Submitted to the Enclyclopedia of Smart Materials, ed. J.A. Harvey, John Wily & Sones, NY, due in Oct. 2000.

TABLE OF CONTENTS

Abstract Introduction Realty that defies non-contact ultrasound Pursuit of Non-contact ultrasonic transducers Piezoelectric transducer for unlimited Non-Contact ultrasonic testing mode Transducer Characterization Scheme Characteristics of Non-Contact Transducers Sensitivity Comparison of Non-Contact and Conventional Contact Transducers Non-Contact ultrasonic analyser Thickness and Velocity Measurements Relative Amplitude Attenuation Measurement Non-Contact Ultrasonic Spectroscopy Applications of Non-Contact ultrasound Materials Characterization and Defect Detection Food, Beverage, and Pharmaceutical Medical High precision level sensing Conclusions Acknowledgements References

Abstract

Nondestructive and Non-invasive ultrasound is commonly used to test materials for defects and properties and to examine human body for diagnostic evaluations. Though the value of ultrasound in industry and medical fields is obvious, its modus operandi is severely limited by the physical coupling of ultrasonic transducers to the test medium by liquids gels, grease, glycerin, etc. For nearly 20 years several attempts have been made to change this mode of coupling to non-contact mode. A variety of ultrasound generation mechanisms have been developed by utilizing piezoelectric, capacitive, laser- based, and electromagnetic phenomena. In this article we describe the development of phenomenally high air/gas transduction piezoelectric transducers in the 100kHz to 5MHz frequency range. We also introduce a dedicated ultrasonic non-contact analyzer system which is characterized by >140dB dynamic range and ±1ns time-of-flight accuracy. Based on the combination of these new transducers and the analytical system, a number of industrial and medical applications in time, frequency, and image domains are also described.

Introduction

Ultrasound is widely used in health care for non-invasive diagnostics and in industry for non-destructive testing. It generates visual images from inside the test medium: the fetus, malignant tissue, stones, etc. in the human body. In the industrial applications, besides defect detection, ultrasound can also be used to determine significant materials characteristics such as density, thickness, mechanical properties, level sensing, and more. Knowledge of ultrasonically analyzed information is important for human health as well as for cost-effective production of quality industrial materials.

Ultrasound operates on the same principle as the other characterization methods also based upon wave- material interaction phenomena. These are: Optical, X-ray, infra-red, Raman spectroscopy, nuclear magnetic resonance, neutron, g-ray, mass spectrometry, etc. By propagating a wave in a given medium, useful information about the medium can be generated by analyzing the transmitted or reflected signals. Ultrasound differs from other wave-based methods because it does not require sample preparation; is non-hazardous; provides the means to determine mechanical properties, microstructure, imaging, & microscopy; is portable; and is cost-effective. Furthermore, ultrasound is applicable to all states of matter, with the exception of plasma and vacuum. Propagation of ultrasound in a material is not affected by its transparency or opacity. Table I provides a comprehensive introduction to ultrasound measurements and to the information revealed either directly or through correlation with measurements.

During the last twenty years, both ultrasound and its applications have grown phenomenally. Uses in industry have gone beyond overt defect detection in metals to include: characterization of elastic and mechanical properties; delaminations in multi-layered materials; proximity and dimensional analysis;
measurements of anisotropy and heterogeneity; surface profiling, chemical corrosion, crystallization and polymerization; liquid and gas flow metering; imaging of surface and internal features of materials; viscosity of liquids; texture and microstructure of granular and cellular materials; applied and residual stresses; high temperature, pressure, and radiation environment applications; robotics, artificial intelligence, and so on. These highly desirable applications of ultrasound have attracted the attention of a wide range of industries such as structural and electronic materials and components manufacturers, aircraft and aerospace, chemical and petroleum, plastics and composites, lumber and construction, highways and aircraft landing strips, bridges and railroads, rubber and tire, food, pharmaceutical, and on and on!

MEASUREMENT CATEGORY	MEASURED PARAMETERS	APPLICATIONS
Time Domain	Times-of-Flight and Velocities of Longitudinal, Shear, and Surface Waves	Density, Thickness, Defect Detection, Elastic and Mechanical Properties, Interface Analysis, Anisotropy, Proximity & Dimensional Analysis, Robotics, Remote Sensing, etc.
Attenuation Domain	Fluctuations in Reflected and Transmitted Signals at a Given Frequency and Beam Size	Defect Characterization, Surface and Internal Microstructure, Interface Analysis, etc.
Frequency Domain	Frequency-Dependence of Ultrasound Attenuation, or Ultrasonic Spectroscopy	Microstructure, Grain Size, Grain Boundary Relationships, Porosity, Surface Characterization, Phase Analysis, etc.
Image Domain	Time-of-Flight, Velocity, and Attenuation Mapping as Functions of Discrete Point Analysis by Raster C-Scanning or by Synthetic Aperture Techniques	Surface and Internal Imaging of Defects, Microstructure, Density, Velocity, Mechanical Properties, True 2-D and 3-D Imaging.
Table 1: Categories of Ultrasonic Measurements and their Applications.

In medical diagnostics, where sophistication of ultrasound has been (and is being) developed to extreme perfection, ultrasound can replace hazardous x-rays in many instances. For example, ultrasound is useful for visualizing the fetus, measurement of the cornea, tissue characterization, general body scanning, imaging of plaque in arteries, gum disease, brain wave measurements, monitoring of the heart beat, skin and breast cancer detection, general body scanning, blood flow metering, etc. And here, too, the list goes on!

Twenty years ago we were content if ultrasound could detect a 1mm defect and 0.5mm resolution in a given test material. Today, thanks to short pulse and high frequency transducers with advanced electronics and signal processing, we are in the micrometer range when it comes to detectability and resolution requirements. Obviously, ultrasound has come a long way since the discovery of piezoelectricity by Jacques Curie in 1876 and its first application by Richardson in 1913 for a sonar.

Nonetheless, this impressive record of ultrasound is marred by the way ultrasound is usually propagated in the test medium. Due to extremely high attenuation of ultrasound by air, its transmission in a test medium is done by physically contacting the transducer to the test medium. Therefore, all conventional ultrasound is based upon the severe limitation of a physical contact between the transducer and the test medium by a liquid gel (1). If this contact could be eliminated, then we could do applications such as diagnosing burnt or malignant skin damage without discomfort to the patient. Similarly, a number of industrial materials sensitive to liquid contact could be tested for the measurements of thickness, density, mechanical properties, defect detection, etc. This is significant in assuring materials quality and process control, and for cost-effective production. Development of Non-Contact Ultrasound (NCU) mode would allow many more useful applications of ultrasound. For example, with NCU we could characterize materials that are porous or hygroscopic materials in the early stages of manufacturing (uncured plastics, green ceramics and powder metals), and those that are continuously rolled on a production line (such as plastics, rubber, paper, construction & lumber, etc.) NCU can also be applied to medical problems where contact with a patient can be harmful, such as for evaluation of wounds and diagnosis of the eyes. During the last few years the major breakthroughs which have occurred now present the possibility of NCU in frequencies familiar to conventional contact-based ultrasonics, i.e., in the MHz range.

For NCU to become a reality, we first need the transducers and electronic systems sensitive enough to transmit and detect ultrasound without using any contact. And herein lies a big problem. Conventional wisdom stipulates that ultrasound (from ~200kHz to >5MHz) cannot be propagated through solids or liquids without a physical contact between the transducer and the test medium. Therefore, NCU has been considered an impossible dream. This is understandably due to the phenomenal acoustic impedance mismatch between the coupling air and the test media. This mismatch can run as high as six orders of magnitude when we consider propagation of ultrasound from air to materials such as steels, super-hard alloys and dense ceramics, cermets, diamond, and diamond-like materials. To realize the NCU mode, this acoustic impedance barrier must be broken. And for this to happen, it is imperative that ultrasonic transducers be characterized by phenomenally high sensitivity. Achieving of NCU is analogous to ``throwing a helium-filled rubber balloon so that it can pierce through a stainless steel wall!''

Realty that defies non-contact ultrasound

The exorbitant mismatch between acoustic impedance of the coupling medium air and that of the test material generates enormous resistance for ultrasound propagation in materials. This, in conjunction with the extremely high attenuation of ultrasound (in MHz region) by air, further compounds the problem of using NCU mode. In simple terms, when ultrasound travels from a medium of low to one of high acoustic impedance, only a fraction of energy is transmitted in the latter. The fraction of ultrasound transmission and energy transferred is given by,

T (Transmission co-efficient in the medium of propagation) = 4Z1Z2 /(Z1 + Z2 )2

(1)


where Z₁ is the acoustic impedance of the medium from which ultrasound is transmitted (or originated), and Z₂ is the acoustic impedance of the medium in which ultrasound is transmitted, and the

Energy transferred in the medium of propagation (dB) = 20 log T

(2)

For more description and significance of plane wave transmission and reflection at a number of interfaces in terms of acoustical pressure and intensity, see ref. (2).

In the case of non-contact transmission ultrasound must propagate from air into the test material and then again into air so that the transmitted wave can be detected by a receiving transducer, Fig. 1. Therefore, the high loss of energy at air-material interface is doubled by further loss at the material-air interface. Table II provides the transmission losses in selected test materials in the non-contact ultrasound mode. As a reference, similar losses for water immersion (contact technique) are also provided. From this data the following conclusions can be drawn:

Fig 1: Interfaces to be crossed by ultrasound (shown by arrow) in non-contact transmission mode in order to propagate ultrasound through a test material. Interface `a' corresponds to air- material (from acoustic impedance Za to Zm) transmission. Interface `b' corresponds to material- air (from acoustic impedance Zm to Za) transmission.

• Transmission losses decrease as the acoustic impedance of the test material comes within the vicinity of the coupling medium, whether it is air or water.
• Total energy loss at various interfaces in the non-contact transmission mode can run as high as six orders of magnitude when compared to similar losses by using water as the coupling medium.

While conducting this exercise, we did not address the issues of interrogating frequency of the transducer and the frequency-dependence of ultrasound attenuation by air. When these factors are combined with the inherent loss of ultrasound energy at various interfaces in the NCU mode, the problem of non-contact transmission in solids is only exacerbated. Relatively speaking, the attenuation of ultrasound in air is intrinsically high compared to that in solids and liquids. And since attenuation in a medium increases as a function of the fourth power of the frequency, transmission of MHz frequencies in air almost becomes incomprehensible. To combat these NCU-defying realities, we must first create ultrasonic transducers that are characterized by high sensitivity (or low insertion loss). Sensitivity is needed not only to overcome interfacial transmission losses, Table II, but also to facilitate transducer excitation by relatively low power voltages. This will avoid the unwanted heating of transducers and their subsequent destruction. Once optimum sensitivity is achieved, then we must increase the transducer frequency so that it is comparable to that used in conventional contact testing. Accomplishment of this task has captivated the imagination of materials and transducer researchers.

MATERIAL Zm (Mrayl)	INTERFACE Fig. 1	TRANSMISSION CO-EFFICIENT equation 1		ENERGY TRANSFER equation 2 (dB)		TOTAL ENERGY LOSS AT INTERFACES a + b (dB)
		In Air	In Water	In Air	In Water	In Air	In Water
Steel 51.0	Air - Steel a Steel - Air, b	0.000034 0.000034	0.11 0.11	-89 -89	-19 -19	178	38
Aluminum 17.0	Air - Aluminum, a Aluminum - Air, b	0.0001 0.0001	0.3 0.3	-79 -79	-10 -10	158	20
Acrylic 3.5	Air - Acrylic, a Acrylic - Air, b	0.0005 0.0005	0.84 0.84	-66 -66	-1.5 -1.5	132	3
Silicone Rubber 1.0	Air - Rubber, a Rubber - Air, b	0.018 0.018	0.96 0.96	-35 -35	-0.35 -0.35	70	<1
TABLE II: Transmission co-efficients and energy transfer in selected materials at various interfaces in the non-contact mode, per Fig. 1. As a reference, similar data for water is also shown. Z (air) = 440Rayl. Z (water): 1.5Mrayl. 1 Rayl = kg/m2.s.

Pursuit of Non-contact ultrasonic transducers

Many have tried to develop non-contact means for materials characterization by utilizing the wave phenomena, which includes optics, thermal, infrared, x-ray, and nuclear magnetic resonance. In pursuit of bulk ultrasonic wave propagation in 1963, White (3) reported the generation of elastic waves in solid materials by the momentary heating of a material surface. This technique eventually led to the development of thermographic method. This has been used for surface and subsurface imaging of composites, metals, etc. by sensing minute temperature fluctuations as a function of material texture, microstructure, defects, and other variables. This method has been applicable to those materials that can sustain heat or those that emanate heat during the process of testing.

Next came laser-induced ultrasound, Bondarenko, et al. (4). It was used to characterize the Rayleigh waves in metals (5) and for subsurface materials evaluation (6). The laser-based method has been applied to those materials, which could withstand the impact of a high power laser beam. Laser-based ultrasound has become acceptable for high melting point metals and ceramics. The non-destructiveness of the laser-based ultrasound method is questionable when analyzing heat and shock sensitive materials such as, polymers, green ceramics and powder metals, pharmaceutical and food products, tissue, etc. Ultrasound generated by electromagnetic acoustic transducers has been used in the NCU mode for non- destructive testing (7). This method is only applicable to ferro-magnetic materials.

The various non-contact analytical methods outlined above do provide useful information about the test materials. However, all of them are limited to specific materials and are partially destructive, complex, or expensive. The difficulty of propagating ultrasound in the test media by the non-contact mode as shown shown in Table II, presents us with limited alternatives for achieving this mode in practical terms. These involve the methods of ultrasound generation based upon true production of ultrasound so that its propagation in the test medium is not affected by its (medium) exclusive properties.

Researchers and transducer experts have been designing piezoelectric devices by manipulating the acoustic impedance transitional layers in front of the piezoelectric element. In the materials industry, one of the early applications of non-contact ultrasound was for the testing of styrofoam blocks by utilizing 25kHz frequency (8). A precursor to high frequency non-contact transducers was the 1982 development of piezoelectric dry coupling longitudinal and shear wave transducers up to 25MHz frequency. Since 1983, these transducers have been commercially available for characterizing thickness, velocity, elastic and mechanical properties of green, porous, and dense materials (9,10,11). Dry coupling transducers feature a solid compliant and acoustically transparent transitional layer in front of the piezoelectric materials such as lead meta-niobates and lead zirconate-lead titanate. These devices, which eliminate the use of liquid couplant, do require contact with the material.

An important by-product of dry coupling devices was the development of air/gas propagation transducers, which utilize less than a 1 Mrayl acoustic impedance matching layer of a non-rubber material on the piezoelectric element. These commercially available transducers have been successfully produced in planar and focused beam configurations for transmitting of ultrasound in air up to ~5MHz frequency and receiving up to 20MHz. Air/gas propagation transducers, between 250kHz to ~5MHz, quickly found applications in aircraft/aerospace industries for imaging and for defect detection in fibrous, low and high density polymers, and composites. For such applications these transducers have been used with high energy or tone burst excitation and signal amplification systems. However, for applications such as level sensing and surface profiling, the low energy spike or square wave transducer excitation mechanism has been sufficient.

Similar transducers of 1MHz and 2MHz frequency were also produced at Stanford University by utilizing silicone rubber as the front acoustic impedance matching layer (12). By using such a transducer at 1MHz, the distance in air could be measured from 20mm to 400mm with an accuracy of 0.5mm. Further improvements in transduction efficiency were shown by planting an acoustic impedance matched layer, that is composed of tiny glass spheres in the matrix of silicone rubber on piezoelectric elements (13,14). Researchers at Strathclyde University (15) have reported air-coupling transducers based upon piezoelectric composites between 250kHz to 1.5MHz frequencies. By utilizing tone burst transducer excitation, they have been successful in producing millivolt level transmitted signals through a composite laminated honeycomb structure at 500kHz.

More recently, piezoelectric transducers featuring perfect acoustic impedance matched layers for optimum transduction in air have been successfully developed from <100kHz to 5MHz (16) [World-wide patent pending and in process]. The sensitivity of these new transducers in air is merely 30dB lower than their conventional contact counterparts. As a result, ultrasound in the MHz region can be easily propagated through practically any material, including even the very high acoustic impedance materials such as steel, cermets, and dense ceramics. This advancement will be discussed in detail along with the various medical and industrial applications in non-contact ultrasound mode.

Air coupled transducers based upon capacitance (electrostatic) phenomena have also undergone substantial developments in recent years. Researchers at Kingston and Stanford Universities have successfully produced micro-machined capacitance air transducers with frequencies as high as 11MHz (17,18). These transducers -- characterized by high bandwidths -- have been used to evaluate composites and other materials. At the time of this writing, there is no reliable method that can produce a reasonable comparison between these capacitance devices and those based upon the reference 16. From experience we have found the capacitance transducers lower in sensitivity and reliability than the new non-contact piezoelectric transducers. In light of this, it is highly desirable to produce a mechanism to compare both new capacitive and piezoelectric transducer developments on a one-to-one basis, based either upon the characteristics and reliability of the transducers or upon their specific applications performance.

Piezoelectric transducer for unlimited Non-Contact ultrasonic testing mode

The efficiency of an ultrasonic device is dependent on the coupling co-efficients and other electro- mechanical properties of the piezoelectric element. It is also based upon the mechanism by which ultrasound is transferred into the first medium of propagation. In the non-contact mode, this medium is air. We cannot overemphasize the significance of the final acoustic impedance matching layer on a piezoelectric element. Since the properties of a given piezoelectric material can be considered constant for a given device, the ultimate transfer of ultrasonic energy in air is entirely controlled by the acoustic characteristics of the final matching layer on the piezoelectric material, Fig. 2. To understand this, we examine the effect of the final acoustic impedance matching layer on the piezoelectric element with respect to transmission of ultrasound in air, as per equations 1 and 2. Table III shows transmission co- efficients of ultrasound in air as well as in water (as a reference) by assuming a number of final acoustic impedance matching materials on the piezoelectric element. From this data, the following conclusions can be drawn:

Fig 2: Schematic of an ultrasonic transducer showing the critical final acoustic impedance matching layer relative to the piezoelectric element and coupling medium air.

• The transmission co-efficient in air increases as the acoustic impedance of the front layer on the piezoelectric material is reduced.
• There is a significantly high transfer of ultrasonic energy in water when compared to that in air due to better acoustic impedance matching of the final layer on the piezoelectric element with that of water.

According to references 12, 13, 14, 15 and the author's 1983 design, the best final piezoelectric matching layer for maximum ultrasound transmission in air is composed of soft polymers. This layer can be non- porous or characterized by open or closed porosity. Such transducer designs yield --58dB to --54dB transfer of ultrasonic energy in air, which is significant for the propagation of 1MHz to 2MHz ultrasound in the non-contact mode. For example, by low energy excitation (such as by broadband spike or square wave pulse), ultrasound can be transmitted in materials that are lower than ~2 Mrayl in acoustic impedance. Propagation in materials between 3 Mrayl to >20 Mrayl by such transducers is arduous, but not impossible. By exciting such transducers with high energy single or multiple pulses, one can obtain low level millivolt transmitted signals through high acoustic impedance materials.

FINAL LAYER ON PIEZOELECTRIC ELEMENT Z (Mrayl)	TRANSMISSION CO-EFFICIENT, T		ENERGY TRANSFERRED, 20 log T (dB)
	In Air	In Water	In Air	In Water
Bare Piezoelectric, PZT, 31.0	0.00006	0.17	-85	-15
Hard Epoxy, 4.0	0.0004	0.79	-68	-2
Silicone Rubber, 1.0	0.001	0.92	-58	-0.7
Porous Rubber, 0.9	0.002	0.94	-54	-0.5
*Pressed Fiber, 0.1	0.018	--	-35	---
TABLE III. Transmission co-efficients and energy transferred in air as a function of the final acoustic impedance matching layer on the piezoelectric element. As a reference, similar data is also shown with water as the coupling medium. Z (air): 440 Rayl. Z (water): 1.5 MRayl.

*World-wide patents pending and in process (16).

It should be said that in this section we have concentrated on the most critical element of a non-contact transducer, i.e., the final layer on the piezoelectric material. However, the ultimate characteristics (frequency, bandwidth, sensitivity, and signal-to-noise-ratio) of a device are produced by the proper selection of the piezoelectric material and the transitional layers between it and the final acoustic impedance matching layer.

We conclude that transducer designs based upon a soft polymer as the final matching layer do demonstrate the feasibility of non-contact ultrasound, but are far from being the most efficient. For example, by successfully planting the final layer composed of fibrous materials, Table III, we are able to increase ultrasonic energy transfer from the transducer to air from --54dB (with porous rubber as the final matching layer) to --35dB. Increment of an order of magnitude in ultrasound transmission from the transducer to air is extremely significant and warrants special attention. For example, the new transducers now make it possible to propagate ultrasound from 200kHz to ~5MHz in the non-contact mode through all materials. Leaving aside transmission in plastics and composites, the sensitivity of the new transducers is high enough for materials with extremely high acoustic impedances such as, steel, dense ceramics and cermets. In the following sections we provide detailed observations of the new non- contact transducers.

Transducer Characterization Scheme
Non-contact transducers, like their contact counterparts, can also be characterized in transmission or in reflection (pulse echo) modes. Fig. 3. shows the set up for characterizing transducers in the transmission mode which is used to analyze the non-contact transducers. Here, the transmitting and receiving transducers are aligned and separated by a 10mm column of ambient air. The transmitting transducer is excited by a pulse of known voltage, V₀ . The output from the receiving transducer is directly fed into a measurement oscilloscope with a mechanism to measure the frequency domain. Frequency and bandwidth are measured directly from the frequency domain envelope, while V x , the received signal amplitude in volts, and the pulse width are measured from the time domain envelope. Signal-to-Noise- Ratio (SNR) is determined by the following relation where measurements are made without signal averaging.

Fig 3: Non-contact ultrasonic transducer characterization scheme in transmission mode. Frequency, Bandwidth, Pulse Width, and Signal-To-Noise Ratio: Directly read from the oscilloscope. Sensitivity = -20Log (Vx /V0). V0 = Excitation volts. Vx = Received signal amplitude volts.

SNR = 20 log Vx /Vn

(3)


where V_x is the amplitude of the received signal in volts, and V_n , the amplitude of the noise or the base line of the oscilloscope.

Sensitivity (loop gain, insertion loss) is determined by,

S (dB) = 20 log Vx/ V0

(4)

where V₀ is the amplitude of transducer excitation in volts.

Characteristics of Non-Contact Transducers
By utilizing the characterization scheme described above, several transducers were analyzed under ambient air environment. Figures 4, 5, and 6 show typical time and frequency domains for 200kHz, 1.5MHz, and 3.0MHz non-contact piezoelectric transducers.

Fig 4: Time and frequency domain of 200kHz, 50mm active area transducers in transmission mode. T-R air separation: 100mm. Bandwidth Center Frequency: 200kHz. Bandwidth @ -6dB: 100kHz (50%). Pulse Width: <20ms. Sensitivity: -46dB. SNR: 46dB.	Fig 5: Time and frequency domain of 1.5MHz, 25mm active area transducers in transmission mode. T-R air separation: 10mm. Bandwidth Center Frequency: 1.4MHz. Bandwidth @ -6dB: 0.92MHz (65%). Pulse Width: <2m s. Sensitivity: -58dB. SNR: 36dB.

Fig 6: Time and frequency domain of 3.0MHz, 12mm active area transducers in transmission mode. T-R air separation: 10mm. Bandwidth Center Frequency: 2.6MHz. Bandwidth @ -6dB: 2.0MHz (75%). Pulse Width: <700ns. Sensitivity: -64dB. SNR: 30dB.

Sensitivity Comparison of Non-Contact and Conventional Contact Transducers
Since sensitivity is the most critical requirement for non-contact transducers, it is important that some kind of a comparison scheme be developed. We have chosen conventional contact transducers as a reference. A number of non-contact transducers were characterized for sensitivity in the transmission mode, Fig. 3. Similar transducers, suitable for conventional contact mode operation, were characterized in water for sensitivity measurements. The set up for contact mode transducer characterization is the same as in Fig. 3, except that in this case the 10mm air column was replaced by a 10mm column of water. Sensitivity was calculated according to equation 4. The sensitivity of non-contact transducers in ambient air was subtracted from the sensitivity of contact water immersion transducers. The difference between the two sensitivities provides a reference indicating the quality of non-contact transducers. This data is provided in Table IV.

FREQUENCY (MHz)	ACTIVE DIAMETER (mm)	SENSITIVITY IN AMBIENT AIR (dB)	SENSITIVITY w.r.t WATER IMMERSION+ (dB)
0.25	50 25	-38 -46	Below 18 Below 26
0.5	50 25	-44 -50	Below 24 Below 30
1.0	25 19 12.5 3.2	-52 -54 -56 -62	Below 32 Below 34 Below 36 Below 38
1.5	12.5	-58	Below 38
2.0	12.5 1.5	-58 -66	Below 38 Below 40
3.0	12.5	-62	Below 40
5.0	12.5	-68	Below44
Table IV. Sensitivity Comparison of Non-Contact and Conventional Contact Transducers.* Mode of Testing: Transmission. Medium of Testing: 10mm Ambient Air for Non-Contact and 10mm Water for Contact Transducers

*Sensitivities reported here were obtained by exciting the transmitting transducer with a broadband and ~15ns pulse. Tone burst excitation sensitivities will be ~12dB higher.
+ For some contact transducers a --20dB sensitivity is assumed.

From this comparison, the following conclusions can be drawn:

• Non-contact transducers are approximately 30dB below their contact counterparts.
• The new non-contact transducers are the highest known sensitivity transducers for ultrasound propagation in air. Air-coupling transducers based upon soft polymer matching layer are at best 50dB below their conventional contact counterparts. Data supporting this argument is shown in Fig. 7. It is an a-scan of a transmitted signal through 25mm carbon steel (acoustic impedance, 51 Mrayl.) This observation was obtained by using 1MHz, 19mm active area diameter, non-contact transducers in the direct transmission mode. The transmitting and receiving transducers were separated by ~10mm of ambient air from the 25mm carbon steel surfaces. The transmitting transducer was excited by a pulser featuring a single burst of 400 V into a 4W square wave pulse. The received signal was amplified by 64dB. Under these settings the transmitted signal through steel is 20mV. No sign of a transmitted signal was observed when we used transducers based on soft polymer as the final impedance matching layer.

Fig 7: A 1MHz non-contact transmitted signal through 25mm carbon steel. Transmitting and receiving transducers are 10mm away from the material surfaces. Transmitting transducer was excited by a 400V into a 4 square wave pulse. The signal from receiving the transducer was amplified by 64dB. The left hand indication corresponds to the directly transmitted signal and the right hand indication to the first reflection corresponding to material thickness. Under these conditions the transmitted signal amplitude is ~20mV.

Non-Contact ultrasonic analyser

Fig 8:Ultrasonic Non-Contact Analysis system, the NCA 1000, shown here with non-contact transducers and monitor screen displaying the thickness and velocity of the test material.

Non-contact transducers can be used with any commercially available pulser-receiver to transmit ultrasound through any material. However, it was desirable to create a special instrument which would fully utilize this transducer development for the true analysis of materials. In 1998, such a system (the NCA 1000) [U.S. patent pending and in process] was developed by Leon Vandervalk and Ian Neeson of VN Instruments in Canada.. It is based on the synthesis of a computer-generated chirp with the best attributes of non-contact transducers. Signal processing in the NCA 1000 is done by synthetic aperture imaging techniques. This system is characterized by >140dB dynamic range and a time-of-flight (tof) measurement accuracy of ±50ns under ambient air conditions and ±1ns under closed conditions. The NCA 1000, Fig. 8, measures tof, thickness, velocity, and integrated response (area underneath transmitted or reflected signals in dB) of materials in time domain. By using the FFT mechanism of this system it is possible to conduct non-contact ultrasonic spectroscopy also. In section 7 we provide a variety of industrial and medical applications of the new transducers and the non-contact analytical system. For further significance of ultrasonic measurement, see Table I. Schemes of thickness, velocity, and frequency-dependence of ultrasound attenuation are given in the following sections.

Thickness and Velocity Measurements
There are several ways by which longitudinal wave velocity in the test materials can be determined by the non-contact ultrasound mode. For example, if multiple reflections corresponding to the thickness of the test material are observed, then one could measure the tof between the two successive peaks to determine the velocity of known thickness material. Appearance of multiple thickness reflections in NCU mode are however, dependent upon the attenuation and acoustic impedance characteristics of the test material. On the other hand, when only the transmission signal is observed (that is without thickness multiples), then one can determine the tof corresponding to that of the test material in a manner analogous to one used with contact delay line transducers.

However, it is best to examine all modes of ultrasound propagation through the test material when the objective is to determine both thickness and velocity by ultrasound. These modes relate to propagation of ultrasound when transmitting and receiving transducers are ``talking to each other'' in air, when ultrasound is transmitted through the test material, and ultrasound reflections in air from transmitting and receiving transducers to the test material surfaces. All possible modes of ultrasound propagation in NCU transmission mode are shown in Fig. 9. Here signals generated by these modes of propagation and their significance are as following:

Mode a is the transmission from transducer 1 to transducer 2 in air -- measures tof, t_a
Mode b is the reflection from transducer 1 to the material surface in air -- measures tof t₁
Mode c is the reflection from transducer 2 to the material surface in air -- measures tof t₂
Mode d is the transmission from transducer 1 to transducer 2 with the test material in between -- measures tof t_c
Mode e is the transmission from transducer 2 to transducer 1 with the test material in between -- measures tof t_c

From these tof measurements, the test material thickness and its velocity are determined according to the following relationships:

Vm = dm/tm	(5)
dm = Va x tam	(6)
tam = ta - (t1 + t2 /2)	(7)
tm = tam- (ta - tc )	(8)

In these equations, d_m is the test material's thickness, V_m is the test material velocity, V_a is the velocity of ultrasound in air (determined from a reference material), t_am is the time of flight in air corresponding to the test material thickness, d_m , and t_m is the time of flight in the test material.

By proper substitutions,

dm = Va x [ta- (t1 + t2)/2]	(9)
Vm = d m / [ ta -( t1 + t2)/2] - (ta- tc)	(10)

Fig.10 shows actual transmitted and reflected signals when a test material is examined in the non-contact transmission mode. Compare with Fig. 9. As can be seen from equation 10, in order to determine the thickness and velocity according to this scheme, one only needs the measurements of four times of flight (t_a,t₁, t₂, t_c ) and the velocity of ultrasound in air. These parameters are measured and calculated by the NCA 1000 computer and are displayed on the monitor. An example is shown in Fig. 11.

Fig 9: Propagation of ultrasound with respect to test material in non-contact transmission mode. The NCA 1000 measures and calculates all times of flight shown in this figure, then displays them on the monitor screen. See text for more details.

Fig 10:Identification and location of various transmitted and reflected signals from a test material in non-contact ultrasound transmission mode. The right hand figure shows the mechanism by which these signals are produced. First Arrival: Ultrasound transmission through air and test material. Next Periodically separated peaks: Reflections corresponding to material thickness. Last Arrivals: Material surface reflections from transducers 1 and 2.

Fig 11: The NCA 1000 screen displaying velocity and thickness of a material. The test material in this case is a 8.8mm polystyrene.

Relative Amplitude Attenuation Measurement
In the time domain display, the NCA 1000 -- besides measuring and displaying the times of flight of the signals -- also shows the Integrated Response (IR) of these signals. It is a measurement of the area underneath a particular peak in power dB units. Due to the very high linear dynamic range of the NCA 1000, the sensitivity of IR relative to the detection of minute changes in test materials can be very useful in characterization. Interpretation of the integrated response with respect to transmitting and receiving transducers in air, IR_a (Fig. 12) and that with the test material, IR_m (Fig. 13), can provide quick and significant information about relative attenuation at a particular frequency, i.e.,

Fig 12: Integrated Response (IR) of a transmitted signal when 1MHz 19mm active area diameter transmitting and receiving transducers are facing each other in a column of 53mm ambient air.

Fig 13:Integrated Response (IR) of a transmitted signal when a 26mm medium-grained SiC grinding wheel is inserted between the transmitting and receiving transducers shown in Fig. 12. IRm = 50dB. Therefore, Relative Attenuation @ 1MHz = 135 -- 50 = 85dB. Variations in relative attenuation at various points in a given material can be related to the defects, texture or the microstructure of the material.

Relative Attenuation (dB)= IRa - IRm

(11)

This technique has been found to be valuable for the analysis of microstructurally complex materials, for minute defect detection, and for phase concentration measurements in liquids and some solids.

Non-Contact Ultrasonic Spectroscopy
By performing the FFT of transmitted or reflected signals, test materials can also be characterized to investigate frequency-dependence of ultrasonic attenuation. Such examinations are important while testing microstructurally complex materials or those materials for which time domain velocity measurements are not sensitive enough. The first step for frequency-dependence of ultrasonic characterization is acquiring the frequency spectrum of a transmitted signal in air, i.e., without the test material. The next step is to do the same with the test material inserted between the transducers. And the final step is to subtract the sample spectrum from that of the reference air spectrum. In the following section, we provide several examples of transmission and reflection spectroscopy of various materials.

Applications of Non-Contact ultrasound

Non-contact transducers have now been successfully produced in the frequency range of 100kHz to >5MHz. While the applications of non-contact transducers greater than 3MHz are so far limited to thin materials, transducers below 3MHz have been extensively used for a number of industrial and medical applications in time, frequency, and image domains (18, 19, 20,21).

Fig 14:Relationship between density and non-contact velocity of green alumina. Transducer frequency: 1MHz.

Fig 15: Relationship between density and non-contact velocity of sintered alumina. Transducer frequency: 1 and 2MHz.

Fig 16:Comparison of ultrasonically and physically determined density of alumina.

Materials Characterization and Defect Detection
Figures 14 and 15, respectively, show the relationships between densities of green and sintered alumina with non-contact velocities of these materials. Fig. 16 shows a correlation between ultrasonically (from reference samples) and physically (apparent) determined densities of green alumina. The Non-contact ultrasonic technique has been successfully applied for characterizing density and defects in a variety of green materials such as ceramics tiles, multi-layer electronic packages, powder metals, cements, concretes, etc. Fig. 17 shows detection of defects in a sheet of aluminum by the non-contact transmission mode.

Fig 17: Detection of defects in an 8mm thick sheet of aluminum by non-contact transmission mode.Top: Defect-free region. Bottom: Region with 1.5mm cylindrical hole.Transducer frequency: 2MHz.

Figures 18 and 19, respectively, show non-contact transmission ultrasonic spectroscopy of space shuttle tiles and packaging foams. For the detection of defects and bubbles and for texture and microstructure characterization, non-contact ultrasonic spectroscopy has also been applied to liquid polymer resins and other viscous liquids, as well as to lumber and construction materials. Fig. 20 shows non-contact reflection spectroscopy for the characterization of material surfaces. Fig. 21 shows the non-contact ultrasonic image of an aircraft composite material.

Fig 18:Non-Contact transmission ultrasonic spectroscopy of extremely porous ceramics (space shuttle tiles) for microstructure characterization. Top: 0.38g/cc. Middle: 0.28g/cc. Bottom: 0.1g/cc. Transducer frequency: 350kHz.	Fig 19:Non-Contact transmission ultrasonic spectroscopy of packaging foam. Top: small cell. Bottom: large cells. Transducer frequency: 350kHz.	Fig 20:Non-Contact reflection ultrasonic spectroscopy for surface characterization of materials. Note that as the surfaces become rough, subsequently the attenuation of ultrasound increases. Transducer frequency: 2MHz.
Fig 21:Non-contact ultrasonic image of an impact damaged 8-ply graphite fiber re-enforced plastic composite 1.5mm thick. Area scanned: 25 x 25mm. Transducer frequency: 2MHz. Observe the central damaged region. This image was produced by scanning a 3mm diameter receiver over the sample with 25mm transmitter underneath it. Both transducers were kept ~5mm away from the sample. A conventional pulser/receiver with 400V into 4 square wave and 64dB gain was used.

Since the NCA 1000 interprets ultrasound reflections from both the transmitting and the receiving transducers from material surfaces, it is now possible to measure thickness of materials that are continuously rolled on a production line and are too wide for micrometers.

Food, Beverage, and Pharmaceutical
Fig. 22 shows transmitted non-contact ultrasound signals from regions with and without almonds in milk chocolate. Fig. 23 is an example of fat content measurement in milk and milk products. Similarly, Fig. 24 shows the measurement of sugar content in water. We have also applied non-contact ultrasound for the detection of beverages and other liquids in plastic, metal, and cardboard cartons. The quality of heat and vacuum seals in pharmaceutical and food packages has also been determined by this technique. Analogous to green ceramics analysis, (Fig. 14), tablets and other powder-based pharmaceutical products have also been characterized as functions of the velocity and frequency-dependence of ultrasound attenuation.

Fig 22: Detection of absence or presence of almonds in milk chocolate. Top: Region without almonds. Bottom: Region with almonds. Transducer frequency: 1MHz.

Fig 23: Estimation of fat content in various types of milk and milk products with distilled water as a reference. Ultrasound measurement here is the relative transmission of 1MHz frequency in samples. Samples in plastic bottles were analyzed in transmission mode with transmitting and receiving transducers separated by ~15mm of ambient air from the bottles.

Fig 24: Estimation of sugar content in water. Ultrasound measurement here is the relative transmission of 1MHz frequency in samples. Samples in plastic bottles were analyzed in transmission mode with transmitting and receiving transducers separated by ~15mm of ambient air from the bottles.

Medical
Among the first uses of non-contact ultrasound in medical applications was the evaluation of burnt skin and bed sores in burn victims (18). By utilizing 2MHz transducers with a prototype portable ultrasonic pulser-receiver, many observations were made at various points on a healthy and a burnt human hand at the Burn Center, University of California, Irvine, under the direction of Joie P. Jones. This work was done by using a single transducer in the reflection (pulse-echo) mode. The collected data was processed to create internal images of two skin conditions, Fig. 25.

Fig 25: Non-Contact ultrasound image cross-section of healthy (top) and burnt (bottom) human hands. Note the damaged region between the epidermis and capillary bed in the burnt hand. The transducer used for this analysis is non-contact 2MHz in reflection (pulse-echo) mode. [These images were generated by Prof. Joie P. Jones, University of California, Irvine and his Burn Center team.]

High Precision level Sensing
By utilizing the high frequency and high SNR characteristics of non-contact transducers and the signal processing of NCA 1000, it is possible to detect distances of liquids or solids from very short to very long ranges within ±0.0003mm accuracy. Fig. 26 shows the reflected signals from a 2MHz 3mm active area diameter transducer to liquid level interfaces from 4.26mm to 112.12mm. Fig. 26.

Fig 26: Reflected signals from non-contact transducer to a liquid in a test tube. Top: Liquid at 4.26mm from transducer. Bottom: Liquid at 112.12mm from transducer. Transducer: 2MHz and 3mm active area diameter.

Conclusions

In this paper we have given the significance of ultrasound for non-destructive characterization of materials and for non-invasive diagnostic applications in the medical field. We have also shown the feasibility of ultrasonic measurements in time, frequency, and image domains analogous to other wave-based methods,

Underscoring the significance of the non-contact ultrasound mode, a detailed discussion on the difficulty of achieving this mode has also been presented along with the struggle of a number of transducer developers. We have also shown that this work ultimately produced very high transduction air propagation transducers, thus making the non-contact ultrasound mode a reality. Applications of these transducers for use in industry and the medical field have been given with documentary evidence. We have also provided an introduction to a dedicated ultrasonic non-contact analyzer and its applications for characterization of industrial materials and food and pharmaceutical products.

We believe that the non-contact ultrasound mode is among the most significant developments for characterization and analysis of all states of matter. Though we have provided selected examples of its applications, there is no doubt that the users of this technology will further enhance the causes of materials quality, process control, and health care in our increasingly complex world.

Acknowlwdgements

The author gratefully acknowledges the assistance of M. Langron in producing the transducers used in this paper. Enthusiastic support of C. Nielson, Dow Chemical USA, and that of Prof. E. Blomme, Katholieke Hogeschool, Belgium is acknowledged in kind. We are also thankful to I. Neeson, VN Instruments, Canada, Prof. B.R. Tittmann, Penn State University, and Prof. J.P. Jones, University of California, Irvine, for their valuable comments and contributions during the development of non-contact ultrasonic testing procedures. The work presented in this paper was supported by the continuing efforts of Ultran for the advancement of industry and medical science through innovative developments in ultrasound.

References

1. Green, R.E., ``Ultrasonic Non-destructive Materials Characterization,'' in Alex Vary, ed., Materials Analysis by Ultrasonics, Noyes Data Corporation, New Jersey, 1987, p. 6.
2. Cho, Z., Jones, J.P., and Singh, M, Foundations of Medical Imaging, Wiley, New York, 1993, pp. 477-486
3. White, R.M., ``Generation of Elastic Waves by Transient Surface Heating,'' J. App. Phys., v. 34, pp. 3559-3567 (1963)
4. Bondarenko, A.A., Drobat, Y.B., and Kruglov, S.V., ``Optical Excitation and Detection of Nanosecond Acoustic Pulses in Non-destructive Testing, `` Soviet J. NDT, v. 12, pp. 655-658 (1976)
5. Ledbetter, H.M. and Moulder, J.C., ``Laser-Induced Rayleigh Waves in Aluminum, J. Acoust. Soc. Amer. V. 65, pp 840-842 (1979)
6. Aindow, A.M., Dewhurst, R.J., Palmer, S.B., and Scruby, C.B., ``Laser-Based Non-Destructive Testing Techniques for Ultrasonic Characterization of Subsurface Flaws, NDT International, v. 17, pp. 329-335 (1984)
7. Allers, G.A., Monitoring Pipe and Tube Wall Properties During Fabrication in a Steel Mill,'' in H.N.G. Wadley, P.A. Parish, B.B. Rath, and S.M. Wolf, editors, Intelligent Processing of Materials and Advanced Sensors, Metallurgical Soc., Pennsylvania, (1986), pp. 17-27
8. Brunk, J.A., Allied Signal, Private communication, (1999)
9. Brunk, J.A., Valenza, C.J., and Bhardwaj, M.C., "Applications and Advantages of Dry Coupling Ultrasonic Transducers for Materials Characterization and Inspection," in Acousto-Ultrasonics, Theory and Applications, John C. Duke, Jr., Editor, Plenum Press, New York (1988), pp. 231-238.
10. Bhardwaj, M.C. and Bhalla, A., "Ultrasonic Characterization of Ceramic Superconductors," J. Mat. Sci. Lett., v. 10 (1991)
11. Kulkarni, N., Moudgil, B. and Bhardwaj, M., ``Ultrasonic Characterization of Green and Sintered Ceramics: I, Time Domain,'' Am. Cer. Soc., Cer. Bull, Vol. 73, No. 6, (1994)
12. Fox, J.D., Khuri-Yakub, B.T., and Kino, G.S., ``High Frequency Wave Measurements in Air,'' 1983 IEEE Ultrasonics Symposium, pp. 581-592 (1983)
13. Yano, T, Tone, M., Fukumoto, A., ``Range Finding and Surface Characterization using High Frequency Air Transducer,'' IEEE Trans. UFFC, v. 34, n. 2, pp. 222-236 (1987)
14. Haller, M.I. and Khuri-Yakub, B.T., ``1-3 Composites for Ultrasonic Air Transducer,'' IEEE Ultrasonics Symposium, pp. 937-939 (1992)
15. Reilly, D. and Hayward, G., ``Through Air Transmission for Ultrasonic Non-destructive Testing,'' IEEE Ultrasonic Symposium, pp. 763-766 (1991)
16. Bhardwaj, M.C., ``Ultrasonic Transducer for High Transduction in Gases and Method for Ultrasound Non-Contact Transmission into Solid Materials,'' world wide patent pending and in process (1997, 98)
17. Schindel, D.W., Hutchins, D.A., Zou, L., and Sayer, M., ``The Design and Characterisation of Micromachined Air-Coupled Capacitance Transducers,'' IEEE Trans. Ultrason. Ferroelect. Freq. Control., v. 42, pp. 42-51 (1995)
18. Ladabaum, I., Khuri-Yakub, B.T., and Spoliansky, D., ``Micomachined Ultrasonic Transducers (MUTS): 11.4MHz Transmission in Air and More,'' App. Phys. Lett., v. 68, pp. 7-9 (1996)
19. Bhardwaj, M.C., ``Innovation in Non-Contact Ultrasonic Analysis: Applications for Hidden Objects Detection,'' Mat. Res. Innovat. (1997) 1:188-196
20. Jones, J.P, Lee, D, Bhardwaj, M., Vanderkam, V., and Achauer, B,, ``Non-Contact Ultrasonic Imaging for the Evaluation of Burn-Depth and for Other Biomedical Applications,'' Acoust. Imaging, V. 23 (1997)
21. Bhardwaj, M.C., ``Non-Contact Ultrasonic Characterization of Ceramics and Composites,'' Proceedings Am.Cer.Soc., V 89 (1998)
22. T. Carneim, D.J. Green & M.C. Bhardwaj, ``Non-Contact Ultrasonic Characterization of Green Bodies,'' Cer. Bull., April 1999