the area of understanding and improving the tribological behavior and durability of joint replacement materials. The chapter begins with an introduction to total joint replacements (TJRs), including the problems being studied and the far-reaching impact of finding potential solutions. Examples of scanning probe methods used and the results obtained are then described. The chapter concludes with a summary and future outlook. 25.1.1 Total Joint Replacements TJR is a procedure in which damaged joints are removed and replaced with an artificial device (prosthesis). Each year, more than a million total replacement surgeries are performed worldwide, with a predominant number being hip and knee replacements. A natural hip joint, shown in Fig. 25.1, is a ball-and-socket joint that
330
S. Sundararajan · K.K.K. Subramanian
Fig. 25.1. A natural hip joint and related components
allows movement in three planes. The head of the femur mates with the socket of the pelvic bone to form the articulating joint. These bearing surfaces are lined with a 1–3-mm-thick layer of human cartilage and operate in the presence of a natural joint fluid known as the synovial fluid. A typical total replacement hip joint is shown in Fig. 25.2a, in which the bearing surfaces are formed by the acetabular liner (usually a polymer) and the femoral head (usually a ceramic or metallic alloy). Most replacement joints are modular in nature as shown in Fig. 25.2a. The acetabular shell and the hip implant are embedded into the bone using cement or using a press-fit. 25.1.2 Social and Economic Significance Researchers have estimated that the demand for TJRs is expected to increase dramatically in the next 25 years [1–3]. A study conducted by the American Academy of Orthopedic Surgeons [2] projects the number of yearly procedures for primary (first-time) total knee replacement to jump by 673% to 3.48 million in 2030. The number of primary total hip replacements performed yearly will increase by 174% to 572,000. The research team based its projections on historical procedure rates from 1990 to 2003, combined with population projections from the US Census Bureau. Reasons for the expected increase include an aging population with arthritis requiring joint replacement; the increasing prevalence of obesity, which puts undue stress on the knee and hip joints; and the trend toward baby boomers remaining physically active later in life, which also places demands on the joints.
25.2 Problems Associated with Total Joint Replacements Though joint replacement procedures are mostly successful, the artificial joints can become loose and unstable during use, which is commonly referred to as aseptic loosening (Fig. 25.2b). The TJRs thus often require revision surgeries to repair or replace the joint. Herbert et al. [4] have reported that a revision surgery costs three to four times more hospital resources than a primary implant. The number of revision surgeries likely will double by 2015 for total knee replacement and by 2026 for total
25 Tribological Properties of Materials for Total Joint Replacement
331
Fig. 25.2. a A typical modular total replacement hip joint showing the unassembled components and the assembled joint. b A new implant and the same implant exhibiting aseptic loosening
hip replacement [2, 5]. Costs of revision joint replacement are particularly great and must be reduced. The best way is to make the original joint implant last longer. For that to happen, there is a need for more research to improve surgical materials and technologies, which in turn require a better understanding of the processes involved in their in-use failure. Continuing education on best practices for surgeons performing the procedures and new medical approaches to prevent loosening of the implant is also important.
332
S. Sundararajan · K.K.K. Subramanian
25.2.1 Tribology Tribological properties of the articulating surfaces in TJRs have been identified as a critical factor affecting their durability and reliability [6–9]. For example, in a total hip replacement there are two primary bearing interfaces that determine the useful life of the implant: the interface between the femoral head (usually a hard material) and the acetabular cup liner (usually a softer material); and the interface between the femoral stem and the femoral head (Fig. 25.2). Sliding wear occurs at the former interface, whereas fretting wear occurs at the latter interface. The modular junctions of hip replacements are especially prone to fretting corrosion [10–13]. The occurrence of wear itself is not catastrophic, but does create wear debris, which causes inflammation of the joint and osteolysis (bone loss). This can cause implantloosening. In order to understand the debris generation it is necessary to understand the mechanism that governs sliding, damage nucleation and subsequent wear at the material interface. Owing to the high contact area between the cup and the liner in hip joints, the friction and wear mechanisms are generally adhesive in nature at the onset of operation, where maximum principal stresses are much less than the yield strength of the ultrahigh molecular weight polyethylene (UHMWPE). However, at later stages, ductile deformation micromechanisms and particulate debris compound the wear. These modes are also predominant in knee joints. Hence, it is of interest to minimize adhesive contributions to friction and wear in TJRs. 25.2.2 Materials There has been an evolution of material pairs for the head–liner interface, beginning with Charnley’s work [14] with stainless steel heads and Teflon liners. Today, the combination of an UHMWPE liner and metallic or ceramic femoral heads is extensively used. However, the average lifetime of artificial hip joints incorporating UHMWPE is only 15–20 years. UHMWPE is a polyethylene with an average molecular mass of greater than approximately 3.1 × 106 g/mol. A considerable amount of research has been devoted to improving the wear life of the UHMWPE-based prosthesis. Medical-grade UHMWPE stock material may undergo a variety of processing techniques during implant manufacturing [15], including molding, extrusion and milling or turning. These manufacturing processes can affect the surface morphology and mechanical properties of the polymer [16–18], which in turn can affect its friction and wear performance. The current state of the art in the orthopedics community is to fully cross-link UHMWPE using radiation and heat treatment in order to minimize sliding wear [19–22]. Cross-linking of UHMWPE is believed to improve the wear resistance by eliminating the modes of plasticity in the semicrystalline polymer. The femoral head and stem form part of the modular total hip replacement (Fig. 25.2a). The stem and head have interlocking tapers. Commonly used materials for these components include titanium alloys (Ti-6Al-4V, ASTM F-136) or cobalt alloys (CoCrMo, ASTM F75 cast and ASTM 562, ASTM F799, ASTM F1537 wrought). In addition, the femoral head may also be composed of ceramic alumina
25 Tribological Properties of Materials for Total Joint Replacement
333
or ceramic zirconia [12,13,23–25]. The use of coatings and other surface treatments on the head have been explored to improve the wear behavior of UHMWPE–metal prostheses [26–32]. 25.2.3 Lubrication in Joints—the Synovial Fluid In a natural joint, synovial fluid lubricates the cartilage. Studies suggest that boundary lubrication is dominant in TJRs [33] and it is generally believed that the synovial fluid acts as a boundary lubricant to the joint interface and helps to minimize friction and wear. Fluid film lubrication, including a squeeze film effect, is believed to play a role in knee joint lubrication as well. It is recognized that the synovial fluid contains various kinds of serum proteins, hyaluronic acid and also various kinds of lipids [34, 35] as shown in Table 25.1. It has been found that albumin constitutes almost 60% of the total protein concentration in serum and synovia [35, 36]. Hyaluronic acid is a type of polysaccharide which is responsible for the viscous nature of the synovial fluid [36] and is believed to help in fluid film lubrication. Phospholipids are believed to play an active role in boundary lubrication of the joints. Synovial fluid was obviously not designed to lubricate engineered TJR materials such as polyethylene and metals/ceramics. The interaction of the TJR materials with the synovial fluid constituents is thus an important factor affecting the tribological performance of the materials [18, 35, 37–44]. The adsorption mechanism of proteins onto the polymeric surface can affect its subsequent friction behavior [45]. Tailoring the hydrophobicity of UHMWPE can affect the protein adsorption and hence the friction behavior of the polymer surface [45, 46]. Clearly, it is of importance to understand how the tribological characteristics of UHMWPE change as soon as the TJR comes in contact with the synovial fluid. As a result of numerous studies to identify an alternative physiological fluid for use in experiments in vitro, the American Society for Testing and Materials has established that the use of bovine serum is acceptable for TJR biomaterials testing. Table 25.1. The constituents of synovial fluid and related information on their function in joint lubrication Constituent
Typical Function concentration (g/l)
Hyaluronic acid 2–4
Viscoelastic properties
Lubricin
–
Mucinous glycoprotein
Phospholipids
2.4
Highly surface active [33]
Proteins
18
Protein adsorbs on foreign surfaces [83]
Reported behavior
Fails to lubricate under load [78, 79] Not active ingredient; carrier for surface-active phospholipids [33, 80] Excellent boundary lubricant [81, 82] May enhance [79] or impede lubrication [44, 45]
334
S. Sundararajan · K.K.K. Subramanian
25.3 Conventional Tribological Testing of Material Pairs for Total Joint Replacements 25.3.1 Wear Tests Joint simulators are the standard for wear testing of TJR materials since they are designed to mimic the orientation and magnitude of joint loads. Joint simulators typically produce wear behavior and particle morphology comparable to those found in clinical cases; however, simulator tests are often difficult to analyze quantitatively and the cost, time and variability of lubrication are some drawbacks to the efficiency of this testing method. Simpler testing geometries such as pin-on-disk and pin-onflat tests are also commonly employed to isolate and study specific mechanisms and articulating conditions. It has been shown that multidirectional paths yield wear behavior and particle morphologies that are more comparable to those found in clinical cases than unidirectional or bidirectional paths [47–49]. Nonetheless, simple geometry tests have proven to be valuable in screening potential material pairs for TJRs and to understand specific tribological mechanisms. 25.3.2 Friction Tests Compared with a wear test, a friction measurement is a more rapid tribological assay. Friction tests can alleviate issues with lubricant volume, time and cost concerns with wear testing. Although low-friction behavior does not necessarily imply low wear, friction measurements can provide useful information regarding the tribological behavior. Frictional work is manifested in energy dissipation processes that can damage the surface. Mechanisms associated with high friction (adhesion, abrasion) can have implications for wear as well.
25.4 Scanning Probe Microscopy as a Tool to Study Tribology of Total Joint Replacements Scanning probe microscopy (SPM) or more specifically atomic force microscopy (AFM) techniques have been extensively used to simulate a single asperity contact and to study the nanotribological properties of material pairs [50]. Techniques to perform adhesion, friction and wear studies using SPM are quite well established. Calibration methods to perform quantitative friction force microscopy are available [51,52]. Techniques to reliably calibrate the normal and lateral spring constants of the cantilevers allow researchers to perform more accurate force measurements [53, 54]. The use of probe shape characterizers and reconstruction algorithms allows precise evaluation and monitoring of the tip shape during experiments [44]. Several contact mechanics theories have been applied to adhesion and friction measurements as well [44, 55, 56]. Scanning probe techniques thus allow accurate determination of the probe and contact dimensions during tribological measurements,
25 Tribological Properties of Materials for Total Joint Replacement
335
which provide means to isolate tribological contributions from specific sources of interest such as adsorbed films. Owing to the small interaction volume, SPM techniques also alleviate concerns with lubricant volume requirements. Despite these advantages, few researchers have utilized scanning probe techniques to study the tribological behavior of TJR materials [18, 40–44, 57–59]. These studies are described in the following sections. 25.4.1 Nanotribology of Ultrahigh Molecular Weight Polyethylene Ho et al. [59] studied the friction behavior of UHMWPE using an atomic force microscope. They evaluated a friction coefficient of 0.22 between a Si3 N4 probe and direct compression molded UHMWPE at normal loads of up to 11 nN. Transients in frictional response of the UHMWPE as a function of normal load range (17– 340 nN) during sliding were observed. The second transient in the frictional force was observed at a normal load of 288 ± 13 nN, corresponding to a contact pressure of 518 ± 13 MPa. Following the second transition in frictional response, plowing of the UHMWPE surface by the probe was observed. They therefore postulated that the deep scratches noticed in retrieved TJR prostheses could be due to the abrasion from protruding hard carbide asperities that exist within the cast CoCr alloy of TJR prostheses. Kanaga Karuppiah et al. [40,44,60] have studied the effect of material processing, and the effect of protein adsorption on the nanotribological behavior of medical-grade UHMWPE. Experiments in contact mode were carried out with a DimensionTM 3100 atomic force microscope (Nanoscope IV, Veeco Instruments, Santa Barbara, CA, USA) in controlled low-humidity (10 ± 4% relative humidity) conditions to minimize effects of adsorbed water vapor. Standard V-shaped silicon nitride probes with a quoted normal spring constant of 0.58 N/m and a tip radius of 10–40 nm were used. The normal spring constants of the cantilevers used were calibrated using the technique described by Torii et al. [53] and were found to have actual values of 0.20– 0.25 N/m. The pull-off (adhesive) force between the Si3 N4 tip and the UHMWPE was measured before and after each test from force–displacement curves. The friction force was calibrated using Ruan and Bhushan’s method [51]. The radius of the tip was characterized before and after the experiments using a commercially available tip characterizer sample TGT01 (Mikromasch). The images were analyzed using commercial software (Image Metrology) to calculate the tip radius for accurate contact area calculations. 25.4.1.1 Effect of Material Processing The effect of UHMWPE processing on its friction response was investigated. Specifically, milled and surface-melted samples of comparable roughness were studied. Figure 25.3 shows representative AFM topography maps of the UHMWPE samples. The melt sample was prepared by heating the polymer surface to above its melting temperature and cooling it at room temperature. The melt sample displayed a lamellar type of structure that is indicative of recrystallization and has been reported in
336
S. Sundararajan · K.K.K. Subramanian
Fig. 25.3. Topography maps of melt and milled ultrahigh molecular weight polyethylene (UHMWPE) samples obtained using atomic force microscopy (AFM) at a 5 µm × 5 µm and b 1 µm × 1 µm scan sizes [44]. The melt sample exhibits lamellar structures that are indicative of recrystallization
previous studies [15, 61]. Figure 25.4, panel a shows the friction response of the materials. The friction force increased with normal load for both samples and exhibited a slight nonlinear behavior. The melt sample showed higher friction than the milled sample. No discernable wear was observed up to loads of 80 nN. Permanent damage (groove depths on the order of 2–10 nm) was observed at loads beyond 80 nN. In the absence of wear, the predominant mechanism during the friction force microscopy experiments can be assumed to be adhesive. The adhesive friction is then given by [62] Fa = τ0 Ar + βL ,
(25.1)
where τ0 is the interfacial shear strength, Ar is the real area of contact, β is a constant that describes the normal load dependency of the shear strength and L is the normal load. The real area of contact can be evaluated using an appropriate contact mechanics theory, while τ0 and β are typically evaluated using curve fits. The generalized transition analysis by Carpick et al. [63] was used to determine the appropriate contact model, which in this case pointed to the Derjaguin–Muller–Toporov (DMT)
25 Tribological Properties of Materials for Total Joint Replacement
337
Fig. 25.4. a Friction response of the processed UHMWPE samples as a function of normal load obtained using friction force microscopy. b Example of a Derjaguin–Muller–Toporov (DMT) model fit to the friction data according to (25.1)
model [64]. According to the DMT model, the real area of contact is given by ∗ 2 3 3R ∗ Ar = π L + 2πR W12 , (25.2) 4E ∗ where R∗ is the composite radius and E ∗ is the composite modulus given by −1 1 − υ12 1 − υ22 . (25.3) E∗ = + E1 E2
338
S. Sundararajan · K.K.K. Subramanian
The contact-depth-independent elastic modulus values obtained from nanoindentation experiments and an assumed value of Poisson’s ratio of 0.3 for silicon nitride and 0.45 for UHMWPE, respectively [59], were utilized for these calculations. The work of adhesion (W12 ) was measured from pull-off forces (FPO ) for various samples from the following equation derived from the DMT analysis: FPO = −2πR∗ W12 .
(25.4)
Since the AFM experiments were carried out in dry conditions, it is reasonable to assume that the adhesion component dominates over capillary contributions and that W12 can be estimated quite reliably from (25.4). It is noted that the DMT analysis for this interface assumes that no wear occurs during contact and that the probe has a parabolic profile. Figure 25.5 shows that the probe did indeed exhibit a parabolic profile which was maintained during experiments. The parameters τ0 and β can then be obtained by fitting (25.1) to the friction force data as shown in Fig. 25.4, panel b. The figure shows that the DMT-based equation gave a good fit, which was characteristic of the friction data for all the samples. The values for τ0 and β obtained are also listed in Table 25.2. The measured interfacial shear strengths of the UHMWPE samples prior to protein adsorption were comparable to those measured using macroscale torsion tests (1–7 MPa) [65]. The values of β obtained were comparable to values reported for high-density and low-density polyethylene (0–0.05) [66]. The β values obtained imply that the melt samples exhibit a slightly higher normal load dependency compared to the milled samples, which displayed values closer to zero (Table 25.2). The melting and reforming process can affect the degree of crystallinity of the sample [61]. The melt sample exhibited lower water contact angles and hence higher surface energy, which is suggestive of a higher degree of crystallinity [67]. The melt sample also exhibited almost twice the elastic modulus and hardness of the milled sample (Table 25.2), which is also indicative of increased crystallinity [66]. The data Table 25.2. Summary of surface and interfacial characteristics for the two ultrahigh molecular weight polyethylene (UHMWPE) samples described in the study [44]. Measured values indicated are average ± standard deviation Elastic modulusb (GPa)
Hardnessb (GPa)
Contact angle (◦ )
τ0c βc (MPa)
UHMWPE sample
RMS roughnessa (nm)
Melt sample Melt sample with proteins Milled sample Milled sample with proteins
3.55 ± 0.90 2.10 ± 0.08 0.151 ± 0.003 66.1 ± 3.57 5.96 3.28 ± 0.28 – 58.3 ± 5.24 12.04
0.03 0.01
6.98 ± 0.85 0.90 ± 0.07 0.073 ± 0.005 96.3 ± 1.43 0.919 7.78 ± 0.75 – 64.3 ± 2.53 3.50
0.005 0
RMS root mean square a Measured using atomic force microscopy over a 1 µm × 1 µm scan area b Measured using a nanoindenter at 15-µN peak load c Calculated according to Derjaguin–Muller–Toporov contact mechanics from friction force data
25 Tribological Properties of Materials for Total Joint Replacement
339
Fig. 25.5. AFM probe profiles before and after experiments obtained using a probe characterization sample and subsequent reconstruction. The profiles are parabolic and there appears to be negligible distortion as a result of experiments [44]
340
S. Sundararajan · K.K.K. Subramanian
therefore suggest that controlling the crystallinity of UHMWPE can affect its surface tribological and mechanical behavior. Surface topography (roughness) can affect the lateral (friction) force response measured using an atomic force microscope [68]. This is especially true in the case of sharp changes in topography and an order of magnitude difference in roughness levels. The difference in roughness levels between the milled and the melt sample is quite small; hence, it is reasonable to assume the observed differences in friction force were caused by material-based effects (at the surface) rather than topography or roughness. Residual stresses (such as the compressive stresses imparted by milling) may also have an effect on the observed interfacial shear strength behavior. 25.4.1.2 Effect of Protein Adsorption Although studies suggest that proteins affect the friction and wear of UHMWPE more significantly than the other constituents [35], the role of proteins in lubrication is not fully understood. Protein adsorption has previously been reported to increase the wear rates of UHMWPE [37,69] as well as to decrease them [70,71]. By studying the friction response of a well-characterized nanoscale contact as a function of protein adsorption, one can better understand the role of proteins in boundary lubrication. Friction response of surface-melted and milled medical-grade UHMWPE as a function of protein adsorption has been investigated using friction force microscopy [44]. Friction response was monitored on indexed areas of the samples before exposure to proteins. The areas were then exposed to protein solution and assessed for protein adsorption before the friction response was measured again. In this study, bovine serum albumin (BSA) tagged with fluorescein isothiocyanate dye was dissolved in phosphate-buffered saline (PBS) solution of pH 7.4 (1×) to make up a concentration of 7.2 mg/ml (BSA) representing 10 vol % dilution. This dilution is comparable to the protein concentration in the human synovial fluid [46]. A 20-µl aliquot of the protein solution was placed on a small indexed area of each UHMWPE sample for 5 min, following which the samples were rinsed with deionized water and dried with nitrogen gas. Protein adsorption onto the samples was verified qualitatively using fluorescence microscopy. Figure 25.6 shows the friction response of the UHMWPE samples before and after exposure to BSA. Both samples exhibited an increase in friction response upon exposure to BSA. The magnitude of this increase was considerably larger in the case of the milled sample. The corresponding interfacial shear strengths of the samples obtained using DMT contact analysis (Fig. 25.7, panel a) showed that exposure to proteins resulted in a 300% increase in the shear strength in the case of the milled sample, whereas the increase in the shear strength was about 100% for the melt sample. In both the milled and the melt samples, the β values (shown in Table 25.2) decreased upon exposure to proteins. Water contact angle data for the two samples are shown in Fig. 25.7, panel b. The milled sample was initially significantly more hydrophobic than the melt sample. In both cases, adsorption of BSA rendered the surfaces more hydrophilic. The reduction in contact angle was very significant in the case of the milled sample compared with that in the case of the melt sample. Table 25.2 lists all the parameters measured for the samples in this study.
25 Tribological Properties of Materials for Total Joint Replacement
341
Fig. 25.6. Friction force microscopy for a melt and b milled UHMWPE samples before and after exposure to protein. The friction response increases upon exposure to proteins in both cases [44]
25.4.1.3 Discussion The data show that protein adsorption tends to increase the friction behavior and interfacial shear strength of UHMWPE. Widmer et al. [45] showed that hydrophobic polyethylene surfaces exhibit higher friction than hydrophilic ones. They used optical waveguide lightmode spectroscopy to show that proteins adsorbed onto hydrophobic surfaces occupy more surface area than those that adsorb onto hydrophilic surfaces. They therefore proposed that proteins denature during adsorption onto
342
S. Sundararajan · K.K.K. Subramanian
Fig. 25.7. Effect of protein adsorption on a interfacial shear strength and b water contact angles of the UHMWPE samples (adapted from [44])
a hydrophobic surface. The denaturing of proteins upon adsorption on hydrophobic surfaces has been reported in relation to other applications [72, 73] as well. Denatured proteins occupy large surface area, with hydrophobic parts undergoing adsorption, exposing hydrophilic regions. The contact angle data support the idea that the proteins denature upon adsorption onto the hydrophobic UHMWPE surfaces or that denatured proteins adsorb preferentially onto the hydrophobic surfaces (Fig. 25.8). The data therefore indicate that a denatured protein layer forms a relatively higher shear strength layer than the virgin material, which increases the adhesive friction response. An increase in the interfacial shear strength may have implications for adhesive wear mechanisms. However, higher friction in most
25 Tribological Properties of Materials for Total Joint Replacement
343
Fig. 25.8. The denaturation of proteins as they adsorb onto hydrophobic UHMWPE. The adsorbed layer exposes hydrophilic regions of the protein molecules. This denatured layer appears to increase the adhesive friction force of the surface. (Adapted from [45])
cases does not necessarily imply higher wear and one should evaluate wear behavior independently from friction experiments. The melt sample showed higher protein adsorption but lower increase in friction upon exposure to protein than the milled sample. This suggests that the conformation of the adsorbed proteins affects the friction performance significantly. The higher surface energy of the melt sample reduces the extent of protein denaturation (or adsorption of denatured proteins) compared with that for the more hydrophobic milled sample and leads to the relatively lower increase in shear strength upon protein adsorption. These studies suggest that changing the crystallinity of UHMWPE can affect its surface energy and protein adsorption and hence the friction behavior in the presence of proteins. 25.4.2 Fretting Wear of Cobalt–Chromium Alloy Mitchell and Shrotriya [58, 74] have investigated the nanoscale wear response of cobalt–chromium (CoCrMo) alloy (ASTM F-75) as a function of contact load and surface residual stress in order to identify the mechanism governing the onset of surface damage in modular implants. CoCrMo samples were cut from a canine femoral stem prosthesis by electric discharge machining, which was chosen to minimize processing-induced residual stress on the machined surfaces. The machined samples were mechanically polished to yield a root mean square roughness of less than 2 nm over a 10-µm range. A unique loading configuration was utilized to apply a range of known in-plane stress states to the specimen surface in order to simulate different residual stress levels. This was achieved through a four-point bending frame as shown in Fig. 25.9. A capacitance gage monitors the deflection of the free end of the c-clamp. Opening of the c-clamp results in application of forces onto the four rollers, leading to four-point bending of the sample. Calibration of the deflection–force relationship allows for the in-plane stress states at any location along the specimen thickness to be determined from the geometry of the sample, the end deflection of the c-clamp, and the distance of the location from the sample mid-plane.
344
S. Sundararajan · K.K.K. Subramanian
Fig. 25.9. The four-point bending setup used for in situ AFM wear testing of samples subjected to an external tensile or compressive stress state. The main components are a c-clamp, a base, a capacitance probe, rollers, and a rectangular sample. Top right: Five contact loading conditions and prestress locations tested. Bottom right: The loading action created by the atomic force microscope tip on a tensile residual stress state. (From [74])
Wear experiments were conducted using a DimensionTM 3100 atomic force microscope (Nanoscope IV, Veeco Instruments, Santa Barbara, CA, USA) with a standard V-shaped silicon nitride probe at different normal loads and at different locations of a loaded specimen which correspond to different residual stresses. The wear tests were conducted at ambient conditions (25 ◦ C, approximately 50% relative humidity) as well as in a simulated corrosive physiological environment of PBS solution, augmented with HCl to a pH of 5.01. This lower pH was chosen owing to the fact that the modular junction of a total hip replacement produces differential aeration, creating local drops in pH within the crevice. Additionally, a pH of around 5 is the thermodynamic stability point of chrome oxide (Cr2 O3 ), the dominant oxide in the passive layer of CoCrMo. The AFM tests in solution were conducted in the same manner as for the ambient conditions except that a fluid cell was used and 0.5–1 ml of solution was added to cover the exposed surface of the sample. Each wear test was performed over a 2 µm × 500 nm area. Seventy cycles at 5 Hz were performed and regions were imaged before and after wear. The normal spring constants and probe radii were determined as described earlier, which allowed for a more precise calculation of Hertzian contact pressures. Normal loads used for calculations of the pressure included the observed adhesive force. In order to minimize the effects of initial surface roughness on the calculation of wear depths, the initial and final images obtained for each test were registered and subtracted to determine the amount of material removed. For each experiment, average trench depth was used to estimate the wear volume.
25 Tribological Properties of Materials for Total Joint Replacement
345
25.4.2.1 Effect of Residual Stresses Figure 25.10 shows the wear volumes as a function of residual stress state and applied normal load (contact pressure) under ambient conditions. The compressive residual stresses have a more pronounced influence on the wear depth. Whereas the tensile residual stress suppressed the wear, the compressive residual stress exacerbates the wear process. For all contact loads, the wear rate increased linearly as the stress state changed from the maximum tensile to maximum compressive stress state. In addition, the wear increases as the load increases for all residual stress states. Even for lower contact loads (2.3–2.6 GPa) the onset of deformation only occurred on the compressively loaded side of the sample. No wear could be determined from the lowest loading condition (1–2 nN, approximately 0.8 GPa, data not shown). It is noted that the same wear depth was achieved with around 2 GPa of contact stress on the compressive side that required approximately 4 GPa of contact stress on the tensile side. The aqueous environment behaves in a slightly different manner.
Fig. 25.10. Average wear volume as a function of residual stress (normalized over yield strength) for different contact loads under ambient conditions [58]
25.4.2.2 Effect of Aqueous Solution in Conjunction with Residual Stresses In contrast to ambient conditions, the rate of wear was increased under both tensile and compressive stresses (Fig. 25.11) in the aqueous solution. The compressive stress still creates larger wear depths than the tensile residual stress; however, in
346
S. Sundararajan · K.K.K. Subramanian
Fig. 25.11. Average wear volume as a function of residual stress (normalized over yield strength) for different contact loads under simulated corrosive physiological solution [58]
all instances the unstressed state produced less wear than the compressive or tensile stressed states. For all contact loads, the wear depth tends to follow a skewed quadratic dependence on the residual stress state (skewed toward the compressive side). The contact stress required to accelerate wear is not as apparent in the aqueous environment as in the ambient environment. However, in contrast to ambient conditions, measurable wear occurred at 0.8–1 GPa of contact stress. 25.4.2.3 Discussion Since the residual stress is 3 orders of magnitude less than the contact loads, the residual stress should not affect the wear depth so drastically. In addition, it is counterintuitive to suggest that compressive stress is deleterious to the wear properties of the metal since it is well known that surface treatments such as shot peening and nitriding often yield enhanced wear performance in metals. The wear depth values however agree very well with the oxide thickness reported in the literature, which is 1.8 nm in ambient conditions [75]. It may then be thought that the film is fractured and needs to repassivate. The difference in the wear volumes may be explained by a change in the repassivation kinetics caused by the residual stress state. This is highly probable since the timescale for repassivation is several orders of magnitude faster than the time required for an AFM scan (milliseconds/seconds compared with minutes) [76]. Since only the scratched region would be expected to undergo fracture/repassivation, the unscratched surface is already in a passive state and the wear depth may be thought of as a reaction rate snapshot. The difference between the residual stressed wear rates and that of the unstressed average is an indication of the effect residual stress plays in the mobility of the
25 Tribological Properties of Materials for Total Joint Replacement
347
reaction rate. Yu and Suo [77] argue that residual stress will affect a surface reaction in either a linear (mobility dependent) or a quadratic (driving force dependent) manner. It was demonstrated that troughs on a wavy surface undergoing chemical reaction and subjected to in-plane stress will have different reaction rates from at the crests. Specifically, a compressively stressed solid gaining mass through a reaction, like repassivation, will experience slower growth at the troughs than at the crests, with the opposite being true for tensile stresses. For the present case this would give credence to the marked linear (mobility) dependence on residual stress. The loading data demonstrate that approximately 2 GPa of average contact stress was required to produce a trench under no residual stress in ambient conditions. Under an aqueous environment, a trench of approximately the same depth is generated with 0.8–1 GPa of contact stress. This suggests that the passive layer may begin to fail by abrasive brittle fracture as would be expected of a ceramic, given that different wear levels were detected for each contact load.
25.5 Summary and Future Outlook Scanning probe methods can thus reveal useful information regarding the tribological phenomena of TJR materials. In particular, the effect of synovial fluid constituents on the tribological response of polymeric materials and the effect of residual stresses on the fretting response of metallic materials have been described. The studies showed that an adsorbed layer of proteins causes an increase in friction behavior of hydrophobic UHMWPE owing to denaturation upon adsorption. Changing the crystallinity of UHMWPE can affect the friction behavior and have ramifications for mechanisms of protein adsorption. Studies on fretting behavior of metallic (CoCrMo) implant materials showed that significant plastic deformation of the metal may not be a prerequisite for oxide fracture. The increased wear depth in ambient compared with aqueous environments suggests that fretting of asperities or wear particles (under dry conditions) are a significant damage mechanism that may precede crevice corrosion. In addition, nearly neutral boundary lubricants may provide significant protection against the onset of fretting. Systematic studies of friction and wear in the presence of the other synovial fluid constituents (i.e., in fluid) can reveal a better picture of how engineered materials will behave in a human joint. These results may have ramifications for design and manufacturing of the implant materials and their surface engineering. Scanning probe methods can also be applied to natural joint materials as demonstrated by Park et al. [57], who studied both macroscopic and microscopic (AFM) friction measurements on bovine articular cartilages. Scanning probe methods can thus be used to explore the role of boundary lubricants in cartilages and provide greater insight into design of biomaterials. Acknowledgements. The authors wish to thank Michael Conzemius (University of Minnesota) for his useful discussions and insight into the area of TJR. He provided the medical perspective to our engineering approach to the research problems.
348
S. Sundararajan · K.K.K. Subramanian
References 1. Dixon T, Shaw M, Ebrahim S, Dieppe P (2004) Ann Rheum Dise 63:825 2. Kurtz S, Mowat F, Ong K, Chan N, Lau E, Halpern M (2005) J Bone Joint Surg Am 87A:1487 3. Mehrotra C, Remington Patrick L, Naimi Timothy S, Washington W, Miller R (2005) Public Health Rep 120:278 4. Hebert CK, Williams RE, Levy RS, Barrack RL (1996) Clin Orthop Relat Res 140 5. Kurtz (2006) http://www6.aaos.org/news/Pemr/press_release.cfm?PRNumber=442 6. Revell PA, Alsaffar N, Kobayashi A (1997) Proc Inst Mech Eng H 211:187 7. McGee MA, Howie DW, Costi K, Haynes DR, Wildenauer CI, Pearcy MJ, McLean JD (2000) Wear 241:158 8. Ingham E, Fisher J (2000) Proc Inst Mech Eng H 214:21 9. El-Warrak AO, Olmstead ML, von Rechenberg B, Auer JA (2001) Vet Comp Orthop Traumatol 14:115 10. Schaaff P (2004) J Appl Biomater Biomech 2:121 11. Hanawa T (2002) Sci Technol Adv Mater 3:289 12. Hallab NJ, Messina C, Skipor A, Jacobs JJ (2004) J Orthop Res 22:250 13. Hallab NJ, Jacobs JJ (2003) Corrosion Reviews 21:183 14. Charnley J (1960) J Bone Joint Surg Br 42:28 15. Kurtz SM, Muratoglu OK, Evans M, Edidin AA (1999) Biomaterials 20:1659 16. Bellare A, Cohen RE (1996) Biomaterials 17:2325 17. Ho SP, Riester L, Drews M, Boland T, LaBerge M (2003) Proc Inst Mech Eng H 217:357 18. Check J, Karuppiah KSK, Sundararajan S (2005) J Biomed Mater Res A 74A:687 19. Shen FW, McKellop HA, Salovey R (1996) J Polym Sci Part B Polym Phys 34:1063 20. Jasty M, Bragdon CR, O’Connor DO, Muratoglu OK, Premnath V, Merrill EW, Harris WH (1997) In: Proceedings of the 43rd annual meeting the of Orthopaedic Research Society, San Francisco, p 785 21. Fisher J, McEwen HMJ, Tipper JL, Galvin AL, Ingram J, Kamali A, Stone MH, Ingham E (2004) Clin Orthop Relat Res 114 22. Shen FW, McKellop H (2005) Clin Orthop Relat Res 80 23. Willmann G (1998) Orthopedics 21:173 24. Katti KS (2004) Colloids Surf B 39:133 25. Cales B (2000) Clin Orthop Rel Res 94 26. Rostami H, Iskandarani B, Kamel I (1992) Polym Composites 13:207 27. Alonso F, Ugarte JJ, Sansom D, Viviente JL, Onate JI (1996) Surf Coat Technol 83:301 28. Lappalainen R, Heinonen H, Anttila A, Santavirta S (1998) Diamond Relat Mater 7:482 29. Raimondi MT, Pietrabissa R (2000) Biomaterials 21:907 30. Onate JI, Comin M, Braceras I, Garcia A, Viviente JL, Brizuela M, Garagorri N, Peris JL, Alava JI (2001) Surf Coat Technol 142:1056 31. Platon F, Fournier P, Rouxel S (2001) Wear 250:227 32. Pompe W, Worch H, Epple M, Friess W, Gelinsky M, Greil P, Hempel U, Scharnweber D, Schulte K (2003) Mater Sci Eng A 362:40 33. Hills BA (1995) Ann Biomed Eng 23:112 34. Brown SS, Clarke IC (2006) Tribol Transa 49:72 35. Sawae Y, Murakami T, Chen J (1998) Wear 216:213 36. Swann DA (1987) Bull Soc Belg Ophtalmol 223(Pt 1):59 37. Chandrasekaran M, Loh NL (2001) Wear 250:237 38. Mazzucco D, Spector M (2004) Clin Orthop Relat Res 17 39. Nakanishi Y, Murakami T, Higaki H, Miyagawa H (1999) JSME Int J Ser C 42:481
25 Tribological Properties of Materials for Total Joint Replacement
349
40. Kanaga Karuppiah KS, Sundararajan (2006) In: Proceedings of IJTC2006 STLE/ASME international joint tribology conference, San Antonio, October 2006 41. Kanaga Karuppiah KS, Sundararajan Z-H, Xu Li X (2006) In: Proceedings of IMECE2006 2006 ASME international mechanical engineering congress and exposition, Chicago, November 2006 42. Kanaga Karuppiah KS, Sundararajan (2005) In: Proceedings of IMECE2005 2005 ASME international mechanical engineering congress and exposition, Orlando, November 2005 43. Kanaga Karuppiah KS, Sundararajan (2005) In: Proceedings of the 4th international surface engineering congress, St Paul, August 2005 44. Karuppiah KSK, Sundararajan S, Xu ZH, Li XD (2006) Tribol Lett 22:181 45. Widmer MR, Heuberger M, Voros J, Spencer ND (2001) Tribol Lett 10:111 46. Heuberger MP, Widmer MR, Zobeley E, Glockshuber R, Spencer ND (2005) Biomaterials 26:1165 47. Bragdon CR, Jasty M, Kawate K, McGrory BJ, Elder JR, Lowenstein J, Harris WH (1997) J Arthroplasty 12:119 48. Joyce TJ, Monk D, Scholes SC, Unsworth A (2000) Biomed Mater Eng 10:241 49. Bennett D, Orr JF, Beverland DE, Baker R (2002) Proc Inst Mech Eng H 216:393 50. Bhushan B (1999) Handbook of micro/nano tribology. CRC, Boca Raton 51. Ruan J-A, Bhushan B (1994) J Tribol 116:378 52. Ogletree DF, Carpick RW, Salmeron M (1996) Rev Sci Instrum 67:3298 53. Torii A, Sasaki M, Hane K, Okuma S (1996) Meas Sci Technol 7:179 54. Sader JE, Chon JWM, Mulvaney P (1999) Rev Sci Instrum 70:3967 55. Schwarz UD, Zworner O, Koster P, Wiesendanger R (1997) Phys Rev B Condens Matter 56:6987 56. Carpick RW, Agrait N, Ogletree DF, Salmeron M (1996) J Vac Sci Technol B 14:1289 57. Park S, Costa Kevin D, Ateshian Gerard A (2004) J Biomech 37:1679 58. Mitchell AS, Shrotriya P (2005) In: Proceedings of the 4th international surface engineering congress, St Paul, August 2005 59. Ho SP, Carpick RW, Boland T, LaBerge M (2002) Wear 253:1145 60. Kanaga Karuppiah KS, Sundararajan Bruck AL, Lin Z (2006) In: Proceedings of MRS fall 2006 meeting, Boston, November 2006 61. Turell MB, Bellare A (2004) Biomaterials 25:3389 62. Carpick RW, Salmeron M (1997) Chem Rev 97:1163 63. Carpick RW, Ogletree DF, Salmeron M (1999) J Colloid Interface Sci 211:395 64. Derjaguin BV, Muller VM, Toporov YP (1975) J Colloid Interface Sci 53:314 65. Park KD, Kim J, Yang SJ, Yao A, Park JB (2003) J Biomed Mater Res B 65B:272 66. Gracias DH, Somorjai GA (1998) Macromolecules 31:1269 67. Yui N, Suzuki Y, Mori H, Terano M (1995) Polym J 27:614 68. Sundararajan S, Bhushan B (2000) J Appl Phys 88:4825 69. Good VD, Clarke IC, Gustafson GA, Downs B, Anissian L, Sorensen K (2000) Acta Orthop Scand 71:365 70. Cooper JR, Dowson D, Fisher J (1993) Clin Mater 14:295 71. Kernick M, Allen C (1997) Wear 203:537 72. Norde W, Giacomelli CE (2000) J Biotechnol 79:259 73. Ying P, Yu Y, Jin G, Tao Z (2003) Colloids Surf B 32:1 74. Mitchell AS, Shrotriya P (2007) Wear 263:1117 75. Milosev I, Strehblow HH (2003) Electrochim Acta 48:2767 76. Goldberg JR, Gilbert JL (1997) J Biomed Mater Res 37:421 77. Yu HH, Suo Z (2000) J Appl Phys 87:1211 78. Kawano T, Miura H, Mawatari T, Moro-Oka T, Nakanishi Y, Higaki H, Iwamoto Y (2003) Arthritis Rheum 48:1923
350
S. Sundararajan · K.K.K. Subramanian
79. Swann DA, Radin EL, Nazimiec M, Weisser PA, Curran N, Lewinnek G (1974) Ann Rheum Dis 33:318 80. Mabuchi K, Obara T, Ikegami K, Yamaguchi T, Kanayama T (1999) Clin Biomech 14:352 81. Hills BA, Butler BD (1984) Ann Rheum Dis 43:641 82. Purbach B, Hills BA, Wroblewski BM (2002) Clin Orthop Relat Res 115 83. Adams GA, Feuerstein IA (1981) Trans Am Soc Artif Intern Organs 27:219
26 Near-Field Optical Spectroscopy of Single Quantum Constituents Toshiharu Saiki
Abstract. The remarkable progress in spatial resolution of near-field scanning optical microscopy offers the possibility of unique interactions between light and matter at the nanoscale. In this chapter, we describe the development of a high-resolution near-field scanning optical microscope with a carefully designed aperture probe and near-field imaging spectroscopy of quantum confined systems. Thanks to a spatial resolution as high as 10–30 nm, we successfully visualize spatial profiles of local density of states and wavefunctions of electrons (excitons) confined in semiconductor quantum dots. Fundamental aspects of localized and delocalized electrons in interface and alloy disorder systems are also clarified through spatial and energy-resolved spectroscopy. Key words: Light–matter interaction, Quantum dot, Wavefunction, Alloy disorder, Localized state
26.1 Introduction Near-field scanning optical microscopy (NSOM) enables us to achieve a spatial resolution far beyond the diffraction limit of light (Fig. 26.1), a restriction which conventional optical microscopy suffers from, by squeezing light through a sharpened optical fiber probe [1, 2]. NSOM has a great advantage over other scanning probe microscopy methods in analytical measurement by combining with a variety of optical spectroscopic techniques, including photoluminescence (PL) spectroscopy, Raman spectroscopy, and ultrafast spectroscopy with femtosecond optical pulses. One of the most attractive applications of near-field spectroscopy is to investigate single quantum constituents, such as semiconductor quantum dots (QDs), trap centers, local minima of a random disorder potential, etc. as illustrated in Fig. 26.2. Semiconductor QDs, where electrons are confined in a nanoscale volume, exhibit a long duration of coherence and a narrow optical spectrum owing to their atomlike discrete energy levels. Thanks to such nature of QDs, they are key elements in semiconductor lasers [3], quantum computing [4], and quantum cryptography [5]. Spatial resolution in far-field optical spectroscopy is unsatisfactory for exploring and addressing these individual QDs among a high-density ensemble of naturally formed QDs. As shown in Fig. 26.3, probing QDs with fixed small apertures or with mesa structures is a useful solution for isolating single QDs, but the imaging ability is sacrificed. This is a preliminary motivation to employ the approach based on NSOM in single QD spectroscopy.
352
T. Saiki
Fig. 26.1. a Diffraction limit of light. b Metal aperture to create a nanoscale light spot far beyond the diffraction limit of light Fig. 26.2. Objects investigated by single-quantumconstituent spectroscopy
Fig. 26.3. Isolation and probing of a single quantum dot by means of mesa etching and metal mask with a small aperture
Recent progress in the fabrication of aperture near-field fiber probe has pushed the spatial resolution of NSOM to less than 30 nm [6–8], which is comparable to or below the sizes of QDs. Such a high spatial resolution allows one to directly map out the spatial distribution of the electron wavefunction. Moreover illumination of QDs with an extremely narrow light source makes it possible to excite optically “dark” states whose excitation is forbidden by symmetry in the far field (breakdown of the usual optical selection rules) [9–14]. These interesting observations and manipulations of electronic states in a quantum confinement system are unique to light–matter interaction at the nanoscale and the essential motivation to utilize the near-field optical method.
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
353
In this chapter, current progress in the instrumentation and measurements of NSOM and its application to imaging spectroscopy of single quantum constituents are described. The most critical element in NSOM is an aperture probe, which is a tapered and metal-coated optical fiber. We examine the design and fabrication of the probe with regard to aperture quality and the efficiency of light propagation. The recent dramatic improvements in spatial resolution and optical throughput are illustrated by single QD spectroscopy, which reveals the intrinsic nature of quantum confined systems. Beyond such an application, real-space mapping of exciton wavefunctions confined in a QD is also demonstrated.
26.2 General Description of NSOM When a small object or an object with tiny substructures is illuminated with light, its fine structures with high spatial frequency generate a localized field that decays exponentially normal to the object’s surface. (The decay length is proportional to the inverse of the spatial frequency.) This localized field, which is referred to as an evanescent field, can be used as a local source of light illuminating and scanning a sample’s surface so close that the light interacts with the sample without diffracting. There are two methods by which a localized optical field suitable for NSOM can be generated. As illustrated in Fig. 26.4a, one method uses a small aperture at the apex of a tapered optical fiber coated with metal. Light sent down the fiber probe and through the aperture illuminates a small area, basically determined by the size of the aperture. The fundamental spatial resolution of so-called aperture NSOM ranges from 10 to 100 nm. In the other method, called apertureless (or scattering) NSOM and illustrated in Fig. 26.4b, a strongly confined optical field is created by external illumination at the apex of a sharpened metal or dielectric tip [15]. Laboratory experiments have yielded spatial resolutions ranging from 1 to 20 nm. A rather large (diffractionlimited) laser spot focused on a tip apex frequently causes an intense background that reduces the signal-to-noise ratio. This contrasts with what is done in aperture NSOM, where the aperture serves as a localized light source without any background.
Fig. 26.4. a Aperture near-field scanning optical microscopy (NSOM) and b apertureless (scattering) NSOM
354
T. Saiki
Fig. 26.5. Standard NSOM setup with a local illumination and local collection configuration
The simplest setup for aperture NSOM, a configuration with local illumination and local collection of light through an aperture, is illustrated in Fig. 26.5. The probe quality and the regulation system for tip–sample feedback are critical to NSOM performance, and most near-field scanning optical microscopes use a method similar to that used in an atomic force microscope, called shear force feedback, the regulation range of which is 0–20 nm (typically) [16]. The light emitted by the aperture interacts with the sample locally. It can be absorbed, scattered, or phaseshifted, or it can induce fluorescence. Which of these occurs depends on the sample and produces contrast in the optical images. In any case, light emerging from the interaction volume must be collected as efficiently as possible. When the sample is prepared on a transparent substrate, signal light is frequently collected with an objective lens arranged in a transmission configuration.
26.3 NSOM Aperture Probe 26.3.1 Basic Process of Aperture Probe Fabrication Great efforts have been devoted to the fabrication of the aperture probe, which is the heart of NSOM. Since the quality of the probe determines the spatial resolution and sensitivity of the measurements, tip fabrication remains of major interest in the development of NSOM. The fabrication of fiber-based optical probes can be divided into three main steps: (1) the creation of a tapered structure with a sharp apex, (2) the coating with a metal to obtain an entirely opaque film on the probe, and (3) the formation of a small aperture at the apex. For sharpening the end of the optical fiber to a desired shape methods based on chemical etching in a hydrofluoric acid (HF) solution [17] are commonly used owing to the advantage of high reproducibility and controllability compared with the heating-and-pulling method. Moreover, to realize the ideal taper structures of the sort that will be discussed later, the chemical-etching method is advantageous because the taper angle can be adjusted by changing the composition of a buffered HF solution. To make a high-definition aperture probe by milling and polishing use of a focused ion beam has been developed [18, 19].
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
355
A simpler method based on the mechanical impact of the tip on a suitable surface is also useful [20]. In both methods, the resulting probe has a flat end and a well-defined circular aperture. Furthermore, the impact method ensures that the aperture plane is strictly parallel to the sample surface, which is important in minimizing the distance between the aperture and the sample’s surface. 26.3.2 Tapered Structure and Optical Throughput Improvement of the optical transmission efficiency (throughput) and collection efficiency of aperture probes is the most critical issue to be addressed for the application of NSOM in spectroscopic studies of nanostructures. In addition to contributing to the high sensitivity of NSOM measurements, optical probes with high throughput will open up attractive new areas of research, such as nonlinear processes and optical manufacturing on the nanometer scale. It is therefore of great importance to understand the limitations and possibilities of aperture probes with respect to light transmission. The tapered region of the aperture probe can be viewed as a metal-clad optical waveguide as viewed in Fig. 26.6. Its mode structure is completely different from that in an unperturbed fiber and is characterized by the cutoff diameter and absorption coefficient of the cladding metal. Theoretical and systematic experimental studies have confirmed that the transmission efficiency of the propagating mode decreases in the region where the core diameter is smaller than half the wavelength of the light in the core (cutoff region). The power that is actually delivered to the aperture depends on the length of the cutoff region, which is determined by the cone angle of the taper. Thus, a fiber probe with a large taper angle is favorable to enhance the power transmission. Further optimization of the tapered structure is needed to maximize the transmission efficiency. However, it is very time-consuming to assess many structure parameters, such as the cone angle and taper length, by trial and error. Numerical analysis is a more reasonable way to attain an optimized structure efficiently. Computational calculation by the finite-difference time-domain (FDTD) method is the most popular and promising method available for this purpose, because it can be easily applied to actual three-dimensional problems [21, 22]. For example, we demonstrated a double-tapered structure, as illustrated in Fig. 26.6, is more suitable
Fig. 26.6. An apertured probe with a double-tapered structure, which is optimized to attain high transmission efficiency
356
T. Saiki
for obtaining high collection efficiency than single-tapered probes [23]. This is because the second tapered region modifies the wavefront of the propagating light to match the guiding mode of the optical fiber. We therefore proposed a double-tapered structure with a large cone angle. This structure is easily realized using a multistep chemical-etching technique, as will be described later. With this technique, the transmission efficiency is much improved by 1–2 orders of magnitude compared with that of the single-tapered probe with a small cone angle. 26.3.3 Fabrication of a Double-Tapered Aperture Probe Here we describe a chemical-etching process with buffered HF solution to fabricate the double-tapered probe. Since the details of probe fabrication with selective etching are detailed in [23], we describe the process only briefly here. The cone angle can be controlled by the buffering condition of the etching solution, which is adjusted by the volume ratio of NH4 F maintaining the proportion of HF to H2 O at 1:1. A two-step etching process is employed to make a double-tapered probe (Fig. 26.6). In the first step, a short tip with a large cone angle of 150◦ is fabricated. Second, the guiding region is obtained using a solution with a different composition. The neck diameter can be controlled by the etching time in the second step. The next step is metal coating and aperture formation. The entire exterior surface of the etched probe was metal-coated with an Au film using a sputtering coating method. A small aperture was created by pounding the metal-coated probe on a sapphire substrate or on the sample itself and squeezing the Au off to the side [20]. Figure 26.7a and b shows scanning electron micrographs of the 50- and 70-nm apertures, taken after conducting PL imaging several times. Not only a smooth and flat end-face but also a round and well-defined aperture was obtained in this fabrication process. As mentioned earlier, the aperture plane is parallel to the sample’s surface to minimize the distance between the aperture and the sample’s surface. The size of the aperture can be selected by carefully monitoring the intensity of light being transmitted from the apex, since the throughput of the probe is strictly dependent on the aperture diameter. Such controllability in aperture formation is demonstrated in the scanning electron micrographs of ultrasmall apertures in Fig. 26.7c.
Fig. 26.7. Scanning electron micrographs of apertured probes with aperture diameters of a 70 nm, b 50 nm, and c 30 nm
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
357
26.3.4 Evaluation of Transmission Efficiency and Collection Efficiency To evaluate the transmission efficiency of aperture probes, researchers often measure the light power emitted by the aperture in the far field. However, in almost all experimental situations the instantaneous electric fields dominate light–matter interaction. Thus, the far-field transmission coefficient (the ratio of the far-field light power emitted by the aperture to the power coupled into the fiber) does not reflect the field enhancement in the near-field region. However, when probes with the same diameter are compared, the far-field transmission coefficient can be valuable information. In Fig. 26.8 the transmission coefficient is plotted against the aperture diameter for conventional single-tapered probes and for double-tapered probes with cone angles of 60◦ and 90◦ . As a light source for the measurement, we used a He–Ne laser (λ = 633 nm). It is clear that the transmission efficiency of the double-tapered structure with a large cone angle is improved by 2 orders of magnitude compared with that of the single-tapered probe with a small cone angle. In the case of a 100nm aperture, the transmission coefficient is as large as 10−2 . The reduction of the throughput with a decrease of aperture diameter is not so severe as predicted by Bethe’s theory, where the throughput must be proportional to (d/λ)−6 (d is the aperture diameter). Thus, we obtain reasonable illumination intensity even in the case of a 10–30-nm aperture. More sophisticated structures, such as a triple-tapered probe, have also been proposed for specific applications. In addition to a high throughput for efficient illumination of light, superior collection of the locally emitted signal is critical for the observation of opaque materials such as semiconductors. We checked the collection efficiency of a 70-nm aperture probe quantitatively by collecting PL from an InGaAs QD buried 70 nm beneath the surface. We compared the PL intensity collected by the aperture probe with that collected simultaneously by an objective with a numerical aperture of 0.8. In Fig. 26.9, the PL signal intensities are plotted as a function of distance between the aperture and the sample’s surface. Within 15 nm of the surface, the amount of light collected by the aperture probe increases rapidly and reaches a value as great as the amount collected by an objective with a numerical aperture of 0.8. This suggests that fabrication of the flat-ended probe makes a considerable contribution to the efficient interaction of the aperture and evanescent field in the vicinity of the sample.
Fig. 26.8. Transmission coefficient of apertured probes as a function of aperture diameter for singletapered and double-tapered probes with two different cone angles
358
T. Saiki
Fig. 26.9. Comparison of collection efficiency of photoluminescence (PL) from a single quantum dot for an objective and for an apertured probe with a 70-nm aperture. NA numerical aperture
The overall throughput (or collection efficiency) of light is determined by various factors such as the wavelength of propagating light and the corresponding dielectric constant of the cladding metal, as well as by the structure of unperturbed fiber. The FDTD simulation of light propagation is a useful and efficient way to obtain the information needed to make an optimized structure. 26.3.5 Evaluation of Spatial Resolution with Single QDs Fluorescence imaging of point-like emission sources is the most straightforward and reliable method for evaluating the spatial resolution of a probe. In particular, semiconductor QDs are most suitable in terms of their sizes, quantum efficiency, and optical stability. We evaluated the spatial resolution for various aperture diameters by PL imaging of single InAs self-assembled QDs [6]. The typical lateral size and height of the dots were 22 and 3 nm, respectively. The thickness of the GaAs cap layer, which determines the achievable spatial resolution, is of critical importance. While a thinner cap layer contributes to higher resolution, it also causes the degradation of the optical quality of the QDs. We used the QD sample with a cap layer of 20-nm thickness, taking into account the balance between the achievable spatial resolution and the optical quality. By illuminating the sample with diode-laser light (λ = 685 nm) through the aperture, we generated most of the carriers in a barrier layer surrounding the QDs. After diffusing in the barrier layer, the carriers were captured in the confined states of QDs. The PL signal from a single QD was collected by the same aperture. All the measurements were performed at 9 K in a cryostat. The tapered structure, defined by a cone angle (θ = 90◦ ) and a neck diameter (D = 1 µm), was optimized from the viewpoint of both the optical throughput and the spatial resolution, based on the results of FDTD simulation.
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
359
Figure 26.10a–c shows the low-temperature PL images of InAs QDs obtained (center photon energy E = 1.33 eV, ΔE = 60 meV) obtained with apertures with diameters of 135, 75, and 30 nm, respectively. The individual bright spots correspond to the PL signal from single QDs. The size of each spot becomes smaller with a decrease of the aperture diameter employed. To evaluate the spatial resolution in Fig. 26.10c, we plotted the cross-sectional profiles of PL signal intensity for the spot indicated by arrows in Fig. 26.10d. The full width at half maximum of the PL signal profile was 38 ± 2 nm. Taking into account the size of the QDs (22 nm), we estimated the spatial resolution of this measurement to be about 30 nm, which corresponds to λ/30. In the illumination-collection mode operation of NSOM, the small aperture plays roles in both excitation and collection of the PL signal. In the case of carrier generation in the barrier layer, however, the excitation process does not contribute to such a high spatial resolution owing to the carrier diffusion in several hundred nanometers. We conclude that high resolution could be achieved solely by the collection process of the PL signal. We carried out the same measurements as described above for other probes and evaluated PL spot size as a function of aperture diameter, as shown in Fig. 26.11. An aperture size of 0 nm means that there is no physical aperture at the apex, which was obtained by stopping the pounding procedure just before opening the aperture. In spite of the lack of a physical aperture, we could detect reasonable intensity of the PL signal and image with a resolution of 75 nm. Except for the 0-nm aperture, the spatial resolution is monotonously enhanced with a decrease of the aperture diameter. The spatial resolution is almost equal to the aperture diameter, whose behavior is also reproduced well by the FDTD simulation.
Fig.26.10.Low-temperature NSOM PL images of single quantum dots, obtained by using aperture probes with aperture diameters of a 135 nm, b 75 nm, and c 30 nm. d Cross-sectional PL intensity profile of the spot indicated by the arrows in c
360
T. Saiki
Fig. 26.11. PL spot diameter as a function of aperture size
26.4 Single-Quantum-Constituent Spectroscopy With the recent progress in the nanostructuring of semiconductor materials and in the applications of these nanostructured materials in optoelectronics, NSOM and near-field scanning optical spectroscopy have become important tools for determining the local optical properties of these structures. In single-quantum constituentspectroscopy, NSOM provides access to individual quantum constituents, such as QD, an ensemble of which exhibits inhomogeneous broadening owing to the distribution of sizes, shapes, and strains. NSOM can thus elucidate the nature of QDs, including the narrow optical transition arising from the atomlike discrete density of states [24, 25]. Single QD spectroscopy has revealed their long coherence times at low temperature and large oscillator strengths of the optical transition. However, to improve these parameters for implementation in quantum computers, accurate information on the wavefunction for individual QDs is of great importance. In addition, in the study of coupled-QD systems as interacting qubits, in which it is difficult to predict the exact wavefunction within theoretical frameworks, an optical spectroscopic technique for probing the wavefunction itself should be developed. By enhancing the spatial resolution of NSOM up to 10–30 nm, which is smaller than the typical size of QDs, local probing allows direct mapping of the real-space distribution of the quantum eigenstate (wavefunction) within a QD, as predicted by theoretical studies [10–14]. In contrast to the well-defined quantum confined systems like QDs, the more common disordered systems with local potential fluctuations still leave open questions. To fully understand such complicated systems, exciton wavefunctions should be visualized with an extremely high resolution less than the spatial extension of the wavefunction. NSOM with a spatial resolution of 10 nm level is the only tool that can be used to obtain such information.
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
361
26.4.1 Light–Matter Interaction at the Nanoscale In this section we summarize a theoretical approach to understand light–matter interaction at the nanoscale [12]. When the nanoscale confined electron system, such as a semiconductor QD, is excited by light with a frequency ω, the absorbed power α(ω) is α(ω) = drE(r)P(r, ω) , (26.1) where E(r) is the spatial distribution of electromagnetic field and P(r, ω) is the induced interband polarization. In the general form the relationship between P(r, ω) and E(r) should be expressed by the nonlocal electrical susceptibility χ(r, r′ ; ω) as P(r, ω) = dr′ χ(r, r′ ; ω)E(r′ ) . (26.2) χ(r, r′ ; ω) can be obtained from the eigenfunction Ψex and the eigenenergy E ex of the exciton state confined in a QD: χ(r, r′ ; ω) ∝
Ψex (r, r ′ )Ψex (r, r ′ ) . E ex − ω − iγ
(26.3)
Here we assume that quasi-resonant excitation at E ex and therefore the contribution of other quantized exciton states are negligible. γ is a damping constant due to phonon scattering and radiative decay of the exciton. By using (26.2) and (26.3), we can write α(ω) in the form
drΨ(r, r)E(r)2 α(ω) ∝ . (26.4) E ex − ω − iγ To illustrate the physical meaning of (26.4), we discuss two limiting cases. For far-field excitation, where the QD is illuminated by a spatially homogeneous electromagnetic field, α(ω) is given by the spatial integration of the exciton wavefunction, 2 (26.5) α(ω) ∝ drΨ(r, r) . From the value of this integral, so-called optical selection rules are derived. If the integral is zero, the corresponding transition is “forbidden” and the exciton state is optically “dark.” In the opposite limit of extremely confined light, E(r) is assumed to be δ(r − R), where R is the position of the nanoscale light source, say, a near-field tip. As a result one can probe the local value of the exciton wavefunction, α(ω) ∝ |Ψ(R, R)|2 .
(26.6)
By measuring α(ω) as a function of the tip position, we can map out the exciton wave function. More interestingly, a dark-state exciton becomes visible by
362
T. Saiki
breaking the selection rule for the optical transition. In the intermediate regime in terms of the confinement of light, Ψex are averaged over an illumination region. Now we try to give an intuitive explanation of the local optical excitation using a classical coupled oscillator model as shown in Fig. 26.12. Each pendulum represents a localized dipole, such as a constituent molecule that makes up a molecular crystal. The dipole–dipole interaction, which forms exciton as a collective excitation of constituent molecules, is taken into account by introducing springs to couple neighboring pendulums. Here we assume the size of the system (the size of molecular crystal) is much smaller than the wavelength of light. Figure 26.12 illustrates the lowest and the second-lowest normal modes of the coupled oscillator. For the far-field illumination, all the pendulums are swung together at the frequency of irradiated light with the same phases; therefore, the second-lowest mode, where the two halves of the pendulums move in opposite directions, cannot be excited by the far-field light, whereas the lowest mode can be. This corresponds to the optical selection rule for far-field excitation of confined exciton systems. For the near-field regime, on the other hand, the situation drastically changes. The trick that the nanoscale confined light plays is to grasp solely a single pendulum and swing it. In this case, if the light frequency matches the eigenfrequency of the individual oscillation mode, any normal mode can be excited regardless of the symmetry of oscillation, which means that the optical selection rule is broken by the near-field excitation. The efficiency of mode excitation is dependent on which pendulum is swung, i. e., the position of the nanoscale light source. By swinging a pendulum in order from the end and observing the magnitude of the mode oscillation for each, we can map out the distribution of the oscillation amplitude of individual pendulums. This illustrates the principle of the wavefunction mapping of exciton states.
Fig. 26.12. Coupled pendulum model to intuitively explain the light–matter interaction at the nanoscale. a The lowest mode and b the second-lowest mode of the coupled oscillator
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
363
26.4.2 Real-Space Mapping of an Exciton Wavefunction Confined in a QD Here we describe PL imaging spectroscopy of a GaAs QD by NSOM with a spatial resolution of 30 nm. This unprecedented high spatial resolution relative to the size of the QD (approximately 100 nm) permits a real-space mapping of the center-of-mass wavefunction of an exciton confined in the QD based on the principle discussed already [26, 27]. A schematic of the structure of a QD sample is shown in Fig. 26.13. We prepared a 5-nm-thick GaAs quantum well (QW), sandwiched between layers of Al(Ga)As grown by molecular-beam epitaxy. Two-minute interruptions of the growth process at both interfaces led to the formation of large monolayer-high islands which localize excitons in a QD-like potential with lateral dimensions on the order of 40– 100 nm [28]. (Hereafter we call this island a QD.) The GaAs QW layer was covered with a thin barrier and a cap layer of 20 nm, which allows the near-field tip to be close enough to the emission source (QD). The experimental setup for low-temperature NSOM imaging spectroscopy is illustrated in Fig. 26.14 [29]. The GaAs QD was excited with He–Ne laser light (λ = 633 nm) through the aperture and carriers (excitons) were created in the barrier layers as well as in the QW layer. Excitons diffused over several hundreds of nanometers and relaxed into the QDs. The PL signal from the QD was collected via the same aperture to prevent a reduction of the spatial resolution due to carrier diffusion. Near-field PL spectra were measured, for example, at 11-nm steps across a 210 nm × 210 nm area and two-dimensional images were constructed from a series of PL spectra. Figure 26.15a shows near-field PL spectra of a single QD at 9 K at excitation densities ranging from 0.17 to 3.8 µW. At low excitation densities, a single emission line (denoted by X) at 1.6088 eV is observed. With an increase in excitation density, the other emission lines appear at 1.6057 eV (XX) and at 1.6104 eV (X∗ ). In order to clarify the origin of these emissions, the excitation power dependence of PL intensities were examined as shown in Fig. 26.15b. The X line can be identified as an emission from a single-exciton state by its linear increase in emission intensity and its saturation behavior. The quadratic dependence of the XX emission with excitation power indicates that XX is an emission from a biexciton state. This identification of the XX line is also supported by the difference in the emission
Fig. 26.13. A GaAs interface quantum dot (quantum island). QW quantum well
364
T. Saiki
Fig. 26.14. Low-temperature NSOM setup for the near-field imaging spectroscopy of GaAs quantum dots
Fig. 26.15. a Near-field PL spectra of a GaAs single quantum dot at 9 K under various excitation power densities. b Excitation power dependence of PL intensities of the X and the XX emission lines
energy of 3.1 meV, which corresponds to the binding energy of a biexciton and agrees well with the values reported previously [30]. The X∗ emission line can be attributed to the radiative recombination of the exciton excited state by considering its energy position (higher-energy side of the single-exciton emission by about 1.6 meV) [31]. Figure 26.16 shows low-magnification PL maps for the intensity of X emissions with three different energies in the same scanning area. These emission profiles were found to differ from QD to QD. The high-magnification PL images in Fig. 26.17 were obtained by mapping the PL intensity with respect to the X (Fig. 26.17, images a, c, and e) and the XX (Fig. 26.17, images b, d, and f) lines of three different QDs. The exciton PL images in Fig. 26.17, images a, c, and e show an elongation along the [−110] crystal axis. The image sizes are larger than the PL collection spot diameter, i. e., the spatial resolution of the near-field scanning optical microscope. The elongation along the [-110] axis due to the anisotropy of the monolayer-high island is consistent with previous observations with a scanning tunneling microscope [28]. We also obtained elongated biexciton PL images along the [-110] crystal axis (Fig. 26.17, images b,
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
365
Fig. 26.16. Twodimensional mapping of the PL intensity for three different X lines at a 1.6088, b 1.6092, and c 1.6033 eV
d, and f) and found a clear difference in the spatial distribution between the exciton emission and the biexciton emission. Here the significant point is that the PL image sizes of biexcitons are always smaller than those of excitons. Figure 26.18, panels a and b show the normalized cross-sectional PL intensity profiles of an exciton (thick lines) and a biexciton (thin lines) along the [110] and [−110] crystal axes. The spreads in the exciton (biexciton) images, defined as the full width at half maximum of each profile, are 80(60) nm and 115(80) nm along the [110] and [−110] crystal axes, respectively. Theoretical considerations can clarify what we see in the exciton and biexciton PL images. The relevant quantity is the optical near field around a single QD associated with an optical transition. This field can be calculated with Maxwell’s equations using the polarization field of the exciton or the biexciton as the source term. The observed luminescence intensity is proportional to the square of the near field detected by the probe. In the following, however, we have calculated the emission patterns simply from the squared polarization fields without taking account of the instrument details. The polarization fields at the position of the probe (rs ) are derived from the transition matrix element from the exciton state (X) to the ground state (0) and from that from the biexciton state (XX) to the exciton state (X) as follows [32]: √ 0| pδ(r − rs ) |X = − 2 pcv ϕ(rs , rs ) X| pδ(r − rs ) |XX = − 3/2 pcv φ(r1 , ra )Φ ++ (r1 , rs ; ra , rs ) r1 ,ra
− 1/6 pcv φ(r1 , ra )Φ −− (r1 , rs ; ra , rs ) , r1 ,ra
366
T. Saiki
Fig.26.17.Series of high-resolution PL images of exciton states (a, c, e) and biexciton states (b, d, f) for three different quantum dots
where φ(re , rh ) stands for the exciton envelope function with the electron and hole coordinates denoted by re , rh , Φ ++ (Φ −− )(r1 , r2 ; ra , rb ) represents the biexciton envelope function with electron coordinates (r1 , r2 ) and hole coordinates (ra , rb ) that is symmetrized (antisymmetrized) with respect to the interchange between two electrons and between two holes, and pcv is the transition dipole moment between the conduction band and the valence band. The spatial distribution of the exciton polarization field corresponds to the center-of-mass envelope function of a confined exciton. For the biexciton emission, the polarization field is determined by the overlap integral, which represents the spatial correlation between two excitons forming the biexciton and is expected to be more localized than the single-exciton wavefunction. Figure 26.18, panel c shows the squared polarization amplitudes of the exciton (thick line) and biexciton (thin line) emission, which were calculated for a GaAs QD with size parameters relevant to our experiments. The calculated profile of the squared polarization amplitude of the biexciton emission is narrower than that of the exciton emission. The spread of the biexciton emission normalized by that of the exciton emission is estimated to be 0.76, which is in good agreement with
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
367
Fig. 26.18. a, b Normalized crosssectional intensity profile of exciton and biexciton PL images estimated from Fig. 26.17, images a and b along the [110] and [−100] crystal axes, respectively. c Spatial distributions of squared polarization fields of the exciton and biexciton emission, which are theoretically calculated for a GaAs quantum disk
the experimental result (0.75±0.08). This theoretical support and the experimental facts lead to the conclusion that the local optical probing by the near-field scanning optical microscope directly maps out the center-of-mass wavefunction of an exciton confined in a monolayer-high island. Furthermore, we can demonstrate a novel powerful feature of the wavefunction mapping spectroscopy. Figure 26.19 shows the PL image and corresponding crosssectional intensity profiles of the exciton ground state X (Fig. 26.19a, b) and the exciton excited state X∗ (Fig. 26.19c, d) from a single QD, which is different from that observed in Fig. 26.6. The exciton PL image exhibits a complicated shape for this QD, unlike the simple elliptical shape shown in Fig. 26.6. This is because the exciton is confined in a monolayer-high island with an extremely anisotropic shape. The significant point is that the exciton ground state image exhibits a single maximum peak in the intensity profile, while a double-peaked intensity profile is obtained from the exciton excited state. This is attributable to the difference in spatial distribution of the center-of-mass wavefunction, which has no node in the ground state, but does have a node in the excited state. 26.4.3 Carrier Localization in Cluster States in GaNAs In contrast to the well-defined quantum confined systems such as QDs grown in a self-assembled mode, the more common disordered systems with local potential
368
T. Saiki
Fig. 26.19. High-resolution PL images and corresponding cross-sectional intensity profile of the lowest exciton state (a, b) and the excited exciton state (c, d). The intensity profiles are taken along the solid line and the dotted line
fluctuations leave unanswered questions. For example, a large reduction of the fundamental band gap in GaAs with small amounts of nitrogen is relevant to the clustering behavior of nitrogen atoms and the resultant potential fluctuations. NSOM characterization with high spatial resolution can give us a lot of important information that is useful in our quest to fully understand such complicated systems, such as details about the localization and delocalization of carriers, which determine the optical properties in the vicinity of the band gap. Dilute GaNAs and GaInNAs alloys are promising materials for optical communication devices because they exhibit large band-gap bowing parameters. In particular, for long-wavelength semiconductor laser application, high temperature stability of the threshold current is realized in the GaInNAs/GaAs QW compared with the conventional InGaAsP/InP QW owing to strong electron confinement. However, GaNAs and GaInNAs with a high nitrogen concentration of more than 1% have been successfully grown only under nonequilibrium conditions by molecular-beam epitaxy and metal–organic vapor-phase epitaxy. The incorporation of nitrogen generally induces degradation of optical properties. To date, several groups of researchers have reported characteristic PL properties of GaNAs and GaInNAs, for example, the broad asymmetric PL spectra and the anomalous temperature dependence of the PL peak energy. More seriously, emission yield drastically degraded with an increase of nitrogen concentration. For improvement of their fundamental optical properties, it is strongly required to clarify electronic states due to single N impurities or N clusters, which interact with each other and with the host states. The interaction gives rise to the formation of weakly localized and delocalized electronic states at the band edge; hence, the
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
369
shape of optical spectra is extremely sensitive to the N composition. In conventional spatially resolved PL spectroscopy, it is easy to detect single impurity emissions in the ultradilute compositional region. However, in order to resolve complicated spectral structures and to clarify the interaction of localized states and the onset of alloy formation, a spatial resolution far beyond the diffraction limit is needed. In this section, we show the results of spatially resolved PL spectroscopy with a high spatial resolution of 30 nm [33]. Spatial inhomogeneity of PL is direct evidence of carrier localization in the potential minimum case caused by the compositional fluctuation. PL microscopy with such a high spatial precision enables the direct optical observation of compositional fluctuations, i. e., spontaneous N clusters and N random alloy regions, which are spatially separated in GaNx As1−x /GaAs QWs. The samples investigated in this study were GaNxAs1−x (x = 0.7−0.8%) epilayers grown on (001) GaAs substrates by using low-pressure metal–organic vaporphase epitaxy. The N concentration was estimated by secondary ion mass spectroscopy. After growth, thermal annealing was performed in ambient H2 to improve the optical properties. Figure 26.20 shows the excitation power dependence of the macro-PL spectrum from a 500-nm GaNAs epilayer sandwiched by a 100-nm thick GaAs buffer layer and a 50-nm thick GaAs cap layer [34]. The measurement temperature was 10 K. Broad and asymmetric PL spectra with a low-energy tail as shown in Fig. 26.20, panel a were observed under low-excitation conditions. With an increase in excitation power density, the PL intensity of the low-energy tail gradually reaches saturation, and the peak energy of the PL spectra (squares) is shifted to the higher-energy side. In the PL spectrum under the highest excitation conditions, we can see the appearance of a shoulder (circles) corresponding to the transition energy of the band edge (1.343 eV) determined by the photoreflectance spectrum. Figure 26.20, panel b shows the excitation power dependence of the macro-PL spectra at room temperature. In contrast to what occurs at low temperature, the peak energy indicated by the circles and the shape of the PL spectra do not change with an increase in
Fig. 26.20. Excitation power dependence of far-field (macroscopic) PL spectra of a GaNAs epilayer at a 10 K and b 300 K
370
T. Saiki
Fig. 26.21. a Far-field PL spectrum of GaNAs QW. b, c Near-field PL spectra obtained with a spatial resolution of 30 nm. d, e Near-field PL mapping with respect to the emission lines indicated by the arrows in spectra b and c, respectively
excitation power. From these results the low-energy tail of low-temperature PL indicates exciton localization due to random fluctuation of alloy compositions. To illustrate the localized nature of the exciton at low temperature, real-space mapping of the PL distribution was performed by using NSOM [33]. The sample was a 5-nm-thick GaNAs single QW with an N composition of x = 0.7% grown on a (001) GaAs substrate. A 30-nm-aperture NSOM probe was employed to locally collect the PL signal from N-related states. PL measurement was performed at 8 K. The thickness of the cap layer was 20 nm, which is thin enough to allow the spatial resolution determined by the size of the aperture as discussed above. Near-field PL spectra exhibit several sharp emission lines in contrast to inhomogeneous broad feature of far-field macroscopic PL spectra (Fig. 26.21, spectrum a). Analyzing several thousand near-field spectra, we found that the PL fine structures were categorized into two groups: sharp emissions with linewidths narrower than 0.2 meV (Fig. 26.21, spectrum b), limited by the spectral resolution of the measurement system) and rather broad emissions with linewidths of several millielectronvolts (Fig. 26.21, spectrum c). For deep understanding of the origins of these emissions, two-dimensional spatial distributions of individual emissions were mapped. We obtained strongly localized profiles (Fig. 26.21, map d), limited by the spatial resolution of NSOM) for the narrow emissions. The spectral and spatial results indicates that the narrow emissions come from QD-like discrete states confined in spontaneous N clusters. On the other hand, the broad emissions show spatially extended behaviors (Fig. 26.21, map e), where typical area sizes are around 100 nm. The result can be interpreted as the excitons being delocalized with a continuous density of states, which is an indication of the formation of random alloy states.
26.5 Perspectives The reliability and reproducibility, as well as the resolution and sensitivity, of NSOM measurements have been largely enhanced in the last several years by the introduction of high-definition fiber probes. Probes having apertures with diameters as small as
26 Near-Field Optical Spectroscopy of Single Quantum Constituents
371
10 nm have been made, and their excellent performance has been demonstrated through the imaging and spectroscopy of semiconductor nanostructures and single molecules. The probes, however, should be further improved with respect to the optical transmission, tip stability, and damage threshold. Microfabricated probes offer a promising solution to these difficulties and are likely to improve experimental reproducibility by providing better-defined conditions. Apertureless (or scattering) techniques will surely improve spatial resolution. In this chapter we focused on the spectroscopic applications of NSOM and covered only a small portion of the research currently being pursued in the area of near-field optics. The utilization of NSOM techniques is not restricted to pure research; it has led to the development of a wide range of nano-optical devices. Real-space visualization of electronic quantum states will contribute to wavefunction engineering, where we deal with complicated potential structures for the design of novel devices. Insight into the nanometer-scale interaction between the mesoscopic electronic system and photons will provide us with the key concepts underlying the operation of nano-optoelectronic devices. For example, the near-field optical coupling of nanoelectronic systems can enhance far-field forbidden transitions. Such alternation of the selection rule for the optical transition will lead to spatiotemporal coherent control of electronic excitation, which can be utilized as the principal function of nanometric devices. Acknowledgements. We are grateful to M. Ohtsu, S. Mononobe, K. Matsuda, N. Hosaka, H. Kambe, K. Sawada, H. Nakamura, T. Inoue, F. Sato, K. Nishi, H. Saito, Y. Aoyagi, M. Mihara, S. Nomura, M. Takahashi, A. Moto, and S. Takagishi for their assistance and fruitful discussions.
References 1. Ohtsu M (1998) Near-field nano/atom optics and spectroscopy. Springer, Tokyo 2. Novotny L, Hecht B (2006) Principle of nano-optics. Cambridge University Press, New York 3. Sugawara M, Hatori N, Ishida M, Ebe H, Arakawa Y, Akiyama T, Otsubo K, Yamamoto T, Nakata Y (2005) J Phys D Appl Phys 38:2126 4. Wu Y, Li X, Duan LM, Steel DG, Gammon D (2006) Phys Rev Lett 96:8740 5. Michler P, Kiraz A, Becher C, Schoenfeld WV, Petroff PM, Zhang L, Hu E, Imamoglu A (2000) Science 290:2282 6. Matsuda K, Saiki T, Nomura S, Mihara M, Aoyagi Y (2002) Appl Phys Lett 81:2291 7. Hosaka N, Saiki T (2001) J Microsc 202:362 8. Hosaka N, Saiki T (2006) Opt Rev 13:262 9. Cho K (2003) Optical response of nanostructures. Springer, Berlin 10. Bryant GW (1998) Appl Phys Lett 72:768 11. Di Stefano O, Savasta S, Pistone G, Martino G, Girlanda R (2003) Phys Rev B 68:165329 12. Simserides CD, Hohenester U, Goldone G, Molinari E (2000) Phys Rev B 62:13657 13. Hohenester U, Goldone G, Molinari E (2005) Phys Rev Lett 95:216802 14. Runge E, Lienau C (2006) Appl Phys B 84:103 15. Inouye Y, Kawata S (1994) Opt Lett 19:159 16. Karrai K, Grober RD (1995) Appl Phys Lett 66:1842 17. Mononobe S, Ohtsu M (1998) IEEE Photon Technol Lett 10:99 18. Muranishi M, Sato K, Hosaka S, Kikukawa A, Shintani T, Ito K (1997) Jpn J Appl Phys 36:L942
372
T. Saiki
19. 20. 21. 22. 23. 24. 25. 26.
Veerman JA, Otter AM, Kuipers L, van Hulst NF (1998) Appl Phys Lett 72:3115 Saiki T, Matsuda K (1999) Appl Phys Lett 74:2773 Nakamura H, Sato T, Kambe H, Sawada K, Saiki T (2001) J Microsc 202:50 Sawada K, Sakai M, Kohashi Y, Saiki T, Nakamura H (2006) J Plasma Phys 72:1019 Saiki T, Mononobe S, Ohtsu M, Saito N, Kusano J (1996) Appl Phys Lett 68:2612 Matsuda K, Ikeda K, Saiki T, Tsuchiya H, Saito H, Nishi K (2001) Phys Rev B 63:121304 Matsuda K, Saiki T, Saito H, Nishi K (2000) Appl Phys Lett 76:73 Matsuda K, Saiki T, Nomura S, Mihara M, Aoyagi Y, Nair S, Takagahara T (2003) Phys Rev Lett 91:177401 Saiki T, Matsuda K, Nomura S, Mihara M, Aoyagi Y, Nair S, Takagahara T (2004) J Electron Microsc 53:193 Gammon D, Snow ES, Shanabrook BV, Katzer DS (1996) Phys Rev Lett 76:3005 Saiki T, Nishi K, Ohtsu M (1998) Jpn J Appl Phys 37:1638 Wu Q, Grober RD, Gammon D, Katzer DS (2000) Phys Rev B 62:13022 Gammon D, Snow ES, Katzer DS (1995) Appl Phys Lett 67:2391 Nair SV, Takagahara T (1997) Phys Rev B 55:5153 Matsuda K, Saiki T, Yamada T, Ishizuka T (2004) Appl Phys Lett 85:3077 Matsuda K, Saiki T, Takahashi M, Moto A, Takagishi S (2001) Appl Phys Lett 78:1508
27. 28. 29. 30. 31. 32. 33. 34.
Subject Index
accumulation VIII 410 acoustic emission X 342 acrylonitrile–butadiene–styrene resin X 350 actin X 303 β-actin IX 164 actuation bimetallic VIII 211 electrostatic VIII 210 piezoelectric VIII 211 thermal bimetallic VIII 210 adhesion VIII 160, IX 228 adhesion force VIII 151, VIII 162, VIII 166, VIII 168, IX 208, IX 225, IX 226, IX 229, IX 230, IX 238, IX 263, IX 278 admittance VIII 380 AFM IX 90, IX 287, X 272, X 274, X 276–278, X 352, X 355–357 micro/nanotribology IX 284 noncontact X 237 AFM probes VIII 429, X 367 conductive VIII 429 conductive coating VIII 431 AFM–SEM systems VIII 283 aging (piezoelectric) VIII 300 Al unmodified IX 254 alkylphosphonic acid IX 244 alkylphosphonic acid SAMs on Al nonperfluorinated IX 248 Amonton IX 285 amplifier VIII 432, X 78 bandwidth VIII 432 current-to-voltage VIII 432 input capacitance VIII 434 integrator–differentiator scheme VIII 433 noise VIII 432 transimpedance VIII 432 amplitude oscillations IX 97 amyloid IX 177–179, IX 181, IX 182, IX 184, IX 186, IX 189–193, IX 199
diseases IX 177, IX 178, IX 180, IX 182, IX 183, IX 189, IX 193 fibrils IX 177–181, IX 183–193, IX 196–199, IX 201 peptides IX 178, IX 192–195, IX 199 proteins IX 178–180, IX 184, IX 185, IX 189, IX 190, IX 193–195, IX 197, IX 198 protofibrils IX 189–191 protofilaments IX 179–181, IX 184–186, IX 188, IX 189, IX 199 amyloidosis IX 178, IX 191, IX 192 angle beveling X 67 anodization X 219–221 anodic oxide X 220 antenna VIII 117, VIII 118, VIII 120, VIII 122, VIII 123 anti-stiction VIII 64 antibody VIII 238 antigen VIII 239 aperture probe VIII 82 chemical etching VIII 98 fiber probes VIII 84, VIII 93 heating VIII 97 metallic tips VIII 84 pulling VIII 97 tube etching VIII 100 aperture SNOM VIII 83 apertureless probes VIII 118 dielectric tips VIII 119 electrochemical etching VIII 119 metallic tip VIII 83 apex radii IX 326 AR-XPS IX 240 area function X 315, X 316, X 323 artifacts VIII 14, VIII 15 atomic force microscopy (AFM) VIII 81, VIII 185, VIII 190, VIII 191, VIII 195, VIII 211, VIII 213, IX 23, IX 76, IX 77, IX 79–81, IX 89, IX 111, IX 334, X 185,
374 X 189, X 192–194, X 199, X 200, X 204, X 286, X 351 attractive regime IX 76 Au X 91 Auger IX 257 Auto Tip Qual VIII 43 Babinet’s reciprocity principle VIII 2 backbone chain IX 278 background removal VIII 17 ballistic electron emission microscopy (BEEM) X 88 Bell–Kaiser formalism X 89 bending experiment VIII 272 bending rigidity IX 199, IX 200 Berkovich indenters X 313–318, X 320, X 332, X 335, X 337, X 341 bias VIII 60 DC X 65 modulating AC X 65 biexciton IX 363 binding energy IX 248 biological samples X 285 biomaterials IX 329, IX 347 BioMEMS IX 283 biomolecules VIII 437, X 285, X 287 bipolar devices net base X 68 blind region VIII 65, VIII 66 blind tip reconstruction VIII 43 boiling water treatment IX 276 boundary element method (BEM) VIII 356 boundary lubrication IX 329, IX 333, IX 347 bow-tie VIII 115, VIII 123–125 bradykinin IX 139 breakdown X 93–95, X 97, X 100 dielectric X 100 extrinsic X 100 field X 97 hard X 93 intrinsic X 100 kinetics X 97, X 99 soft X 93 spots X 97, X 98 weak X 98 brush model IX 263 buffer bootstrapped VIII 383 c-fos IX 166 CAFM X 90, X 98
Subject Index calibration dC/ dV X 70 N X 70 radiation pressure VIII 292 calibration grating IX 326 calibration samples X 81 GaAs X 81 InP X 81 CAM analysis IX 254 PFMS/Cu IX 261 cantilevers VIII 290, IX 89, IX 334, IX 335 effective mass VIII 296 force sensors VIII 257 mass sensors VIII 255, VIII 258, VIII 276 silicon nitride IX 338, IX 344 spring constants VIII 290, IX 334, IX 335 tips IX 91 V-shaped VIII 292 vibration IX 316 capacitance VIII 379, X 64 apex VIII 441 attofarads X 64 dC/ dV X 65 gain X 65 local VIII 440 measurement VIII 439 resolution VIII 440 sensor X 65 stray VIII 440 capacitance–distance curves VIII 440 capacitor structure VIII 53 capillary forces VIII 72 carbon X 225 amorphous carbon X 225 diamond X 225, X 240 graphite X 219, X 225 carbon nanotubes VIII 437 carrier density X 67 intrinsic X 67 spilling X 69 CD AFM VIII 31, VIII 32 CD mode VIII 34, VIII 52, X 362 CD probes overhang VIII 72 cell endothelial IX 127 epithelial IX 127 cell adhesion IX 89, IX 98, IX 102 cell mechanics IX 50, IX 89 cell membrane IX 217, IX 227, IX 228, IX 230
Subject Index cell surfaces IX 111 cell viscoelasticity IX 93 cell–cell junctions IX 132 characterizer VIII 48 charge density VIII 358 charge transfer IX 15 charge trapping X 94 charge writing VIII 202 charges X 95 fixed X 66 interface VIII 403 space VIII 404 trapped X 66, X 96 charging X 349 chemical bond IX 12 chemical contrast X 248 chemical manipulation X 248 chemical mechanical planarization (CMP) X 132, X 134 chemical stability IX 268 chemical state IX 273 chemical vapor deposition (CVD) VIII 140, VIII 142 chemisorbs IX 236 cholesterol IX 207, IX 220–222, IX 224 chromatin fibers X 290 chromium oxide IX 344 chromosomes X 290, X 295, X 296, X 299, X 305 cilium IX 132 circuit lumped element VIII 385 resonant VIII 384 clamped beams IX 323 CMOS gate VIII 68 CMP X 132, X 389 CMP pre-metal dielectric X 386 CNT VIII 46, VIII 48, VIII 73 CNT probes VIII 69 CNT-across-trench VIII 47, VIII 48 cobalt–chromium (CoCrMo) IX 343 CODY mode IX 26, IX 48, IX 49 collagen IX 191, X 289, X 290 collection efficiency IX 355, IX 357 collector X 88 colloidal probe IX 151 complex compliance X 337–340 composite modulus IX 337 compositional imaging IX 24 compression experiments VIII 274
375 compression-free force spectroscopy IX 156, IX 164 condensation dropwise IX 238, IX 279 condensed phase IX 5 conductive AFM (CAFM) VIII 422, VIII 434, X 63, X 189, X 190, X 205–208, X 211 applications VIII 435 experimental setup VIII 434 conical indenters X 311, X 313–315, X 330, X 333–335 contact ohmic X 78 permanent VIII 154, VIII 156 contact angle IX 338, IX 340, IX 342 hydrophilic IX 340–342 hydrophobic IX 329, IX 333, IX 340–343, IX 347 contact CMP X 389 contact etch X 387 contact image sensor X 406 contact mechanics IX 336 contact mode VIII 291, VIII 432, IX 335, X 78 contact models IX 95 Derjaguin–Muller–Toporov (DMT) IX 29, IX 38 Dugdale IX 37 Dupré’s work of adhesion IX 38 Hertz model IX 33, IX 36 Johnson–Kendall–Roberts (JKR) IX 29, IX 38 Maugis–Dugdale (MD) IX 37, IX 38 Sneddon’s extensions IX 36 contamination IX 245 continuous stiffness measurement X 309 contrast VIII 392 converter current-to-voltage VIII 382 convolutions VIII 354 copper oxide IX 240 correction factor (CF) X 84 correlation coefficient VIII 58 correlation spectrum analyzer VIII 427 resolution VIII 427 correlator VIII 410 corrosion inhibition IX 238 cost of ownership VIII 39 Coulomb IX 285 covalent bond IX 171 Si–C bond IX 172
376 CPD VIII 353 creep compliance X 337, X 340, X 341 critical dimension (CD) modes X 362 critical surface tension IX 241, IX 255, IX 261, IX 277, IX 278 cross-linker IX 170 crystallinity IX 329, IX 338, IX 340, IX 343, IX 347 cross-linking IX 332 recrystallization IX 335 semicrystalline IX 332 CS-AFM VIII 422 cube-corner indenter X 314 current X 91 nanometric localization X 91 current transport X 93 direct tunnel X 93 Fowler–Nordheim X 93 percolation X 93 current-sensing atomic force microscopy (CS-AFM) VIII 422 conductive probes VIII 429 current detection instrumentation VIII 432 experimental setup VIII 429 operating modes VIII 431 cutoff VIII 86, VIII 93, VIII 95, VIII 109, VIII 115 CVD hot filament VIII 144 1D nanostructures IX 311, IX 312 2D SPM Zoom VIII 39 3D imaging IX 25 deep level transient spectroscopy (DLTS) VIII 408 deep trench (DT) X 362 deformations plastic X 79 delocalization IX 368 depletion VIII 401, X 66 Derjaguin–Muller–Toporov (DMT) IX 336 Dexel VIII 49, VIII 51 diamond-like carbon VIII 67 dielectric X 94 dielectric constant nonlinear X 108 diffusion VIII 204–207, X 75 digital light processor mirror device X 408 digital pulsed force mode (DPFM) IX 23, IX 32 dilated scan VIII 47
Subject Index dimensional metrology X 361 AFM X 361 dual beam X 361 optical profiler X 361 scatterometry X 361 SEM X 361 stylus profiler X 361 TEM X 361 X-SEM X 361 dimyristoylphosphatidylcholine (DMPC) IX 72 phase separation IX 73 dip-pen nanolithography (DPN) VIII 191, VIII 193–195, VIII 200, VIII 204, VIII 206, VIII 207, VIII 209 dipalmitoylphosphatidylcholine (DPPC) IX 76 dipole VIII 356 dishing, erosion, and total indicated run-out X 395 disordered systems IX 360, IX 367 dissipation VIII 259 DNA VIII 184, VIII 186, VIII 188, VIII 191, VIII 198, VIII 200–202, VIII 213, VIII 437, IX 64, IX 71, IX 72, IX 76, IX 78, X 287 elastic behaviour IX 72 dopant VIII 403 DOPC IX 209, IX 210, IX 212, IX 216–218 doping VIII 430, IX 8 Doppler velocimetry VIII 303 DP SAM IX 271 DPFM IX 23, IX 39, IX 50 DPN electrochemical (e-DPN) VIII 203 DPPC IX 76, IX 209–220, IX 224 DRAM DT capacitor X 397 drift VIII 299 DT mode VIII 34, VIII 52 dual-indenter method X 330, X 333–336 dynamic force microscopy (DFM) X 237 dynamic force spectroscopy (DFS) IX 89, IX 100 dynamic mechanical analysis (DMA) X 309, X 337 dynamic methods (of calibration) VIII 292, VIII 295 dynamic mode VIII 291 dynamic range X 64 dynamic technique IX 315
Subject Index electrostatic resonant-contact AFM IX 315 mechanical resonant-contact AFM IX 315 effective modulus IX 36 effective radius IX 34 eigenmodes higher-order IX 24 elastic modulus IX 338 elasticity VIII 151, VIII 158, VIII 160, VIII 168, IX 89, IX 94 elastoplastic X 79 electric dipole VIII 89 electric field built-in X 68 electrical activation X 74 electrical noise cross-correlation spectrum VIII 445 electrical noise microscopy (ENM) VIII 422, VIII 443 experimental setup VIII 444 thermal noise VIII 444 electrical transport VIII 424 electrochemical reaction IX 80 electroless plating X 242 electron band IX 6 electron correlation effect IX 7 electron injection IX 80 electron mean free path X 88 electron microscopy IX 177, IX 181, IX 183, IX 184 electrophotographic X 345, X 347 electrophotographic process X 347, X 353, X 355 charging X 348, X 351 development X 349 exposure X 348 fusing X 349, X 350, X 355, X 356 organic photoconductors (OPC) X 348, X 352–355 phase image X 356 photoconductors X 348–351 transfer X 349 electrophotographic systems X 347–349 electrospinning IX 318 electrostatic IX 208, IX 212, IX 216–219, IX 222–225 electrostatic discharge (ESD) VIII 62 electrostatic force microscopy VIII 351 eletrophotographic processes X 349 ellipsometry IX 65
377 elongation IX 323 energy gap IX 6 environmental control X 226 CO2 X 228 ethyl alcohol X 228 EtOH X 228 hydrocarbon liquid X 229 enzyme-linked immunosorbent assay (ELISA) VIII 188–190, VIII 213 epithelial cells IX 224, IX 225, IX 229, X 293, X 301 equipotential VIII 356 erosion VIII 42 erosion algorithm VIII 49 ESD VIII 62 etching X 242, X 243, X 247, X 248 evanescent field IX 353 exciton IX 361, IX 362 exciton excited state IX 364 failure probability X 100 far-field VIII 92 fast dither tube actuation VIII 71 fast scan control VIII 71 F/D curve IX 313 F/D curve nonlinear IX 325 FDTD VIII 116 ferroelectric data storage X 114 FFM analsyis IX 261 FIB VIII 104, VIII 113 FIB-SEM VIII 54 fiber-based optical probes IX 354 fibroblast X 303 fibroblasts living X 300 fibronectin (FN) IX 156 fiducial mark VIII 57 field enhancement VIII 84, VIII 93, VIII 120 filter membranes IX 139 FinFET VIII 32 FinFET gate formation X 383 fingerprinting VIII 54 finite element analysis (FEA) VIII 63, VIII 293, X 309 finite-difference time domain VIII 88 finite-element modeling VIII 223 flash memory X 109 flat-band X 65, X 94 fluid IX 137 fluorescence VIII 78, VIII 88
378 fluorocarbon IX 237, IX 252, IX 257, IX 265, IX 278 focused electron/ion beam induced processing deposit density VIII 282 deposition VIII 251, VIII 253, VIII 280 etching VIII 251, VIII 253 milling VIII 253, VIII 281 principle VIII 251 process control VIII 276 force IX 89, IX 209, IX 210, X 78 electrostatic VIII 380 lateral VIII 309 force curve IX 92, IX 153 force mapping mode X 296 force measurements IX 92 force mode analog pulsed IX 32 force mode pulsed IX 26, IX 47 force modulation mode X 298 force spectroscopy IX 113, IX 116, IX 187, IX 199, IX 200 force standards VIII 292 force volume imaging IX 24 force–distance curve IX 27 force–extension curve IX 153, IX 172 Fowler–Nordheim injection X 96 tunnel X 97 fracture VIII 269 strain VIII 269, VIII 272 stress VIII 269 frequency domain IX 97 frequency modulation (FM) VIII 154, VIII 291 damping VIII 154, VIII 156 frequency shift VIII 154, VIII 155 frequency modulation atomic force microscopy VIII 315, VIII 318, VIII 327 applications VIII 326 biological systems VIII 327 instrumentation VIII 318 liquids VIII 315 magnetic activation VIII 316 nonbiological systems VIII 326 phase detuning VIII 343 theoretical foundations VIII 317 fretting IX 329, IX 332, IX 347 friction IX 80, IX 329, IX 332–337, IX 340, IX 342, IX 343 adhesive IX 332, IX 336, IX 342, IX 344 friction force IX 262, IX 263
Subject Index friction mapping IX 301 friction measurement IX 47 functionalization IX 99 GaNAs IX 368 gate etch X 379 gate resist pattern X 378 gate sidewall spacer X 385 geometrical filter VIII 41 graphite-like layer IX 303 gray-scale VIII 47 green fluorescent protein (GFP) X 303
IX 161,
Hall mobility X 86 scattering factors X 86 halothane IX 211–215, IX 218, IX 219 hardness IX 315, IX 338 height fluctuations IX 136 Hertz IX 94 Hertzian contact model IX 155, X 311, X 312, X 324, X 325, X 328 spherical indenter IX 155 heterodyne VIII 1, VIII 3, VIII 8, VIII 19, VIII 20, VIII 23 pseudo-heterodyne VIII 8, VIII 13 heterostructure Si/SiGe/Si X 76 high-resolution IX 252, IX 255, IX 257, IX 259 higher bias IX 4 higher harmonics IX 24 higher-order SNDM (HO-SNDM) X 105, X 110 highest occupied molecular orbital (HOMO) VIII 370 hollow cantilevers VIII 85, VIII 106, VIII 127 homodyne VIII 1, VIII 3, VIII 5, VIII 7, VIII 8, VIII 12, VIII 13, VIII 20 humidity VIII 193, VIII 204, VIII 206, VIII 209, VIII 210, X 226, X 242 adsorbed water X 226 humidity effect X 226, X 228 hydrodynamic drag IX 98 hydrophilicity X 226, X 228, X 242 hydrophobicity VIII 67, IX 255, IX 261, X 226, X 228 hydroxylated IX 251 hysteresis (piezoelectric) VIII 300
Subject Index ICTS VIII 416 image VIII 174 imaging IX 112, IX 113, IX 119, IX 123 immobilization VIII 184, VIII 189 impedance VIII 386 electrical VIII 425 measurement VIII 425 resolution VIII 426 spectroscopy VIII 425 impingement IX 324 in-line metrology VIII 39 indentation IX 89, IX 95 indentation measurements IX 25, IX 29 integrins IX 157, IX 159 interaction mechanism IX 10 interface states X 66 interfacial shear strength IX 336, IX 342 interfacial surface energy IX 241 interferometric techniques VIII 3, VIII 13 intermittent contact VIII 154, VIII 156, VIII 170, IX 24 intermittent-contact AFM (IC-AFM) X 237 intrinsic noise VIII 263 IQTS VIII 413 ITRS VIII 37 IVPS VIII 46, VIII 47 joint simulators IX 334 JT distortion IX 10 jumping mode VIII 432, IX 26, IX 51 junction X 68 EJ X 68 MJ X 68 Kelvin VIII 388 Kelvin probe force microscopy (KFM) VIII 351, X 231, X 236 Kelvin probe microscopy (KPM) X 147 Langmuir–Blodgett IX 56, IX 60, IX 61, IX 72, IX 73, IX 76, IX 84 amphiphiles IX 60 horizontal deposition IX 61 Langmuir–Schaefer IX 61 vertical deposition IX 61 lateral force microscopy (LFM) X 234 lateral resolution X 67 lateral stiffness VIII 62 layers epithelial IX 129 LER X 397
379 leukocytes IX 144 lift-off X 242 light–matter interaction IX 361 line edge roughness (LER) VIII 55 line width variation (LWV) VIII 55 living cells IX 113, IX 123, X 292, X 302 loading rates IX 101 local chemical reactions IX 79 local material properties IX 26 local mechanical properties IX 23, IX 50 localization IX 368 lock-in amplifier X 65 lock-in detection VIII 426 loop energy hysteretic X 311, X 325, X 326, X 329 loss compliance X 338 loss modulus X 337, X 339 lubrication IX 278 LWR X 397 macromolecule material X 345, X 350 PFA X 351, X 356 photoconductor X 350 polycarbonate (PC) X 350 polyester X 351 polytetrafluoroethylene (PTFE) X 349, X 351 styrene–acrylate X 351 tetrafluoroethylene perfluoroalkoxy vinyl ether copolymer (PFA) X 349 macromolecule materials X 348 macromolecules X 350–352, X 355 magnetic thin-film recording head X 401 manipulation X 304 mass method (of calibration) VIII 296 mathematical morphology VIII 40 dilation VIII 40 erosion VIII 40 Mayer’s law X 312, X 331 mechanical properties IX 311, IX 312 elastic modulus IX 312 elongation IX 311 tensile strength IX 311 toughness IX 311 Young’s modulus IX 311, IX 312 median filter VIII 51 membrane proteins IX 112, IX 124 membranes IX 139, IX 193, IX 195–197, IX 225 cell IX 191–193 model IX 177, IX 194
380 MEMS VIII 64 MEMS/NEMS IX 283 meniscus VIII 204, VIII 206, VIII 207 meniscus bridge IX 302 meniscus force VIII 61 16-mercaptohexadecanoic acid (MHA) VIII 192 messenger RNAs (mRNAs) IX 150 metal X 223 Al X 240 Mo X 243 Nb X 240 Ni X 240 NiFe X 240 Ti X 223, X 240, X 243 metal–insulator–semiconductor nanoMIS X 64 metal-coating VIII 103 metal nanostructures VIII 18 metal trench etch X 390 metal trench photo pattern X 390 metallic nanostructures VIII 4, VIII 19 metrology VIII 31 MHA VIII 194, VIII 199, VIII 200, VIII 206, VIII 207 micro-contact-printed IX 50 microarrays VIII 184, VIII 185, VIII 188– 190 microbead IX 151 carboxylated polystyrene beads IX 153 microcontact printing VIII 185 microelectromechanical system (MEMS) VIII 61 microelectronic device IX 238 microfabricated devices VIII 302 microfluidics VIII 197, VIII 198 microscopy VIII 379, IX 64 scanning near-field optical microscopy (SNOM) IX 65 microspotters VIII 196 microsteps VIII 34 microtribology/nanotribology IX 284 middle line width (MCD) VIII 58 miniaturization IX 237 Minkowski addition/subtraction VIII 50 mobility X 85 drift X 85 Hall effect X 85 modal analysis VIII 258 modelling VIII 403 modification IX 80
Subject Index molecular dynamics (MD) simulations X 149 molecular orientation IX 12, IX 323 molecular packing IX 250 molecular recognition IX 112, IX 116, IX 122 molecular-beam epitaxy X 242 Moore’s law VIII 37 morphology X 184, X 187, X 188, X 190, X 205 MOS X 94 MOSFET X 239 effective channel length X 68 mRNAs IX 164 multi wall nanotube (carbon) (MWCNT) VIII 141 multiple-bond attachments IX 103 N clusters IX 368, IX 370 nanoelectrochemical techniques oxidation IX 79 nanoelectronics IX 62, IX 64 nanofibers polymeric IX 318 nanofibers oriented IX 319 nanofountain probes (NFP) VIII 191, VIII 199, VIII 200, VIII 202, VIII 209, VIII 211, VIII 212 nanografting VIII 202 nanoindentation IX 313, IX 338, X 309 nanolithography IX 56 addition IX 56 constructive IX 56 elimination IX 56 nanopatterning IX 56 soft lithography IX 67 substitution IX 56 nanomanipulation VIII 256 nanomanipulator IX 144, IX 164 nanomechanics VIII 268 nanoMOS X 98 nanoparticle VIII 83 nanopatterning IX 76 nanoscale capacitance microscopy VIII 423 nanoscale impedance microscopy (NIM) VIII 422, VIII 437 bandwidth VIII 440 calibration VIII 443 experimental setup VIII 438 parasitic contribution VIII 437 stray contribution VIII 437
Subject Index nanoscale spectroscopy VIII 24 nanoscope VIII 45 nanoscratch technique X 342 nanosource VIII 83, VIII 84 nanosystem IX 55 nanotechnology IX 55, IX 62, IX 283 nanotemplates IX 84 nanotube VIII 135, VIII 138, VIII 149 mechanical properties VIII 149, VIII 150, VIII 158, VIII 160, VIII 166 nanowear mapping IX 308 near-field VIII 78, VIII 89 resolution VIII 92 reverse tube etching VIII 102 spatial frequency VIII 91 spectrum VIII 91 near-field optical microscopy (SNOM) X 189, X 201–205 near-field scanning optical microscopy (NSOM) IX 351, X 144 Nioprobe VIII 44 NIST-traceable standards VIII 57 nitric acid treatment IX 257, IX 273 nitride X 223, X 244 TiN X 223 ZrN X 223 noise VIII 359, VIII 382, VIII 383, VIII 443 correlated instrument VIII 428 electrical VIII 426 measurement VIII 427 nanoscale VIII 444 Nyquist formula VIII 444 resolution VIII 427 spectroscopy VIII 427 noise spectrum IX 316 nondestructive VIII 56 notching VIII 68 NSOM IX 353, X 144 NSOM aperture IX 353 nuclear magnetic resonance (NMR) IX 177, IX 181, IX 185 OCD VIII 35 ODP SAM IX 270 OP SAM IX 272 optical efficiency VIII 84, VIII 85 throughput VIII 78, VIII 86, VIII 88, VIII 94, VIII 99, VIII 106 optical lever VIII 290 calibration VIII 290 optical microscope IX 182
381 optical proximity correction X 398 optical resolution VIII 78, VIII 102 optical selection rules IX 361 optical waveguides VIII 19 optimal estimator VIII 360 organic thin-film transistors (OTFTs) VIII 368 oscillation amplitude VIII 168 oscillator VIII 380 oscillatory measurements IX 98 oxidation X 218, X 219 anodic X 98 cathodic oxidation X 219, X 222 electrochemical oxidation X 219, X 220 field-enhanced oxidation X 222 low-temperature X 72 UV/ozone X 72 wet X 72 oxide thickness X 70 parallel beam model VIII 293 PBS IX 344 PCR IX 164 PDMS VIII 185, VIII 194 percolation X 96, X 100 perfluoroalkyl chain IX 256 perfluoroalkylphosphonate IX 237 perfluoroalkylsiloxy IX 237 perfluorodecyldimethylsiloxy–copper IX 255 perfluorosiloxy IX 252 permeability IX 127 permittivity VIII 378 persistence length IX 197, IX 198 PFDP SAM IX 244, IX 270 PFM IX 50, IX 51 PFMS SAM IX 252, IX 271 PFMS-based SAM on Cu IX 255 phase measurements X 357 phase transformation IX 303, IX 308 phase transition IX 209–212, IX 218, IX 219 phase-locked loop (PLL) VIII 232 phosphate-buffered saline (PBS) IX 340 phospholipids IX 209, IX 210, IX 217–220 phosphorus IX 250 photolithogaphy X 248 photolithography VIII 186–188 photomask defect review and repair X 400 photomask pattern and etch X 399 physisorbtion IX 240, IX 257
382 piezo actuator VIII 71 piezo stage IX 287, IX 288, IX 292 large amplitude IX 288 ultrasonic linear piezo drive IX 292 piezoresistive cantilever VIII 227 piezoresistive detection VIII 267 piezoresistive sensor VIII 227 pin-on-disk IX 334 pin-on-flat IX 334 pinocytotic organelles IX 140 pipette VIII 198, VIII 199 planarization efficiency (EFF) X 141 point-mass model VIII 259 Poisson X 85 Poisson equation VIII 355, IX 338 Poisson’s ratio VIII 293 polariton resonances VIII 3, VIII 18, VIII 22 polarization VIII 108, VIII 110, VIII 114, VIII 116 poly(dimethylsiloxane) VIII 185 polymer IX 252 polymer nanofibers IX 318 polymer solar cells X 182–184, X 189 polymers binary mixture IX 42 mixing entropy IX 43 poly(methyl methacrylate) (PMMA) IX 29 styrene–butadiene rubber (SBR) IX 29, IX 31, IX 41, IX 42, IX 45, IX 46 surface energy IX 43 process control VIII 256 procollagen X 289 profile carrier X 85 dopant X 85 resistivity X 85 prostaglandin (PG) IX 162 proteins VIII 437, IX 60, IX 75, IX 76, IX 78, IX 177–179, IX 182, IX 187–189, IX 191–193, IX 195, IX 196, IX 329, IX 333, IX 338, IX 340–343, IX 347 bovine serum albumin (BSA) IX 340 height reduction IX 71 plastic deformation IX 75 streptavidin IX 71, IX 72 proteins denatured IX 342 PSF VIII 358 pull-off IX 335
Subject Index quality factor VIII 236 quantum dots IX 351 quantum wells X 76 quercetin-3-O-palmitate (QP) IX 72 fibrelike IX 73 self-organized spiral IX 73 radius VIII 391 RAM capacitor ferroelectric X 398 Raman VIII 78, VIII 119, VIII 121, VIII 128, VIII 129 raster scanning VIII 38 real area of contact IX 336, IX 337 receptor IX 150 reconstruction VIII 359 reduction X 230 cathodic reaction X 230 reentrant image reconstruction VIII 41 reference metrology VIII 39 reference metrology system (RMS) VIII 35 reinforcement IX 321 reliability IX 323 repassivation IX 346 repeatability VIII 47, VIII 299 repulsive regime IX 76 residual stress IX 343, IX 345–347 resist VIII 55, X 244, X 247, X 248 resistance X 78 electrical VIII 424 measurement VIII 424 probe Rtip X 78 spreading X 78 tip–semiconductor nanocontact X 78 resistivity X 81, X 91 bulk X 91 polycrystalline Au films X 91 resolution VIII 386 effective contact radius X 81 energy X 91 spatial X 81, X 91 resonance experiment VIII 274 clamping VIII 276 vibration amplitude VIII 274 resonance frequency VIII 221–223 resonant imaging IX 24 resonator VIII 384 RMS VIII 31, VIII 32 rotational degrees of freedom IX 5 roughness IX 340, IX 343, IX 344, X 396 rupture force IX 100
Subject Index SA IX 84 SAM coatings VIII 67 SAM formation IX 240 sample cross section VIII 60 sample preparation X 69 passivation X 69 polishing X 69 sample properties adhesion force IX 28 deformation ratio IX 45 detachment energy IX 31 elasticity IX 26 electrical double layer IX 47 electrostatic interactions IX 30 energy diffusion IX 46 hysteretic loss IX 30 indentation depth IX 44 local adhesion IX 26, IX 29 local stiffness IX 29 plastic deformation IX 26 Poisson’s ratio IX 33, IX 46 postflow IX 31, IX 32 real surface topography IX 45 relaxation time IX 31 sample compliance IX 29 stiffness IX 30 topography IX 28, IX 30, IX 44 viscoelastic IX 26 Young’s modulus IX 33, IX 46 SAMs comparison between IX 273 satellite IX 242 scan mode CD mode VIII 33 contact mode VIII 33 DT mode VIII 33 noncontact mode VIII 33 tapping mode VIII 33 scanning capacitance VIII 352 scanning capacitance microscopy (SCM) VIII 377, VIII 423, VIII 443, X 63, X 236 scanning capacitance spectroscopy (SCS) X 69 scanning electron microscopy (SEM) X 286 scanning force microscopy (SFM) IX 56, IX 68, IX 73, IX 80 attractive regime IX 56, IX 69, IX 71–73 DSFMA IX 84 dynamic IX 68, IX 69, IX 84 dynamic SFM IX 56 fibres IX 74
383 force modulation IX 76 hard tapping IX 73–75 phase-lag IX 76 Q-control IX 70, IX 76 quality factor IX 69, IX 71 repulsive regime IX 71 tapping mode IX 68–71, IX 73, IX 76, IX 84 scanning Kelvin probe microscopy (SKPM) VIII 388 scanning near-field optical microscope (SNOM) VIII 78, X 294 aperture VIII 78 apertureless VIII 78 illumination mode VIII 82 scanning nonlinear dielectric microscopy (SNDM) X 105 scanning probe lithography (SPL) IX 67, IX 76, IX 79, IX 82, IX 84 addition nanolithography IX 76 constructive nanolithography IX 76 dip-pen nanolithography (DPN) IX 76, IX 77, IX 84 elimination nanolithography IX 76 nanoelectrochemical IX 84 nanopipet IX 76 nanotemplates IX 82 scanning probe microscopy (SPM) IX 56, IX 76, IX 78, IX 80, IX 84, IX 89, IX 334, X 63, X 64, X 257, X 261, X 348, X 360 atomic force microscopy (AFM) X 360 dip-pen nanolithography (DPN) IX 78 elimination nanolithography IX 78, IX 79 elimination substitution IX 79 force modulation microscopy IX 72 scanning capacitance microscopy (SCM) X 64 scanning spreading resistance microscopy (SSRM) X 64 substitution nanolithography IX 78 scanning spreading resistance VIII 352 scanning spreading resistance microscopy (SSRM) VIII 402, X 63 scanning tunnelling microscopy (STM) VIII 366, VIII 377, IX 67, IX 68, IX 78, IX 79, X 88 scattering interface charge X 85 interface roughness X 85 local strain fluctuations X 85
384 scatterometer VIII 35 Schottky barrier X 88 schottky barrier height (SBH) X 88 SCM VIII 377, VIII 443 secretion IX 134 self-assembled IX 84 self-assembled monolayers (SAM) VIII 67, IX 57–59, IX 62, IX 64, IX 65, IX 67, IX 76, IX 78–81, X 225, X 234, X 245, X 247, X 249 1-alkenes IX 57, IX 60 alkylthiols IX 57, IX 59, IX 60 alkyltrichlorosilane IX 60 alkyltrichlorosilanes IX 80 1-alkynes IX 57, IX 60 amino-terminated SAM X 231 aminosilane SAM X 226 chemical maps IX 81 chemisorption IX 57, IX 58, IX 60 coatings IX 62 friction IX 81 multifunctional nanoparticle IX 63 1-octadecene IX 60, IX 80, IX 81 organosilanes IX 57, IX 60, IX 80 physisorption IX 60 sensor IX 62 SH-terminated SAM X 230 1-terminated alkenes IX 80 thiolated IX 62 thiols IX 68, IX 78, IX 79 trichlorosilane IX 78 self-assembly IX 56, IX 79, IX 82 self-organization IX 76, IX 82, IX 84 self-replacement IX 79 SEM VIII 35, VIII 36 semiconductors VIII 401, X 218, X 223, X 236, X 239 a-Si X 244, X 247 GaAs X 218, X 223, X 239, X 240 III–V X 67 InP X 223 n-type X 66 p-type X 66 Si X 218, X 237, X 239 Si–H X 218, X 222, X 241, X 242, X 248 SiC X 67, X 223 SiGe X 67, X 223 silicon X 218 sensitivity VIII 361 separation work IX 155 sessile drop IX 241, IX 252
Subject Index SFA micro/nanotribology IX 284 SFM tip IX 73 shake-up IX 242 shallow trench isolation (STI) VIII 52, X 136, X 370 shape factor X 100 shear force feedback IX 354 shear modulus IX 89, IX 96, IX 199, IX 200 SI measurement system VIII 303 sidewall angle VIII 54 sidewall roughness VIII 53, VIII 55, VIII 69 SiGe source/drain recess X 385 silanization IX 242, IX 257 silanols IX 238, IX 259 silicon nitride VIII 65 silicon probe VIII 69 silicon-on-insulator (SOI) X 239 siloxane IX 252, IX 257, IX 259 siloxy IX 278 siloxy–copper IX 259, IX 261, IX 278 SIMS VIII 403, X 75 single molecules VIII 436, IX 89 single wall nanotube (carbon) VIII 141, VIII 144 single-electron transistor (SET) X 240 single-molecule measurements IX 100, IX 116, IX 120 size dependence IX 312, IX 322 skin depth VIII 103 slider for hard drive X 405 sliding velocity IX 278 SNDM (NC-SNDM) noncontact X 105, X 111 Sneddon’s stiffness X 310, X 311, X 313, X 316 SNOM apertureless VIII 83 SNOM/AFM X 295 SOCS VIII 45, VIII 46 sodium hydroxide treatment IX 277 solid-supported monolayer (SSM) IX 56, IX 57, IX 61, IX 63–65, IX 67–69, IX 76, IX 84 phenomena IX 61 properties IX 61 technological needs IX 61 space charge region (SCR) VIII 355 spatial resolution IX 358 spectroscopy VIII 83, VIII 127, VIII 408, IX 64 IR IX 65, IX 80
Subject Index IR spectroscopy IX 80 lasers IX 65 Raman IX 65 sampling depth IX 64 secondary ion mass spectrometry (SIMS) IX 65 X-ray photoelectron spectroscopy (XPS) IX 65 spherical indenter X 312, X 314, X 324, X 326, X 329, X 335, X 337, X 340 spherical indenters X 310 SPM X 347, X 352, X 353, X 357 force–distance curve X 355 phase imaging X 347, X 356 surface potential microscopy X 352 spreading resistance X 70, X 81 spring constants VIII 222, VIII 257, IX 92, IX 325 calibration VIII 281, VIII 292 SSRM X 78 resolution X 80 stability IX 257 stage drift VIII 70 step height reduction ratio (SHRR) X 141 STI X 142 STI CMP X 375 STI etch X 372 stick–slip IX 285, IX 302 stiffness VIII 150, VIII 166 STM IX 2 STM interpretation IX 3 storage compliance X 338, X 340 storage modulus X 337, X 339 strain distribution VIII 223 strength VIII 270, VIII 272 bending VIII 272 tensile VIII 269 Weibull statistic VIII 268 streptavidin IX 71, IX 72 stress IX 347 compressive IX 344, IX 346 constant voltage X 94 electrical X 93, X 94 ramped voltage X 94 tensile IX 344–346 time X 98 stress concentration VIII 63 stress–strain curve X 309, X 311, X 314, X 318, X 330, X 331, X 333, X 334, X 342 structuring element VIII 50
385 substrates IX 241 superlensing VIII 25, VIII 26 surface (heterogeneous) VIII 174 surface band bending VIII 362, VIII 363, VIII 365, VIII 366 surface charge region VIII 365 surface energy IX 263, IX 338, IX 343 surface local density of states IX 3 surface modification VIII 66 surface patterning tool (SPT) VIII 195, VIII 196, VIII 208 surface photovoltage spectroscopy (SPS) VIII 363, VIII 366 surface plasmons VIII 121, VIII 124, VIII 129 surface potential VIII 355, X 145 surface potential image X 353 surface properties IX 116, IX 117 surface reconstruction IX 14 surface tension IX 323 surfactant IX 207, IX 208, IX 219, IX 220, IX 222, IX 225, IX 229, IX 230 SWR X 397 synovial fluid IX 330, IX 333, IX 340, IX 347 hyaluronic acid IX 333 lubricin IX 333 phospholipids IX 333 serum IX 333 synovia IX 333 system noise VIII 70 tangent plane VIII 49 tangent slope algorithm VIII 49 taper angle VIII 88, VIII 106, VIII 110 tapping X 362 technology node VIII 73 TEM VIII 36, VIII 60 temperature stability VIII 265 tensile experiment VIII 269 tensile test IX 316 thermal fluctuations IX 93 thermal method VIII 292, VIII 297 thermal noise VIII 236 thermoionic emission X 91 thin oxides VIII 435 three-point bending IX 314 threshold voltage X 96 time-lapse AFM IX 189 tip–sample interaction VIII 60 TipCheck VIII 44
386 tips
VIII 135, VIII 148, VIII 391, IX 89, X 72 apex VIII 362 artifacts IX 324 aspect ratio VIII 72 characterizer VIII 42 Co/Cr X 72 conductive diamond X 72 diamond X 80 geometries IX 94 lateral stiffness VIII 72 lifetime X 370 mechanical properties VIII 149, VIII 158, VIII 160, VIII 166 on aperture VIII 125, VIII 126 overhang VIII 44 Pt/Ir X 72 qualification X 370 radius X 67, X 70 tip reconstruction in situ VIII 41 shape characterization X 370 shape profile VIII 65 wear VIII 61, VIII 64, X 370 width VIII 44 tissue section X 292 TJRs IX 334 tool-to-tool matching VIII 37 total joint replacements (TJRs) IX 329, IX 330, IX 332 alumina IX 333 ceramic IX 330 femoral head IX 330, IX 332 hip joint IX 329 metallic alloy IX 330 revision surgeries IX 330 UHMWPE IX 329, IX 338, IX 347 zirconia IX 333 transducer VIII 379, VIII 382 transient VIII 378 transition IX 209, IX 210, IX 214 transmission coefficient IX 357 transmission efficiency IX 355, IX 357 transmission-line VIII 386 trap density critical X 96 traps X 93 trench structure IX 319 Tresca criterion X 310, X 324 tribology IX 332 trident probe VIII 68 tunnel transmission probability X 89
Subject Index tunneling current IX 3 TUT VIII 36 type-B measurement error
VIII 47
UHMWPE IX 332 ultrahigh vacuum (UHV) VIII 361 ultrastructure IX 116 ultrathin film IX 56 uncertainty VIII 56, VIII 60, VIII 70 vacuoles IX 142 vacuum levels X 94 van der Waals IX 208, IX 212–214, IX 225, IX 250, IX 273 variational principle VIII 151 VEH VIII 44 vesicles IX 136 Vickers indenter X 313, X 314, X 320, X 341, X 342 VideoDisc pickup VIII 379 villi IX 130 viscoelasticity IX 89, IX 96, X 296, X 309–311, X 337, X 339, X 341, X 342 vitronectin (VN) IX 159 voltage VIII 198 volume IX 134 von Mises criterion X 310, X 330 W-mapping IX 158, IX 161, IX 162 water balance IX 127 water molecule adsorption VIII 233 wavefunction IX 360 waveguide VIII 85, VIII 89, VIII 103, VIII 116 wear VIII 65, IX 332, IX 334, IX 336, IX 340, IX 343–347 adhesive IX 332, IX 342 depth IX 347 groove depths IX 336 mapping IX 303 trench depth IX 344 wear depth IX 347 wear volume IX 344 wear-mechanism mapping IX 304 wear resistance IX 267 wedge method VIII 311 Weibull distribution X 99 Wiener filter VIII 360 work function VIII 362, X 94 work of adhesion IX 103
Subject Index
387
work-hardening X 329, X 330, X 333, X 335, X 336 X-ray diffraction IX 185
IX 177, IX 180, IX 181,
yield stress X 309–311, X 314, X 318, X 324–330, X 333, X 335, X 336, X 342
Young’s modulus VIII 62, VIII 73, VIII 269, VIII 272–274, VIII 293, IX 89, IX 94, IX 134, IX 198, IX 200, IX 227–229, IX 231, X 296 Zisman method IX 261 Zisman plot IX 241, IX 252