Expanding the Vision of Sensor Materials (1995)

History has shown that advancements in materials science and engineering have been important drivers in the development of sensor technologies. For instance, the temperature sensitivity of electrical resistance in a variety of materials was noted in the early 1800s and was applied by Wilhelm von Siemens in 1860 to develop a temperature sensor based on a copper resistor. The high resonance stability of single-crystal quartz, as well as its piezoelectric properties, have made possible an extraordinarily wide range of high performance, affordable sensors that have played an important role in everyday life and national defense. More recently, a new era in sensor technology was ushered in by the development of large-scale silicon processing, permitting the exploitation of silicon to create new methods for transducing physical phenomena into electrical output that can be readily processed by a computer. Ongoing developments in materials technology will permit better control of material properties and behavior, thereby offering possibilities for new sensors with advanced features, such as greater fidelity, lower cost, and increased reliability.

As noted in the preface, the Committee on New Sensor Technologies: Materials and Applications was asked to identify novel sensor materials that could benefit the manufacture and operation of advanced systems for the Department of Defense and the National Aeronautics and Space Administration and to identify research and development (R&D) efforts that could accelerate the development and incorporation of these emerging sensor materials in particular applications with potentially high payoff. To provide a foundation for its recommendations in these areas, the committee began by assessing the current status of sensor technologies. Early in this assessment, the committee found that applications, not materials, drive new sensor development. Therefore the committee identified a conceptual framework that could relate sensor materials to application needs within which the importance of particular sensor materials could be determined.

Given the extensive body of published work relating to the broad, multidisciplinary subject of sensor technologies, the committee prepared a summary bibliography drawn from the recent literature (Appendix A). The bibliography includes review articles, books, and monographs relating to the wide range of sensor technologies. These references can form a basis from which a more detailed study of any particular sensing technology, principle, or application can be initiated. Several key journals dealing with sensing have been included in the bibliography; they are suggested as starting points for investigating the most recent developments and trends in sensor technologies. Additional information is available from the reference list at the end of each chapter.

Despite the extensive published literature that treat the fundamentals of sensor technology, considerable ambiguity exists in sensor definition and classification, as illustrated by a recent buyer's

guide for sensors in which two lists of sensor suppliers are provided, one based on properties sensed and the other on technologies used (Sensors, 1992). The latter list includes both physical phenomena (for example, acoustic, electrochemical, Hall effect and infrared sensors), and material types (such as bimetallic, fiberoptic, thick-and thin-film, and zirconium oxide sensors).

Understanding the physical or chemical effects that yield useful transduction is important in selecting and designing sensors. However, these effects by themselves are usually not sufficient to establish an unambiguous sensor classification, since typical sensors may use more than one effect. A simple example is a diaphragm pressure gauge. The diaphragm uses one form of mechanical energy to create another (pressure generates displacement and strain); however, the creation of an electrical signal from the displacement or strain can be accomplished using many approaches. The diaphragm could be made of a piezoelectric material, in which the air would induce an electrical charge; an inductive or capacitive effect could be employed to measure the charge related to the strain and the deflection and thereby infer the pressure. Thus understanding all of the possible field effects and features of transducer materials behavior provides the most complete set of sensor design options.

In order to accelerate the incorporation of emerging sensor materials in new applications, it is critically important that the sensor materials community be able to readily identify sensing needs that candidate materials could fulfill.

The formal study of sensor technology is plagued by ambiguity in definitions and terminology. This evolving field of endeavor is extraordinarily broad with nearly every scientific and technical discipline playing an important role. Thus, it should not be surprising that there is no unanimous concept of a sensor. Given the impossibility of presenting a universally accepted definition for sensors, the committee used terms and definitions that are generally accepted in the current technical literature to provide the basis for discussion in this report.¹ (A complete tutorial on sensors and their transduction principles is beyond the scope of the present report.)

The terms "sensor" and "transducer" have often been used as synonyms. The American National Standards Institute (ANSI) standard MC6.1 defines a transducer as "a device which provides a usable output in response to a specific measurand" (Instrument Society of America, 1975). An output is defined as an "electrical quantity," and a measurand is ''a physical quantity, property, or condition which is measured." In 1975, the ANSI standard stated that "transducer" was preferred to "sensor." However, the scientific literature has not generally adopted the ANSI definitions, and thus currently "sensor" is the most commonly used term. Therefore, the term "sensor" will be used throughout this report.

The committee recognizes that, for the purpose of this report, the output of a sensor may be any form of energy. Many early sensors converted (by transduction) a physical measurand to mechanical energy; for example, pneumatic energy was used for fluid controls and mechanical energy for kinematic control. However, the introduction of solid-state electronics created new opportunities for sensor development and control, with the result that sensors today almost exclusively produce an electrical output for use in such applications as computer-based controls, archiving/recording, and visual display. This need for electrical interfacing is causing a broadening in the definition of a sensor to include the systems interface and signal conditioning features that form an integral part of the sensing system. With progress in optical computing and information processing, a new class of sensors, which involve the transduction of energy into an optical form, is likely. Also, sensors based on microelectromechanical systems may enable fluidic elements to operate as controls and actuation devices in the future. Thus the definition of a "sensor" will continue to evolve.

The definition of a sensor does not precisely define what physical elements constitute the sensor. For example, what portion of a thermocouple is the sensor? Is it solely the bimetallic junction? Does it include the wires used for transmission purposes? Does it include any packaging or signal processing? On the basis of information in the

current technical literature, the committee chose to adopt the following definitions:

Sensor element: The fundamental transduction mechanism (e.g., a material) that converts one form of energy into another. Some sensors may incorporate more than one sensor element (e.g., a compound sensor).

Sensor: A sensor element including its physical packaging and external connections (e.g., electrical or optical).

Sensor system: A sensor and its assorted signal processing hardware (analog or digital) with the processing either in or on the same package or discrete from the sensor itself.

In order to describe and characterize the performance of a sensor, a large and specific vocabulary is required. Several excellent references, which provide a basic review of transducer characteristics,

are cited in the bibliography (Appendix A). Table 1-1 lists some of the characteristics important for describing a sensor and its static and dynamic performance. (Most of the characteristics listed under "static" are also important for dynamic measurements.) Sensor characteristics will be discussed in greater detail in Chapter 2 in the context of a set of "descriptors" used by the committee to provide a common framework for sensor technologists and users. Appendix B contains a definition for each of the sensor descriptors used in this report.

Lion (1969) introduced a classification of principles according to the form of energy in which sensor signals were received and generated, which yielded a matrix of effects. Table 1-2 lists the six energy forms or signal domains generally encountered with examples of typical properties that are measured using those energy forms.

Table 1-3 (Göpel et al., 1989), contains the most common transduction principles, excluding biological and nuclear effects, and illustrative physical phenomena. The table demonstrates some interesting complexities in definitions. For example, a device that converts electrical energy into mechanical energy, such as by piezoelectricity (which may be considered a sensor by definition), is more generally termed an output transducer or an actuator rather than a sensor. Clearly then, the appropriate use of "sensor" or "actuator" is not based on physics but instead on the intent of the application.³ Classifying the signal domains in the manner shown in Table 1-3, while not precise, demonstrates that understanding the physics of the application is vital to selecting the appropriate sensor scheme, materials, and design. It is one method of visualizing the transduction principles involved in sensing.

A rigorous attempt at classifying sensors was undertaken by Middlehoek and Noorlag (1982), in which they represented the input and output energy

only as the transduction principle and disregarded any "internal" or compound transduction effects that may have taken place. In addition, they included two other types of sensors: self-generating and modulating. They referred to self-generating and modulating as fundamental transduction principles to be included in a chart such as Table 1-3, thereby creating a third dimension. A sensor based on a modulating principle requires an auxiliary energy source; one based on a self-generating principle does not. No standard convention has been established in the technical literature as to whether a modulating sensor should be classified as "passive" or "active"; both terms are used in the literature. Therefore, the committee adopted the terms "self-generating" and "modulating'' to avoid any confusion that could arise from the use of "passive" and "active."

In order to more clearly depict the sensor taxonomy

FIGURE 1-1  Self-generating sensor.

approach adopted by the committee, Appendix C contains several simple examples that depict the sensor taxonomy. They were drawn from previous National Materials Advisory Board reports on materials processing. The examples in the appendix include thermocouple, transducers, scale of measured properties, and typical constraints.

A comparison of Figures 1-1 and 1-2 illustrates schematically the difference between a self-generating sensor and a modulating sensor. An example of a self-generating sensor is a piezoelectric pressure sensor. In this case, the mechanical energy form (strain or pressure) creates electrical signal (an electrical charge) as a result of the fundamental material behavior of the sensor element. An example

FIGURE 1-2   Modulating sensor.

FIGURE 1-3  Self-generating piezoelectric pressure sensor.

of a modulating sensor is a fiberoptic magnetic-field sensor in which a magnetostrictive jacket is used to convert a magnetic field into an induced strain in the optical fiber. The resulting change in the gauge length of the fiber is measured using interferometry (i.e., the strength of the magnetic field is inferred). Schematic representations of a piezoelectric pressure sensor and a fiberoptic magnetic-field sensor are depicted in Figures 1-3 and 1-4, respectively.

Often sensors incorporate more than one transduction principle; thus, sensors can be conveniently classified simply by their input energy form or signal domain of interest. The committee adopted a sensor taxonomy for this report that is based on the input energy form or measurand as a practical engineering-oriented approach that provides insight into selecting sensors technologies for applications. This approach, however, does not emphasize the underlying mechanisms to the extent that a more science-based taxonomy would; this limitation is particularly telling when multiple response interactions occur. Nor does this approach lead to rapid identification of low-cost sensors, sensors that exploit a particular type of material, etc. Therefore, alternative sensor taxonomies are also useful.

In addition, research efforts should be directed at improving the understanding of multiple physical responses to a sensing phenomena. For instance, it has been shown that reaction of certain gases on a surface can be effectively measured with a novel sensing approach that uses the differential thermal expansion of a bimetallic material and changes in heat capacity and thermal conductivity of the sensor elements (Gimzewski et al., 1994).

Other possible classification methods for sensors include:

• physical or chemical effect/transduction principle;

• measurand (primary input variable);

• material of the sensor element;

• application;

• cost;

• accuracy; and

• output signal domain.

Sensors, in their most general form, are systems possessing a variable number of components. Three basic components have already been identified: a sensor element, sensor packaging and connections, and sensor signal processing hardware. However, there are additional components to certain sensors. The fiberoptic magnetic-field sensor illustrated schematically in Figure 1-4 is an example of a common sensor that uses "compound" sensors to transduce a magnetic field into an electric signal. There are numerous technologies available to convert a magnetic signal into an electrical signal; however, application constraints (cost, environmental effects, packaging, etc.) strongly influence the actual physical design of a sensor and the selection of sensor materials and technologies.

The anatomy of a complete sensor system is shown in Figure 1-5. Technological components in current sensor systems include:

• sensor element(s) and transduction material(s);

• interconnection between sensor elements (electrical and/or mechanical) input "gate";

FIGURE 1-4  Modulating fiberoptic magnetic field sensor.

FIGURE 1-5   Anatomy of a sensor system.

• output "gate" and interconnection;

• packaging;

• modulating input interconnects;

• calibration device;

• calibration input/outputs;

• output signal modifying device (amplifier);

• output signal processing (for smart sensors); and

• actuators for calibration

The scope of hardware elements is indicative of widening definition of a sensor attributable to advances in silicon micromachining, micropackaging, and microelectronics. It is clear from the preceding discussion that modern sensors are much more than a transduction material. Opportunities for introducing new materials in sensors thus arise from three areas: (1) sensor transducer mediums (material); (2) sensor packaging materials; and (3) electronic (signal processing) devices and readouts. This report focuses attention on the sensor transducer medium but recognizes the importance, and in some instances dominance, of materials requirements for the other portions of a sensor system.

Many recent advances in sensors have not come from the synthesis of new transduction materials (except perhaps for chemical sensors) but rather from microelectronic innovations in low-cost, large-scale manufacturing of interconnections, microelectronics, and micromachining that have allowed more complex sensor systems to be formed using well-known sensor elements.

One of the most important advances in sensor technology in the last ten years has been the focused development of smart sensors. The definition of "smart" and "intelligent" sensing can be debated. In general, it is difficult to identify any features in a smart sensor that parallel intelligence in natural systems; however, the terms have become cemented in the technical jargon. The basic tenet of smart sensors is that the sensor complexities must be concealed internally and must be transparent to the host system. Smart sensors are designed to present a simple face to the host structure via a digital interface, such that the complexity is borne by the sensor and not by the central signal processing system. This report does not address specific technologies associated with smart sensing but instead presents the concept and identifies where and why opportunities exist for new sensor materials as well as for the utilization of existing materials that have not traditionally been used for sensing applications.

The basic requirement for a smart sensor is that

it be a system with dedicated "on-chip" signal processing. Realization of this concept simply means that electronic (or optical) signal processing hardware is dedicated to each sensor and miniaturized to the point that it becomes a part of the sensor package. Figure 1-6 provides a schematic representation of a smart sensor that employs "on chip" signal processing within a sensor package. With reference to Figure 1-5, a smart sensor would include the sensor, interface circuit, signal processing, and power source.⁴ The subsystems of a smart sensor include:

• a primary sensing element;

• excitation control;

• amplification (possibly variable gain);

• analog filtering;

• data conversion;

• compensation;

• digital information processing;

• digital communications processing; and

• power supply.

The primary sensor element within a smart sensor may not be made of a conventional transducer material. Nonlinear and hysteretic materials, previously discarded as being too unreliable or unstable for sensing applications, may now be applied in a sensor that contains its own dedicated microprocessor; the need to burden a central processor with a complex constitutive model or filtering algorithm is thereby avoided. Applications can be envisioned that exploit the inherent memory or hysteresis of nonlinear materials to reduce the signal processing workload for example, "record" peak temperature.

The principal catalyst for the growth of smart-sensing technology has been the development of microelectronics at reduced cost. On-chip actuators for self-calibration and mechanical compensation may be created using micromachining techniques or thin-film technologies. Many silicon manufacturing techniques are now being used to make not only sensor elements but also multilayered sensors and sensor arrays that are able to provide internal compensation and increase reliability. It is difficult to predict the future in smart sensing, as the new applications will be driven by the availability of new sensing materials, an improved understanding of the transduction characteristics of "old" materials, and manufacturing techniques for microactuators, microsensors, and microelectronics. It is clear, however, that the smart-sensing concept creates new opportunities for using novel materials for sensors. The smart-sensing concept makes it possible to avoid the constraint of the paradigm

FIGURE 1-6   Schematic representation of a smart sensor.

that sensor elements must be linear and noise-free; however the cost of the added electronics must be considered in the sensor system design analysis.

Potential advantages of the smart-sensor concept include:

• lower maintenance;

• reduced down time;

• higher reliability;

• fault tolerant systems;

• adaptability for self-calibration and compensation;

• lower cost;

• lower weight;

• fewer interconnections between multiple sensors and control systems; and

• less complex system architecture.

These advantages of smart sensors are application specific. There is certainly justification for many applications in distributing the signal processing throughout a large sensor system so that each sensor has its own calibration, fault diagnostics, signal processing, and communication, thereby creating a hierarchical system. Innovations in sensor technology have generally allowed a greater number of sensors to be networked or more-accurate sensors to be developed or on-chip calibration to be included. In general, new technology has contributed to better performance by increasing the efficiency and accuracy of information distribution and reducing overall costs. However, these performance enhancements have been achieved at the expense of increased complexity of individual sensor systems. Currently, the practical utility of smart sensors seems to be limited to applications that require a very large number of sensors.

The field of sensor technology is extremely broad, and its future development will involve the interaction of nearly every scientific and technical discipline. The basic definitions and terminology in this chapter have been presented to establish some consistency in discussions of sensor applications and technologies, since considerable ambiguity exists in sensor definitions and classifications. In the remainder of the present report, a sensor classification system based on the measurand, or primary input variable, is used. The committee acknowledges that alternative systems of sensor taxonomy may be useful in particular circumstances, but for the purposes of the present study, the aforementioned scheme was adopted as the most practical option. In order to accelerate the incorporation of emerging sensor materials in new applications, it is critically important that the sensor materials community be able to readily identify sensing needs and to target those physical phenomena that candidate materials could sense.

The definitions of the terms "sensor," "sensor element," and "sensor system" given above have been adopted by the committee in order to facilitate coherent and consistent analysis of sensor technologies. Many modern "sensors" are in fact sensor systems, incorporating some form of signal processing. Integration of sensor functions into a ''black box" system in which the technical complexity is effectively hidden from the user is a growing trend in sensor development. Of particular interest is the smart sensing concept, which creates new opportunities for using novel materials in sensors, for instance by removing the constraint that sensor elements be linear and noise-free (although the cost-effectiveness of such an approach would depend on the application). Since modern sensors encompass much more than a transduction material, there are many opportunities for introducing new materials in sensor systems, although this report focuses on transducer materials.

Gimzewski, J.K., C. Gerber, and E. Meyer. 1994. Observations of a chemical reaction using a micromechanical sensor. Chemical Physics Letters 217(5/6):589.

Göpel, W., J. Hesse, and J.N. Zemel, eds. 1989. Sensors: A Comprehensive Survey, Vol. 1. New York: VCH.

Instrument Society of America. 1975. Electrical Transducer Nomenclature and Terminology. ANSI Standard MC6.1. Research Triangle Park, North Carolina: Instrument Society of America.

Lion, K.S. 1969. Transducers—problems and prospects. IEEE Transactions on Industrial Electronics 16(1):2–5.

Middlehoek S., and D.J.W. Noorlag. 1982. Three-dimensional representation of input and output transducers. Sensors and Actuators 2(1):29–41.

Sensors. 1992. 1993 Buyer's Guide. Sensors: The Journal of Machine Perception 9(12).