Encyclopedia of Smart Materials (Vols 1 and 2) - M. Schwartz (2002) WW Part 9 pps

54 J.M Paros and L Weisborg, Mach Design 27: 151–156 (1965) 55 Y Bellouard, Ph.D Thesis, Lausanne, EPFL, n◦ 2308 (2000) 56 W Nix, Scripta Materialia, 39(4/5): 545–554 (1998) 57 R.D James, Int J Solids Struct 37: 239–250 (2000) 58 R Gorbet, Ph D Thesis, University of Waterloo, 1997 MICROTUBES WESLEY P HOFFMAN Air Force Research Laboratory AFRL / PRSM Edwards AFB, CA PHILLIP G WAPNER ERC Inc Edwards AFB, CA INTRODUCTION Background Microtubes are very small diameter tubes (in the nanometer and micron range) that have very high aspect ratios and can be made from practically any material in any combination of cross-sectional and axial shape desired In smart structures, these microscopic tubes can function as sensors and actuators, as well as components of fluidic logic systems In many technological fields, including smart structures, microtube technology enables fabricating components and devices that have, to date, been impossible to produce, offers a lower cost route for fabricating some current products, and provides the opportunity to miniaturize numerous components and devices that are currently in existence In recent years, there has been tremendous interest in miniaturization due to the high payoff involved The most graphic example that can be cited occurred in the electronics industry, which only 50 years ago relied exclusively on the vacuum tube for numerous functions The advent of the transistor in 1947 and its gradual replacement of the vacuum tube started a revolution in miniaturization that was inconceivable at the time of its invention and is not fully recognized even many years later Miniaturization resulted in the possibility for billions of transistors to occupy the volume of a vacuum tube or the first transistor, and it was not the only consequence The MEMS was the accelerometer that was manufac sensor for air-bag actuation On the market toda microfluidic devices, mechanical resonators, bios glucose, and disposable blood pressure sensors th serted into the body The vast majority of microsystems are made a clusively on planar surfaces using technology de fabricate electronic integrated circuits The fabr these devices takes place on a silicon wafer, an vice is formed layer-by-layer using standard c techniques that include electron beams or phot phy, thin-film deposition, and wet or dry etch isotropic and anisotropic) Three variations of th tional electronic chip technology can be used, for to make three-dimensional structures that hav pect ratios and suspended beams These include (lithographie, galvanoformung, abformung) pro the Hexsil process (3), and the SCREAM (sing reactive etching and metallization) process (4) nique most employed, the LIGA process, which veloped specifically for MEMS-type applications struct and metallize high-aspect-ratio microfeat is done by applying and exposing a very thick X sitive photoresist layer to synchrotron radiation up to 600 microns high that have aspect ratios 1 can be fabricated by this technique to make tr dimensional objects The Hexsil process uses a has a sacrificial layer of silicon dioxide to form con structures that are released by removing t dioxide film A third approach is the SCREAM cromachining process that can fabricate high-as single-crystal silicon suspended microstructures icon wafer using anisotropic reactive ion etch however, that like the conventional technique use electronic circuits, all of these variations use a la proach that starts on a flat surface In addition, there are some disadvantages of th tional electronic chip fabrication technique and cations, even though there have been numerous innovative successes using these silicon wafer-b nologies This is due to the fact that these tec require building up many layers of different ma well as surface and bulk micromachining whic some very difficult material science problems to be solved These include differential etching a down one material without damaging any previ In addition, there are the problems of interconne ers in a chip that have different functions An e this is a microfluidic device in which there are b 54 J.M Paros and L Weisborg, Mach Design 27: 151–156 (1965) 55 Y Bellouard, Ph.D Thesis, Lausanne, EPFL, n◦ 2308 (2000) 56 W Nix, Scripta Materialia, 39(4/5): 545–554 (1998) 57 R.D James, Int J Solids Struct 37: 239–250 (2000) 58 R Gorbet, Ph D Thesis, University of Waterloo, 1997 MICROTUBES WESLEY P HOFFMAN Air Force Research Laboratory AFRL / PRSM Edwards AFB, CA PHILLIP G WAPNER ERC Inc Edwards AFB, CA INTRODUCTION Background Microtubes are very small diameter tubes (in the nanometer and micron range) that have very high aspect ratios and can be made from practically any material in any combination of cross-sectional and axial shape desired In smart structures, these microscopic tubes can function as sensors and actuators, as well as components of fluidic logic systems In many technological fields, including smart structures, microtube technology enables fabricating components and devices that have, to date, been impossible to produce, offers a lower cost route for fabricating some current products, and provides the opportunity to miniaturize numerous components and devices that are currently in existence In recent years, there has been tremendous interest in miniaturization due to the high payoff involved The most graphic example that can be cited occurred in the electronics industry, which only 50 years ago relied exclusively on the vacuum tube for numerous functions The advent of the transistor in 1947 and its gradual replacement of the vacuum tube started a revolution in miniaturization that was inconceivable at the time of its invention and is not fully recognized even many years later Miniaturization resulted in the possibility for billions of transistors to occupy the volume of a vacuum tube or the first transistor, and it was not the only consequence The MEMS was the accelerometer that was manufac sensor for air-bag actuation On the market toda microfluidic devices, mechanical resonators, bios glucose, and disposable blood pressure sensors th serted into the body The vast majority of microsystems are made a clusively on planar surfaces using technology de fabricate electronic integrated circuits The fabr these devices takes place on a silicon wafer, an vice is formed layer-by-layer using standard c techniques that include electron beams or phot phy, thin-film deposition, and wet or dry etch isotropic and anisotropic) Three variations of th tional electronic chip technology can be used, for to make three-dimensional structures that hav pect ratios and suspended beams These include (lithographie, galvanoformung, abformung) pro the Hexsil process (3), and the SCREAM (sing reactive etching and metallization) process (4) nique most employed, the LIGA process, which veloped specifically for MEMS-type applications struct and metallize high-aspect-ratio microfeat is done by applying and exposing a very thick X sitive photoresist layer to synchrotron radiation up to 600 microns high that have aspect ratios 1 can be fabricated by this technique to make tr dimensional objects The Hexsil process uses a has a sacrificial layer of silicon dioxide to form con structures that are released by removing t dioxide film A third approach is the SCREAM cromachining process that can fabricate high-as single-crystal silicon suspended microstructures icon wafer using anisotropic reactive ion etch however, that like the conventional technique use electronic circuits, all of these variations use a la proach that starts on a flat surface In addition, there are some disadvantages of th tional electronic chip fabrication technique and cations, even though there have been numerous innovative successes using these silicon wafer-b nologies This is due to the fact that these tec require building up many layers of different ma well as surface and bulk micromachining whic some very difficult material science problems to be solved These include differential etching a down one material without damaging any previ In addition, there are the problems of interconne ers in a chip that have different functions An e this is a microfluidic device in which there are b In addition to the processing problems mentioned before, there are other limitations inherent in conventional lithographic techniques that are based on planar silicon For example, in some applications such as those that involve surface tension in fluidics, it is important to have a circular cross section However, it is impossible to make a perfectly round tube or channel on a chip by conventional technology Instead, channels on the wafer surface are made by etching a trench and then covering the trench by using a plate (16,17) This process can produce only angled channels such as those that have a square, rectangular, or triangular cross section Because of the limitations already mentioned, we heartily agree with Wise and Najafi in their review of microfabrication technology (18) when they stated, “The planar nature of silicon technology is a major limitation for many future systems, including microvalves and pumps.” In the literature, there are at least two technologies in addition to microtubes that remove microfabrication from the flatland of the wafer One uses “soft lithography,” and the other uses laser-assisted chemical vapor deposition (LCVD) “Soft lithography,” conceived and developed by Whitesides’ outstanding group at Harvard, encompasses a series of very novel related technologies that include microcontact printing, micromolding, and micromolding in capillaries (11) These technologies can fabricate structures from several different materials on flat and curved surfaces By example, structures can be fabricated using microcontact printing by first making a stamp that contains the desired features This stamp, which is usually made from poly(dimethylsiloxane) (PDMS) has raised features placed on the surface by photolithographic techniques The raised features are “inked” with an alkanethiol and then brought into contact with a gold-coated surface, for example, by rolling the curved surface over the stamp The gold is then etched where there is no self-assembled monolayer of alkanethiolate Features as small as 200 nm can be formed by this technique However, the microstructures produced by this technique are the same as those produced by standard techniques, except that the starting surface need not be flat By using these techniques, submicron features can be fabricated on flat or curved substrates made of materials, such as metals (19), polymers (20), and carbon (21) In addition, these technologies can be used to make truly three-dimensional free-standing objects (22,23) Another step away from the standard planar silicon technology is the LCVD process (24,25) which can “write in space” to produce three-dimensional microsystems In trast to these two technologies, microtube techno offers the ability to make microdevices from pract material because the technology is not limited by e position or the availability of CVD precursor mat addition, in contrast to these other technologies, m technology provides the opportunity to make tu also to make it in a variety of cross-sectional shapes that can be used to miniaturize systems components, and fabricate components or system not currently possible to produce Microscopic and Nanoscopic Tubes and Tubules Commercially, tubing is extruded, drawn, pult rolled and welded which limits the types of mate can be used for ultrasmall tubes as well as their internal diameters In addition, it is not curren ble to control the wall thickness, internal diamet surface roughness of the inner wall of these tubes tion of a micron by these techniques Using con techniques, ceramic tubes are currently availabl small as 1 mm i.d Copper tubing can be obtained as 0.05 mm i.d., polyimide tubing is fabricated as 50 µm i.d., and quartz tubing is drawn down as 2 µm i.d This means that quartz is the only tub mercially available that is less than 10 µm i.d Th tubing is used principally for chromatography There are, however, other sources of small tu are presently at various stages of research and ment For some time, several groups have been us as templates (26–28) to fabricate submicron diam ing These tubes are made by using electroles tion to metallize a tubular lipid structure forme Langmuir–Blodgett film Lipid templated tubes uniform in diameter, which is fixed at ∼0.5 µm by structure Lengths to 100 µm have been obtaine technique which is extremely expensive due to t the raw materials Other groups are making submicron diamete using a membrane-based synthetic approach Th involves depositing the desired tubule material w cylindrical pores of a nanoporous membrane Co “track-etch” polymeric membranes and anodic a oxide films have been used as the porous s Aluminum oxide, which is electrochemically et been the preferred substrate because pores of diameter can be made from 5–1000 nm Martin polymerized electrically conductive polymers fro uid phase and electrochemically deposited met in a single membrane are uniform in diameter, length, and thickness But in contrast to the lipid process, the diameter of the tubules can be varied by the extent of oxidation of the aluminum substrate Although diameters can be varied in this process, it should be clear that these tubules are limited in length to the thickness of the porous membrane In addition, the wall thickness is also limited in that the sum of the inside tubule diameter and two times the wall thickness is equal to the starting pore diameter Using a sol-gel method, tubules can be made in about the same diameter range as in the membrane approach By hydrolyzing tetraethlyorthosilicate at room temperature in a mixture of ethanol, ammonia, water and tartaric acid, Nakamura and Matsui (36) made silica tubes that had both square and round interiors The tubules produced by this technique were up to 300 µm long, and the i.d of the tubes ranged from 0.02 to 0.8 µm By introducing minute bubbles into the sol, hollow TiO2 fibers that have internal diameters up to 100 µm have also been made (37) by using the sol-gel approach On an even smaller scale, nanotubules are fabricated using a number of very different techniques The most wellknown tube in this category is the carbon “buckytube” that is a cousin of the C60 buckyball (38–42) Since carbon nanotubes were first observed as a by-product in C60 production, the method of C60 formation using an arc-discharge plasma was modified to enhance nanotube production The process produces tubules whose i.d is in the range of 1–30 nm These tubules are also limited in length to about 20 microns Similar nanotubes of BN (43), B3 C, and BC2 N (44) have been made by a very similar arc-discharge process In addition, nanotubes of other compositions (45,46) have been prepared using carbon nanotubes as a substrate for conversion or deposition An alternative technique for manufacturing carbon tubes that have nanometer diameters has been known to the carbon community for decades from the work of Bacon, Baker, and others (47–50) The process produces a hollow catalytic carbon fiber by pyrolyzing a hydrocarbon gas over a catalyst particle The fibers, which vary in diameter from 1 nm to 0.1 µm have lengths up to centimeters, can be grown either hollow or has an amorphous center that can be removed by catalytic oxidation after a fiber is formed Other nanoscale tubules whose diameters are slightly larger and smaller than buckytubes have been made from bacteria and components of cytoskeletons and by direct chemical syntheses Chow and others (51) isolated and molecular self-assembly Some of the nanotubule into this category are made from cyclic peptide clodextrins (54), and bolaamphiphiles (55) Cycl nanotubules have an 0.8 nm i.d: and can be mad microns in length Other self-assembled nanotu range from 0.45 to 0.85 nm i.d have been synthes cyclodextrins (54,56) in lengths in the tens of na Although it is clear that individual nanotubul rently useful for certain applications, such as e tion, reinforcement, or as scanning probe micro (57), it is not obvious how individual nanotubu observed and economically manipulated for use other than by using a scanning probe microscope ( this problem is solved, the future of individual n in devices is uncertain However, this problem c cumvented if the nanotubules are part of a larger as in an array If oriented groups or arrays of submicron to m meter tubes or channels perpendicular to the sur wafer or device are desired, there are at least fo available to make them Using the technique des fore for making anodic porous alumina, a two s cation process (59) can be used to fabricate a dered honeycomb nanohole array from gold or The metal hole array is from 1–3 micron thick and 70 nanometers in diameter For smaller tubes or a technique (60) has recently been developed to d bundles of quartz tubes to form an array This pr duces a hexagonal array of glass tubes each as 33 nm in diameter This translates to a density channels per square centimeter Even smaller r rays of channels can be synthesized by a liqu template mechanism (61,62) In this process, a silicate gels are calcined in the presence of surf produce channels 2–10 nm in diameters Finally of ∼4 nm in cross section can be produced (63) pe larly to the surface of an amorphous silica film b hematite crystals in a Fe–Si–O film and then etch the hematite crystals Finally, several technologies exist to make ch layers of channels of desired orientation in soli These technologies are another spin-off of the p graphic process used for integrated circuits O dimensional plane, channels that range in size to hundreds of microns in width and depth have ricated (16,17) on the surface of silicon wafers dard microphotolithographic techniques Formi croscopic channels and holes in other materials o an electrochemical cell where a nontoxic salt solution is the electrolyte Sheet architecture technology (67) developed at Pacific Northwest National laboratory is used to fabricate numerous microscopic chemical and thermal systems, such as reactors, heat pumps, heat exchangers, and heat absorbers These devices may consist of a single photolithographically etched or laser-machined laminate that has a cover bonded to seal the channels, as described before, or may consist of multiple layers of plastic or metal laminates bonded together It is quite apparent from this brief and incomplete review, that a number of very novel and innovative approaches have been used to make microsystems as well as tubes and channels whose diameters are in the range of nanometers to microns In the next section, the basics of microtube technology which complements these other technologies are discussed AFRL MICROTUBE TECHNOLOGY Properties and Production of Microtubes Except for self-assembled tubules, the microtube technology developed at the Propulsion Directorate of the Air Force Research Laboratory (AFRL) can produce tubes in the size range of those made by all of the other techniques cited In contrast to tubing currently on the market and the submicron laboratory scale tubing mentioned before, microtubes can be made from practically any material (including smart materials) and will have precisely controlled composition, diameter, and wall thickness in a great range of lengths In addition, this technology can produce tubes in a great diversity of axial and cross-sectional geometries For most materials, there is no upper diameter limit, and for practically any material, internal diameters greater than 5 µm are possible In addition, for materials that can survive temperatures higher than 400◦ C, tubes can be made as small as 5 nanometers by using the same process To date, tubes have been made from metals (copper, nickel, aluminum, gold, platinum, and silver), ceramics (silicon carbide, carbon, silicon nitride, alumina, zirconia, and sapphire), glasses (silica), polymers (Teflon), alloys (stainless steel), and layered combinations (carbon /nickel and silver/sapphire) in sizes from 0.5–410 µm Like many of the techniques described before, microtube technology employs a fugitive process that uses a sacrificial mandrel, which in this case is a fiber High-quality coating should be sufficient to demonstrate that practi cross-sectional shape imagined can be fabricated in Fig 2, the wall thickness of the tubes can be uniform around the tube It is also possible to co wall thickness accurately along the length of th ual tubes and among the tubes in a batch or a co process It can be seen in Fig 2 that the walls can nonporous It will be shown later that the microst the walls and extent of porosity that the walls co also be controlled In addition to the possibility sectional tube shapes, using a fugitive process al fabricating tubes that have practically any axial as is shown later The maximum length in which these tubes can has yet to be determined because it depends on m ables, such as the type of tube material, the com of the sacrificial tube-forming material, and the porosity in the wall It is possible that there is no l in length for a tube that has a porous wall For n wall tubing, the maximum length would probably meter range because there is a direct relationship the tube i.d and the maximum possible length for most applications conceived to date, the len only be of the order of a few centimeters Based o calculation, it is apparent that even “short” tub tremendous aspect ratio For instance, a 10-µm 25 cm long has an aspect ratio of 2500 Using microtube technology, there is no upper l in wall thickness for most materials To date, free tubes have been made whose wall thickness ra 0.01–800 µm (Fig 3a) Most of the microtubes date have demonstrated surprising mechanical In fact, preliminary studies of both copper and sil whose wall thickness is in the micron range ha that microtubes can have up to two times th strength of an annealed wire of the same mater same cross-sectional area Besides precise cont tube wall thickness and composition, the interior these tube walls can have practically any desired t degree of roughness In addition, the walls can ra nonporous to extremely porous, as seen in Fig 4 interior or exterior surfaces of these tubes can be one or more layers of other materials (Fig 5), In addition to free-standing microtubes, solid m structures that have microchannels can be fabr making the tube walls so thick that the spaces bet tubes are filled (Fig 6) The microchannels can be oriented, or they can have a predetermined or (c) (d) Figure 1 Examples of microtubes: (a) 10-µm silicon carbide tubes; (b) 410-µm nickel tubes; (c) 26-µm silicon nitride tube; and (d) 0.6-µm quartz tube Any desired orientation or configuration of microtubes can be obtained by a fixturing process Alternatively, composite materials can be made by using a material different from the tube wall as a “matrix” that fills in the space among the tubes The microtubes imbedded in these monolithic structures form oriented microchannels that, like free-standing tubes, can contain solids, liquids, and gases, and as act as waveguides for all types of electromagnetic energy Microtube Applications Discrete thinner walled microtubes are useful in areas as diverse as spill cleanup, encapsulation of medicine or explosives, insulation that is usable across a very wide range of temperature, and as lightweight structural reinforcement similar to that found in bone or wood The cross-sectional shape of these reinforcing tubes can be tailored to optimize mechanical or other properties In addition, thinner walled tubes are useful as bending or extension actuators when fabricated from smart materials Thicker walled tubes (Fig 3b: nickel and SS) that are just as easily fabricated are needed in other applications, such as calibrated leaks and applications that involv or external pressure on the tube wall The ability to coat the interior or exterior these tubes with a layer or numerous layers of terials enlarges the uses of the microtubes and a fabricating certain devices For example, apply tion or corrosion protection layers on a structur cialty tube material will greatly enlarge its uses can be coated on the inner and /or outer tube sur hance chemical reactions The catalytic activity o can also be enhanced by increasing the porosity in as shown before in Fig 4 Multiple alternating c and insulating layers on a tube can provide a mul microcoaxial conductor or a high-density microc As stated before, the interior surface of these t can have practically any desired texture or degree ness This control is highly advantageous and all microtubes in many diverse applications For exa tical waveguides require very smooth walls, wh alytic reactors would benefit from rough walls (B the fabrication technique, the roughness of the interior can be quantified to a fraction of a micron scanning probe microscopy techniques on the m (c) (d) Figure 2 Tubes larger than 1 µm i.d can be made in any cross-sectional shape such as (a) 17-µm star, (b) 9 × 34-µm oval, (c) 59-µm smile, and (d) a 45-µm trilobal shape (a) (b) Figure 3 Tubes can be structurally sound and have (a) very thin walls or (b) thick walls Figure 4 Microtube that has a porous tube wall Figure 6 Solid nickel structure that has oriented micr Microtubes can be made straight or curved (Fig 7), or they can be coiled (Fig 8) Coiled tubes whose coils are as small as 20 µm can be used, for example, as flexible connectors or solenoid coils For the latter application, the coils could be of metal or of a high temperature superconductor where liquid nitrogen flows through the tube Another application for coils is for force sure measurement No longer are we limited to q crosprings Using microtube technology, the diam wall thickness of the tube, the diameter of the tube material, and the coil spacing can be very (a) (a) (b) Figure 5 (a) Sapphire tube that has a silver liner (b) Nickel tube that has a silver liner (b) Figure 7 Examples of curved silver tubes: (a) s (b) multiple tubes Figure 8 (a) Section of “large” coiled tube (b) Open end of coiled tube controlled to give whatever spring constant is needed for the specific application In addition, these microcoils can be made from a variety of smart materials and used as actuators or sensors For example, the length of a spring made from Nitinol ® can easily be changed by applying heat It is also possible to wrap one or more coiled spring tubes around a core tube (Fig 9) Applications for this kind of device range from a counterflow heat exchanger to a screwdrive for micromachines (For the screw application, the wrapped coil cross section could be made rectangular.) Like coiled spring tubes, bellows can be used as microinterconnects, sensors, and actuators and can be made in practically any shape imaginable Figure 10a shows a bellows that has a circular cross section, and the bellows in Fig 10b has a square cross section and aligned bellows segments The bellows in Fig 10c is square and has a twist A slightly more complex bellows shown in Fig 10d is a tapered-square camera bellows that has a sunshade to demonstrate the unique capability of this technology It demonstrates the ability to control cross section and axial shape and to decrease and increase the cross-sectional Figure 9 A coiled tube wrapped around a tube or fiber that can be used as a heat exchanger or as a microscopic screwdrive dimension in the same device Bellows fabricat crotube technology can have a variety of shaped connections to systems for use, for example, as fin exchangers, hydraulic couplings for gas and liquid mixers for multiple fluids The bellows in Fig 1 thicker transitional region and a dovetail on th connection to a device machined on a silicon wafe male dovetail to mate with this bellows is a com available trench design (68) on a silicon wafer vides a way to attach the bellows to the wafer, wh pressurized by using proper sealing (No other te available can join a fluidic coupling to a wafer for ization to relatively high pressures.) If one end of the bellows is sealed, an entirely n of applications becomes possible For example, if end is sealed, the bellows can be extended hyd or pneumatically In this configuration, a bellows used as a positive displacement pump, a valve ac for micromanipulation As a manipulator, a singl could be used for linear motion, three bellows co thogonally placed for 3-D motion, or three bello be attached at several places externally along t (Fig 11) and differentially pressurized to produc ing motion This bending motion would produce a ger, and several of these fingers would make up a h large forces and displacements possible by using nique far surpass those currently possible by ele or piezoelectric means and fulfill the need e by Wise and Najafi (18) when they stated tha area of micro-actuators, we badly need drive me capable of producing high force and high disp simultaneously.” For most applications, it is necessary to inte crotubes and the macroworld This is possible i ber of ways For example, a tapering process can in which the diameter is gradually decreased to m mensions Alternatively, the tubes and the macro be interfaced by telescoping or numerous types of ing schemes (Fig 12) An example of a thin-walled tube telescoped to a 250-µm o.d tube is shown i A tube of this type could be used as a micropitot tu course, could be made more robust by thickening 1.5 2.0 3.0 5.0 3.2 12.5 12.5 1.5 12.5 12.5 −62 −58 −58 −66 −62 −68 Below 38 Below 38 Below 38 Below 40 Below 40 Below 44 a Sensitivities reported here were obtained by exciting the transmitting transducer using a broadband and 15-ns pulse Tone burst excitation sensitivities will be 12 dB higher b For some contact transducers, a −20-dB sensitivity is assumed mode through a solid material by using AC and NC transducers Figures 8 and 9 present observations in support of this Both observations correspond to NCU transmission through 20-mm thick aluminum by 1-MHz and 20-mm active area diameter transducers in the direct transmission mode In both cases, transducers are separated from the test material surfaces by an ∼5-mm air column Furthermore, in Figs 8 and 9, the transmitting transducer was excited by a high-energy 400-volt (into 4 input impedance) pulser, and the signal received from the Figure 8 A 1-MHz noncontact transmitted signal through 20-mm aluminum by using transducers based on a soft, porous, polymer matching layer Excitation of the transmitting transducer: 400 V into 4- input impedance Receiving transducer amplification: 64 dB Transmitting and receiving transducers are 5 mm away from the material surfaces Under these conditions, the transmitted signal amplitude is 13.1 mV Compare with Fig 9 Figure 9 A 1-MHz noncontact transmitted signa 20-mm aluminum by using new transducers based pressed fiber matching layer Excitation of the transmit ducer: 400 V into 4- input impedance Receiving trans plification: 64 dB Transmitting and receiving trans 5 mm away from the material surfaces Under these cond transmitted signal amplitude is 111.1 mV Compare w receiving transducer was amplified by a 64-dB key difference is that Fig 8 was obtained by AC ers and Fig 9 by NC transducers Under these co the amplitude of the transmitted signal throug aluminum by AC transducers is 13.1 mV, whe 111.1 mV for NC transducers This clearly estab superiority of the new noncontact transducer de the other air-coupled transducers described in Figure 10 A 1-MHz noncontact transmitted signa 20-mm aluminum by using new transducers based on c fiber matching layer Excitation of the transmitting t one burst 16-V sine wave Receiving transducer am 64 dB Transmitting and receiving transducers are 5 from the material surfaces Under these conditions, the ted signal amplitude is 3.26 mV No other air-coupled t can transmit ultrasound under similar conditions in hig impedance materials by very low level excitation Figure 11 Ultrasound transmission through 25-mm carbon steel (Z, 51Mrayl) by using the new noncontact transducers (1-MHz, 20-mm active area diameter) with a single burst of 16-volt sine wave excitation and 64-db amplification of the received signal by an equivalent transducer No other air-coupled transducer can transmit ultrasound under similar conditions in high acoustic impedance materials by very low level excitation including this author’s air/gas propagation transducers that are commercially available from Ultran Laboratories since 1983 To demonstrate further the exceptionally high sensitivity of NC transducers, we decided to conduct an experiment that would normally be considered impossible! An experiment analogous to that described before was performed, except that in this case the NC transmitter was excited merely by a single burst of a 16-volt sine wave and 64-dB amplification of the received signal Figure 10 presents the observation from this, showing a 3.26-mV signal transmitted through 20-mm aluminum in the noncontact mode AC transducers (based on soft polymer matching layers, that is, porous or nonporous or have hollow spheres) were unsuccessful in generating ultrasound transmission through 20-mm aluminum by 16-volt excitation, despite high signal averaging! It is also interesting to note that by low energy excitation, using NC transducers, we propagated megahertz frequencies even in steel, whose acoustic impedance is 51 Mrayl, six orders of magnitude higher than in air! We show an example of this startling conclusion in Fig 11 It is important to note that the purpose of observations shown in Figs 8–11 was to demonstrate the high sensitivity of NC transducers relative to any other similar device The purpose of these experiments is not to recommend or suggest the usage of low-energy transducer excitation for testing materials NONCONTACT ULTRASONIC ANALYZER As is evident from the preceding sections, NC transducers can be used with any suitable commercially available pulser-receiver to transmit and to detect ultrasound through any material However, our ultimate goal was to generate an NCU mode that would rival the performance Figure 12 Ultrasonic noncontact analysis system 1000, using noncontact transducers and a monitor scre ing the thickness and velocity through the test materi of conventional contact or immersion ultrasound transduction of NC transducers is not enough plish this task For example, we still need to ove natural barrier of acoustic impedance mismatch the coupling air and the test medium From Table that losses due to this mismatch are formidable T vent this and not jeopardize our objective of equa performance with that of conventional contact-ba sound, a new mechanism of transducer excitatio nal amplification was needed Nevertheless, this impossible task, too, was overcome In 1997–199 ultrasonic system was conceived and produced as the NCA 1000,2 this instrument was develope Vandervalk and Ian Neeson of VN Instruments, C is based on synthesizing a computer-generated c bined with the best attributes of noncontact tra Signal processing in the NCA 1000 is done by aperture imaging techniques The NCA 1000 is c ized by a dynamic range of >150 dB and a tim (TOF) measurement accuracy of ±10 ns in am and better than ± lns in closed conditions The N (Fig 12) measures the TOF, thickness, velocity, grated response (area underneath transmitted o signals in dB) of materials in the time domain the FFT mechanism of this system, it is also p conduct noncontact ultrasonic spectroscopy Fur by raster scanning the transducers or the test we can generate surface or internal images of th terial Such images can be representative of the surface roughness, TOF, transmission attenuat city, or thickness REFLECTION AND TRANSMISSION IN NONCONTACT MODE Analogous to conventional contact or immers sound, ultrasound in the non-contact mode is also and transmitted at various interfaces as well a a test medium In this section, we provide ex 2 U.S patent pending and in process various paths of ultrasound reflection and transmission as functions of test material’s interfaces and its volume Non-contact transducers T Single-Transducer Operation (Pulse-Echo) Separate Transmitter and Receiver Operation on the Same Side (Pitch-Catch) By using two noncontact transducers, one a transmitter (T) and the other a receiver (R), on the same side of the test material, (Fig 14), it is possible to launch and measure the characteristics of longitudinal, shear, and surface waves in practically all types of material Generation of these waves Ambient air Surface wave Longitudinal and/ or shear waves Figure 14 Schematic of a setup for same side operati rate transmitting (T) and receiving (R) transducers, als various types of wave propagation in the test material is determined by Snell’s law, Sin i/Sin r = Va /Vm , where i is the incident angle in air, r the refraction the test material, Va is the ultrasound velocity in Vm the velocity in the test material By manipulating the incident angle in air, a wave types can be produced in a test material F shows the far side thickness reflection of a lon 0.08 REL AMP (power units) By operating one transducer simultaneously as a transmitter and receiver (analogous to the pulse-echo technique), it is very easy to generate reflection from an air–material interface due to the extremely high reflection coefficient at this interface However, in this mode, it is nearly impossible to produce a far side reflection corresponding to the test material thickness in ambient air This difficulty stems from several factors, such as the extremely small transmission of ultrasound in the test material, the extremely high beam spread on the surface of the material, and the inherent electrical noise associated with single transducer operation from the initial pulse To a degree, the adverse effects of these factors can be minimized by a focused transducer, which will reduce the beam spread and focus ultrasound (thus intensify the reflected energy) within the test material Figure 13 shows pulse-echo reflection signals from a 9 mm thick silicone rubber sample This observation was generated by using a 1-MHz focused NC transducer Similar results have also been observed for other plastic materials However, at this time, we have no concrete proof of generating these observations from high acoustic impedance materials, such as metals and ceramics Because reflection from an air–material interface is extremely strong in single transducer operation, ultrasonic reflectivity can be used to characterize the surface characteristics of the material Such applications include surface acoustic impedance, surface roughness, particle size measurement, surface texture and microstructure, distance, proximity, and level sensing, and any other surface conditions that are sensitive to noncontact ultrasound R Peak @ 106 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 60 70 80 90 100 110 Time (microsec) 120 Figure 15 Thickness reflection of longitudinal wave tion in 12-mm thick aluminum per Fig 14 setup Di tween the transducers and the material surface: ∼15 ent air Distance between T and R: ∼50 mm Incident a ∼3.0◦ Figure 16 Thickness reflections of longitudinal and shear waves in 32-mm thick transparent polystyrene per Fig 14 setup Distance between the transducers and the material surface: ∼13 mm ambient air Distance between T and R: ∼50 mm Incident angle in air: ∼6.0◦ wave from 12-mm thick aluminum Figure 16 shows the far side thickness reflection of longitudinal and shear waves from 32 mm thick transparent polystyrene Figure 17 is a longitudinal wave refracted surface wave in aluminum produced by an incident angle equal to the first critical angle (i.e., total reflection of a longitudinal wave) which is 3.16◦ for aluminum Figure 18 shows a shear wave refracted surface wave in aluminum, generated by an incident angle equal to the second critical angle (i.e., total reflection of a shear wave), which is 6.3◦ for aluminum It is important to note that while performing these experiments, distances— corresponding to transducers and the test materials and angles of transmitting and receiving transducers—were not measured accurately The primary function of this exercise is to show the feasibility of various types of bulk and surface wave generation by noncontact ultrasound Applications of such measurements include NCU evaluation of materials from one side, defect detection, anisotropy measurements, and relationships of ultrasonic velocities to test material elastic and mechanical properties flection of shear wave when the incident angle is eq second critical angle Setup for this is shown in Fig 1 between the transducers and the material surface: ∼15 ent air Distance between T and R: ∼100 mm Incident a ∼6.5◦ Direct Transmission When a test material is inserted between two n transducers facing opposite each other in air, th sound is transmitted and reflected from all inter responding to air and the material Details o shown in Figs 19 and 20 Direct transmission is the easiest technique in noncontact ultrasoun fore, it has been quite extensively studied and d Applications of this technique are numerous: and velocity measurements, defect detection, tex T1 tc 0.2 Rel amp (power units) t1 Air 2tm Test material 0.16 0.12 Air 0.08 0.04 0 100 t2 Longitudinal surface wave @ 202.7µs T2 125 150 175 200 Time (microsec) 225 250 Figure 17 Surface wave in aluminum produced by total reflection of longitudinal wave when the incident angle is equal to the first critical angle Setup for this is shown in Fig 14 Distance between the transducers and the material surface: ∼15 mm ambient air Distance between T and R: ∼100 mm Incident angle in air: ∼3.2◦ Figure 19 Propagation of ultrasound in the direct tr noncontact mode Here, tc is the complete time of flight responding to the propagation of ultrasound in air a material, 2tm is the round-trip TOF through the tes thickness, t1 and t2 are, respectively, TOFs from tra to the top surface of the material, and transducer T2 to surface of the material For the further significance Fig 20 0.005 0 100 125 150 175 200 225 Figure 20 Direct transmis contact ultrasound propagatio a test material (7-mm thick c Fig 19 setup 250 Time (microsec) microstructural evaluation, transmission, velocity, thickness, and TOF imaging, detection of the presence or absence of liquids in containers, and many more metals, and fibrous and particulate materials As the magnitude of transmitted signals through creases substantially when examined under high sures Figure 22 shows the propagation of 4-MHz sion through 8-mm thick steel under 60 bars ni using a single transducer in the pulse-echo techn VERY HIGH FREQUENCY NCU PROPAGATION IN MATERIALS NONCONTACT ULTRASONIC MEASUREMENTS By using NC transducers in ambient air, we have amply demonstrated that frequencies as high as 2 MHz can be easily propagated through a variety of materials, including fibrous and particulate, plastics, ceramics, metals, and composites However, frequencies even higher than 2 MHz have been successfully investigated for propagation through solids Figure 21 shows an example of 4-MHz propagation through 4.5-mm thick multilayer graphite fiber-reinforced plastic composite in ambient air Similar observations have also been made for soft polymers, thin Because ultrasound can be reflected and tra through a test material and its surfaces, one ca respective signals to make significant measure the time and frequency domains Such measurem be further related to important test material c istics, such as velocity, thickness, defects, inte surface texture or microstructure, and other ult dependent parameters 1 0.9 Rel amp (power units) 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 10 11 12 13 14 15 16 Time (microsec) 17 18 19 20 Figure 21 Noncontact u direct transmission throug thick multilayer graph reinforced plastic composit in ambient air per Fig Transducers are 4-MHz an active area diameter Dista transducers to test mate faces are ∼3 mm First pea transmitted signal throug the test material Subsequ test material thickness refl its multiples tance from transducer to the surface of test material is ∼25 mm First peak: gas– materials interface Subsequent peaks: thickness and thickness multiples corresponding to the test material Velocity and Thickness Measurements the attenuation, the lower the acoustic impedan thicknesses greater than one wavelength in the terial, material thickness reflections are obser the other hand, when only the transmission sig served (i.e., without thickness multiples for at materials), one can determine the TOF of the kno ness of the test material similarly to using con line transducers In such a case, the solid dela the transducers are replaced by air columns in fr transmitting and receiving transducers as funct distances between the test material surfaces As seen in the aforementioned technique, to d the test material velocity, its thickness must b However, in the NCU mode, the thickness of th terial can also be measured For simultaneous ments of material thickness and its velocity in mode, we must examine all paths of ultrasound sion and reflection to and from the test mater paths relate to propagation of ultrasound relativ mitting and receiving transducers “taking to each the air column, ultrasound transmitted throug There are several ways to determine longitudinal wave velocity in test materials by noncontact ultrasound For example, if multiple reflections corresponding to the thickness of the test material are observed (Fig 23), then one can measure the TOF, tm , between the two successive peaks to determine the velocity of a material of known thickness The TOF measured this way corresponds to the round-trip TOF in the test material Therefore, Vm = 2dm /t m (8) For example, tm measured between any two successive peaks from Fig 23 is 10.4 µs for a 13.5-mm thick material (in this case, isotropic graphite); thus, the velocity is 2595 m/s It is important to note that the appearance of multiple thickness reflections in the NCU mode depends on the attenuation and acoustic impedance of the test material and the frequency of transducers For example, the lesser 0.02 0.018 Figure 23 Velocity in material determined by direct transmission NCU propagation in a test material (in this case, 13.5-mm thick isotropic graphite) characterized by multiple thickness reflections Note that the TOF (round-trip TOF through the material) between any two successive peaks is 10.4 µs; thus, the velocity in the material is 2595 m/s Rel amp (power units) 0.016 Transmission peak @ 61.9 µs 1st thickness reflection peak @ 72.3 µs 0.014 0.012 2nd thickness reflection peak @ 82.7 µs 0.01 3 rd thi reflect @93.1 0.008 0.006 0.004 0.002 0 50 55 60 65 70 75 80 Time (microsec) 85 90 material, and ultrasound reflections from the test material surfaces in air These paths of ultrasound propagation in the NCU transmission mode—needed to determine the test material thickness and velocity—are shown in Fig 24 The signals generated by these paths of propagation and their significance are as follows: Path (a) is the transmission from transducer 1 to transducer 2 in air—measures TOF, ta If ultrasound is propagated from transducer 2 to transducer 1, the same TOF is measured Path (b) is the reflection from transducer 1 to the material surface in air—measures TOF, t1 Path (c) is the reflection from transducer 2 to the material surface in air—measures TOF, t2 Path (d) is the transmission from transducer 1 to transducer 2, and the test material is in between—measures TOF, tc If ultrasound is propagated from transducer 2 to transducer 1, the same TOF is measured From these times of flight measurements, the test material thickness and its velocity are determined according to the following relationships: Vm = dm /t m (9) dm = Va tam (10) t am = ta − (t1 + t2 )/2 (11) t m = tam − (ta − tc ) (12) In these equations, dm is the test material’s thickness, Vm , the velocity of ultrasound in the test material Va , the velocity of ultrasound in air (determined from a reference material), tam , the time of flight in air corresponding to the test material thickness dm , and tm is the time of flight in the test material By proper substitutions, dm = Va [ta − (t1 + t2 )/2] (13) Vm = dm /tam − (ta − tc ) (14) As an example, Fig 20 shows actual transmitted and reflected signals when a test material is examined in the noncontact transmission mode Identification and location of these signals with respect to the test material are shown in Fig 19 As can be seen from Eqs (13) and (14), to determine the thickness and velocity according to this scheme, one only needs the measurements of four times of flight (ta , t1 , t2 , tc ) and the velocity of ultrasound in air These parameters were measured and calculated by the NCA 1000 and are displayed with the velocity and thickne test material on the monitor screen (Fig 25) Integrated Response, Transmissivity, and Reflectivity Measurements In the time domain, the NCA 1000 measures and the times of flight of the signals and also shows grated response (IR) of these signals IR is a mea of the area underneath a particular peak in power Due to the very high, >150-dB, linear dynamic ra NCA 1000, the IR can be used to measure the a ultrasonic energy transmitted (transmissivity) or (reflectivity) from a test material and relate it (IR changes in the material For example, IRm = IRc − IRa , where IRm is the amount of ultrasonic energy ted in the test material, IRc is the ultrasound sion through air and the material (between the ting and receiving transducers), and IRa is the a ultrasound energy transmitted only through air IRa are measured directly by the NCA 1000 If a given test material is homogeneous, then surement of IRm , it has been found, is related to t mission coefficient [Eq (1)] To illustrate this, we e a flat polished specimen of polystyrene Figure the IRc (−21.72 dB) of the transmitted peak of u from air into the specimen, and Fig 27 shows peak, but only through air, corresponding to IRa (+ thus yielding IRm −63.42 dB for the specimen portant to note that this measurement correspo closely to I Rm = 20 log T, where T, the transmission coefficient, is defined b For example, the calculated value for ultrason transmitted in polystyrene [Eqs (1) and (4)] is − which is very close to −63.42 dB determined by m integrated response peaks It should be pointed ou transmission coefficient is assumed to be independ trasonic attenuation and the thickness of the test In reality, this is not absolutely true For example thicknesses of polystyrene samples at different fre yield different values of T Though these varia Figure 25 The NCA 1000 screen displaying velocity of ultrasound and thickness of a material The test material is a 22.5-mm porous sintered ceramic very small, yet they are measurable On the other hand, if the transmission coefficient can be measured with a very high degree of certainty and precision, then it should also be possible to measure the absolute density of the test material by first determining the acoustic impedance Zm of the test material: Z1 [(2 − T) + 2(1 − T)1/2 ], T ρm = Zm /Vm Zm = (17) (18) density of the test material For accurate determ these characteristics, factors such as ultrasound tion (analogous to absorption coefficient in X-ra tion) and material thickness must also be consid T = I2 /I0 = Z1 Z2 /(Z1 + Z2 )2 = exp (−µρx), where T is the transmission coefficient, I2 the u energy transmitted into the material (acoustic im Rel amp (power units) Measurement of I Rm and solving Eqs (17) and (18) provide approximate ideas about the acoustic impedance and Rel amp (power units) 0.06 IRc = −21.72 dB 0.05 0.04 0.03 0.02 0.01 50 45 40 35 30 25 20 15 10 5 0 IRa = +41.7dB 20 0 20 25 30 35 40 45 Time (microsec) 50 55 60 Figure 26 Noncontact transmission through a 20-mm thick flat polished polystyrene sample showing the integrated response IRc of the transmitted peak 25 30 35 40 45 Time (microsec) 50 Figure 27 Noncontact transmission through air col ing the integrated response IRa of the transmitted that while making this measurement, the distance b transmitting and receiving transducers was compensa 20 mm thickness of the specimen in Fig 26 Figure 28 Procedure for noncontact ultrasonic spectroscopy Top: FFT magnitude spectrum of ultrasound transmission through air as a reference Bottom: FFT magnitude spectrum of ultrasound transmission where the test material (composite rubber) is between the transmitting and receiving transducers in air −60 Attenuation (dB) −62 −64 −66 −68 −70 300 sonic velocity in low-density green alumina Transduc 12.5-mm active area diameter Z2 ) from air (acoustic impedance Z1 ), I0 t ultrasound energy, µ the material ultrasound att coefficient, ρ the material density, and x the mater ness At this time, the development of these rela and techniques for measuring T, Z, µ, and ρ by n ultrasound are in progress (21) Because the measurement of the transmissi cient still needs to be validated, it is best to refer transmissivity (when propagating ultrasound th material in the direct transmission mode) or re (when ultrasound is reflected from the surface o terial) Such measurements are useful in evalu test material’s internal and surface characteris as defects, texture, microstructure, and roughnes Noncontact Ultrasonic Spectroscopy 500 700 900 1100 1300 1500 1700 1900 Frequency (kHz) Figure 29 Frequency dependence of ultrasonic attenuation by subtracting the air reference from that of the sample spectrum (Fig 28) By performing the fast Fourier transformation transmitted or reflected time domain signals, tes als can also be characterized to investigate the f dependence of ultrasonic attenuation Such exam are important while testing microstructurally materials or those for which time domain velo surements are not sensitive enough The first 12000 y = 3472.4x - 2533.2 R2 = 0.9787 11000 Velocity (m/s) 10000 9000 8000 7000 6000 5000 4000 2.5 2.7 2.9 3.1 3.3 3.5 Density (g/cc) 3.7 3.9 4.1 Figure 31 Relationship betw sity and noncontact velocity i alumina Transducers: 1 MHz f less dense than 3.5 g/cc and samples denser than 3.5 g/cc crostructure, and roughness 1.75 1.75 1.8 1.85 1.9 1.95 2 2.05 Physically determined density (g/cc) 2.1 2.15 APPLICATIONS OF NONCONTACT ULTRASOUN Figure 32 Comparison of ultrasonically and physically determined density of green alumina Noncontact transducers have now been success duced in the frequency range of 100 kHz to 2650 Velocity (m/s) 2600 y = 2734x - 4285.7 R2 = 0.9517 2550 2500 2450 2400 Figure 33 Relationship between velocity and density for isostatically pressed, high-density green alumina Transducers: 1-MHz, 12.5-mm active area diameter 2350 2.44 2.45 2.46 2.47 2.48 2.49 2.5 2.51 2 Density (g/cc) 1100 Velocity (m/s) 1000 y = 232.01x - 1120.7 R2 = 0.9949 900 800 700 600 Figure 34 Relationship between velocity and density for pressed green tungsten carbide Transducers: 500-kHz, 12.5-mm active area diameter 500 7.00 7.25 7.50 7.75 8.00 8.25 8.50 Density (g/cc) 8.75 9.00 9 0.06 0.05 0.01 0 110 115 120 125 130 135 Time (microsec) 140 145 150 Rel amp (power units) 0.07 Rel amp (power units) 0.02 95 100 0.06 120 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 50 60 70 Time (microsec) 80 50 60 70 Time (microsec) 80 0.05 0.04 0.03 0.02 0.01 0 110 105 110 115 Time (microsec) Figure 36 Defect detection in a sample of 20-mm porous, low thermal expansion sintered ceramic Top: u transmission through a defect-free region Bottom: u transmission through a region that has 1.5-mm diam drilled cylindrical hole Compare the amplitudes of the ted ultrasound intensity of the two regions Transduce and 12.5-mm active area diameter 0.04 0.03 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 90 115 120 125 130 135 140 145 150 Time (microsec) Figure 35 Defect detection in a sample of 14-mm thick green porcelain Top: ultrasound transmission through a defect-free region Bottom: ultrasound transmission through a region that has 1.5-mm diameter side-drilled cylindrical hole Compare the amplitudes of the transmitted ultrasound intensity of the two regions Transducers: 1-MHz and 12.5-mm active area diameter Rel amp (power units) Rel amp (power units) 0.07 Rel amp (power units) and physically determined densities of green alumina The noncontact ultrasonic technique has been successfully applied to characterize density and defects in a variety of green materials such as ceramics tiles, multilayer electronic packages, powder metals, cements, and concrete Figures 33 and 34, respectively, show the velocity–density relationships for isostatically pressed high-density green alumina and tungsten carbide Examples of defect detection in green and sintered ceramics are shown in Figs 35 and 36, and similar observations for aluminum are shown in Fig 37 Figures 38 and 39, respectively, show trend plots of direct transmission and same side reflection in graphite fiber-reinforced plastic (GFRP) composites bonded to a honeycomb structure The same side observations (Fig 39) 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 Figure 37 Detection of defects in an 8-mm thick sh minum by noncontact transmission mode Top: defect-f Bottom: region that has a 1.5-mm cylindrical hole Tr 2-MHz and 12.5-mm active area diameter Figure 38 NCA 1000 trend plot showing direct transmission through a GFRP composite bonded to a honeycomb structure at various points Regions showing the sharp drop in transmission are indicative of defects, such as delaminations Transducers: 500-kHz and 12.5-mm active area diameter, separated from the material surfaces by ∼40 mm ambient air correspond to the bond between the GFRP and the honeycomb Examples of textural and microstructural analysis by noncontact ultrasonic spectroscopy are shown in Figs 40 and 41 Figure 40 shows the frequency dependence of ultrasound attenuation by three specimens of extremely porous ceramics (in this case, space shuttle tiles), and Fig 41 shows similar observations from two samples of packaging foam whose cell dimensions vary Figure 42 shows very high frequency noncontact transmission spectroscopy of two samples of paper towels Similar observations have also been made to detect bubbles and pores in li other materials To evaluate surface characteristics by noncon sonic spectroscopy, several grinding discs of SiC v particle size were chosen These discs were placed air distance of 10 mm from a 2-MHz, 12.5-mm a diameter transducer Reflection from a polished steel was assumed as a reference signal Reflect from the reference and test materials were analyz forming FFT, and the frequency dependence of u was measured by subtracting the sample FFT spe Figure 39 NCA 1000 trend plot showing the T–R reflection from same side in a GFRP composite bonded to a honeycomb structure at various points Regions showing the sharp drop in reflected ultrasound from composite to honeycomb bond are indicative of defects, such as delaminations Transducers: 500-kHz and 12.5-mm active area diameter, separated from the test material surface by ∼10 mm ambient air Figure 40 Noncontact transmission ultrasonic spectroscopy of extremely porous ceramics (space shuttle tiles) for microstructural characterization Top: 0.38 g/cc Middle: 0.28 g/cc Bottom: 0.1 g/cc Transducers: 250-kHz, 25-mm active area diameter −40 1400 1600 1800 2000 2200 Frequency (kHz) Figure 43 Noncontact reflection ultrasonic spectrosco face characterization of materials Note that as the su come rough, the attenuation of ultrasound increases T 2-MHz and 12.5-mm active area diameter 0 Ultrasound reflectivity (%) Attenuation (dB) −20 −40 −60 −80 −100 160 200 240 280 320 Frequency (kHz) 360 400 Figure 41 Noncontact transmission ultrasonic spectroscopy of packaging foam Top: small cell Bottom: large cell Transducer: 250-kHz and 25-mm active area diameter 100 90 80 70 60 50 40 30 20 10 0 y = −0.752x + R2 = 0.99 0 20 40 60 80 100 1 Particle size (micron) Figure 44 Noncontact reflectivity measurement for su acterization of materials Compare the reduction in and high ultrasound attenuation (Fig 36) as functions ing particle size Transducer: 2-MHz and 12.5-mm a diameter −10 Attenuation (dB) −20 −30 −40 −50 −60 1700 1950 2200 2450 2700 2950 3200 3450 3700 3950 4200 4450 Frequency (kHz) Figure 42 Very high frequency noncon mission ultrasonic spectroscopy of two paper towels Top: 0.2-mm thick, relati shallow dimpled texture Bottom: 0.4relatively soft and deeply dimpled textu Figure 45 Partial contact ultrasonic image of an impactdamaged eight ply graphite-fiber-reinforced plastic composite 1.5 mm thick Area scanned: 25 × 25 mm This image was generated by using a 2-MHz and 25-mm active area diameter transmitter in contact with the material surface, and a 2-MHz, 3-mm diameter receiver was raster scanned in noncontact mode across the other surface of the material A conventional pulser that sent 400 V into a 4- input impedance square wave and 64-dB receiver gain was used that of the steel reference (Fig 43) It is quite evident that as the particle size increases, the frequency-dependent attenuation also increases A similar experiment was performed in which the integrated response (IR) of reflected ultrasound from the test and reference materials surfaces was measured The reflectivity of ultrasound (as a function of particle size) was determined by comparing the sample an impact-damaged 1.5-mm thick multilayered reinforced plastic composite The test materia case was placed on a large stationary transmitt ducer, and a small receiver in the noncontact scanned across the other surface of the material shows noncontact transmission images of a th impact-damaged glass-fiber-reinforced plastic by monitoring signals corresponding to transmi thickness reflection through the material To dem the analytical ability of the NCA 1000 system a thick sample of an iron powder compact wa by monitoring the transmission integrated resp the material velocity These images are shown i Figure 48 shows an image of defects in alumin lar images for materials, such as steel welds, fi cheese, meats, wood, and other materials have erated by using the NC transducers with the N and other commercial instruments Because the NCA 1000 interprets ultrasound r of both the transmitting and the receiving tr from material surfaces, it is now also possible to Figure 46 Noncontact ultrasonic imaging of a mild-impact-damaged 6.4-mm thick multilayered glass-fiber-reinforced plastic composite by using NCA 1000 and 1-MHz, 12.5-mm active area diameter transducers with a 1-mm aperture Left: Direct transmission image Right: First thickness reflection image −69.2188 −68.4375 −67.6563 −66.875 −66.0938 1470 1520 1570 1620 1670 Figure 47 Noncontact imaging of a green 14-mm thick iron powder compact by using the NCA 1000 and 500-kHz, 12.5-mm active area diameter transducers with a 3-mm aperture in direct transmission mode Left: Relative attenuation of integrated response (dB) Right: Velocity (m/s) Area scanned: 50 × 50 mm Note that the outer high velocity region is also characterized by high attenuation (low IR) and the inner region of low velocity by lower attenuation the thicknesses of materials that are continuously rolled on a production line and are too wide for micrometers Food, Beverage, and Pharmaceutical 0.4 Rel amp (power units) Figure 49 shows transmitted noncontact ultrasound signals from regions with and without almonds in milk chocolate Figure 50 is an example of fat content measurement in milk and milk products Similarly, Fig 51 s measurement of sugar content in water We h applied noncontact ultrasound to detect bever other liquids in plastic, metal, and cardboard co The quality of heat and vacuum seals in pharm 100 90 80 70 0.3 0.2 0.1 0 300 60 310 320 330 340 350 Time (microsec) 360 310 320 330 340 350 Time (microsec) 360 40 30 20 10 0 0 10 20 30 40 Figure 48 Noncontact ultrasonic image of a 8-mm thick aluminum sheet in transmission mode showing 1.5-mm (top) and 2.0-mm (bottom) side-drilled cylindrical holes Also compare with Fig 37 Transducers: 2-MHz and 12.5-mm active area diameter Image provided by E Blomme, Katholiieke Hogeschool, Kortrijk, Belgium Rel amp (power units) 50 3 0.4 0.3 0.2 0.1 0 300 Figure 49 Detection of absence or presence of almon chocolate Top: region without almonds Bottom: regio almonds Transducer: 1-MHz, 12.5-mm active area dia ... 70 80 90 10 0 99 .7745 99 . 012 1 0. 099 5 0.7 411 1. 91 4 3 3.2 316 4.5225 5.7424 6.8786 7 .92 31 8.8866 9. 715 0 0 .10 02 0.7440 1. 92 30 3.2475 4.5434 5.7 695 6 .90 83 7 .95 97 8. 91 9 0 9. 7835 of applications of ANN to... Brittain, and G.M Whitesid and Actuators A 72: 12 5? ?13 9 ( 19 99) Soc 11 2: 897 6– 897 7 ( 19 90) 30 C.R Martin, Science 266: 19 61? ? ? 19 66 ( 19 94) 31 J.C Hulteen and C.R Martin, J Maters Chem 7: 10 75? ?10 87 ( 19 97)... Chem 19 0: 17 27? ?17 35 ( 19 89) 18 J Mathew-Krotz and K.J Shea J Am Chem Soc 11 8: 815 4– 815 5 ( 19 96) 19 S Marx-Tibbon and I Willner J Chem Soc Chem Commu 12 61? ? ?12 62 ( 19 94) 20 D.K Robinson and K Mosbach