10.1109/ULTSYM.2012.

0502

Experimental Analysis of the Particle

Oscillations in Acoustic Levitation
Marco A. B. Andrade1, Flávio Buiochi1, Sebastian Baer2, Cemal Esen2, Andreas Ostendorf2, and Julio
C. Adamowski1

1
Mecatronics Engineering Department, University of São Paulo, Brazil,
2
Laser Applications Technology, Department of Mechanical Engineering, Ruhr-Universität Bochum, Germany

position and the acoustic radiation force that acts on the

Abstract—The objective of this paper is to investigate the levitated object. These two effects could lead to a negative
spontaneous particle oscillations that occur in a single-axis damping in the equation of motion, which causes the particle
acoustic levitator. A high speed camera was used to record the to oscillate instead of damping its oscillations.
particle oscillations of spheres and a tracking algorithm was
implemented in Matlab to obtain the sphere position as a function
The objective of this work is to study the parameters of
of time. A FFT analysis was applied to obtain the frequencies of acoustic levitation that affect the spontaneous oscillations of
the particle oscillations in both axial and radial directions. The the levitated particle. For this purpose, a high-speed camera is
frequencies of the particle oscillations were compared with the used to record the spontaneous oscillations of different types
theoretical frequencies, which were obtained by a theoretical of spheres in a single-axis acoustic levitator. A tracking
model that assumes that the oscillations of the levitated sphere algorithm is applied to obtain the axial and radial positions of
can be described by a spring-mass system.
the sphere as a function of time. The dependence of each
Keywords— acoustic levitation, Langevin transducer, particle parameter on the levitation instability is investigated by an
oscillations FFT analysis.

I. INTRODUCTION II. ACOUSTIC LEVITATOR

Acoustic levitation in combination with spectroscopy In this work, a single-axis acoustic levitator, consisting of a
methods has many potential applications in the areas of 20.3 kHz Langevin transducer with a concave radiating surface
analytical and bioanalytical chemistry [1, 2]. However, and a concave reflector is used to levitate small spheres. The
spontaneous oscillations and rotation of the levitated particle complete description of this acoustic levitator is presented in
limit the applicability of acoustic levitation. In some cases, [7].
small oscillations rapidly grow until the levitated particle is Before conducting the experimental analysis of the particle
ejected from the pressure node of the standing wave field. oscillations, the displacement amplitude at the center of the
Despite the recent progress in the design of acoustic levitators, transducer face was obtained by using a single point Laser
it is not completely understood how to design an acoustic Doppler Vibrometer (OFV-534 Sensor Head with an OFV-
levitator able to avoid oscillations of the levitated particle. 5000 controller, Polytec GmbH, Germany). It was obtained a
Recently, we designed a new single-axis acoustic levitator with displacement amplitude of 6.6 μm when the transducer is
a concave radiating surface and a concave reflector, which has excited with a electrical voltage amplitude of 10 V at the
a higher restoring force in the radial direction when compared resonance frequency (20342 Hz).
to a levitator with a flat transducer. Then we measured the After obtaining the transducer resonance frequency, the
sphere oscillations with a laser Doppler vibrometer and it was Finite Element Method was used to determine the potential for
shown that the new levitator design allows to reduce the lateral the acoustic radiation force that acts on a small sphere as a
oscillations of the levitated object [3]. function of the distance between the transducer and the
Although there is an extensive literature related to design reflector, following the same procedure of [7]. All the
and applications of acoustic levitation, few papers discuss the simulations were performed with the Langevin transducer
origin of the spontaneous oscillations of the levitated particle operating at 20342 Hz with a transducer face displacement
[4-6]. According to Rudnick and Barmatz [4], the spontaneous amplitude of 1 μm. This analysis allows obtaining the optimal
oscillations are caused by the dependence of the levitator separation distances between the transducer and reflector. One
resonance frequency with the position of the levitated object of the best separation distances for levitation corresponds to
and also due to a time delay between the change of the object 28.6 mm. The potential for the acoustic radiation force that
acts on a small sphere for this separation distance is presented
This work was funded by the Brazilian Agencies FAPESP, CNPq and in Fig. 1. The cross marks in the figure correspond to the
Petrobras/ANP and by the German Research Association (DFG) under No. points of minimum potential, where levitation can occur. The
Sch 184/43-1.

978-1-4673-4562-0/12/$31.00 ©2012
Authorized licensed use limited IEEE del Valle. Downloaded on2006
to: Universidad 2012 IEEE
November 12,2024 International
at 19:46:02 UTC from Ultrasonics
IEEE Xplore. Symposium Proceedings
Restrictions apply.
potential was obtained by using the Gor’kov theory [8]. In The middle levitation point of Fig. 1 has a relative elastic
order to show the potential that is independent of the sphere constant of 4.1 x 106 N/m4 in the radial direction and 5.2 x 107
radius, it was plotted the relative potential [9] in Fig. 1, which N/m4 in the axial direction, which gives an elastic constant
is defined by [7]: ratio of 12.7 between the axial and the radial constants. These
two relative elastic constants were obtained by using equations
~ ⎛ p2 ρ u 2 ⎞⎟ (4) and (5) for a transducer face displacement amplitude of 1
U =⎜ − (1) μm. Due to the quadratic dependence of the acoustic potential
⎜ 3 ρc 2 2 ⎟
⎝ ⎠ with the amplitude displacement of the transducer face, the
elastic constants increase four times when the displacement
where p 2 and u 2 are the mean square amplitudes of the amplitude is increased twice.
pressure and velocity, respectively, ρ is the air density and c is
the propagation velocity in air. The potential for the acoustic
radiation force that acts on a sphere of radius R is related with
the relative potential through the following equation:

~
U = 2πR 3U (2)

The relation between the acoustic potential U and the force

F that acts on the levitated sphere is:

F = −∇U (3)

For a separation distance of 28.6 mm between the

transducer and the reflector, there are three points of minimum
potential. In the experiments with a high-speed camera, the
sphere is placed at the middle levitation point, that is located at
x = 0 mm and z = -14.5 mm. The relative acoustic potential for
this levitation point in both x and z directions are presented in
Fig. 2. Both curves are very similar to the potential of a spring-
mass-system around the position of minimum relative acoustic
potential. Therefore a relative elastic constant is defined in
both directions:
Fig. 2. Relative potential for the acoustic radiation force: (a) along the x-axis
~2
for z = -14.5 mm; (b) along the z-axis.
~ ∂U
kx = 2 (4)
∂x III. ANALYSIS OF THE PARTICLE OSCILLATIONS

~ The experimental apparatus used to study the spontaneous

~ ∂ 2U oscillations of the levitated sphere in an acoustic levitator is
kz = 2 (5)
∂z presented in Fig. 3. In this apparatus, a single-axis acoustic
levitator, consisting of a 20.3 kHz Langevin transducer with a
concave radiating surface and a concave reflector is used to
levitate small spheres of different diameters and materials
(polypropylene, glass and steel). During the experiment, the
Langevin transducer is excited by a harmonic signal of 20342
Hz and different voltage amplitudes. This signal is generated
by a 33250A function generator (Agilent Technologies Inc.,
Santa Clara, CA) and amplified by an 800A3A power
amplifier (Amplifier Research Corp., Souderton, PA). In order
to find experimentally the best separation distance between the
transducer and reflector, a piezoelectric disk was bonded to the
bottom of the reflector, and its electrodes was connected to an
oscilloscope. In the experiment the transducer is maintained
fixed and an automatic positioning system is used to move the
Fig. 1. Relative potential for the acoustic radiation force for a 28.6 mm reflector. For optimal separation distances between the
separation distance, 1 μm displacement amplitude at the transducer face and transducer and the reflector, the amplitude of the electrical
operation frequency of 20342 Hz.

2007
Authorized licensed use limited to: Universidad del Valle. Downloaded on 2012 IEEE
November 12,2024 International
at 19:46:02 UTC from Ultrasonics
IEEE Xplore. Symposium Proceedings
Restrictions apply.
signal measured at the piezoelectric disk is maximum. The FFT for the three spheres are presented in Fig. 5. The
A high-speed camera (Fastec Inline 1000 monochrome, frequency of each peak of Fig. 5 is presented in Table I.
Fastec Imaging Corp., San Diego, CA) is used to record the According to this table, the ratio between the oscillation
spontaneous oscillations of the spheres. During the frequency in the axial and the radial directions obtained
experiments, it was used a frame rate of 250 or 500 Hz to numerically has good agreement with that obtained
obtain the videos of the particles oscillations. In order to experimentally for the steel sphere. However, the results
increase the contrast between the levitating sphere and the obtained experimentally for the polypropylene and glass
background, a backlight with a light diffuser sheet is use to spheres present significant difference when compared with the
illuminate the sphere. A tracking algorithm was implemented numerical simulation.
in Matlab to obtain the particle trajectory in both axial and According to the Gor’kov theory, the elastic constant is
radial directions. independent of the sphere density. Maintaining constant all the
parameters of the levitator except the density of the levitated
sphere, we can determine its influence on the spontaneous
oscillation frequency. Assuming a spring-mass system, the
ratio between the natural frequencies f1/f2 of the oscillations of
two spheres of densities ρ1 and ρ2 for a fixed value of elastic
constant is given by:

f1 ρ2
= (6)
f2 ρ1

Fig. 3. Experimental apparatus. Equation (6) was applied to determine the ratio between the
oscillation frequencies of spheres with different densities,
IV. RESULTS AND DISCUSSION which are present in Table II. In this table, fp corresponds to
The spontaneous oscillations of levitating spheres was the natural frequency of the polypropylene sphere, fg
analyzed by changing different parameters of the levitator, corresponds to the natural frequency of the glass sphere and fs
such as the distance between the transducer and reflector, corresponds to the steel sphere. The ratio between frequencies
electrical voltage amplitude, and the density of the sphere. A was determined both in the axial and radial directions. The
typical graphic of a sphere oscillation, which was obtained results of Table II shows that there are good agreement
through the tracking algorithm, is shown in Fig. 4. To obtain between the theoretical values calculated with (6) and the ratio
this graphic, the position of the reflector was altered until the between frequencies for spheres vibration in the axial
measured voltage amplitude at the attached piezoelectric disk direction. However, there are significant differences between
was maximum. The maximum was obtained when the the experimental results for the oscillations in the radial
separation distance between the transducer and reflector was direction from those predicted theoretically. These differences
29.9 mm, which is similar to that obtained numerically (28.6 for the radial direction can explain the results present in Table
mm). The radial and axial positions of the spheres as a I, where the experimental values of fa/fr for glass and
function of time are shown respectively in Figs. 4(a) and 4(b). polypropylene spheres differ significantly from that obtained
These positions were obtained for a 3-mm-diameter steel numerically.
sphere when it was placed at the middle levitation position.
During the levitation, the Langevin transducer was excited by
a sine wave of 20342 Hz with voltage amplitude of 30 V.
Assuming that the sphere oscillations can be described by
a three-dimensional harmonic oscillator [9], we can use the
ratio between the elastic constants in axial and radial
directions to obtain the ratio between oscillation frequencies in
both directions. According to the simulation, the ratio between
the elastic constants in the axial and radial direction is 12.7.
The ratio between the oscillation frequencies in axial and
radial directions is given by the square root of 12.7, which is
equal to 3.56.
The ratio between the natural frequencies in the axial and
radial directions is experimentally measured by determining
the FFT of the spheres displacements. The FFT analysis was
conduct for three spheres made of polypropylene (ρp = 875 Fig. 4. Position as a function of time of a 3-mm-diameter steel sphere for a
transducer excitation amplitude of 30 V: (a) radial position; (b) axial.
kg/m3), glass (ρg = 2475 kg/m3) and steel (ρs = 7850 kg/m3).

2008
Authorized licensed use limited to: Universidad del Valle. Downloaded on 2012 IEEE
November 12,2024 International
at 19:46:02 UTC from Ultrasonics
IEEE Xplore. Symposium Proceedings
Restrictions apply.
 why the frequency is not increasing with the transducer
amplitude in Fig. 6.

Fig. 6. Frequency of the spheres oscillations as a function of the transducer

excitation amplitude: (a) radial; (b) axial.

Fig. 5. FFT of the oscillations for 3-mm-diameter spheres with different Fig. 7. Indirect pressure measurement by using a laser Doppler vibrometer.
materials for a transducer excitation amplitude of 30 V: (a) radial; (b) axial.
V. CONCLUSION
Table I: Frequency of the oscillations in the radial direction (fr) and axial
direction (fa) and the ratio between the natural frequency in both directions. This paper presented the experimental study the particle
material fr (Hz) fa (Hz) fa/fr oscillation in acoustic levitation and it was proposed a model
polypropylene 11 52 4.73 based on the assumption that the frequency of oscillations can
glass 7.4 29 3.92 be described by a spring-mass system, where the elastic
steel 4.4 16 3.64 constant is determined by Gor’kov theory. The experimental
results show that this simple model can explain the radial and
Table II: Analysis of the oscillations frequencies of the levitated sphere with axial oscillation frequencies of the particle.
density.
radial axial theoretical REFERENCES
fp/fs 2.5 3.2 ρ s ρ p = 3.0 [1] S. Santesson, J. Johansson, L. S. Taylor, L. Levander, S. Fox, M.
Sepaniak, and S. Nilsson, “Airborne Chemistry Coupled to Raman
fp/fg 1.5 1.8 ρ g ρ p = 1.7 Spectroscopy”, Anal. Chem., vol. 75, pp. 2177-2180, 2003.
[2] S. Santesson, S. Nilsson, “Airborne Chemistry: Acoustic Levitation in
fg/fs 1.7 1.8 ρ s ρ g = 1.8 Chemical Analysis”, Anal. Bioanal. Chem., vol. 378, pp. 1704-1709,
2004.
[3] S. Baer, M. A. B. Andrade, C. Esen, J. C. Adamowski, G. Schweiger,
and A. Ostendorf, “Analysis of the particle stability in a new designed
The experimental setup of Fig. 3 was also applied in the ultrasonic levitation device”, Rev. Sci. Instr., vol. 82, art. no. 105111,
determination of the frequencies of the spheres oscillations as 2011.
a function of the transducer excitation amplitude. The obtained [4] J. Rudnick, M. Barmatz, “Oscillational Instabilities in single-mode
oscillation frequencies are presented in Fig. 6. According to acoustic levitators”, J. Acoust. Soc. Am., vol. 87, pp. 81-92, 1990.
[5] E. A. Torres., A. Santillan, “Analysis of the Effects of the Oscillations
Gor’kov theory, the elastic constant increases with the sound
of a Rigid Sphere Inside a Cylindrical Cavity Containing a Standing
pressure, and therefore, the oscillations frequencies should Acoustic Wave”, Proceeding of Acoustics’08-Paris, June-July, pp.
increase with the transducer excitation amplitude. However, it 4379-4384, 2008.
can be observed in Fig. 6, that the frequencies are not [6] A. Santillan, “Analysis of Oscillational Instabilities in Acoustic
increasing with the transducer excitation amplitude. This result Levitation Using the Finite-difference Time-domain Method”, Proc.
IEEE Ultrason. Symp., Orlando, U.S.A., pp. 1552-1555, 2011.
can be explained by the fact that the sound pressure does not [7] M. A. B. Andrade, F. Buiochi, and J. C. Adamowski, “Finite Element
increase linearly with the transducer excitation voltage, as it Analysis and Optimization of a Single-Axis Acoustic Levitator”, IEEE
can be observed in Fig. 7. In this figure, the sound pressure Trans. Ultrason. Ferroelectr. Freq. Control, vol. 57, pp. 469-479,
was measured indirectly by a laser Doppler vibrometer. It can 2010.
[8] L. P. Gor’kov, “On the Forces Acting on a Small Particle in an Acoustic
be seen that for small amplitudes of transducer excitation, the Field in an Ideal Fluid”, Sov. Phys. Dokl., vol. 6(9), pp. 773-775, 1962.
output of the vibrometer increases linearly with the amplitude. [9] M. Barmatz, P. Collas, “Acoustic Radiation Potential on a Sphere in
However, when the transducer amplitude reaches a voltage Plane, Cylindrical and Spherical Standing Wave Fields”, J. Acoust. Soc.
around 10 V, the pressure stops to increase, which can explain Am., vol. 77, pp. 928-945, 1985.

2009
Authorized licensed use limited to: Universidad del Valle. Downloaded on 2012 IEEE
November 12,2024 International
at 19:46:02 UTC from Ultrasonics
IEEE Xplore. Symposium Proceedings
Restrictions apply.