Vogel JFM 1989

J. Fluid Me&. (198Y),1101.206, p p .
299-338 299
Printed in Great Britain
Optical and acoustic investigations of the dynamics

of laser-produced cavitation bubbles
near a solid boundary
By A. VOGELt, W. L A U T E R B O R N I A N D R. TIMMI1
Drittes Physikalisches Institut, University of Gottingen, Federal Republic of Germany
(Received 11 May 1988 and in revised form 18 November 1988)
The dynamics of laser-produced cavitation bubbles near a solid boundary and its
dependence on the distance between bubble and wall are investigated experimentally.
It is shown by means of high-speed photography with up to 1 million frames/s that
jet and counterjet formation and the development of a ring vortex resulting from the
jet flow are general features of the bubble dynamics near solid boundaries. The fluid
velocity field in the vicinity of the cavitation bubble is determined with time-
resolved particle image velocimetry. A comparison of path lines deduced from
successive measurements shows good agreement with the results of numerical
calculations by Kucera & Blake (1988). The pressure amplitude, the profile and the
energy of the acoustic transients emitted during spherical bubble collapse and the
collapse near a rigid boundary are measured with a hydrophone and an optical
detection technique. Sound emission is the main damping mechanism in spherical
bubble collapse, whereas it plays a minor part in the damping of aspherical collapse.
The duration of the acoustic transients is 2 6 3 0 ns. The highest pressure amplitudes
at the solid boundary have been found for bubbles attached to the boundary. The
pressure inside the bubble and at the boundary reaches about 2.5 kbar when the
maximum bubble radius is 3.5mm. The results are discussed with respect to the
mechanism of cavitation erosion.
1. Introduction
Historically, the interest in the dynamics of cavitation bubbles in liquids mainly
arose from their destructive action on solid surfaces, which has been observed on ship
propellers and in hydraulic machinery. As recently summarized by Tomita & Shima
(1986), cavitation erosion is attributed to the action of acoustic transients emitted
during bubble collapse and to the impingement of the high-speed liquid jet that
develops when a bubble collapses in the vicinity of 8 solid boundary. To study the
jet formation, Lauterborn (1974) and Lauterborn BE Bolle (1975) have produced
single cavitation bubbles by focusing &-switched laser pulses into the liquid under
investigation. Thereby plasma formation occurs at the laser focus leading to emission
of an acoustic transient and generation of a cavitation bubble (Brewer & Rieckhoff
1964; Carome, Moeller & Clark 1966). Since laser-produced bubbles can be made
t Present address: H. Wacker Laboratory for Medical Laser Applications at the Eye Clinic of
the University of Munich, Mathildenstr. 8, 8000 Miinchen 2, FRG.
$ Present address : Institute for Applied Physics, Technische Hochschule Darmstadt,
SchloBgartenstr. 7, 6100 Darmstadt, FRG.
1) Present address: MBB/Erno, Hiinefeldstr. 1-5, 2800 Bremen 1, FRG.
300 A . Vogel, W . Lauterborn and R.Timm
highly spherical and are free from mechanical distortions, they meet the requirements
necessary for comparing the results of experimental work and numerical investi-
gations. During the few last years, laser-produced cavitation bubbles have also
become important in the medical field of ophthalmology and urology. In ocular
surgery by means of photodisruption, as introduced by Fankhauser (Fankhauser et
al. 1981) and Aron-Rosa (Aron-Rosa et a2. 1980), and in laser-induced lithotripsy (see
Steiner 1988), laser pulses are used to produce a plasma with subsequent bubble
formation. Tissue disruption and stone fragmentation are related to the mechanism
of cavitation erosion, because they are probably due to the combined effects of the
plasma, the bubble dynamics and the acoustic transients emitted during bubble
generation and collapse (Vogel et al. 1986; Reichel et al. 1987). Recently, Coleman
et al. (1987) have demonstrated that cavitation bubble dynamics may also contribute
to renal calculi disintegration in extracorporeal shock-wave lithotripsy.
Naud6 & Ellis (1961), and Tomita & Shima (1986) have shown that the dynamical
behaviour of a bubble strongly depends on the dimensionless distance y = s/R,,
between cavitation bubble and wall, with R,, being the maximum bubble radius,
and s denoting the distance between the location of bubble formation and the wall.
This paper summarizes the results of detailed investigations of jet formation and
sound emission as a function of y . The jet formation has been studied by means of
high-speed photography with a framing rate of up to 1 million framesls. It was
shown that over a wide range of the parameter y a ring vortex emerges from the jet
flow, as theoretically predicted by Benjamin & Ellis (1966), and that in a more
limited y-interval a counterjet appears as regularly as the main jet itself. Formerly,
only data about the bubble shape had been obtained both experimentally (see
Lauterborn & Hentschel 1985) and in numerical model calculations. However, a
more detailed investigation of jet formation should include the velocity and pressure
field around the bubble. Recently, theory has succeeded in calculating pathlines and
pressure contours in the neighbourhood of collapsing bubbles (Blake, Taib &
Doherty 1986). By combining particle image velocimetry (PIV) with high-speed
photography, we are now able to investigate the temporal development of the
unsteady fluid flow around collapsing cavitation bubbles experimentally and to
compare it with the numerical results obtained by Kucera & Blake (1988).
In earlier experimental and theoretical work Radek (1972),Hinsch & Brinkmeyer
(1976) and Ebeling (1978) found that the pressure pulse emitted upon spherical
cavitation bubble collapse has a duration between 10 and 40 ns. Similar values have
been reported by Ebeling (1978) and Stepp et al. (1985) for the duration of the
acoustic transients following laser-induced plasma formation. This is much shorter
than the rise time of most calibrated pressure transducers commercially available.
Exposing these transducers to the pressure pulses one records merely the impulse
response of the transducer, and the indicated pressure values are lower than the
actual ones. Transducers with a rise time of a few nanoseconds are, on the other hand,
very insensitive and have to be placed close to the centre of the spherical pressure
waves. They are not well suited for measuring the sound emission during spherical
cavitation bubble collapse, because every boundary in the vicinity of the bubble
distorts its dynamics. Therefore, we developed a technique to determine the pressure
of short acoustic pulses incorporating a sensitive pressure transducer with a
frequency bandwidth smaller than that of the acoustic pulse. For this, the recording
of the acoustic signal was combined with a synchronously performed fast optical
measurement of the temporal profile of the pressure pulse. From these data, the
actual pressure value can be calculated if the transfer function of the transducer is
Dynamics of laser-produced cavitation bubbles near a solid boundary 301
known. Further processing of the data allows an estimation of the energy of the
acoustic transients.
2. Experiments
2.1. High-speed photography
Three different experimental arrangements have been used for high-speed
photography of laser-produced cavitation bubbles. The basic set-up is outlined in
figure l ( a ) . The bubbles are generated in the cuvette filled with water by using a
&-switched ruby laser which delivers pulses with an energy of 1 W m J . The
focusing lens has a focal length of 25 mm, and the cone angle of the focused laser
beam is 19.5' (in water). The large cone angle was chosen to reduce the probability
of generating multiple plasmas by limiting the volume with suprathreshold light
intensity. The bubble dynamics is investigated with a drum camera (Impulsphysik
model Strobodrum) at 20000 frames/s using diffuse illumination by a flash lamp
through a ground glass. The exposure time is set by the spark duration of about 1 ps.
The comparatively low framing rate allows the recording of the whole life cycle of
the cavity. To study the formation of a ring vortex during bubble collapse, the
bubbles were photographed in side view and top view as shown in figure 1 ( b )and 1 (c),
respectively. In the arrangement of figure l ( b ) a brass block was used as the solid
boundary, whereas in 1 (c) the cuvette wall served as a transparent rigid boundary.
Figure 2 shows an arrangement for high-speed photography with a schlieren
technique. For the determination of the maximum jet velocity one has to know
whether it penetrates the opposite bubble wall during collapse or if it merely pushes
the wall ahead during rebound. Only in the latter case can the jet velocity be
identified with the velocity of the tip of the bubble protrusion observed during
rebound by Benjamin & Ellis (1966), and Lauterborn & Bolle (1975). A ring vortex
predicted by Benjamin & Ellis (1966) can, on the other hand, only be formed during
first collapse if the jet penetrates the bubble wall immediately upon impingement. To
clarify these questions, we produced the cavitation bubbles within a stationary
temperature gradient generated with two Peltier elements as depicted in figure 2 ( b ) .
The jet penetrating the opposite bubble wall has a different temperature, and
therefore a different refractive index than the surrounding liquid. Thus, it becomes
a weak phase object and can be visualized by means of a schlieren technique as shown
in figure 2 ( a ) ,though it would be invisible by normal photographic techniques. I n the
schlieren set-up a flashlamp spark is imaged into the plane of a vertical slit with a
width of 40 pm. The illuminating light beam is then collimated and - after passing
the cuvette -blocked out with a vertical wire. The wire serves as a high-pass filter
for the optical system which images the cavitation bubbles into the film plane of the
drum camera. The sensitivity of the apparatus was tested with a model jet from a
syringe with a diameter of 0.2mm (see Vogel & Lauterborn 1985). When the
surrounding liquid has a temperature of 273K ( 2 0 ° C ) , a jet with a temperature
difference of 0.5K, corresponding t o an optical phase shift of A/50,could still be
detected. Since the temperature gradient in the cuvette can be adjusted t o more than
1 K/mm and the maximum cavitation bubble diameter amounts to some millimetres,
the apparatus is sensitive enough to visualize the jet if it pierces the opposite bubble
wall.
For studying the bubble dynamics during first collapse with high temporal
resolution, we applied a rotating mirror camera (Beckmann and Whitley, model 330)
and an image converter camera (Hadland Imacon 790) with a maximum framing
302 A . Vogel, W. Lauterborn and R . Timm
(4 (b)
Water
Flash lamp
-
Line of
,
Slght
* Laser
focus
///A'//////
Ground glass
(4
Cuvette with Cuvette wall
water
Laser
focus
Drum camera
Line of sight
FICIJRE 1. ( a ) Basic experimental arrangement for high-speed photography of cavitation bubbles.
(b, c) Principal configurationsfor the photography of bubble dynamics with the line of sight parallel
to the solid wall (b) and perpendicular to the solid wall (c).
Peltier element 1
"Ji
I 1
Spatial
; j filter
@?
FIGURE 2. (a) Experimental arrangement for flow visualization by optical high-pass filtering
(schlieren photography). (b) Apparatus for the generation of a stationary temperature gradient.
rate of 1 million frames/s. The experimental arrangement is shown in figure 3. To

synchronize the photographical recording with the bubble collapse, the camera is
triggered by the collapsing cavity itself (Lauterborn & Timm 1980). This is
accomplished by letting the cavity modulate the intensity of a helium-neon laser
beam which is picked up by a fast photodiode. The laser beam is blocked out when
the cavity is expanded. During collapse the light intensity at the diode rises until the
cavity reaches its minimum volume. The sudden rise of the diode signal during the
final collapse stage triggers the oscilloscope which is used to derive the trigger signal
for the camera.
2 . 2 . Time-resolved particle image velocimetry
I n particle image velocimetry, the fluid is seeded with microparticles serving as light
scatterers, the specific gravity of which is chosen t o match that of the fluid. A plane
in the fluid is illuminated by a thin light sheet, and the particle distribution in this
plane is imaged onto photographic film. Since the scattering particles follow the fluid
flow, they are recorded as a chain of particle images when the film is multiply
Flash
(.7 lamps
n
mirror camera
I I t
Trigger signal
-
3. Experimental arrangement for high-speed photography with triggering from the
FIGURE
cavity itself.
exposed. The projection of the fluid-velocity vector onto the observation plane can
be calculated from the direction of each chain, the distance between the particle
images and the time between successive exposures. The temporal evolution of the
flow field can be monitored by taking a series of multiple exposure photographs. For
this, particle image velocimetry and high-speed photography with a drum camera
were combined in the experimental arrangement shown in figure 4 (Vogel &
Lauterborn 1 9 8 8 ~ )The
. light source is a 3.6 W argon-ion laser. The laser beam is
chopped by an acousto-optic deflector controlled by a cascade of two burst
generators. Thus, pulse trains of variable length, pulse separation and repetition rate
can be generated. Each frame was exposed with a sequence of five pulses with 1 ps
duration and 3 ps distance between successive pulses. The argon laser beam is
expanded vertically by a cylindrical lens and focused horizontally by a convex lens
to produce a light sheet of 15 mm height and 140 pm thickness in the plane where the
centre of the bubble is located. Since the energy of each argon laser pulse is only 3.5
pJ, an imaging system with a large aperture (f-No. 1.25) was chosen to achieve
optimum utilization of the scattered light. The water in the cuvette is seeded with
Vestamid particles which have an average diameter of 25 pm and a density of
1.01-1.02 g/cm3. Simultaneously with high-speed photography, both the output
signal of the second burst generator and the acoustic signal of the cavitation bubble
are recorded with a transient recorder. The first peak of the acoustic signal represents
the shock wave emitted during bubble generation, and the following peaks indicate
the successive bubble collapses. The output of the burst generator is the driving
signal for the illuminating light pulses and marks the instants of film exposure. By
comparing both signals, the picture sequence of time-resolved particle image
velocimetry can be exactly related to the life cycle of the cavitation bubble. For
evaluation, the photographs were magnified and the prints were analysed by
individual inspection. With the apparatus described above, fluid velocities can be
determined within a range of from less than 2 m/s to 30m/s. The maximum
304 A . Vogel, W. Lauterborn and R. Timm
AOD
driver
-
~
yulllll
8
Burst -
generator I1
generator I
/ l . Order Trigger signal I
Drum camera
FIQURE
4. Arrangement for time-resolved particle image velocimetry of the flow around laaer-
produced cavitation bubbles.
temporal resolution achieved is 10000 frames/s and the average density of

measurement values is two points/mm2.
2.3. Measurement of acoustic pulses
For calculating the actual values of short pressure pulses from the output signal and
the transfer function of the pressure transducer, the temporal profile of the pressure
pulse has to be known. As confirmed by the experimental results described below, the
profile of the acoustic transients from laser-produced cavitation bubbles can be
approximated by an exponential impulse. Vogel & Lauterborn (19883) have shown
that the peak amplitude of an exponential pressure pulse is given too low by a
factor of ,
I
I
.
m=-
t, In 2
when the pulse duration T is much shorter than the rise time t, of the pressure
transducer. In deriving relation ( l ) ,the assumption was made that the transfer
function of the transducer is constant up to the cutoff frequency v, = itAand zero for
higher frequencies. If 7 and t, are known, (1)yields a correction of the indicated
pressure values. The rise time t A is most conveniently measured when the transducer
is hit by a laser-induced acoustic transient with duration T < t,. The pulse duration
T can be obtained through the optical detection of the temporal profile of the pressure
pulses which is a part of the experimental arrangement outlined in figure 5 . The
probe laser beam from an argon laser is expanded and collimated, and it illuminates
the circular aperture 1. The diameter of the aperture is adjusted so that only the
central part of the beam with nearly constant intensity can pass. This part is focused
Transient recorder 2
storage ~ l l o s c o p c
5. Experimental arrangement for the investigation of acoustic transient emission during

FIGURE
cavitation-bubble generation and bubble collapse.
into the cuvette close to the emission centre a t a radius r0 of the acoustic transients
and the focus is imaged into a small aperture directly in front of a fast photodiode.
When an acoustic pulse passes the probe beam waist, a transient angular deflection
q5 of the probe is produced due t o the refractive index change caused by the pressure
pulse. As a result, the probe beam is partly blocked out by aperture 2 and the
intensity of light arriving a t the photodiode is modulated. The deflection of the probe
beam leads to a reduction of the photodiode voltage U independent of the direction
of the deflection so that the voltage is a function of the modulus of the deflection. The
U(#)-characteristic is shown in figure 6 for the theoretical case of a probe beam with
a rectangular intensity profile as well as for the experimental apparatus used. I n the
nearly linear part of the U(q5)-characteristic, the change of the photodiode voltage
caused by the pressure wave p ( r , t ) passing the probe beam focus at ro can be
approximated by
where U,, denotes the voltage for q5 = 0. Thus, the experimental probe beam
deflection signal is proportional to the modulus of the time-derivative of the pressure
pulse at the probe beam waist. The temporal resolution of the apparatus described
is 10 ns, limited by the time necessary for the pressure pulses to travel through the
probe beam focus.
The pressure amplitude of the acoustic transients is determined with hydrophone
1 (Celesco LC5-2) a t a distance of about 20 mm from the emission centre (see figure
5). The sensitivity of the transducer is 500mV/bar and its rise time 33311s. The
transducer signal is registered with 10 MHz sampling rate with transient recorder 1
(Maurer ADAM TC 1008). This instrument has the ability to store 64 k samples with
8 bit resolution, which allows for tracking the sound emission throughout the whole
lifetime of the cavitation bubbles. For the optical detection of acoustic transients,
the storage oscilloscope is externally pretriggered by a reference pulse from the ruby
306 A . Vogel, W . Lauterborn and H.Timm
0 1 2 3 4 5 6 I
191(degrees)
FIGURE6. The normalized photodiode signal U/U0 as a function of the angular probe beam
deflection 4 : - - - -,theoretical curve for homogeneous intensity distribution acrosu the probe beam ;
-, curve determined experimentally for the apparatus in figure 5.
laser and then - after an adjustable delay -triggered internally. This method may
fail for the relatively weak acoustic pulses emitted after bubble collapse near a solid
boundary. Therefore, the correct triggering is checked by coupling the recording of
the optical signal with the registration of the acoustical signal by a second pressure
transducer (Imotec PVDF pressure gauge) with a rise time of 145ns, and a
sensitivity of about 2.5 mV/bar. Its signal is recorded with 60 MHz sampling rate
through transient recorder 2 (Sony/Tektronix 390 A/D). The short rise time of
hydrophone 2 allowed, moreover, a less distorted observation of the profile of the
acoustic transients than achievable with the more sensitive hydrophone 1.
3. Results
3.1. Jet formation
Figure 7 shows several high-speed photographic series of jet formation which
illustrate the changes in bubble dynamics with varying distance between bubble and
boundary. When the bubble is relatively far away from the solid boundary ( y = 2.3,
figure 7 a ) , it migrates toward the boundary during collapse and a jet is developed
which leads to the protrusion of the lower bubble wall. The jet is clearly visible in the
bright centre of the bubble images. The jet flow creates a ring vortex which is
manifested by the continuous migration of the bubble ring toward the boundary (see
$3.2).I n the second picture series ( y = 1.56, figure 7 b ) , a counterjet appears before
the main jet is visible, manifesting itself in the protrusion on the upper bubble side
(see 93.3). When y is smaller than one, no counterjet is observed. With decreasing
y , the jet becomes thicker and touches the lower bubble wall in an earlier stage of the
collapse (see also Tomita & Shima 1986).When y is further reduced, the jet diameter
diminishes again as can be seen in figure 7 ( d )with y = 0.34.
Fundamental considerations about the mechanism of jet formation were published
FIGURE7. Jet formation at different dimensionless distances y between the site of bubble
generation and boundary. R,, is the maximum bubble radius and E is the bubble elongation before
collapse (see text). The series were taken with the set-up shown in figure 1 at 20000 framesls. The
boundary is given by the darker part below the bubbles in the individual frames. The scale can be
read from the maximum bubble radius noted.
308 A . Vogel, W . Lauterborn and R. Timm
as early as 1966 by Benjamin & Ellis. In the following their ideas are applied to
bubble collapse near a solid boundary. The boundary (which for the sake of definition
is assumed to be located below the bubble) retards the fluid flow towards the
collapsing cavity. This leads to a difference in the pressure in the liquid above and
below the bubble, causing a stronger acceleration of the upper bubble wall than of
the lower one, and thus a migration of the bubble centre towards the boundary. This
means that the bubble achieves a Kelvin impulse due to the Bjerknes force which is
created by the pressure gradient normal to the boundary. The Kelvin impulse can be
regarded as linear momentum of ‘the bubble ’ if a virtual mass induced by the fluid
motion is attributed to the cavity. The Bjerknes force almost vanishes in the final
stage of the collapse because of the reduction of bubble volume, and the Kelvin
impulse approaches a constant value. Since the induced mass of the bubble
diminishes during collapse and the Kelvin impulse remains constant, the trans-
lational velocity of the bubble centre has to increase. For this reason, the velocity
of the upper bubble wall has to grow much faster than that of the lower wall. This
leads to the formation of a liquid jet directed toward the boundary. The jet hits the
opposite bubble wall in the final stage of the collapse and penetrates the bubble
during rebound, causing the well-known protrusion on its lower side, if the bubble is
not attached to the boundary. Since the Kelvin impulse needs to be conserved during
collapse and rebound, the jet flow leads to the formation of a vortex ring migrating
toward the boundary. It becomes clearly visible after the decay of the protrusion.
Lauterborn (1982) has already pointed out that the argument using the Kelvin
impulse cannot describe the deformation of the bubble surface in detail, because the
Kelvin impulse is an integral value and can only determine aspects of the gross
bubble motion. The questions of how and a t what time during collapse the upper
bubble wall changes its shape to form the tip of the jet can only be answered by
experiments or theoretical investigations that track the bubble dynamics in detail.
A starting point of heuristic reasoning is given by the observation that during
collapse the bubble first becomes elongated in the direction normal to the rigid wall
(see figure 7), because the collapse of the lower bubble wall is retarded by the solid
boundary. The extent of the elongation E depends mainly on y. It is defined by E =
h/Rho,,where Rhordenotes the horizontal bubble radius at the instant of maximum
bubble elongation, and h is the distance between the level of Rhorand the top of the
bubble. The relation between jet formation and elongation can be understood by
taking into account that the collapse time of a bubble is proportional to its maximum
radius. Therefore, more highly curved parts of an elongated bubble corresponding to
a smaller bubble radius collapse faster than less curved parts. The top and bottom
of the elongated part, i.e. the parts of the bubble with the highest curvature, collapse
faster. Thereby, the velocity vectors connected to the collapse of the top elements
become much more focused than those connected to the side elements, and the jet and
- under certain circumstances - a counterjet are formed (see Lauterborn 1982). Since
this process starts earlier when the bubble is more elongated and since the elongation
increases with decreasing y, the jet is developed earlier with lower y. At large y , the
bottom flow is not inhibited by the wall and an elongation does not develop, i.e. the
collapse is spherical. A counterjet (see figure 7 b ) can appear at moderate y-values,
but only for y > 1 when the lower bubble wall is not attached to the boundary and
features a high curvature. The counterjet is always weaker than the jet owing to the
inhibition of the collapse of the bottom flow by the solid boundary.
A t bubble collapse, potential energy of the expanded cavity is converted into
kinetic energy of the liquid around the collapsing bubble. The driving force for that
FIGURE 8. Visualization of the jet flow into a cavitation bubble collapsing near a solid boundary.
The picture series is taken with the arrangement of figure 1 at 20000 frames/s, y = 1.25, R,,, =
4.5 mm. The upper dark band in each frame is a coloured layer of silicone oil on water (brighter
background). The solid boundary is visible as the dark stripe in the lower part of each frame.
is the difference between the pressure in the liquid and the pressure in the bubble. As
a result of the continuous reduction of the volume available for the fluid inflow,
however, the bubble wall is still accelerated for some time in a collapse stage when
the pressure in the bubble already exceeds the static pressure in the liquid. This
‘geometrical effect ’ relies on the inertia of the fluid flow and is most pronounced a t
the upper side of the cavity where the bubble wall has already achieved the highest
velocity in the early stages of jet formation. Thus, the jet flow essentially derives its
force in the same way as the spherical collapse: it is the concentration of a finite
amount of energy into a very small volume. This process can be nicely observed in
figure 8, where the jet flow has been made visible by stratifying a coloured layer of
silicone oil with the viscosity of water above a 7.5 mm thick water layer covering a
solid boundary. In this arrangement, the jet develops earlier than after bubble
generation in pure water, because the density of silicon oil ( p = 0.773 g/cm3) is much
lower than the density of water, so that the effect of the interface between both
liquids resembles that of a free surface (see Blake & Gibson 1981).
Figure 9 shows the results of the evaluation of a photographic series taken with
time-resolved particle image velocimetry a t a framing rate of 10 kHz. As expected,
the largest flow velocities are found on the side of the bubble where the jet is
developing toward the solid boundary. Jet formation is further indicated by the fact
that a t the end of the collapse the velocity vectors do not point toward the centre of
the bubble, but toward a spot close to its surface on the side opposite to the solid
boundary (see also figure lob). The fluid flow is thereby focused and strongly
accelerated. Figure 9 ( 9 )shows that during the second collapse a vortex ring is formed
as a consequence of jet flow through the bubble.
As indicated by figure 9, no satisfactory temporal resolution of the velocity field
evolution in the vicinity of the bubble could be achieved even with the highest
possible framing rate of 10 kHz. I n addition, the spatial resolution could not be made
to exceed 1.5 data points per mm2 because of the small size of the photographic
negative which had t o be used due t o the high framing rate. Therefore, a different
approach was attempted to increase spatial resolution. The image quality was
improved by using a framing rate of 5 kHz and a larger image scale. The velocity field
dynamics was now followed by taking pictures of different series photographed a t
different stages of the bubble collapse. Figure 10 contains two velocity field
FIGURE 9. Evolution of the velocity field around a laser-generated cavitation bubble close to a solid
boundary (hatched area). The first collapse occurs shortly after (d), the second collapse shortly
after (e). y = 1.9, R,,, = 2.5 mm. Time t is normalized by Rayleigh's collapse time T, of a spherical
bubble: T = t/T,, with T = 0 corresponding to the instant of bubble generation and T = 2.0
corresponding to the instant of spherical collapse. The length of the arrows represents the fluid
velocity at their respective starting point. Figure adapted from Vogel & Lauterborn (1988a).
representations for y = 2.4 achieved by this procedure. I n this case, spatial resolution
is two to three data points per mm2. The dynamic range of the velocity measurement
is about 15:1.
Interpolation of the velocity fields in figure 10 results in the pathline portrait in
figure 11 (a).Here, the solid boundary is drawn at the bottom of the diagram in order
to facilitate comparison with the numerical results in figure 11( b ) .The pathlines up
to bubble shape 1 were obtained from figure lO(a). Figure l O ( b ) allowed their
continuation up to shape 2. The pathlines within shape 2 are speculative
(a) T = 1.85
R,,, = 2.2 m m
% 1.1
,z.z
\2.9
(b) T = 2.08
R,, = 1.8 mm
$2
I 1 mm
51
.1.0
$3)
FIGURE 10. Velocity field in the vicinity of a cavitation bubble collapsing near a solid wall (hatched
area at the top) at times (a)T = 1.85 and ( b )T = 2.08, y = 2.4. The velocity values are given in m/s.
The diagrams are taken from two different series. Figure adapted from Vogel & Lauterborn ( 1 9 8 8 ~ ) .
continuations of the flow pattern in the bubble vicinity. Figure 11( a ) gives a good
visualization of the fluid flow concentration during bubble collapse which results in
jet formation. For a comparison of the experimental results with theory, in figure
11 (b) a pathline portrait for several points on the bubble surface is shown which was
numerically calculated by Kucera & Blake (1988) using the boundary integral-
method developed by Blake et al. (1986). There is remarkably good agreement
between experiment and theory.
312 A . Vogel, W. Lauterborn and R. Timm
FIGURE 11. ( a ) Experimentally obtained pathline portrait of the flow around a collapsing
cavitation bubble near a solid wall at y = 2.4. Bubble shape 0 represents the bubble a t maximum
expansion, shape 1 corresponds to T = 1.85 and shape 2 to T = 2.08. (6) Calculated pathline
portrait for y = 2.4 from Kucera & Blake (1988) for several points on the wall of the collapsing
bubble. The dots in (a) and (6) indicate the location of bubble formation. Figure adapted from
Vogel & Lauterborn ( 1 9 8 8 4 .
3.2. Ring vortex dynamics
Figures 12 and 13 show in side view and top view the bubble dynamics for relatively
large distances between bubble and boundary (y = 2.55 in figure 12a and y = 2.15 in
figure 13a). A ring vortex is formed after the first collapse, the core of which is the
cavitation bubble. The ring vortex which conserves the Kelvin impulse of the cavity
becomes clearly visible only after the decay of the bubble protrusion preceding the
second collapse. After taking a toroidal shape, the bubble distintegrates into various
parts which collapse separately but nearly a t the same time and coalesce again
during rebound. The ring vortex persists during several oscillations of its toroidal
bubble core (figure 12). I t migrates toward the solid boundary and starts to expand as
soon as it reaches the boundary (figure 13). After the second collapse, the vortex core
takes a polygonal shape which probably can be attributed to instability waves on the
vortex. This phenomenon has also been observed by Maxworthy (1977) on turbulent
vortex rings in single-phase flows. The jet is visible in the bubble centre as a dark bar
in aide view and as a dark spot in top view. The migration of the bubble or ring
vortex, respectively, is shown in figure 14 for the picture series from figures 12 (a)and
13(a). During first and second bubble collapse, the migration is accelerated, which
reflects the conservation of the Kelvin impulse as described in $3.1. The Kelvin
impulse is then carried by the ring vortex moving with nearly constant velocity
towards the boundary. The mean velocity is 6.4 m/s for the vortex in figure 12 ( a )and
5.2 m/s in figure 13(a). The smaller velocity in the second case results from the
smaller size of the bubble involved.
Figure 15 contains some pictures taken immediately after first bubble collapse
with the apparatus depicted in figure 2. In each picture the shock wave emitted
during bubble rebound is visible as a bright ring surrounding the bubble. In the lower
right picture one can see two shock waves indicating an irregular bubble collapse due
to deviations from sphericity of the expanded bubble. The images of the shock waves
FIGURE 12. Ring vortex formation near a solid boundary ( a )in side view according to figure 1 (b),
and (6) in top view with line of sight as in figure figure 1 ( c ) . Picture series taken a t 20000 frames/s.
(a)y = 2.55 and R,,, = 2.5 mm, ( b )y = 2.3 and R,,, = 2.1 mm. In ( a )the solid boundary is visible
as a dark stripe in the lower part of each frame. I n (6) the boundary is parallel to the plane of the
paper and transparent for photographic reasons. Here the dark lines at the lower and upper border
of some frames are the actual picture edges.
are blurred because of the relatively long exposure time of about 1 ps. The vertical
bar through the shock-wave image in the left picture is due to the directional
sensitivity of the high-pass filtering with a vertical wire. It can be inferred from the
photographs that the lower bubble wall opposite to the jet is not pierced by the jet
during collapse but pushed ahead during rebound. Otherwise the jet flow would have
been visible as a bright bar below the bubble because of the sensitive spatial filtering
technique applied. The temperature gradient of 1 K/mm around the site of optical
breakdown leads to EL difference of several K between the temperature of the liquid
close to the upper and lower wall of the expanded bubble. This is because during
expansion the liquid volume surrounding the laser focus is distributed into a thin
layer around the bubble without affecting the temperature distribution slightly
further away. The visualization of the jet flow in figure 8 shows that during jet
formation liquid that had been more than 1 mm away from the upper bubble wall is
314 A . Vogel, W.Lauterborn and R . Timm
FIQURE 13. Ring vortex formation near a solid boundary in ( a )side view and ( b )top view. Framing
rate ZOO00 s-l, (a)y = 2.15 and R,, = 2.05 mm, (b)y = 1.9 and R,,, = 2.6 mm. The explanation
of the dark borders of the frames is as given for ( a ) and (a) in figure 12.
entrained into the bubble. The temperature of this liquid certainly differs from that
of the liquid below the bubble by several K-much more than the temperature
difference of 0.5 K necessary for detecting the jet (see 52.1).
The bubble content of vapour and gas possibly acts as a cushion which in
conjunction with the surface-tension forces maintains the liquid surfaces of jet tip
and bubble wall intact. This prevents the penetration of the lower bubble wall
immediately after impingement of the jet. I n fact, the jet is observable as a bubble
protrusion in normal photography only owing to this gas and vapour layer
surrounding it. This means on the one hand that the tip of the jet can be identified
with the tip of the bubble protrusion, and on the other hand that, although the
vorticity has already built up in the course of the first collapse (see figure 11 a ) , the
vortex ring develops only after the disintegration of the protrusion shortly before the
second collapse. This process is visible in figure 7 (a)and shown with high temporal
resolution in figure 16. Induced by the ring vortex in the liquid surrounding the
cavity, there is probably another vortex generated inside the bubble which becomes
stronger during collapse owing to the reduction of bubble volume and compression
of its content. If the ring vortex inside the bubble behaves similarly t o turbulent ring
vortices in single-phase flows, a wake is formed upstream behind the ring (Maxworthy
1974). This may be the cause of the deformation of the upper bubble side, the
y = 2.55
R,,, = 2.5 m m +
'0
Second collapse
\O
' 0
*'
\.
'OrtA*
y=2.15
R,,, = 2.05 mm
I I I I
0 0.2 0.4 0.6 0.8 1.0 1.2
t (ms)
14. Development of distance 1 between bubble centre and boundary in the photo-
FIQURE
graphic series of figures 12 (a) and 13 (a).
FIGURE 15. Schlieren pictures taken immediately (i.e. 1-3 ps) after the first collapse of three
cavitation bubbles with y = 2.2 and (a)R,,, = 1.5mm, (b) R,,, = 1.6 mm, (c) R,,, = 2.5 mm. The
solid boundary is below the bubbles outside the frame format. The frame sizes are ( a ) 6 mm x
6.8 mm, (b) 4 mm x 10 mm and (c) 6.5 mm x 16 mm.
development of which can be observed while the jet keeps flowing through the
bubble. This interesting feature is not an exception but is almost typical for y-values
higher than 2 when no or only a weak counterjet appears during first collapse.
With decreasing y some changes are observed in the scenario described above. In
figure 17, where y is 1.66,the bubble touches the boundary during rebound after first
collapse so that a ring vortex can only exist for a very short period during second
collapse when the lower bubble side is again detached from the boundary (third and
fourth frame). In the fourth frame instability waves can be seen on the toroidal
I1 FLM 206
316 A . Vogel, W. Lauterborn and R.Timm
FIQURE16. Development of a ring vortex and formation of a wake during second bubble collapse.
Series taken with the rotating mirror camera at a framing rate of 300000 s-l, y > 2. The solid
boundary is visible as the dark stripe at the lower edge of each frame. Frame size 4.5 mm x 3 mm.
FIGURE17. Second bubble collapse at y = 1.66 and R,,, = 3.3 mm. Series taken with spatial
filtering at 20000 frrtmes/s. The location of the solid boundary is indicated by the white horizontal
bar in the left frame. Frame size 4.5 mm x 5 mm.
cavity. They are visible with the schlieren technique because it allows one to
distinguish between the principal cavity which blocks out the light and the cloud of
microbubbles surrounding the cavity which appear bright owing to their high
scattering efficiency. The microbubbles generally arise during the first bubble
collapse, since deviations from sphericity of the expanded bubble lead to instabilities
growing during collapse which finally cause the separation of parts of the principal
bubble (see Strube 1971). Figure 18 shows the first and second collapse for y = 1.37.
After the first collapse, no ring vortex forms but merely a radial outward flow on the
solid boundary arises from the jet decelerating the second collapse in horizontal
direction. Apparently, the toroidal bubble disintegrates into several separately
collapsing parts. This is gathered from the observation of two acoustic transients in
frame 15 emitted from different sites during second collapse (the acoustic transient
a t the left side of the picture is weak and only faintly visible). The protrusion of the
upper bubble side, which shortly after formation appears as a cloud of microbubbles,
indicates that a counterjet is formed during first collapse (see $3.3 below).
The bubble dynamics is especially interesting when y is at about unity. In this
case, a ring vortex has already formed before the first collapse. Figure 19 shows a
bubble with y = 0.96 at the stage of maximum expansion, a t an intermediate stage
and about 30 ps before collapse. In the last picture, the upper bubble side has already
FIGURE 18. First and second bubble collapse at y = 1.37 and R,,, = 4.0mm. Series taken with
spatial filtering a t 20000 frames/s. The boundary is indicated by the white bar in the left frames.
Frame size 4.8mm x 5.6 mm.
FIGURE 19. Schlieren pictures of bubble collapse a t y = 0.96 and R,,, = 3.9 mm. The left picture
shows the stage of maximum expansion (T = 1.03),the middle frame was taken at T = 1.80and the
right picture a t T = 2.19, corresponding to an instant about 30 ps before collapse. The location of
the solid boundary is given by the white dotted line of microbubbles below the laser-produced
cavitation bubble. Height of the frames 5 mm.
involuted to form the tip of the jet and the lower bubble side has detached from the
rigid boundary. The jet touches the opposite bubble wall before the bubble reaches
its minimum size and most probably pierces the wall because it does not give way to
the jet as in the case of large y , where the collision between jet and bubble wall
coincides with the beginning of the rebound. Thus, a ring vortex develops and during
the final stage of the collapse the bubble has a toroidal form as already observed by
Shutler & Mesler (1965).
The first numerical investigation of the aspherical collapse a t y = 1.0 was done by
Plesset & Chapman (1971).Their work has set a historical milestone which was only
surpassed 15 years later by Blake et al. (1986). I n figure 20, the results from both
papers are compared at different times T = t/T, with the experimental data from
figure 19. Here T, is Rayleigh’s collapse time, and t starts with the instant of bubble
generation. In order to facilitate the comparison, the elongation 8 of the upper part
of the bubble and the ratio p = Rv/Rhor of vertical and horizontal bubble radius is
11-2
318 A . Vogel, W . Luuterbom and R.Timm
T = 1.000 E = 1.00 T = 1.885 E = 1.10 T = 2.119 E = 0.85

p = 1.00 p= 1.21 p = 1.15
T = 1.047 E = 1.00 T = 1.856 E = 1.21 T = 2.164 E = 0.63
T = 1.030 E = 1.04 T = 1.800 E = 1.16 T = 2.190 E = 0.67

/3 = 0.96 p= 1.03 p = 0.53
FIGURE 20. Comparison of numerically calculated bubble shapes taken (a)from Plesset & Chapman
(1971) and (b) from Blake et al. (1986) with experimentally determined curves (c) obtained from
the picture series in figure 19. y = l.Oin the numerical calculations, and y = 0.96 in the experiment.
T, e and p are explained in the text. Note that ( a ) and ( 6 ) show a cross-section of the bubble,
whereas in (c) the bubble shape is depicted in side view.
given for each diagram. The comparison reveals that the results from the model of
Blake e l al. (1986) agree better with the actual bubble dynamics than those of Plesset
& Chapman (1971). This is very reasonable, because Plesset & Chapman only
considered the collapse phase from a sphere, whereas Blake et al. also included the
growth phase.
In the parameter range around y = 1.0, the fluid motion directed radially towards
t,he bubble centre is partly transformed into the rotational movement of the ring
vortex. The consequence is, as discussed in a theoretical paper by Chahine & Genoux
(1983), that the bubble implosion is decelerated and the bubble remains relatively
large during collapse. I n figure 21, the collapse is compared for the cases of y = 2.2
and y = 0.96. The minimum extension of the bubble is 3 mm horizontally and 1 mm
vertically at y = 0.96 but only 0.6 mm horizontally and about 80 pm vertically at
y = 2.2. Therefore, the pressure rise inside the bubble is relatively low a t y = 0.96
resulting in a weak sound emission, as will be shown in $3.4.3below.
FIGURE 21. Comparison of bubble collapse at different distances from the boundary. (a)y = 2.2 and
R,,, = 2.7 mm, ( b ) y = 0.96 and R,,, = 3.9 mm. In each case, the bubble is shown in the stage of
maximum expansion in the left frame and at the instant of minimum size in the right frame. In (a)
the middle frame was taken about 1 ps before collapse and in (b) about 30 ps before collapse. The
pictures in (a)were taken with the apparatus from figure 1 and those in ( b ) with that from figure
2. The solid boundary is given in (a)by the dark stripe a t the lower edge of each frame, and in ( b )
by the white dotted line of microbubbles below the principal bubble. Height of the frames: ( a )
4.3mm and (b) 5 mm.
3.3. Counterjet formation
In figures 7 ( b ) and 18, a protrusion on the upper bubble side can be seen immediately
after the first collapse. This phenomenon has been interpreted by Lauterborn & Bolle
(1975) as a consequence of a counterjet, i.e. a jet flowing in the opposite direction to
the principal jet. Two typical cases of bubble collapse with and without the
development of a counterjet are compared in figure 22. Figure 23 contains a
summary of the conditions of counterjet formation which has been obtained by
evaluation of high-speed photographic series. It appears that the occurrence of a
counterjet is correlated with the distance y between bubble and boundary as well as
with the bubble elongation e. Although the elongation is mainly a function of y, it
can also be affected by the form of the optical breakdown used for bubble generation :
an elongated plasma causes a greater elongation of the bubble during the final
collapse stage than a spherical plasma. To facilitate the discussion of figure 23, the
limits of the parameter ranges with weak, pronounced and strong counterjets are
sketched. It is clearly visible that the strength of the counterjet increases with
decreasing y (as long as y is larger than one) and with increasing e. As pointed out
in $3.1, a small y-value and a large e-value also lead to early jet formation. It can
therefore be assumed that the occurrence of a counterjet is related to the instant of
jet formation.
The following heuristic explanation of counterjet formation takes up the ideas of
jet formation presented in $3.1. Although the driving force for bubble collapse is the
difference between the pressure in the surrounding liquid and that inside the bubble,
the bubble wall is still accelerated when the pressure inside the bubble exceeds the
320 A . Vogel, W . Lauterborn and R . Timm
FIGURE 22. First bubble collapse (a) without and (6) with counterjet formation. (a) y = 2 . 3 ,
R,, = 2.0 mm, 6 = 1.04; ( b ) y = 1.56, R,, = 3.2 mm, e = 1.13. Pictures taken with the apparatus
from figure 1 at 20000 frames/s. The solid boundary is at the lower edge of each frame. Frame
size in (a)and (6) 7.3 mm x 5.6 mm.
0 1
I
I
I
I
0 1
I
I
0 1
I
I
I
0
1 0
I
1.00 I I I I I O l O
I n 1
0 0.5 1.o 1.5 2.0 2.5 3.0

Y
FIGURE 23. Strength of the counterjet as a function of the dimensionless distance y between bubble
and boundary and the bubble elongation E . The measurement points 1 and 2 correspond to the
photographic series of figures 22 ( a )and 22 (b).
static pressure in the liquid, because the fluid flows into a cavity with continuously
decreasing volume. Obviously, this acceleration due to inertia forces is effective only
as long as the curvature of the bubble wall is concave. Once the convex tip of the jet
is formed, its velocity remains fairly constant. This explains why the jet achieves a
higher velocity a t y = 1.5 than a t y = 1.0 when the tip of the jet appears earlier than
in the former case (see Plesset & Chapman 1971; Blake et al. 1986). Since the
curvature of the lower bubble side is concave during the whole collapse time, it is
FIGURE 24. Comparison of bubble collapse (a) without and with (b) counterjet formation. Series
taken with the rotating mirror camera at 940000 frames/s in (a)and 880000 frames/s in (b). The
solid boundary is below the bubble outside the frame format. Frame size in (u)and (b) 3 mm x
2.5 mm.
accelerated until the very end of collapse. Therefore, the velocity of its middle part
can exceed the jet velocity when y is low enough to cause a relatively slow jet. This
defines the condition of counterjet formation.
The above considerations are illustrated by figures 24 and 25, Since the
photographic record of the bubble dynamics with a rotating mirror camera a t very
high framing rates can only cover a small part of the collapse process, neither the
maximum bubble radius nor y are known exactly. However, it can be inferred from
the diagram in figure 23 that y must be larger than 2.3 in figure 24(a) and smaller
than 2.0 in figure 24(b).Figure 25 shows the movement of the top, bottom and centre
\ t 500
-15 -10
I \ Bottom
1 dbm)
I- 1000
Tip of counterjet
_._._.-. -.-.-
-15 -10 -5 20 f(W)
FIGURE
25. Distance d from the collapse centre for different points of the bubble wall versus
time. (a) is evaluated from figure 24(a) and (6) from 24(6).
of the bubble and of the tip of the counterjet. The respective velocities and
accelerations can be determined from the first and second derivatives of the curves.
Figure 24(b) shows that during collapse with counterjet formation the top of the
bubble is already involuted 6 ps before collapse, and from figure 2 5 ( b ) one can see
that it is no longer accelerated after this time. The bottom of the bubble, however,
is accelerated until the end of collapse so that its final velocity (140 m/s) exceeds that
of the upper bubble side (100 m/s). This is the origin of the counterjet which
causes a protrusion in the upper bubble wall even before the main jet deforms
the lower wall. At the beginning, the upward velocity of the protrusion is 35 m/s,
thus being almost as high as the velocity difference between upper and lower wall.
After only l o p s it has decreased to 8 m / s and after 20ps to almost zero. The
diameter of the counterjet is smaller than that of the jet owing to the longer lasting
focusing of the fluid flow creating the counterjet. It penetrates the centre of the jet,
thereby pushing ahead a part of the compressed gas and vapour inside the bubble by
which it becomes visible on the photographs. The protrusion disintegrates into a
cloud of microbubbles which remain behind while the bubble centre is migrating
towards the rigid boundary. After a short time, the jet overcomes the distortion by
the counterjet due t o its larger volume and impulse. However, the jet flow has
become extremely turbulent, is thicker than in the case without a counterjet and
causes a broader protrusion of the lower bubble wall, The whole surface of the
rebounding bubble is rough owing t o the interaction between jet and counterjet (see
figures 22 b and 24 b).
When y is larger than 2, the mechanism described above is usually not strong
enough to cause a counterjet, but it is responsible for the flattening of the bubble
during the ultimate collapse phase which can be seen in figure 25 (a).I n this case, the
upper bubble wall features no involution until the end of collapse, and its velocity
(110 m/s) remains higher than that of the lower wall (90 m/s). When y is smaller
than unity, no counterjet can be developed in spite of early jet formation, because
the bottom of the bubble is retained by the solid boundary after the bubble has
reached the stage of maximum expansion.
The experiments reveal that a t y = 1.5 a counterjet is always formed. It is
remarkable that this phenomenon does not emerge in jmy of the numerical
calculations performed with this y-value published to date. Lauterborn & Bolle
(1975)compared experimentally determined bubble shapes at y = 1.5 and R,,, = 2.6
mm with curves numerically obtained by Plesset & Chapman (1971).The agreement
was very good until the end of the investigated period (about 2 ps before collapse),
although the experimental bubble developed a strong counterjet visible after
collapse. The middle picture in figure 21(a) taken about 1 ps before collapse also
closely resembles the diagrams of the latest stage in bubble collapse that could be
calculated with the models of Plesset & Chapman (1971) and Blake et al. (1986) for
y = 1.5. Hence, the counterjet formation takes place in the ultimate collapse phase
which can no longer be modelled by the numerical calculations. This is also confirmed
by figure 25 ( b ) showing that the velocity of the lower bubble wall exceeds that of the
upper wall only during the final microsecond before collapse. I n order to reflect the
whole diversity of phenomena observed experimentally, including counterjet and
ring vortex formation, the numerical calculations ought to be able to track the
bubble dynamics after the jet reaches the opposite bubble wall, and they ought to
take into account the bubble content of vapour and gas. These are certainly very
important during the final collapse stage, as well as surface-tension forces and the
effects of compressibility.
3.4. Sound emission
3.4.1. Projile of the acoustic transients
The acoustic transients emitted after cavitation-bubble generation and after
bubble collapse have been investigated with the experimental arrangement outlined
in figure 5 . Figure 26 shows typical signals from hydrophone 2 together with the
corresponding results of the optical detection. The optical measurement reveals that
the acoustic transients are much shorter than the rise time of the pressure
transducer. For evaluating the optical measurements, first the distortion of the
signal resulting from the nonlinearity of the U(4)-characteristic of the apparatus (see
figure 6) is corrected. Then p(t)/P,,, is calculated by numerical integration. Two
pressure profiles obtained in this way are given in figure 27. The duration T (width
a t half-maximum) of the acoustic transients is about 25 ns both after laser-induced
plasma formation and after bubble collapse. The same duration wits observed for the
acoustic pulses after the second bubble collapse. The pressure pulse caused by the
expanding plasma features a steep shock front with a rise time of 10 ns. Since this
324 A . Vogel, W. Lauterbomz and R.Timm
FIGURE 26. Typical signals of the PVDF-transducer (top) and of the photodetector (bottom) ( a )
after laser-induced bubble generation and ( 6 ) after cavitation bubble collapse. The distance
between the emission centre of the acoustic pulses and the probe was 1.2 mrn for the optical and
8 mm for the acoustic measurement. Figure adapted from Vogel & Lauterborn (19886).
value equals the temporal resolution of the apparatus, the actual shock front is
probably even shorter. During cavitation-bubble collapse, a relatively slow pressure
rise occurs within 1-2 ps followed by a shock front similar to that after breakdown.
This result can be explained by the theoretical work of Hickling & Plesset (1964)and
Fujikawa & Akamatsu (1980).They found a strong pressure rise near the bubble wall
during the last stages of the collapse and the evolution of a shock front during the
rebound of the bubble. Behind the shock front, the pressure decreases nearly
exponentially.
The duration of the acoustic transients was measured as a function of their
distance r from the emission centre in the range from 0.6 mm to 6.0 mm. The pulse
duration was found to be nearly constant both for the transients after bubble
generation and after bubble collapse. A broadening of the pressure profile predicted
by the propagation theory for finite-amplitude waves of Cole (1948) and Esipov &
Naugol’nykh (1972) was not observed. This surprising result is probably due to the
spherical form of the acoustic waves which leads to a rapid decrease of the pressure
amplitude so that the domain of nonlinear sound propagation is limited to a small
region around their emission centre.
The duration of the acoustic transients emitted during bubble collapse near a solid
boundary is the same as during spherical collapse when y d 0.75 or y 2 1.05. In the
range 0.75 < y < 1.05 the acoustic pulses are too weak to be detected with the optical
technique. The signals from the pressure transducer, however, often have a duration
of several microseconds, i.e. longer than the impulse response of the measuring
instrument. It is not clear whether this is due to one long pressure pulse or to a series
of short pulses which could not be resolved by the transducer. The latter is suggested
by the schlieren image in figure 21 ( b ) .
1.o
I I I I I I I I I I 1 I I
(4
0.8 -
0.6 -
-
P
Pmax
0.4 -
0.2 -
- I / I I I I I I I I
0 10 20 30 40 50 60 70 80 90 100 110 1201!130
I I I I I I l l I I I I I I
1.o -
0.8 - -
0.6 -
p
Pmax 0.4 -
0.2 -
I I I I I l l
0 10 20 30 40 50 60 70 80 90 100 110 120 130 !140
t (ns)
FIGURE
27. Pressure profiles ( p ( t ) / p m a xof) the acoustic transients (a) after laser-induced bubble
generation and ( b ) after cavitation bubble collapse: -, experimentally determined profile ;
_ _ _ _ , exponential decrease. The measurement distance from the emission centre of the acoustic
pulses was 1.2 mm. ( a )and ( b ) correspond t o the photodiode signals (a)and ( b ) in figure 26. Figure
taken from Vogel BE Lauterborn (1988b).
3.4.2. Spherical-bubble collapse

The investigation of the sound emission of spherical cavitation bubbles is of no
direct interest for the explanation of cavitation erosion, because bubbles close
enough t o a solid boundary t o cause damage will always collapse aspherically.
Nevertheless, it provides the basis for interpretation of the data obtained for the
asymmetrical collapse and is to date the only means of comparing experimental
results with theory.
The collapse times Tcl,T,, and Tc3for the first, second, and third spherical bubble
collapse are proportional to the maximum bubble radius R,,, before the respective
collapse. The bubble radius can be deduced from the time distance 2T, between the
peaks of the pressure signal making use of the relationship
derived for spherical collapse by Rayleigh (1917).p is the density of the liquid, pStat
the static pressure and p , the vapour pressure of the liquid. The energy E , of the
cavitation bubble is given by
E, = d @stat -~ v%at.
) (4)
326 A . Vogel, W. Lauterborn and R. Tirnm
The energy loss during bubble collapse is obtained by comparing the different values
of the bubble energy before and after collapse. The peak pressure amplitude of the
acoustic transients is calculated from the transducer signal, and the pulse duration
with the help of (1). I n cases where 7 could not be determined synchronously with the
acoustic measurement, we calculated the peak pressure assuming a pulse duration
r = 20 ns a t the site of the transducer. The energy of a spherical acoustic transient
as derived by Cole (1948) is
E, = -
47cr2/ p z dt, (5)
PC
where c denotes the sound velocity in the liquid and r the distance between the
pressure transducer and the emission centre of the transient.
I n figure 28 ( a )the peak pressure of the acoustic transients emitted during the first
and second cavitation bubble collapse is plotted versus the maximum bubble radius.
The pressure values refer t o a distance r of 10 mm between the bubble centre and the
site of measurement. The acoustic signals used for evaluation have been selected to
exclude as far as possible bubbles that deviate from the spherical form. Those
bubbles are indicated by multiple pressure pulses emitted during both bubble
generation and collapse, which are easily visible with the optical detection technique.
The shock waves measured after the first collapse of a spherical bubble generally
have approximately the same amplitude as those emitted by the laser-induced
plasma generating the bubble, Figure 2 8 ( b ) shows that the energy of the acoustic
transients increases proportionally to the energy of the cavitation bubbles as long as
the bubble energy does not exceed 30 m J , corresponding to a radius of 4 mm. Larger
bubbles can only be produced with a high laser pulse energy of more than 300 mJ.
Therefore, they probably frequently feature deviations from the spherical form
because multiple plasmas are generated in the region surrounding the ruby laser
focus. Since these deviations grow during bubble collapse (see Strube 1971 ; Plesset
1974), the bubble implodes less violently than during spherical collapse and the
sound emission is diminished.
The average energy loss of the cavitation bubbles during their first collapse is
84%. It was found that the major part of it - a t most 90% and on average 73Y0 -
is due to the emission of sound. Heat conduction and viscosity seem to play only a
minor part in damping the oscillations of transient spherical bubbles. These findings
are in good agreement with the theoretical work reported by Ebeling (1978),
Nishiyama & Akaizawa (1979) and Fujikawa & Akamatsu (1980) as well as with the
experimental work of Teslenko (1980).The results of Hentschel & Lauterborn (1982),
who claimed that only about 1.2% of the bubble energy is converted into acoustic
energy, were obtained without properly taking into account the limited bandwidth
of the pressure transducer.
In the domain of linear sound propagation - when the pulse broadening and
dissipation can be neglected - the amplitude of a spherical acoustic wave is inversely
proportional to the distance r . Hickling & Plesset (1964) as well as Fujikawa &
Akamatsu (1980) have shown by numerical calculations that the 1/r law is also
approximately true for the nonlinear domain in the direct vicinity of the collapse
centre. I n this region, the effects due to the dissipation of sound energy are
compensated for by the development of the shock front. As shown in the next
paragraph, we found that the minimum radius of laser-produced cavitation bubbles
during first collapse is about 50 pm. Taking the experimentally observed pressure
amplitudes as a basis and using the l / r law, the pressure at r = 50 pm is calculated
500 I I I I I
200 -
0 0 0
150 -
100 - . .. ...*..
50 - ,'.. .-
1 I I I I 0-
0 1. Collapse
' 2. Collapse . a
20
0 10 20 30 40 50 60
4 (mJ)
FIGURE 28. ( a )Peak pressure p of the acoustic transients emitted during first and second spherical
bubble collapse as a function of the maximum bubble radius Itmax. The pressure values are
normalized to a distance r = 10 mm from the bubble centre. ( b ) Energy E, of the acoustic transients
plotted versus the bubble energy E,. Figure adapted from Vogel t Lauterborn (1988b).
t o be 60 kbar when R,, = 3.5 mm. Thus, the maximum pressure developed inside
the bubble during collapse is about 60 kbar.
The minimum bubble radius during spherical collapse was measured with the help
of the high-speed photographic series shown in figure 29, which were taken with the
image converter camera at 1 million frames/s. For each series, a radius-time curve
obtained through numerical calculations based on the model of Gilmore (1952) was
fitted to the experimentally determined radius values. The Gilmore model takes into
account the effects of compressibility of the liquid, the surface tension and the bubble
content of permanent gas. The bubble radius from the first picture was in each case
taken as the initial condition for the numerical calculation and the best fit was
(4
800
700
600
500
R (Pd
400
300
200
100
500
400
300
200
100
0
-5 0 5
t (PSI
FIGURE 29. High-speed photographic series of spherical bubble collapse far from boundaries taken
with the image converter camera at 1 million frame/s, and comparison of the corresponding radius
values with numerical calculations based on Gilmore’s theory. t , is the instant of collapse obtained
from the position of the acoustic transients on the photographs.
achieved by varying the associated velocity of the bubble wall, and the equilibrium
radius of the bubble, i.e. the radius, where the pressure of the permanent gas inside
the bubble equals the static pressure of the liquid. The instant t, of collapse was
experimentally obtained from the position of the acoustic transients on the
photographs using the normal velocity of sound in water (1490 m/s; the exact
velocity of the acoustic transient is not known, because of its shock-wave properties
in the initial stage of propagation). Since in figure 29 ( a )the moment of exposure of
the last frame is almost identical with t, as well as with the time of collapse resulting
from the numerically calculated curve, the radius value of 50 pm taken from the
photograph is identified with the minimum bubble size. An uncertainty is left,
however, because the exposure time of 200 ns may cause a blurred image if during
a short time the bubble is smaller than 50 pm. I n figure 29 ( b ) t , does not correspond
to the instant of collapse given by the theoretical curve, and in both figures 29 ( a )and
29 (b) the minimum bubble size calculated numerically (4.9 pm in a and 6.8 pm in b )
is much smaller than the experimental value. Obviously, the Gilmore theory does not
properly model the final stage of the collapse. This might be due to its neglect of non-
equilibrium condensation of the vapour inside the bubble and of heat conduction.
Fujikawa & Akamatsu (1980) have taken these two factors into consideration, and
have shown that in this case the bubble contracts slowly in the final stages of the
collapse in comparison with the adiabatic case assumed in the Gilmore model.
Moreover, the perfect sphericity assumed in theory is never completely fulfilled in
reality, which leads to a weaker collapse than theoretically predicted.
3.4.3. Bubble collapse near a solid boundary

The solid boundary leads to a decrease of the amplitude of the pressure pulse
caused by the first bubble collapse and to an increase of the pressure produced at the
second collapse. In addition, the boundary causes a prolongation of the collapse time
so that (3) does not hold any longer. The maximum bubble radius may still be
determined from the collapse time, if the prolongation factor against Rayleigh’s
collapse time for the spherical collapse is known. The knowledge of the maximum
bubble radius before the first collapse and after rebound makes it possible to
calculate the bubble energy and the energy loss during collapse. The dependence of
the prolongation factor k on the dimensionless distance y between bubble and
boundary was experimentally investigated by evaluating a large number of high
speed photographic series of bubble dynamics. The result is shown in figure 30
together with a theoretical curve derived from a perturbation analysis by Rattray
(see Plesset & Chapman 1971).
Figure 31 shows the pressure amplitude of the acoustic transient after the first
bubble collapse as a function of y , The pressure was normalized to a distance of
10 mm from the collapse centre. Since the sound emission during aspherical bubble
collapse may be anisotropic, two series of measurements were performed with the
pressure transducer placed above the bubble and the boundary in one series and
beside the bubble in the other series. The average bubble radius was R,,, = 3.5 0.6
mm in both cases. Thus, the pressure values for a certain y-value may belong to
bubbles of slightly different amplitudes. This partly masks the law governing the
relationship between sound emission and the normalized distance y. The importance
of sound emission in damping the bubble oscillations for aspherical collapse can be
estimated by relating the energy of the acoustic transient produced during the
collapse to the energy loss of the bubble as shown in figure 32. The curves have been
fitted to the measurement values in order to facilitate the comparison of both
measurement series with different microphone positions.
Looking at figures 31 and 32, the strong decrease in sound emission for y-values of
about y = 0.9 attracts attention. This behaviour was also observed by Naudd & Ellis
(1961) and by Tomita & Shims (1986). It is due to the modification of the bubble
collapse by jet and ring vortex formation discussed in ss3.1 and 3.2. For large y-
values, the collapse is almost spherical and the liquid jet does not emerge before the
rebound of the bubble. Therefore, the liquid moves almost radially towards the
collapse centre, leading to a violent collapse with strong sound emission. I n the range
around y=O.9, the jet appears before the bubble has achieved its minimum
1.6
1
1.5
1.4
1.3
k,
1.2
1.1
1.o
0 0.5 1.o 1.5 2.0 2.5 3.0 3.5

Y
1 A L \
V.,
I I I I I I I
0 0.5 1.o 1.5 2.0 2.5 3.0 3.5
Y
FIGURE 30. Prolongation factors k, and k, against the time for asymmetrical bubble collapse as a
function of y (a)for the first collapse and ( b )for the second collapse : - - - -, theoretical curve derived
by Rattray (see Plesset & Chapman 1971); -, curve fitted to the values obtained experimentally.
Figure taken from Vogel & Lauterborn (19886).
extension. The jet penetrates the bubble wall at the side of the solid boundary, and
during the last stages of the collapse a vortex ring with a hollow toroidal core
consisting of gas and vapour is formed which moves towards the boundary. The
kinetic energy of the radial flow into the bubble is thereby partly transformed into
energy of rotation. Hence, the bubble contents becomes less compressed than in the
case of large y-values (see the comparison of minimum bubble sizes for both cases in
figure 21a and b ) , and the sound emission is diminished. For y+O, the sound
emission increases again, although the jet is still formed very early during the
collapse, similar to the case of y = 0.9. However, if y is very small, the bubble is
nearly hemispherical and its centre can hardly move during collapse. Therefore, no
pronounced vortex ring is developed in spite of the outward flow on the surface of the
solid boundary following jet impact. The flow is directed towards the bubble centre
for most parts of the bubble surface, as in the spherical collapse. This causes a strong
compression of the bubble contents and the generation of an intense pressure pulse.
Figure 32 shows that for y < 2.0 the damping of the bubble oscillation is
predominantly caused by mechanisms other than the emission of sound. Instead,
damping is probably mainly due to viscosity and heat transport, which are
,
I 1 I
- . . I . I
1
250 - (a) * .
:,. ... . *. ....- .

. .- ., .
:
200 ' '
..
... ' . ..... .
. .. - -.
.. .
-.
*-.
.. .. .*
I I I 1 I I I
0 0.5 1.0 1.5 2.0 2.5 3.0

Y
. . '.
200 I - ..
. .. .
-*
..- :.. ..
I.
*
'.'.:
. .
100 i. 7
.;
..;
50
. -. . 5
I I I I
0 0.5 1.0 1.5 2.0 2.5 3.0

Y
FIGURE 31. Pressure amplitude p after the first bubble collapse as a function of y . The pressure
values refer to a distance of 10 mm from the collapse centre. R,,, = 3.5k0.6 mm. The hydrophone
was placed (a)above and ( b )beside the bubble. Figure adapted from Vogel & Lauterborn (1988b).
significant within this parameter range because of the turbulent jet flow and the ring
vortex.
For y < 1.1, this jet hits the solid boundary before the bubble reaches its minimum
size, so that the pressure pulses due to jet impact and bubble rebound are generated
a t different times. Nevertheless, only one strong transient is registered by the
hydrophone during bubble collapse. It is very likely that this is the transient emitted
by the rebounding bubble. The pressure pulse produced a t the site of the jet impact
has to pass the cavitation bubble with a low acoustic impedance until it reaches the
hydrophone. It is therefore strongly attenuated a t the side where the liquid is,
whereas the coupling to the solid wall is much better. The pulse from the rebounding
bubble, however, is directly transmitted t o the transducer by the liquid surrounding
the bubble. This view is supported by holographic interferograms of shock-wave-
%
1 I 1 I I
--I
70 -
Y
FIGURE 32. Fraction of the energy loss AEB during the first bubble collapse which is converted into
acoustic energy E,, plotted versus y. The average value for the spherical collapse is marked by the
horizontal bar. -, hydrophone placed above the bubble : - - - -, hydrophone placed beside the
bubble. Figure taken from Vogel & Lauterborn (1988b).
induced bubble collapse on a transparent solid wall taken by Sanada et aE. (1986).The
interferograms show the pressure field in the liquid surrounding the bubble and
in the solid wall, indicating strongly anisotropic sound emission. Our acoustic
measurements thus only allow the determination of the pressure inside the collapsed
cavitation bubble but not the impact pressure of the jet, even though the latter may
locally be higher than the former (see also $3.5 below).
The highest pressure amplitudes a t the solid boundary due to the compression of
the bubble contents are achieved when y is very small. One can infer from figure 31
that for y = 0.2 and R,,, = 3.5 mm the pressure amplitude after the first bubble
collapse is about 150 bar a t a distance of 10 mm from the collapse centre. This value
is the average of the data from each hydrophone position. It is estimated from figure
7 ( d ) that the minimum volume of the collapsed bubble corresponds to that of a
bubble with a radius of about 0.6 mm. Calculation of the pressure value a t this radius
proceeding from the measured amplitude yields the value of 2.5 kbar for the
maximum pressure inside the bubble and at the boundary. This result agrees well
with the findings of Jones & Edwards (1960) (10 kbar for y = 0 and R,,, = 4.9 mm),
but it is almost 17 times higher than the value reported by Tomita & Shima (1986)
(150 bar for y = 0.2 and R,,, = 3.5 mm). The deviation can be explained by the fact
that Tomita & Shima employed a pressure transducer with a rise time of 1 ps without
considering the much shorter duration of the pressure pulses to be measured, whereas
Jones & Edwards performed a dynamic calibration of their transducers.
3.5. Cavitation erosion

Our measurements have shown that the highest pressure amplitudes at the solid
boundary due to the compression of the bubble contents are achieved for y+O. I n
this case, moreover, there is no water film between bubble and boundary cushioning
the impact of the liquid jet. Bubbles attached to the boundary thus have the largest
damage potential. Nevertheless, from the measured and extrapolated jet velocities
and pressure values some reasonable doubt arises about whether the action of a single
isolated cavitation bubble collapsing in a liquid without additional flow will lead to
erosion of hard metals. As discussed below, there is some evidence that only
modifications of the bubble collapse by shock wavebubble interaction or by a flow
carrying the bubble will produce liquid jets with a velocity high enough to damage
such materials.
The highest jet velocity observed in our study was 156 m/s for y about 2.3. This
value is obtained from figure 25(a) by identifying the tip of the jet with the lower
bubble wall. I n figure 25(b) with y < 2 the velocity is 118 m/s. A reduction of jet
velocity with decreasing y was also found in the numerical studies by Plesset &
Chapman (1971) (170 m/s for y = 1.5,130 m/s for y = 1.0)and by Blake et al. (1986)
(161 m/s for y = 2.0, 110 m/s for y = 1.5, 86 m/s for y = 1.0, when a collapse in
water under atmospheric pressure is assumed). We have shown in figure 20 that the
second numerical model is a better approximation to real bubble dynamics than the
first. Therefore, 86 m/s is a more realistic value for the jet velocity a t y = 1.0 than
130 m/s. Only for y-values smaller than about 0.9 does the jet hit the solid wall
without being decelerated by a water layer or being transformed into a ring vortex
before the end of the first collapse (see $3.2). For estimating the damage capability
of the jet we therefore have to presume a jet velocity of 100 m/s or less.
The water hammer pressure corresponding to a jet velocity of 100 m/s is about
1.5 kbar. Since, according to Lesser & Field (1983) and to Lush (1983), the pressure
due to the impact of a liquid jet with a conical or round tip can be up to 2.9 times
the water hammer pressure, it may reach about 4.5 kbar. This is in the same range as
the pressure of 2.5 kbar achieved inside a collapsed cavity attached to the wall. Both
values are close to the (static) Brine11 hardness of, for example, 99 % pure aluminium
which is about 3 kbar (Wellinger & Gimmel 1963), but Lush (1983) found the
dynamic hardness of 99% pure aluminium measured with a falling ball test to be
about 13 kbar, and Tomita & Shima (1986) observed pitting of an indium specimen
hit by a pressure-wave-induced jet only after the water hammer pressure was more
than ten times the yield strength of indium. An initially spherical cavitation bubble
should therefore probably hardly be able to erode aluminium or harder metals
neither by jet impact nor by the pressure inside the collapsed bubble. On first
sight, this seems to be contradictory to the results of Naud6 & Ellis (1961) and Lush,
Wood & Carpanini (1983) who observed small pits in aluminium below the centre of
spark-produced bubbles. Their bubbles, however, were very large (R,,, = 5 mm),
and pits could only be produced in a very small interval around y = 0.2. Shutler &
Mesler (1965) and Tomita & Shima (1986) found no central pits, but only a ring-
shaped damage pattern located a t the periphery of the collapsed bubble. Tomita &
Shima pointed out that this could be explained by the interaction of the outward-
flowing jet with the contracting cavity surface as a result of which microbubbles are
produced. One can see in figure 7 (c) with y = 0.75 and in 7 ( d ) with y = 0.34 that for
small y-values the cavitation bubble reaches its minimum size about 50 ps after the
jet has hit the boundary. This means that the microbubbles produced by the jet flow
shortly after their generation are struck and collapsed by a pressure pulse of about
2.5 kbar arising from the collapse of the main cavity. Thereby, probably much higher
pressures are produced locally a t the solid boundary. They may be well above the
dynamic hardness of aluminium or other metals.
A similar interaction between bubbles and pressure pulses occurs on a larger scale
during the collapse of cavity clouds in hydraulic cavitation and during extracorporeal
shock wave lithotripsy. It has been reported by Tomita & Shima (1986) that a
pressure pulse with an amplitude of 50 bar hitting a gas bubble produces jets with
a velocity of 200-370 m/s. Dear & Field (1988) have observed jets with 400 m/s after
a disk-shaped cavity was struck by a shock wave with a strength of 2.6 kbar. The
respective impact pressures of the jet are 3-6 kbar, or even 9-18 kbar, if the tip of
the jct has the appropriate shape. It is likely that the jets can be even faster when
the acoustic transients interact with already collapsing cavitation bubbles rather
than with gas bubbles. A pressure pulse with an amplitude of 50 bar necessary to
induce the fast jets can be produced by the spherical collapse of a cavitation bubble
with R,,, = 3.5 mm at a distance of 60 mm from the first bubble (see figure 28). This
distance decreases to 30 mm if a slightly asymmetric collapse ( y + 1.5)of the second
bubble is assumed (see figure 31). Bubble-pressure pulse interactions as described
above are very likely to occur in bubble clusters in hydrodynamic cavitation. They
are probably also responsible for the damage due to liquid-jet impact observed
recently by Coleman et al. (1987) and Delius et al. (1988) in experiments with an
extracorporeal shock wave lithotripter. Each pressure pulse from a lithotriptor
consists of a peak with an amplitude of several hundred bar followed by a part with
a negative pressure of about 100 bar. The negative pressure produces Cavitation with
both transient and stable bubbles. The stable bubbles are then collapsed by the
leading edge of the following pressure pulse. Delius et al. have found damage
structures explainable only by simultaneous jet impact at several adjacent sites. This
has given evidence that the pits observed have not been caused by simple cavitation
bubble collapse but by the simultaneous interaction of preformed bubbles with a
pressure pulse.
Another mechanism that may influence the jet velocity is the deformation of the
initial bubble shape by a flow around the bubble. We have discussed in $3.3. how
the jet is linked to the bubble elongation developing during collapse. We found that
a strong elongation of the bubble in the vertical direction, i.e. perpendicular to the
solid boundary, causes a low velocity which for y > 1 allows a counterjet to develop.
Voinov & Voinov ( 1976) have theoretically investigated the consequences arising
when a bubble detached from the wall is elongated in the horizontal direction a t the
stage of maximum bubble expansion. They found that at an elongation of only 10YO
the jet velocity is twice that in the case of an initially spherical bubble. Larger
elongations lead to even higher jet velocities but also to a strong decrease of the
liquid volume forming the jet. First experimental evidence for this theory can be
found in the work of Kling & Hammitt (1972).They generated cavitation bubbles in
a flow through a venturi with a velocity of 25 m/s under an ambient pressure of
2 bar. This led to an elongation of 15% parallel to the venturi walls (measured a t
maximum bubble expansion) for a bubble at y = 1.14 and R,, = 2 mm. The
maximum jet velocity (taken from their figure 5 ) was about 300m/s, and they
reported an impact velocity of 120 m/s resulting in pits in an aluminium foil in spite
of the water layer between bubble and boundary. The numerical calculations of
Blake et al. (1986) for the no-flow case give a maximum jet velocity of 122 m/s for
y = 1.0 and an ambient pressure of 2 bar, and our measurements with y < 2 (figure
2 5 b ) deliver a jet velocity of 167 m/s when the results are transformed to an ambient
pressure of 2 bar. The jet velocity observed by Kling & Hammit (1972) for a bubble
elongated 15 % in the horizontal direction is thus twice as high as it would have been
for an initially spherical bubble. A further confirmation of the acceleration of the jet
due to bubble elongation is given by the results of Ellis & Starrett (1983). They
reported that the impact pressure of liquid jets from bubbles elongated parallel to the
solid boundary in a stagnation flow of 6 m/s is more than an order of magnitude
higher than the pressure produced in the no-flow situation.
Thus there are two mechanisms that may enhance the damage capability of
collapsing bubbles beyond the threshold of damage for stronger materials : (i)
pressure pulses from adjacent collapsing bubbles ; and (ii) a fluid flow leading to an
elongation parallel to the boundary at the stage of maximum bubble expansion.
4. Conclusions
The dynamics of laser-produced cavitation bubbles near a solid boundary has been
investigated by high-speed photography, time-resolved particle image velocimetry
and acoustic measurements. The main results are as follows:
(i) Bubble collapse near a solid boundary is accompanied by the formation of a
high-speed liquid jet toward the boundary. The jet flow leads to ring vortex
formation after the first collapse for a relative distance y between bubble and
boundary larger than 1.5, and to ring vortex formation before the first collapse for
y-values of about 0.9. The ring vortex formation before the first bubble collapse
causes a strong reduction of the sound emission during collapse.
(ii) From successive measurements of the velocity field of the fluid flow in the
vicinity of the bubble, pathlines have been derived agreeing well with the results of
numerical calculations. It was shown that the dynamical behaviour of the bubble is
more accurately described by the numerical model of Blake et al. (1986) than by the
model of Plesset & Chapman (1971) which does not include the growth phase of the
bubble.
(iii) Counterjet formation disturbing the development of the principal jet has
been found to be a common feature of bubble dynamics in the parameter range
2.0 2 y 2 1.0.
(iv) Sound emission is the main damping mechanism in spherical bubble collapse
in water, where on average 73 Yo of the energy loss of the bubble is due to the emission
of an acoustic transient. It plays a minor part in the damping of aspherical bubble
collapse, especially in the range of 1.3 2 y 2 0.6, where it contributes less than 10 %
t o the energy loss of the bubble.
(v) The pressure profile of the acoustic transients has the form of an exponential
impulse with a rise time of less than 10 ns. The duration of the pressure pulses (their
width a t half-maximum) has been found to be 20-30 ns a t from 0.6 mm to 6 mm from
the emission centre of the pulses.
(vi) During spherical bubble collapse, a maximum pressure of about 60 kbar is
developed inside a bubble with R,,, = 3.5 mm. The highest pressure amplitudes
a t the solid boundary are produced when y approaches zero. For y = 0.2 and
R,,, = 3.5 mm, the maximum pressure inside the bubble is about 2.5 kbar. This is
about the same value as the estimated water hammer pressure resulting from jet
impact on the boundary.
(vii) J e t formation seems to be the mechanism mainly responsible for cavitation
erosion - but to fully realize the damage potential of the jet, one has to consider the
acceleration of jet formation by acoustic transients from adjacent bubbles and/or the
change of jet velocity due to deviations from the spherical bubble form which may
occur in a fluid flow.
The work reported here was financially supported by the Fraunhofer-Gesellschaft.

The experiments with the rotating mirror camera were done at the Institut fur den
Wissenschaftlichen Film in Gottingen.
REFERENCES
ARON-ROSA, D., ARON,J. J., GRIESEMANN, M. & THYZEL,R. 1980 Use of the neodymium:YAG
laser to open the posterior capsule after lens implant surgery: a preliminary report. Am.
Intraocul. Implant SOC.J . 6 , 352-354.
BENJAMIN, T. B. & ELLIS,A. T. 1966 The collapse of cavitation bubbles and the pressure thereby
produced against solid boundaries. Phil. Trans. R. SOC.Lond. A 260, 221-240.
BLAKE,J. R. & GIBSON,D. C. 1981 Growth and collapse of a vapour cavity near a free surface.
J. Fluid Mech. 111, 123-140.
BLAKE,J. R., TAIB,B. B. & DOHERTY, G. 1986 Transient cavities near boundaries. Part 1 . Rigid
boundary. J. Fluid Mech. 170,47%497.
BREWER,R. G. & RIECKHOFF, K. E. 1964 Stimulated Brillouin scattering in liquids. Phys. Rev.
Lett. 13, 334-336.
CAROME,E. F., MOELLER, C. E. 6 CLARK,N. A. 1966 Intense ruby-laser-induced acoustic impulse
in liquids. J. Acoust. SOC.A m . 40,1462-1466.
CHAHINE,G. L. & GENOUX,P. G. 1983 Collapse of a cavitating vortex ring. Trans. ASNE I:
J . Fluids Engng 105, 40M05.
COLE, R. H. 1948 Underwater Explosions. Princeton University Press.
COLEMAN,A. J., SAUNDERS, J. E., CRUM,L. A. & DYSON,M. 1987 Acoustic cavitation generated
by an extra corporeal shockwave lithotripter. Ultrasound Med. Biol. 13, 69-76.
DEAR,J. P. & FIELD, J. E. 1988 A study of the collapse of arrays of cavities. J. Fluid Mech. 190,
409425.
DELIUS,M., MULLER, M., VOQEL,A. & BRENDEL, W. 1988 Shock waves and cavitation. In Biliary
Lithotripsy (ed. J. Ferrucci, J. Burhenne & M. Delius). Chicago: Yearbook.
EBELINQ, K. J. 1978 Zum Verhalten kugelformiger, lasererzeugter Kavitationsblasen in Wasser.
Acustica 40, 229-239.
ELLIS,A. T. & STARRET,J. E. 1983 A study of cavitation bubble dynamics and resultant
pressures on adjacent solid boundaries. I n Proc. 2nd Intl Conf. on Cavitation, paper C 190/83.
I. Mech. E, London.
ESIPOV,I. B. & NAUQOL’NYKH, K. A. 1972 Expansion of a spherical cavity in a liquid. Soviet
Phys. Acoust. 18, 194-197.
FANKHAUSER, F., ROUSSEL,P., STEFFEN, J.,VANDER ZYPEN,E. & CHRENKOVA, A. 1981 Clinical
studies on the efficiency of high power laser radiation upon some structures of the anterior
segment of the eye. Intl Opthlmol. 3, 12S139.
FUJIKAWA, S. & AKAMATSU, T. 1980 Effects of the non-equilibrium condensation of vapour on the
pressure wave produced by the collapse of a bubble in a liquid. J. Fluid Mech. 97, 481-512.
GILMORE, F. R. 1952 The growth or collapse of a spherical bubble in a viscous compressible fluid.
Cal. Tech. Inst. Rep. 2 M .
HENTSCHEL, W. & LAUTERBORN, W. 1982 Acoustic emission of single laser-produced cavitation
bubbles and their dynamics. Appl. Sci. Res. 38, 225-230.
HICKLINQ, R . & PLESSET, M. S. 1964 Collapse and rebound of a spherical bubble in water. Phys.
Fluids 7,7-14.
HINSCH,K. & BRINKMEYER, E. 1976 Investigation of very short cavitation shock waves by
coherent optical methods. High Speed Photography, SPIE Vol. 97, pp. 166-171. Toronto.
JONES, I.R. & EDWARDS,D. H. 1960 An experimental study on the forces generated by the
collapse of transient cavities in water. J. Fluid Mech. 7,59M09.
KLINQ,C. L. & HAMMITT, F. G. 1972 A photographic study of spark-induced cavitation bubble
collapse. Trans. ASME D : J . Basic Engng 94, 825-833.
KUCERA, A. & BLAKE,J. R. 1988 Computational modelling of cavitation bubbles near boundaries.
In Computational Techniques and Applications CTAC-87 (ed. J. Koyce & C. Fletcher),
pp. 391400. North-Holland.
LAUTERBORN, W. 1974 Kavitation durch Laserlicht. Acustica 31, 51-78.
LAUTERBORN, W. 1982 Cavitation bubble dynamics - new tools for an intricate problem. Appl.
Sci. Res. 38, 165-178.
LAUTERBORN, W. & BOLLE,H. 1975 Experimental investigations of cavitation-bubble collapse in
the neighbourhood of a solid boundary. J. Fluid Mech. 7 2 , 391-399.
LAUTERBORN, W. & HENTSCHEL, W. 1985 Cavitation bubble dynamics studied by high speed
photography and holography : Part One. Ultrasonics 23, 26&268.
LESSER,M. B. & FIELD,J. E. 1983 The impact of compressible liquids. Ann. Rev. Fluid Mech. 15,
97-122.
LUSH,P. A. 1983 Impact of a liquid mass on a perfectly plastic solid. J . Fluid Mech. 135,373-387.
LUSH,P. A., WOOD,R. J. K. & CARPANINI, L. J. 1983 Pitting in soft aluminium produced by
spark-induced cavitation bubbles. In Proc. 6th Intl Conf.on Erosion by Liquid and Solid Impaet
(ed. J. E. Field & Corney), paper 5. Cavendish Laboratory, Cambridge.
MAXWORTHY,T. 1974 Turbulent vortex rings. J. Fluid Mech. 64, 227-239.
MAXWORTHY, T. 1977 Some experimental studies of vortex rings. J . Fluid Mech. 81, 465-495.
NAUDE,C. F. & ELLIS,A. T. 1961 On the mechanism of cavitation damage by nonhemispherical
cavities collapsing in contact with a solid boundary. Trans. ASME D: J . Basic Engng 83,
648-656.
NISHIYAMA, T. & AKAIZAWA, M. 1979 Pressure waves produced by the collapse of a spherical
bubble. Technol. Rep. Tohoku Univ. 44, 579-602.
PLESSET, M. S. 1974 Bubble dynamics and cavitation erosion. In Proc. Symp. on Finite-Amplitude
Wave Effects in Fluids, IPC Science and Technology, Guildford (ed. L. Bjmncl), pp. 203-209.
PLESSET, M. S. & CHAPMAN,R. B. 1971 Collapse of an initially spherical vapour cavity in the
neighbourhood of a solid boundary. J. Fluid Mech. 47, 283-290.
RADEK,U. 1972 Kavitationserzeugte Druckpulse und Materialzerstorung. Aczcstica 26, 27&-283.
RAYLEIGH, LORD 1917 On the pressure developed in a liquid during the collapse of a spherical
cavity. Phil. Mag. 34, 94-98.
REICHEL,E., SCHMIDT-KLOIBER, H., SCH~FFMANN, H., DOHR,G. & EHERER,A. 1987 Interaction
of short laser pulses with biological structures. Optics Laser Technol. 19, 4 0 4 4 .
SANADA, K.,IKEUCHI,J., TAKAYAMA, K. & ONODERA, 0. 1986 Interaction of an air bubble with
a shock wave generated by a micro-explosion in water. Proc. Intl Symp. on Cavitation. Sendai,
Japan, pp. 67-72.
SHUTLER, N. D. & MESLER, R. B. 1965 A photographic study of the dynamics and damage
capabilities of bubbles collapsing near solid boundaries. Trans. ASME D : J . Basic Engng. 87,
51 1-51 7 .
STEINER, R. (ed.) 1988. Laserlithotripsy. Proc. 1st Intl Symp. on Laser Lithotripsy Oct. 1987 in Ulm.
Springer.
STEPP,H., HOFBAUER, E., SEEBERQER, M. & U N S ~ L D E., 1985 Fast pressure measurements on
shock waves intraocularly generated by picosecond Nd :YAG-laser pulses. Laser in der Medizin
und Chirurgie 1, 151-154.
STRUBE, H. W. 1971 Numerische Untersuchungen zur Stabilitat nicht-spharischer schwingender
Blasen. Acustica 25, 28S303.
TESLENKO, V. S. 1980 Experimental investigation of bubble collapse at laser-induced breakdown
in liquids. In Cavitation and Inhomogeneities in Underwater Acoustics (ed. W. Lauterborn),
pp. 3@34. Springer.
TOMITA,Y. & SHIMA,A. 1986 Mechanism of impulsive pressure generation and damage pit
formation by bubble collapse. J . Fluid Mech. 169, 535-564.
VOQEL,A,, HENTSCHEL, W., HOLZFUSS, J. & LAUTERBORN, W. 1986 Cavitation bubble dynamics
and acoustic transient generation in ocular surgery with pulsed Er’d : YAG-lasers. Ophthalmology
93, 1259-1269.
338 A . Vogel, W. Lauterborn and R.Timm
VOGEL,A. & LAUTERBORN, W. 1985 Ein Raumfrequenzfilterungsverfahren zur Untersuchung der
Stromungsverhaltnisse in der Umgebung von Kavitationsblasen. Optik 69, 17G-179.
VOGEL,A. & LAUTERBORN, W. 1988a Time resolved particle image velocimetry applied to the
investigation of cavitation bubbles. Appl. Opt. 27, 1869-1876.
VOGEL,A. & LAUTERBORN, W. 1988 b Acoustic transient generation by laser-produced cavitation
bubbles near solid boundaries. J . Acoust. SOC.Am. 84, 71S731.
Vomov, 0. V. & Vomov, V. V. 1976 On,the process of collapse of a cavitation bubble near a wall
and the formation of a cumulative jet. Soviet Phys. Dokl. 21, 133-135.
VYAS, B. & PREECE, C . M . 1976 Stress produced in a solid by cavitation. J . Appl. Phys. 47,
5133-5138.
WELLINGER & GIMMEL1963 Werkstofftabelle der Metalle, p. 182. Stuttgart : Alfred Kroner.

Vogel JFM 1989

Uploaded by

Copyright:

Available Formats

Vogel JFM 1989

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vogel JFM 1989

Uploaded by

Copyright:

Available Formats

J. Fluid Me&. (198Y),1101.206, p p .

Optical and acoustic investigations of the dynamics

(Received 11 May 1988 and in revised form 18 November 1988)

rate of 1 million frames/s. The experimental arrangement is shown in figure 3. To

/ l . Order Trigger signal I

temporal resolution achieved is 10000 frames/s and the average density of

5. Experimental arrangement for the investigation of acoustic transient emission during

T = 1.000 E = 1.00 T = 1.885 E = 1.10 T = 2.119 E = 0.85

T = 1.047 E = 1.00 T = 1.856 E = 1.21 T = 2.164 E = 0.63

T = 1.030 E = 1.04 T = 1.800 E = 1.16 T = 2.190 E = 0.67

0 0.5 1.o 1.5 2.0 2.5 3.0

-15 -10 -5 20 f(W)

0 10 20 30 40 50 60 70 80 90 100 110 1201!130

3.4.2. Spherical-bubble collapse

3.4.3. Bubble collapse near a solid boundary

0 0.5 1.o 1.5 2.0 2.5 3.0 3.5

:,. ... . *. ....- .

0 0.5 1.0 1.5 2.0 2.5 3.0

0 0.5 1.0 1.5 2.0 2.5 3.0

3.5. Cavitation erosion

The work reported here was financially supported by the Fraunhofer-Gesellschaft.

You might also like