Functional Thin Films and Nanostructures for Sensors Synthesis Physics and Applications

Functional Thin Films and Nanostructures for Sensors

Integrated Analytical Systems Series Editor: Dr. Radislav A. Potyrailo GE Global Research, Niskayuna, NY

For other titles published in this series, go to www.springer.com/series/7427

Anis Zribi

•

Jeffrey Fortin

Editors

Functional Thin Films and Nanostructures for Sensors Synthesis, Physics, and Applications

Editors Anis Zribi United Technologies Corporation Fire and Security Kidde Detection Technology Research Development and Engineering Colorado Springs, CO USA

Jeffrey Fortin GE Global Research Center Micro and Nano Structures Technologies Niskayuna, NY USA

ISBN: 978-0-387-36229-8 e-ISBN: 978-0-387-68609-7 DOI: 10.1007/978-0-387-68609-7 Library of Congress Control Number: 2008944096 © Springer Science+Business Media, LLC 2009 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper springer.com

To our families and parents. Olena, Nadia Michelle, Abi, Libby

Foreword

In recent years, there has been a convergence of fundamental materials science and materials processing methods. This convergence, although highly interdisciplinary in nature, has been brought about by technologies such as bandgap engineering and related techniques that have led to application-specific devices such as lab-on-achip and system-on-a-chip. The demand for reduced device size, device portability, and low power dissipation coupled with high speed of operation continues to dictate terms and conditions for the evolution of nanotechnology. The present trend in approaches to systems manufacturing continues to focus on integration of multifunctionalities on the same chip. These functionalities include, for example, onboard laser sources, sensors, and amplifiers. Both the military and civilian markets continue to drive the research and development component. In recent years, the emergency preparedness guidance systems have added excitement and curiosity to this expanding industry. The outgrowth of technologies of interest for emergency preparedness includes the development of terahertz sources and detectors and systems for detection of explosives and concealed weapons, among others. Sensors made from bulk materials have been around for a long time. Enormous advances in the processing technologies of thin films have led to the ability to manufacture application-specific functional thin films. These include transparent electrodes and antireflection films such as indium tin oxide, which serve as interface components between humans and electronic devices, or optical circuit elements used in optical communication networks, or as contacts and antireflection coatings in solar cells. Products are also being developed with magneto-optical, electrochromic, or UV material for their use as functional thin films in optics. Photonic crystals contain a variety of functional thin films; they require processing of thin films under very stringent control of their structure and properties. For microelectromechanical systems (MEMS), in addition to silicon-based technology, ferroelectric thin films are being used in the fabrication of microactuators and micromotors, capacitors, and other thin-film devices. Functional thin films are being used in the manufacture of devices such as surface acoustic wave (SAW) devices for high-frequency telecommunications filtering, infrared detectors, pressure sensors, accelerometers, force sensors, vibration, thickness, and chemical sensors and biosensors. The reduction in size from bulk to micro- and nanostructured transducers, while promising high sensitivity, high speed, and increased selectivity, vii

viii

Foreword

requires new design considerations that should consider factors such as integration with other devices and device lifetime. Functional thin films offer an enormous infrastructure for a highly interdisciplinary integration of inorganic/semiconducting, organic/bio, and electronic/optoelectronic sensor systems. The field is constantly evolving and will continue to do so by absorbing novel materials approaches such as carbon nanotubes, high Tc superconductors, ferroelectrics, and thermoelectrics. The chapters in this book are designed to give the reader the big picture, from the design phase to the implementation and realization of a transducer. Every effort has been made to include the state-of-the-art in each chapter. The intended audience is scientists, researchers, and engineers, however, graduate students will find the book to be very useful in their research and understanding of sensors and beyond. The editors and contributors are leading researchers in industry and academia in their subject areas. Newark, New Jersey February 2008

N. M. Ravindra

Series Preface

In my career I’ve found that “thinking outside the box” works better if I know what’s “inside the box.”

Dave Grusin, composer and jazz musician Different people think in different time frames: scientists think in decades, engineers think in years, and investors think in quarters.

Stan Williams, Director of Quantum Science Research,Hewlett Packard Laboratories Everything can be made smaller, never mind physics; Everything can be made more efficient, never mind thermodynamics; Everything will be more expensive, never mind common sense.

Tomas Hirschfeld, pioneer of industrial spectroscopy

Integrated Analytical Systems Series Editor: Dr. Radislav A. Potyrailo, GE Global Research, Niskayuna, NY The book series Integrated Analytical Systems offers the most recent advances in all key aspects of development and applications of modern instrumentation for chemical and biological analysis. The key development aspects include: (i) innovations in sample introduction through micro- and nanofluidic designs; (ii) new types and methods of fabrication of physical transducers and ion detectors; (iii) materials for sensors that became available due to the breakthroughs in biology, combinatorial materials science, and nanotechnology; and (iv) innovative data processing and mining methodologies that provide dramatically reduced rates of false alarms. A multidisciplinary effort is required to design and build instruments with previously unavailable capabilities for demanding new applications. Instruments with more sensitivity are required today to analyze ultratrace levels of environmental pollutants, pathogens in water, and low vapor pressure energetic materials in air. ix

x

Series Preface

Sensor systems with faster response times are desired to monitor transient in vivo events and bedside patients. More selective instruments are sought to analyze specific proteins in vitro and analyze ambient urban or battlefield air. For these and many other applications, new analytical instrumentation is urgently needed. This book series is intended to be a primary source of both fundamental and practical information of where analytical instrumentation technologies are now and where they are headed in the future. Looking back over peer-reviewed technical articles from several decades ago, one notices that the overwhelming majority of publications on chemical analysis has been related to chemical and biological sensors and has originated from departments of chemistry in universities and divisions of life sciences of governmental laboratories. Since then, the number of disciplines has dramatically increased because of the ever-expanding needs for miniaturization (e.g., for in vivo cell analysis, embedding into soldier uniforms), lower power consumption (e.g., harvested power), and the ability to operate in complex environments (e.g., whole blood, industrial water, or battlefield air) for more selective, sensitive, and rapid determination of chemical and biological species. Compact analytical systems that have a sensor as one of the system components are becoming more important than individual sensors. Thus, in addition to traditional sensor approaches, a variety of new themes has been introduced to achieve an attractive goal of analyzing chemical and biological species on the micro- and nanoscale.

Preface

Anyone with the most cursory knowledge of sensors must have had a chance to use such devices at some point in their life or career. Whether to collect data in a lab course, to automate an otherwise tedious process, to improve the efficiency of a delicately tuned process, or to do something as mundane as taking a family picture, sensors have become an integral part of our environment and our daily lives. Charge-coupled devices also known as CCD photodetector arrays, for example, have revolutionized photography, astronomy, spectroscopy, X-ray diffraction, and medical imaging to name but a few. A number of scientific discoveries have been enabled by CCDs including the possibility to determine molecular and lattice structures at intermediate stages of a chemical synthesis or a structural transformation. At the core of the widespread adoption of sensors are their rapidly decreasing footprint and cost and increased functionality. The miniaturization of solid-state devices in general and sensors in particular was made possible thanks to significant transformations and a large number of incremental and disruptive inventions in the area of thin-film and nanostructure science and fabrication technologies. Thin films and nanostructures can play multiple roles in a sensor including structural support, reliability enhancement, filtering, and transduction. Thin films and nanostructures are called functional when they fulfill a function other than structural support. These micro- and nanostructured materials have applications that extend far beyond sensing to data storage, lighting, displays, hydrophobic coatings, decoration, and a large number of other fields that are outside the scope of this book. In this book, these materials are discussed in the context of transduction and how they contributed to the current sensor revolution. Sensor design and fabrication are multidisciplinary and require broad and deep knowledge in diverse areas of science and engineering such as materials science, physics, chemistry, biology, and mechanical and electrical engineering. Covering a subject with so many roots in diverse scientific and engineering disciplines is undoubtedly a daunting task and any author who attempts it will do so with significant trepidation. Aware of the challenge at hand, the editors of this book attempted, ambitiously, to cover in one volume an account of general sensor theory, design considerations related to the use of functional thin films and nanostructures, and specific case studies of functional thin films and nanostructure applications in sensing. Part of our motivation in taking on this task is that no such work, to our xi

xii

Preface

knowledge, has been published. Having said this, we are strongly familiar with the large body of publications in this area that we refer to in this book and we are keenly indebted to the works of many authors in putting this book together. This book is devoted to teaching the new sensor designer the key steps involved in developing sound transducer technology from materials selection, to design for performance, to process development, and finally to integration. Throughout the chapters, the authors emphasize and highlight the important role played by functional thin films in solving problems and discuss how to take advantage of such materials to build superior devices. The book is also intended to provide the more experienced designers with a condensed summary of sensor design methodology and excellent references that will prove useful in future sensor design endeavors. To put all of the shared design and fabrication knowledge into perspective and add a touch of reality to the concepts discussed in Chapters 1 through 4, Chapters 5 through 8 are completely dedicated to putting the theory into practice and demonstrating the whole design process using a number of concrete applications. February 2008

Anis Zribi Jeffrey Fortin

Editor Biographies

Anis Zribi is the manager of the Detection Technology Research, Development and Engineering group at Kidde UTC Fire and Security. Prior to joining UTC, he was a senior scientist and a principal investigator at the Global Research Center (GRC) of General Electric where he (leads) led research in the area of Microsystems and microfluidics for chemical and biological detection. He received an M.S.E. in physics from the Polytechnic Institute of Engineering (1996 France), an M.S. in materials physics from Chalmers University of Technology (1998 Sweden) and a Ph.D. in materials science from the State University of New York (2002 NY). Since joining GRC in 2002, Dr. Zribi has contributed to and led several projects including the Nanotechnology Advanced Technology Program, the Photonics Advanced Technology Program, and a number of MEMS sensors and actuators projects. His research interests and activities at GRC include MEMS spectrometers, chemical and biological sensing, magnetic field sensing, medical parameters sensing, fouling detection, and micro- and nanotransducers. Dr. Zribi holds 7 patents and over 40 pending patent applications in MEMS, photonics, and sensors. He authored or co-authored more than 32 articles in peer-reviewed journals and conference proceedings and two book chapters. Jeff Fortin is the manager of the Microsystems and Microfluidics Lab at GE Global Research in Niskayuna, NY. His team’s charter is to develop and deliver innovative micro- and nanosystems and microfluidics via the development and integration of MEMS and NEMS sensing, actuation, and microfluidic technologies, driving miniaturization, increased performance, portability, and low cost. Jeff holds a Ph.D. in engineering science from Rensselaer Polytechnic Institute, an M.S. in physics from RPI, and a B.A. in physics from the University of Southern Maine. He has over ten years of experience in semiconductor technology, MEMS, and microsensors. He joined GE GRC in 2000 and since this time his research has focused on MEMS and microsystems design and fabrication for a variety of microsensor and microactuator applications for GE. He holds ten patents and has co-authored over 12 refereed journal articles in the area of MEMS, thin polymer film development, and chemical vapor deposition as well as eight conference publications. He is also the co-author of a text on chemical vapor deposition polymerization of parylene and is a member of the MRS. xiii

Contents

1

Sensor Design Guidelines ........................................................................ Anis Zribi

1

2

Transduction Principles........................................................................... Jeffrey Fortin

17

3

Growth and Synthesis of Nanostructured Thin Films ......................... Yiping Zhao

31

4

Integrated Micromachining Technologies for Transducer Fabrication ..................................................................... Wei Cheng Tian

65

Applications of Functional Thin Films and Nanostructures in Gas Sensing ....................................................... Audrey Nelson

85

5

6

Chemical Sensors: New Ideas for the Mature Field ............................. Radislav A. Potyrailo

103

7

Applications of Functional Thin Films for Mechanical Sensing ......... Chang Liu

145

8

Sensing Infrared and Terahertz Regions by Functional Films ........... Magnus Willander, Victor Ryzhii, and Qingxiang Zhao

167

Index ................................................................................................................

211

xv

Contributors

Jeffrey Fortin GE Global Research Center Micro and Nano Structures Technologies Chang Liu Northwestern University Department of Mechanical Engineering Audry Nelson GE Sensing, TelAire Radislav Potyrailo GE Global Research Center Vicktor Ryzhii University of Aizu Computer Solid State Physics Laboratory Wei-Cheng Tian GE Global Research Center Micro and Nano Structures Technologies Magnus Willander Linköping University Institute of Science and Technology Qingxiang Zhao Linköping University Institute of Science and Technology Yiping Zhao University of Georgia Department of Physics and Astronomy Anis Zribi United Technologies Corporation Fire and Security, Kidde Detection Technology Research, Development and Engineering Colorado Springs, CO, USA xvii

Chapter 1

Sensor Design Guidelines Anis Zribi

Abstract This chapter focuses on introducing fundamental design principles of transducers, familiarizing readers who are new to this field with the common vocabulary used in describing transducer performance, and providing a succinct historical background about the implementation of thin films and nanostructures in sensors and analytical instruments. A systematic methodology and a sequence of guiding steps to follow in designing a transducer beginning with a concept, through materials selection, and transducer design and fabrication are presented. These steps are covered in more detail in subsequent chapters with concrete examples.

The Big Picture Sensors are ubiquitous in our environment and play essential roles in our everyday life. Our own view of the world is defined by our senses that enable us to perceive stimuli from the environment through a network of biological sensors. Tiny hairs in our inner ears detect the deflection of a membrane as it vibrates in response to acoustic waves and make it possible for us to hear; photoreceptors in our eyes enable us to see objects and discern their colors; chemical receptors on the tongue (known as taste buds) allow us to differentiate between salty, sweet, bitter, and sour. This fascinating network of biological sensors caters to our organs’ needs to control certain biological processes and our needs for security and safety. Driven by the need to better understand our world, to increase the productivity of industrial processes and machines, and to improve our quality of life, scientists and engineers constantly seek to develop the necessary measurement tools. These tools are sensors and instruments that are often inspired by biological sensors and their functioning principles. Such devices have become essential for advancing A. Zribi United Technologies Corporation Fire and Security, Kidde Detection Technology Research, Development and Engineering 4820 Centennial Blvd, Suite 145, Colorado Springs 80919, CO, USA e-mail: [email protected] A. Zribi and J. Fortin (eds.), Functional Thin Films and Nanostructures for Sensors, Integrated Analytical Systems, DOI: 10.1007/978-0-387-68609-7_1, © Springer Science + Business Media, LLC 2009

1

2

A. Zribi

metrology and science, optimizing and controlling processes, providing security against known and unknown threats, law enforcement, and health monitoring. The miniaturization of sensors and analytical instruments is a continuing trend that finds its roots in nature and its beginnings in miniaturizing mechanical, optical, and recently electronic devices. The driving forces for sensor miniaturization are numerous and keep increasing as we encounter and develop new applications for devices that have become virtually invisible and intangible. Sensor cost and portability are major incentives for miniaturization but additional reasons include faster response, integration of multiple functions, reduction of occupied space, and lower device-to-device variability. Electronic miniaturization, starting with the invention of the transistor in 1947, played a key role in developing the fabrication processes that stand at the heart of the new leap in sensor miniaturization. In fact tremendous progress has been made in the past 20 years in high vacuum technologies, ultra purification processes for raw materials, self-assembly, and material deposition and etching processes with various degrees of selectivity, high precision, and small feature-patterning techniques. This progress produced dramatic advances and control over device quality and induced a technological evolution from the bulk crystalline age to the age of thin films, thin film multilayers, and nanostructured materials. In this new age, the properties and performance of submicron devices are dominated by surface and interfacial phenomena. This paradigm shift produced materials with novel or enhanced transduction properties known as functional thin films and functional nanostructures. Thin films and nanostructures play an increasingly important role in state-of-the-art sensors and actuator technologies both as transducers (functional materials) and structural materials. microelectro mechanical (MEM) systems provide a good example of the growing use of materials confined to submicron dimensions to fulfill numerous and versatile functions in advanced devices. Doped silicon/polysilicon thin films, for example, have been implemented as strain gauges in MEM pressure sensor devices (Petersen 1982), temperature-sensing elements in micromachined thermopiles (Petersen 1982), ultrasound transmitters in capacitive ultrasound micromachined transceivers (cMUT; Jin et al. 1998), active alignment actuators in high-accuracy fiberoptic aligners (Petersen 1982), and the list goes on. The growing interest in functional thin films and nanostructures is not only driven by device miniaturization but also by the novel and unique set of physical and chemical properties that materials confined to submicron dimensions exhibit. In his famous talk before the audience of the 1959 annual meeting of the American Physical Society, Richard Feynman predicted many of the now proven advantages of “manipulating and controlling things on a small scale” (Feynman 1959). Since Feynman’s talk, the advantages of scaling devices and materials to submicronic dimensions have been proven to go beyond a dramatic increase in data storage density, faster and more intelligent computing, faster heat removal, and higher natural resonance frequency. Today, functional thin films and nanostructures are being used and are under development for use in a wide range of devices, sensors, and actuators. Such devices include: accelerometers (air bag devices), force sensors, shutters, optical switches,

1

Sensor Design Guidelines

3

optical computation, micromotors, chemical and biological sensors; MEMS devices such as piezomicroactuators and sensors; very small scale microreaction vessels for chemical and biological (lab-on-a-chip) sensing; substrates for plasmon resonancebased signal enhancement and amplification, electro-optical devices for thermal imaging based on the pyroelectric effect, display technologies; pressure mats (piezoelectric thin film); surface acoustic wave (SAW) devices for telecommunications, chemical and biological detection, and more. These new applications for functional thin films and nanostructures are enabled by newly discovered physical and chemical phenomena that underlie the transduction mechanisms in these materials. In fact, size confinement and the associated symmetry breaking, interfacial interactions, high surface-to-volume ratio, structural disorder, and induced entropy have been extensively covered in recent publications. The fundamentals of these effects and their applications in transduction are the topic of numerous current investigations. Quantum dots, for example, have been demonstrated to exhibit a lateral quantum confinement that enables direct coupling of normally incident light with the intraband electronic excitations (Kuo et al. 2001). The geometry of these nanostructures (height and radius) can be tuned to ensure that the lowest energy of electronic transition from ground state to first excited state falls within a specific optical spectrum band. It has also been demonstrated that the confinement significantly reduces the electronic tunneling rate of these nanostructures. These properties indicate the potential of quantum dots to be used as low dark current and hence high signal-tonoise ratio photodetectors especially in the near infrared to infrared part of the spectrum where current detectors are prone to thermally induced electronic tunneling. This and other transduction benefits that emanate from the physical confinement of material structures are discussed in later chapters of this book.

Sensor Architecture The basic function of a sensor is to selectively identify and measure a physical, chemical, or biological parameter such as pressure, light intensity, gas concentration, or the presence and concentration of a biological analyte. The typical architecture of a sensor encompasses a transducer or multiple transducers (operating in series or in parallel) directly exposed to the measurand, acquisition and conditioning electronics, a power source, a processor, a storage medium, and a display. The transducer plays a central role in the operation of the sensor: it is essentially an energy converter where the input energy (mechanical, optical, chemical, biological, electrical) is converted into an electrical signal most of the time. The electrical signal is then acquired by the electronics, conditioned, and noise filtered out before processing (e.g., interpolation using a calibration curve) and finally the data (typically the magnitude of the measurand) are either stored in memory and displayed or routed for action (e.g., alarm) or simply displayed. Fig. 1.1 shows a general block diagram that highlights the main subsystems of a sensor system and the flow of data among the various blocks. The focus of this

4

A. Zribi

Measurand M Sensing Material

X

Transducer

Y

Y2

Y1 Amplifier

Filter

A/D

Numerical Value N

Power Block

Processor

Storage

Display

Fig. 1.1 Sensor architecture block diagram

book is on the transducer block and the role that functional thin films and nanostructures are increasingly playing in high-performance transduction. Depending on the type of transducer, two basic types of sensors can be distinguished: quantitative (analog or digital) and threshold or binary. The two are quite different in function and in application. A quantitative sensor produces an output value that is a direct and continuous function of a measurand value. For example, a thermocouple might have a potential differential of 10 mV at room temperature and a potential differential of 20 mV 10° above room temperature. Any differential potential value between these two is possible depending on the particular temperature to which the sensor is exposed. Threshold sensors, on the other hand, have only two states, often called “on” and “off”. Perhaps the most familiar example of a threshold sensor is a smoke detector which is triggered to the on position if a fire erupts and the signal is used to trip an alarm.

Sensor Figures of Merit/Performance Attributes The wealth of sensor and transducer technologies available to measure the same measurands (Mi) increases the challenge of evaluating and comparing the performance of sensor devices. Therefore, it is critical to define a set of performance criteria that the designers can use to develop sensors that meet customer specifications and the user can use to appraise and contrast the various options. Table 1.1 summarizes the list of performance attributes that are most commonly used to assess sensor technologies and a more detailed description of these parameters is provided later in this chapter.

1

Sensor Design Guidelines

5

Table 1.1 Sensor Performance Attributes Attribute Denotation Input dynamic range Response curve Sensitivity Response time Resolution Accuracy Precision Hysteresis and drift Selectivity

DM N = f(M) S t n

H Se

N Mimin

Mimax

ng

asi

re Inc

Mi

Mij

δMi

δN

sing M i

Decrea

Mi Input Dynamic Range

Fig. 1.2 Typical response curve of a sensor/transducer to a measurand Mi

Input Dynamic Range The dynamic range of a sensor is the span of measurands that constitute the overall operating domain for the device. Within this interval, the sensor is supposed to maintain its properties and reliability characteristics.

Response Curve Every quantitative sensor is characterized by a response curve that represents the output of the sensor N versus the measurand M applied to its input. The transducer response can be linear or nonlinear as shown in Fig. 1.2, but in most cases the sensor electronics are designed to linearize the response curve of the overall sensor in

6

A. Zribi

order to simplify the calibration procedures. Also, it is worth noting that customarily the response of the sensor is normalized with respect to reference values (e.g., resistance at room temperature, resonance frequency of the membrane under known conditions). Considering all the possible measurands (Mi) and key noise parameters that can affect the sensor output, the sensor response N can be expressed as N = ∑ ΔMi i

∂N ∂M i

(1.1)

given that the different measurands affect the output independently of each other.

Sensitivity Although seemingly obvious, the definition of sensitivity can be confusing at times. Looking at the block diagram of the sensor in the first section, each one of the building blocks of the sensor introduces a term in the overall sensitivity of the sensor. These are called, following the diagram from input to output: • Transducer sensitivity • Amplification • Analog filter sensitivity The sensitivity of transducer j to measurand Mi is defined by i Stransducer j =

∂N j ∂M i

(1.2)

,

where Mi are all the possible measurands and key noise parameters that can affect the transducer sensitivity such as temperature and pressure. Nj is a quantity resulting from the energy transformation performed by transducer j. Transducers in a detector can operate in parallel or in series. Typically, transducers that operate in parallel are used for imaging applications, which we are not concerned with within this manuscript. Assuming the p transducers that make up the transducer block operate in series, the total transduction block sensitivity to measurand Mi can be expressed by i Stransducer

∂N p ∂M i

=

∂N p ∂N p − 2 ∂ M p −1 ∂ M p −1

...

p ∂N1 = Π S kj . ∂M i k = 1

(1.3)

The amplification A is an intrinsic property of the amplifier and the analog filter sensitivity is defined in a similar manner to the transducer block by Equation (1.3): Sfilter =

∂Y2 . ∂Y1

(1.4)

1

Sensor Design Guidelines

7

The overall sensitivity of the device to measurand Mi is defined by the following equations, ∂Y2 ∂M i

(1.5)

∂Y2 ∂Y1 ∂N . . ∂Y1 ∂N ∂Mi

(1.6)

Si total = Si total =

Si total = Sfilter * A * Stransducer .

(1.7)

Response Time The response time of the device is a cumulative quantity incorporating the response times of the respective building blocks. It is defined as the transient response of the sensor when it experiences a step change in the measurand. If ti are the response times of the various blocks, then the total response time of the sensor is: t total = t transducer + t filter + t A / D

(1.8)

t total = t transducer + t electronics .

(1.9)

In most diffusion-based chemical and biological sensors, tchemical transducer is the dominant response time in Equations (1.7) and (1.8). Therefore, we typically can ignore the contributions from electronics and physical transducers. However, this may not be the case in thin-film or nanostructure-enabled sensors as the response time of the material is considerably reduced because of the designed nanomorphology of the material. This simplification is also not an option for physical sensors where the response time of the transducer is on the same order of magnitude as the other sensor subsystems. The response time of the whole device is usually estimated as the time required for a transient output signal to reach a fraction (e.g., 70%) of its steady-state change.

Resolution The sensor resolution is the smallest change in input that leads to a detectable change in output. The most general definition is given by Equation (1.10): Re solution =

N N = noi Lim M i → noise Si total Si tot

(1.10)

8

A. Zribi

Clearly for nonlinear sensors, the resolution is dependent on the operating point and can vary if operating conditions vary. Maximum resolution is attained at maximum sensitivity and minimum noise.

Accuracy The accuracy of a transducer is the maximum deviation of its output from the value of the unknown measurand as determined by a gold standard technique. It is a quantitative indication of the degree of conformity of the sensor to a standard. Accuracy represents a systematic error or a bias in the sensor and measuring it and correcting for it are difficult. Transducer calibration is required to account for this type of error and partially correct it. Long-term changes in the performance of the transducer because of materials’ aging and wear will affect its accuracy.

Precision The second type of measurement uncertainty associated with a transducer is random errors or what is often referred to as precision. Precision is often mistakenly confounded with accuracy. It quantifies statistical fluctuations in the measurement and is attributed to variability in the measurement conditions and the limitations of the selectivity of the sensor towards the measurand of interest. Precision represents the repeatability of the measurement given the same sample.

Hysteresis and Drift Hysteresis characterizes the lagging of the sensor response behind the variation of the measurand. This can be attributed to the sensing material memory and/or to the transducer properties. As a performance specification, hysteresis is defined as the maximum difference between the upscale and downscale readings on the same artifact during a full-range traverse in each direction. It is often reported as the ratio (or percentage) of the difference between the upscale and downscale readings to the full scale. Drift can be defined as the slow unpredictable change of the sensor output at constant input. Drift can affect both the signal and noise levels and it can emanate, for example, from residual stress relaxation, residual diffusion, material aging, and degradation. Drift in sensors is defined for a specific time interval of interest.

Selectivity It is typical of a sensor designed to detect variations of a given measurand M to exhibit secondary internal sensitivities to other measurands (Mi) or noise factors of

1

Sensor Design Guidelines

9

different natures, physical or chemical, as well as biological. In general, selectivity may be quantified by p

Se = ∏ j =1 j ≠1

Stjransducer , St1ransducer

(1.11)

where Sj is the sensitivity of transducer j to measurand Mi. In many cases, the sensitivities to the various measurands are interdependent therefore second-order and cross-terms should be included in the expression of the selectivity.

Sensor Design Considerations Sensor design is one of the most interdisciplinary technical areas requiring both breadth and depth of knowledge in materials science, physics, chemistry, biology, mathematics and statistics, electronics, and packaging. The successful design of a sensing system requires very good communication between scientists and engineers of different backgrounds. Much as in any design activity, it is very difficult and undesirable to bind creativity by a set of design rules or a design methodology. However, past a first stage, focused on brainstorming, transduction mechanism and materials down-selection and feasibility analysis, a successful design team must have a clear objective and a guiding design methodology to steer their effort. The typical sensor design steps include: • • • • • • • • • • •

Collection and analyses of the device specifications Transducer selection/invention Materials selection/invention Sensor design and modeling Prototyping Measurement of materials properties Prototype testing and model calibration Design iteration(s) Final device fabrication Technology transition to manufacturing Fabrication process scaleup

The most challenging and least regulated steps are the first three and they are least covered by the literature. In the next sections, we provide some guidance, concepts, and ideas to help the new sensor designer make faster progress towards his or her ultimate goal.

Selection and/or Invention of the Transduction Mechanism Assuming that the design team is armed with a clear set of specifications for the desired sensor performance and the operating and storage conditions, the next steps

10

A. Zribi

ought to target the selection or invention of the transducer(s) that can convert the measurand into an electrical/optical signal. The questions that need to be answered by the team at this stage are: 1. What specific physical, chemical, and biological properties does the measurand possess that differentiate it from other potential sensor inputs? 2. For each of the identified characteristics, what are the potential confounding inputs, which in the future will be identified as noise sources? 3. What type of energy conversion is required to transform the input signal (mechanical, chemical, biological, electromagnetic) to a device output signal (optical or electrical)? 4. What are the candidate transduction schemes that can be used for the target measurand? The final transduction scheme may involve one or multiple transducers operating in series or parallel (array of detectors). Considering the following criteria can further refine the list of candidate transduction mechanisms. 1. 2. 3. 4.

Simplicity of the transducer and robustness to failure Materials requirements dictated by operating environment Development time Cost

Selection of Transducer Material The transducer material properties are important from performance and sensor reliability perspectives. The suite of functional materials available to sensor and micro-/nanoinstrument designers is rapidly expanding and numerous techniques have been developed to integrate a large variety of organic and inorganic materials and their alloys. The wealth of options puts the designer face to face with the challenge of selecting the best-suited materials given a transducer design concept. Numerous considerations come into play when selecting a transducer material and they fall into three categories: performance, process compatibility, and reliability. Although it is very difficult to discuss the performance selection criteria of a transducer in general terms, process compatibility and reliability criteria are common to all transducers and they are of chemical (chemical resistance, photo definability, adhesion) and physical nature (rheology, mechanical, dielectric). Depending on the transducer length scale, whether it falls in the bulk (>100 mm), micro (> wl) for adjacent sectors. Because wh is much higher than wl, within the sector with rotational speed wh, the sample will receive a smaller amount of vapor per radian than that with wl. Thus, the shape of the nanorods can be tuned by changing the ratio of wh/wl as well as the phase sector N. Fig. 3.32 shows a program sequence for a twofold symmetry rotation GLAD. Different slopes in different phase sectors indicate the difference in rotation speeds. Fig. 3.33 shows the resulting Si nanorod structures with two-, three-, and fourfold symmetry rotation (Zhao et al. 2002b). The examples show that different symmetrical rotations alter the lateral arrangement of the nanorod arrays while also changing the shape or aggregation of the nanorods. Due to rotational symmetry, all the nanorods are aligned with each other and are perpendicular to the substrate surface.

56

Y. Zhao

Fig. 3.32 The program sequence for twofold symmetry rotation

Fig. 3.33 The SEM top view images of different GLAD samples with different rotational symmetries. The upper row lists films deposited on bare Si(100) substrates, and the lower row lists films deposited on the colloid substrates which have point defects (Zhao et al. 2002b)

Phase Modulation: Controlling the Orientation of the Nanorods If the phase sectors are not divided symmetrically, different sectors will receive different amounts of vapor; one should expect that the tilting angle of the nanorods can be modulated (Ye et al. 2002). One simple case is to divide one revolution into two different phase sectors as shown in Fig. 3.34. In one sector, the substrate rotates

Growth and Synthesis of Nanostructured Thin Films

Azimuthal Rotation Angle

3

57

2π

π

φ 0

0

T/2 Program Time

T

Fig. 3.34 The program sequence for asymmetry rotation

at speed wl for an angular range of f, and in the remaining sector, the substrate rotates at speed wh for an angular range of 2p – f . Because the asymmetry in the f direction is intentionally introduced, there will be a net F|| after a complete revolution, and there will be an effective vapor incident angle a different from q. The apparent lateral flux F¢|| will be changed according to the sizes of sectors and the ratio of the rotational speeds: −f / 2 ⎡ f / 2 F|| cos ϕ F|| cos ϕ ⎤ 2 F|| ⎛ 1 1⎞ f F¢ || = ⎢ ∫ dϕ + ∫ dϕ ⎥ / t = − ⎟ sin . ⎜ wh t ⎝ wl wh ⎠ 2 ⎢⎣ −f / 2 w l ⎥⎦ f /2

(3.1)

The total time t for a complete revolution is: t=

f 2p − f + . wl wh

(3.2)

Therefore the effective flux incident angle a can be expressed by

tan a =

F¢ || F⊥

=

⎛ 1 1⎞ f 2 F sin q ⎜ − ⎟ sin 2 ⎝ wl wh ⎠ ⎛ f 2p − f ⎞ F cos q ⎜ + w h ⎟⎠ ⎝ wl

= 2 tan q sin

f 2

⎛ wh ⎞ ⎜⎝ w − 1⎟⎠ l ⎛w ⎞ 2p + ⎜ h − 1⎟ f ⎝ wl ⎠

.

(3.3)

Because the experimental incident angle q is fixed, the column density should be constant whereas the effective incident angle a is changed by changing f and wh/wl. We expect the tilt angle b will change continuously; the continuous tuning of

58

Y. Zhao

the tilt angle has been demonstrated experimentally (Ye et al. 2002). A special case is if the substrate is rotated uniformly, wh/wl = 1; then tan a= 0. The deposited columns have no preferential deposition angle, and the columns stand up along the normal of a substrate.

Formation of Regular Array of Nanorods by Templates Almost all the nanocolumnar films shown above are deposited onto bare flat Si substrates. In these cases, the random nucleation is the deterministic mechanism for the lateral distribution of the nanocolumns. However, if we can use one of the major growth mechanisms of GLAD, the shadowing effect, properly, we can grow a high aspect ratio regular array of nanocolumns. To achieve this goal, we can use a substrate with a proper two-dimensional nanotemplate as shown in Fig. 3.35a. The features on the template will act as shadowing centers, and the deposition particles will only accumulate onto the shadowing centers under the proper geometric deposition conditions especially when the vapor incident angle q is larger than the critical angle qc. The critical angle qc, as shown in Fig. 3.35a, is determined by the geometry of the template, tan q c = L / h . Fig. 3.35b also shows an example of a regular helical Si spring grown onto regular W plugs fabricated by photolithography. Clearly, the use of a template provides a lateral control of the nanocolumnar growth.

Control of NSTF Film Porosity All the above fabrication methods concentrate on a single degree of freedom for the substrate manipulation which is azimuthal rotation. In fact, there are two other rotational degrees of freedom one can utilize: the polar rotation to change the particle incident angle q, and the nutation to rotate the substrate to face a different

(a) D

L h

q>qc

q