·Home ·Table of Contents ·Methods and Instrumentation

Electromagnetic Acoustic Transducers (EMATs) Design Evaluation of their Performances S. ALIOUANE, M. .HASSAM, A. BADIDI BOUDA & A. BENCHAALA Centre scientifique et technique en Soudage et Contrôle B.P.64, Route de Dely Ibrahim, Chéraga, Algiers ALGERIA Tél/Fax: 213.2.36.18.50 e-mail: s_aliouane@hotmail.com,abadidi@hotmail.com Contact

Abstract:

The non-destructive testing (NDT) by ultrasounds is increasingly asked by manufacturers in the framework of the products quality insurance and preventive maintenance. It offers ,as compared to the radiography many advantages among others: the security, the flexibility and the cost. However, this technique requires some precautions during a testing, namely :
• the use of a couplant.
• the necessity to have a good state of surface,
• The method involve mechanical contact between the power source and the pieces tested.

To improve the flexibility and the rapidity of this method, it would be useful to generate and receive acoustical waves with a contactless electromagnetic probe. The contactless electromagnetic-acoustic transducers (EMATs) offer several advantages, which have an essential importance for the non-destructive testing with ultrasound
This work consists of EMATs conception and construction for the contactless excitation and detection of different wavetypes. We have conceived and realised receivers and transmitters EMATs.
This paper is a report of experiments results with these different types of EMATs. We examine the different parameters that influence them in order to optimise them and to know the degree of the applicability of this technique in the non-destructive testing of materials in Algeria.
keywords: EMATs, Electrodynamics effect,Magnetostriction, phased array transducer.

Introduction

Besides piezoelectric, an other physical effects can be used for generating and receiving ultrasonic waves. It is the electromagnetic acoustic effect [7]. Although it produces weaker signals than obtainable by the piezoelectric effect, it nevertheless offer a number of advantages. In this effect the energy is transmitted by electromagnetic fields which avoid mechanical contact with the test piece. The conversion into or from acoustic energy takes place in the surface of the work piece concerned. Compared with the piezoelectric oscillator, which is coupled to the work piece, the surface of the work piece forms in the case of this direct method a part of the acoustic transducer. This method thus require no coupling medium, so avoiding a number of difficulties. Indeed, wet coupling can introduce various disturbances [1], [2]. Due to the interference of the waves reflected at the two interfaces, the permeability of a liquid layer depends to a great extent on the thickness. Consequently, it is necessary to reduce the thickness of the liquid layer to a minimum and to keep it constant. When testing hot work pieces, the difficulty of finding a suitable couplant increases with the temperature. Finally, the electromagnetic acoustic transducer not contains mechanically oscillating components[13]. These components have a natural frequency influencing the overall frequency response.

Principles of electromagnetic acoustic excitation -Basic theory

Fig 1: Principle of electromagnetic acoustic generation

The electromagnetic acoustic excitation of ultrasound is based upon two physical mechanisms: the Lorentz forces and the magnetostriction in ferromagnetic materials.
The theory we expose here describes the acoustic electromagnetic phenomenon only using Lorentz forces. A coil, placed near the surface, transmits an r.f electromagnetic signal (Figure1). The response of the metal takes place in the classic skin depth d , where eddy current j(z, t) and their associated electrical field E(z, t) are induced.
When we apply a static magnetic field, we disturb the balance giving birth to a force that acts on electrons. It is the Lorentz force given by the following relation :

(1)

where n₀ is the electrons density. For B₀ oriented according Ox the Lorentz force is longitudinal, resulting in a variation of the electron charge density along Oz. An internal electrical field E(z) oriented following Oz has to exist to maintain charge neutrality. Its ions come in an compressional oscillations of amplitude x _z. B₀ is to Oz, the currents j(z, t) are to Oy direction and the Lorentz force act in the Ox direction. The Lorentz Forces act to produce a shearing force along Ox arise shear oscillations of amplitude x _x . Thus for these two polarisations , the acoustic wave equation is :

(2)

d is the density of the metal and s the speed of ultrasonic wave

Dobbs [6] demonstrates that when d <<l and at distances z >> d we have the following wave solution: ,where b = q²d ²/2 and q is the ultrasonic wave number

The conversion coefficient h of the electromagnetic acoustic transducer is the ratio of the generated acoustic power P to the electromagnetic power Q entering the surface. It is done by the following relation[11], [12].

(4)

At 10MHz, for compressional waves in aluminium, B₀= 1T we have h = 5.7x10^-5 .
In non- ferromagnetic materials the Lorentz force is the one contribution to the generation of sound. In a ferromagnetic material, additional forces are produced by magnetostrictive stresses. Ultrasonic wave is no more a linear function of the applied magnetic field. According to these principles an EMAT then consists of a means for producing a static bias field, i.e., a permanent magnet or electromagnet, plus a coil of wires carrying a dynamic drive current.

Realisation

EMATs prototypes realised, are all set in a metallic compact cylindrical case. Electromagnet is installed there. The electromagnetic insulation is completed by the connection USK7-coil by a BNC-LEMO cable . The electromagnet is shod with a sole made of Plexiglas which fixes the electromagnet yoke and the exciting coil (figure 2,3).

Fig 2: EMAT prototype

Fig 3: EMAT and Piezoelectric Transducer

Exciting and receiving coils are realised on printed circuits. The electromagnet used have U-form for longitudinal waves EMATs , and E-form For shear waves.

Experiments

Influence of the excitation coil geometry
We want to determine the coil geometry influences on the efficiency for the emission and its sensitivity to the detection. For this purpose, several set of coil of different designs are tested: single flat spirals, double spirals and meander coil. Four sets of flat spirals was fabricated as shown in figure 4

Fig 4: spiral coil as printed circuit

Fig 5: meander, double and single coil for 21mm gap

We have therefore conceived a set of different characteristic coil (number of turn, distance between turns). Coils were printed on the same rectangles surface 35 X 11 mm² (Fig.4). We were limited by 9 turns. Small width of the gap plays an important role in generation phenomenon. Indeed, the developed static field is as important as the gap is smaller. Thereby, both influence of these two parameters namely: the turn number and the gap width, present a contradictory character. Figure 5 shows an other set of coils: double spiral, single and meander coil. These coils are destined for operate in a 21mm-gap while precedents operated in a 14 mm-gap. Table 1 gives the best results obtained by each category of coil.

Coil type		6 turn-single Spiral	2 x 3 turn-double spiral	8 period-meander	9 turn-single Spiral	10 turn-single Spiral
Gain (dB)	transmission	85.5	90	99	90	86
	reception	105	108	110	100	102
Rejection (%)	transmission	0	0	22	5	0
	reception	35	40	40	40	40
Width of gap (mm)		14	21	21	14	21
Table 1: coil geometry influence on the echo high


Influence of the static magnetic induction
Figure 7 shows that for the case of aluminium, the transmitting and receiving height echo is practically linear with the electromagnet current . We observe the same aspect in copper and in stainless steel. Unlike the ferromagnetic steel (figure 6). indeed, amplitude of the longitudinal wave, increases rapidly, reach a saturation around a low intensity (I=0.05 A) and remains constant with increasing current.

Fig 6: magnetic induction influence in ferromagnetic steel

Fig 7: magnetic induction influence in aluminium

Influence of the lift off
Influence of the lift-off (coil-material surface) is investigated on the shear and longitudinal echoes magnitudes. For both polarisation, the height of transmitting and receiving echoes are inversely proportional to the lift-off value. We obtained exponential tendency curves (figure 8, 9) with a determination coefficient R²=0.995.

Fig 8: lift off influence in aluminium

Fig 9: lift off influence in stainless steel

Optimisation of parameters

Static magnetic field
We established the table 2 where are data optimal intensity values.

Material	mode	Optimal Induction (Ampère)
		Transmission	Reception
Ferromagnetic steel	longitudinal	0.05	0.06
	Shear	0.05	0.02
Aluminium	Longitudinal	0.21	0.21
	Shear	0.21	0.21
Copper	Longitudinal	0.21	0.21
	Shear	0.21	0.21
stainless steel	Longitudinal	0.21	0.21
	Shear	0.21	0.21
Table 2 : optimal induction

Curves (Fig.6, and 7) representing the evolution of the amplitude according to the magnetic induction indicates us that for non ferromagnetic materials, the efficiency of the transduction is linear with the static magnetisation. So, It is imperative to dope electromagnet to the maximum by providing a strong electrical current in the other hand, it is useless to increase the current for ferromagnetic materials especially steel since the response reaches the saturation for the longitudinal mode at relatively low intensity and around the same intensity for the shear mode.

Coil Design
The 6 turn-spiral gave the best result in transmission, nevertheless, it is not only coil form that causes good efficiency recorded. Indeed, the gap in where coil is placed also is a parameter that affects significantly the dispersion of magnetic field lines. The six-turn-coil is printed on a 14 mm-gap while the meander occupies 21mm width. The dispersion of the magnetic field lines increases with the gap width and then diluting the static magnetic power provided by the electromagnet current.

The meander is suitable for generating angled shear waves. One must bear in mind that such coil possesses essential characteristics of a phased array with the limitation that the phases of adjacent array elements are a constant 180° [8],[10]. We will see a study that describes the behaviour of a such coil and the way whose one optimise it by acting on the spatial period of the coil. The table 1 shows the best results obtained for each category of coil. We have noted gain, rejection of the ultrasonic unit and the gap width for a 100 % dynamics in the ferromagnetic steel. We observe that in the case of the transmission, the lowest gain and the lowest rate of rejection have been recorded for the single six turn-spiral printed in a 14 mm gap. For the reception, 9 turn-spiral was the most efficient.

Fig 10: relative efficiency (in dB) of electromagnetic generation of shear waves in variety of metals as function of frequency.

Nature of the material and excitation frequency
We was not able to make an experimental study on the influence of the exciting frequency. The theoretical study shows the great influence of this parameter on the amplitude of the generated signal. Curves that illustrate the behaviour of the transduction according to the frequency are traced for several materials [9]. Materials are matched to their mechanical and electrical characteristics namely: sound velocity (s) and the skin depth parameter (b ).
We suppose therefore that propagating mediums are electrically and mechanically isotropic. Curves give the amplitude of generated stress as the frequency . According to relation (4), the relative efficiency of the pulse-echo method for generating shear waves in various metals room temperature as a function of frequency is shown in figure10 ,where it is assumed that B and B₀ are constant and any attenuation in the metal is negligible. The efficiency in aluminium at 2 MHz is taken as the reference point, positive dB values indicating a higher efficiency (in magnesium, f) and negative dB values a lower efficiency (a-d). it is apparent that the electromagnetic transducer is most efficient in light metals (Al, Mg) and its efficiency is strongly frequency dependent in metals with small acoustic velocities and so large values of b (Sn, a). The calculation for carbon steel with µ=100 (c) is only a crude estimate, since it ignores any ferromagnetic interactions in the generation process.

Wave Types
The excitation coil design and the electromagnet configuration determine modes of the transduction. Being the seat of the excitation current pulse this coil and this magnet can be arranged in many configurations to produce a variety of wave types such as longitudinal and shear bulk waves excited normal to a sample surface, angled shear waves and guided modes (including the difficult to excite horizontal polarisation) [8]. Identification of these waves is made by the measurement of the propagation velocity in a known sample. Longitudinal waves are naturally generated by the spiral coil and parallel configuration of the static magnetic field. We therefore, produce longitudinal waves in ferromagnetic, stainless steel and in the aluminium. the greatest efficiency is recorded for the ferromagnetic steel with a peak corresponding to an current of 0.05 A in the electromagnet. This is caused by the magnetostriction phenomenon that disappears if the material is saturated. In the aluminium, there is a perfect linearity between amplitude and magnetisation. Cylindrical yoke with E-form was made to generate shear waves. Nevertheless, we know that the flat spiral combined to U-form yoke also generates shear and longitudinal modes. Experiment shows that only E-form yoke generates a pure shear mode, while the U-form yoke produces a quasi-compressional waves. With a double spiral coil (figure 5, middle) and meander coil combined to a U-form magnetic yoke, surface waves are generated. The experiments consists to transmitting with an EMAT on a well polished surface and receiving generated surface waves with a conventional transducer with mechanically adjustable angle, beforehand adjusted (at pulse-echo mode) to the second critical angle. The surface waves echoes were important as compared to bulk waves echoes obtained previously. This was observed only on the ferromagnetic steel. Same experiment on the aluminium gave no result. We have concluded that generated stress are from magnetostriction. This hypothesis has been confirmed when a great sensitivity of the Rayleigh wave echo to the static magnetic field was observed.

Phased array EMATs.
Meander coil transducer possesses essential characteristics of a phased array with the limitation that the phases of adjacent array elements are a constant 180°. This explains his angled radiation [10]. Its directional pattern can be compiled from the diagram of the array element and the directivity coefficient of the group by the following relation [14]:



where a_m is the multiplicatif coefficient connecting the current of excitation I_m (of the source A_m) aware I₁ (exciting the source A₁) as I_m=a_mI₁. the vector localises the source A_m by taking as references the position of the source A₁. is a unit vector oriented in the direction of the ultrasonic wave propagation. In the simplest case, namely when all array element diagrams are the same, the combination in the far zone is multiplicative.
The diagram of the far field can be therefore made more directive simply by acting on the excitation amplitude weight of each array elements. This gives us an efficient means for controlling orientation, polarisation and focalisation of the generated ultrasonic beam. This control can be done on the one hand, by acting on the configuration of the static magnetisation and excitation. In the other hand, by modifying geometrical characteristics of the coil namely the special period of the meander.
The most important is that this orientation and focalisation control can be performed electronically by acting on the phases and amplitudes of array elements in coil.
To obtain this control, windings of the EMAT must be separated up into individual, separately operated loops which (in the case of transmission) are sequence-timed respectively whose reception voltage is step-delayed (figure 11). Radiation diagram of phased array EMAT can be controlled by spatial configuration (linear, circular..),distance between tow adjacent elements, excitations amplitudes and phases of each element and finally nature and number of the used elements [14].

Fig 11: linear phased array for orientation and focalisation controlling

Conclusion

This investigation shows the importance of the electromagnetic acoustic method because demonstrates that the former opens large perspective in NDT&E area. We have conceived and realised EMATs prototypes that generate and receive longitudinal, shear, and surface waves. EMAT permits very easy control of orientation and polarisation of the ultrasonic beam. Adequate weighting reduces side lobes and refine the main lobe. A solution to eliminate the problem of the bi-directional radiation of the EMAT can be performed using the phased array properties of meander coil. We identified parameters that influence the transduction. The optimisation of these parameters has allowed to increase the efficiency of EMATs and to avoid useless energy losses, notably current in the electromagnet. Excitation was adapted with an current exciting electronics. Ultrasonic units being conceived to operate with piezoelectric probes, they based on tension excitation electronic. Finally, we have demonstrated the usefulness and the interest of these devices that function without contact up to 2 mm lift-off . this method can be very useful for accurate measurements like residual stress measurement[1],[3],[4] and speedy testing like automated testing.

Reference

1. D. Husson, S. D. Bennett, et G. S. Kino, "Measurement of Stress with Surface Waves" Materials Evaluation, Vol 43 / N°, Janvier 1985, pp. 92-100.
2. Moon Ho Park, In Sup Kim & Young Ku Yoon, "Ultrasonic Inspection of Long Steel Pipes Using Lamb Waves", NDT&E, Vol.29/ N°1, Février 1996, pp.13-20:
3. Raymond E. Schramm, Jacek Szelazek et A. Van Clarrk, Jr.,"Ultrasonic Measurement of Residual Stress in the Rims of Inductively Heated Railroad Wheels", Materials Evaluation, Vol 54 / N°8, Août 1996, pp. 929-934.
4. Riichi Murayama, Kazuo Fujisawa, Hidekazu Fukuoka et Masahiko Hirao, "Development of an on-line Evaluation System of Formability in Cold-Rolled Steel Sheets Using Electromagnetic Acoustic Transducers (EMATs)", NDT&E International, Vol, 29, No. 3, pp. 141-146, 1996.
5. H. Kwun et A. E. Holt, NDT&E international, volume 28 n° 4 (1995), "Etude de faisabilité de la détection des corrosions des tubes en acier enrobés par capteurs magnétostrictifs".
6. E. Rolland Dobbs.(1973) "Electromagnetic Generation of Ultrasound", Research Techniques in Nondestructive Testing, Vol. 2, pp. 419-441. Academic Press London and New York.
7. J. Krautkrämer,. H. Krautkrämer.(1977) "Other Methods for Transmitting and Receiving Ultrasound", Ultrasonic Testing of Materials, pp. 150-172. Springer-Verlag Berlin Heidelberg Nrew York.
8. H. Salzburger,. W. Mohr.(1979). "Electromagnetic-Acoustic Generation of Ultrasound", 2^nd Seminar on Characterization of ultrasonic Equipment. October 9-12, 1979. I.Z.S.P, Saarbrücken, Germany.
9. Djillali Kalleche,. (1986). "Génération d'Ondes Ultrasonores Sans Contact par Voie Électromagnétique", Thèse de Magister, Institut de physique, U.S.T.H.B, Alger.
10. M. G. Silk.(1984). "Electromagnetic Acoustic Transducers", Ultrasonic Transducers for Nondestructive Testing., pp. 111-119. Adam Hilger Ltd, Bristol.
11. Dobbs, E. R. (1973). In Physical Acoustics (W. P. Mason and R. N. Thurston, Eds.) Vol. 10. Academic press, New York.
12. Dobbs, E. R. and Llewellyn, J. D. (1971). Non-destruct. Test. 4, 49-56.
13. Tarek Boutkedjirt. , «Etude théorique et expérimentale d'un transducteur ultrasonore de référence à disque piézoélectrique épais» Thèse de Magister, Institut de physique, U.S.T.H.B, (1993).
14. Akssas, R. «théorie des réseaux rayonnants» ,cours magistral , Ecole National Polytechnique d'Alger. (1999).