1JA

-*

montb:

NAVAL POSTGRADUATE SCHOOL

Monterey ,
California

~d&3r

THESIS

AN INVESTIGATION OF ACOUSTIC CAVITATION PRODUCED BY PULSED ULTRASOUND

by
Bruce, Jr. and Robert D. Middleton, Jr.
L.

Robert

December 1987
Thesis Advisor:

Anthony A. Atchley

Approved

for public release; distribution is unlimited.

T238708

SEC-''

~v

C-ASS

CA"

*H

-AGE

REPORT DOCUMENTATION PAGE

a

REPORT SECURITY CLASSIFICATION

5ECwf:' v C.ASS
:

1b

RESTRICTIVE

MARKINGS

2a

CA T ON AuTriQRiTY

DISTRIBUTION 'AVAILABILITY OF REPORT

2d

DEC.ASS

CA'ON DOWNGRADING SCHEDULE

Approved for public release; distribution is unlimited.

5

PERFORM. \G ORGAN. ZA", ON REPORT NUMBER(S)

MONITORING ORGANIZATION REPORT NUMBER(S)

6a

\AME OF PERFORM. \G ORGANIZATION

6b OFFCE
(If

SYMBOL

7a

NAME OF MONITORING ORGANIZATION

applicable)

Naval Postgraduate School

6c.

61

Naval Postgraduate School

7b

ADDRESS

Cty. State,

and ZIP Code)

ADDRESS

(Cty, State,

and ZIP Code)

Monterey, California
8a

93943-5000
8b OFFICE
(if

Monterey, California
SYMBOL
9

93943-5000

NAME 0= FUNDING SPONSORING ORGANIZATON

ADDRESS
(Cry, State,

PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER

applicable)

8c.

and ZIP Code)

10

SOURCE OF FUNDING NUMBERS

PROJECT

PROGRAM
ELEMENT NO

TASK

NO

NO

WORK UNIT ACCESSION NO

11

TITLE (Include Security Classification)

AN INVESTIGATION OF ACOUSTIC CAVITATION PRODUCED BY PULSED ULTRASOUND

12

PERSONAL AUTHOR(S)
,

BRUCE. Robert L.. and MIDDLETON

13a.

Robert P., Jr
14

TYPE OF REPOR T

3b

Master's Thesis
16

ME COVERED "ROM TO

DATE OF REPORT

(Year,

Month, Day)

15

PAGE COUNT

1987 December

95

SUPPLEMENTARY

\jO T ATlON

COSATi CODES

18

SU8JECT TERMS {Continue on reverse

if

necessary

and

identify by block

number)

iROUP

SUB-GROUP

Microcavitation, Cavitation Threshold, Threshold Detection, Transducer Calibration

and
identify by block

19

ABSTRACT (Continue on

reverse

if

necessary

number)

Transient cavitation produced by pulsed ultrasound in the MHz range First, a theoretical background which develops has been investigated. bubble motion in general and focuses specifically on transient caviSecond, the development of an apparatus which tation is presented. is designed specifically for studying this type of cavitation is Third, the use of the apparatus to make measurements discussed. of cavitation thresholds at various frequencies, pulse durations, Finally, results of the and pulse repetition rates is discussed. cavitation threshold measurements are interpreted using theories developed earlier.

20

DISTRIBUTION 'AVAILABILITY OF ABSTRACT

21

ABSTRACT SECURITY CLASSIFICATION

QfJNCLASSlFiED/UNUMITED
22a

SAME AS RPT

DTIC USERS

NAME OF RESPONSIBLE INDIVIDUAL

1473, 34MAR
83

UNCLASSIFIED
22b TELEPHONE (Include Area Code)
^2c OFFICE

SYMBOL

Pv-nf^^nr A.A. A1-rhlev. Code 61AV

DD FORM
APR
edition
All

408-646-2848
until

61Ay
SECURITY CLASSIFICATION OF THIS PAGE
t*

may be used

exhausted

other editions are obsolete

U.S.

Government

Printing Office

1986606-24.

Approved

for public release; distribution is unlimited.

An

Investigation of Acoustic Cavitation

Produced by Pulsed Ultrasound

by
Robert L. Bruce, Jr. Lieutenant Commander, United States Navy B. A., University of Texas, 1975

and
Robert D. Middleton, Jr. Lieutenant, United States Navy
B.S.,

Tulane University, 1980

Submitted in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

IN

ENGINEERING ACOUSTICS

from the

NAVAL POSTGRADUATE SCHOOL

December 1987

ABSTRACT
Transient cavitation produced by pulsed ultrasound in the
range has been investigated.
First,

MHz

a theoretical background which

develops bubble motion in general and focuses specifically on transient

cavitation
is

presented.

Second, the development of an apparatus

is

which

is

designed specifically for studying this type of cavitation

Third, the use of the apparatus to

discussed.

make measurements

of

cavitation thresholds at various frequencies, pulse durations,

repetition rates is discussed.
old

and pulse

Finally, results of the cavitation threshearlier.

measurements are interpreted using theories developed

TABLE OF CONTENTS
I.

INTRODUCTION

II.

THEORY

10 10
11

A
B.

BACKGROUND
CAVITATION REGIMES

BUBBLES UNDER QUASI-STATIC EQUILIBRIUM

CONDITIONS
11

D.

BUBBLES UNDER THE INFLUENCE OF A SOUND FIELD

STABLE BUBBLE GROWTH DUE TO AN ACOUSTIC

...

17

E.

PRESSURE FIELD
1.

21

TheAreaEffect

22 22

2. 3.

The Shell

Effect

Threshold Equation

23

F.

BUBBLES DRIVEN TO UNSTABLE GROWTH AND

VIOLENT COLLAPSE
24
28
31

G.

SONOLUMINESCENCE
ACOUSTIC SCATTERING FROM BUBBLES

H.

III.

EXPERIMENTAL APPARATUS

34
34

A
B.

BACKGROUND
CAVITATION ENCLOSURE AND LIGHT-TIGHT

ENCLOSURE

34

C.

FLUID

MANAGEMENT SYSTEM
....

37
39
42

D.

CAVITATION GENERATION AND DETECTION SYSTEM

E.

TRANSDUCER CALIBRATION APPARATUS

IV.

EXPERIMENTAL PROCEDURE

45 45

A
B
C

BACKGROUND
RECIPROCITY CALIBRATION

46

AXIAL

BEAM PATTERN MEASUREMENT OF SOURCE TRANSDUCERS

46

D.

TRANSVERSE BEAM PATTERN MEASUREMENT OF SOURCE TRANSDUCERS

1.

48 48 50

Theory
Procedure

2.

E.

DETERMINATION OF TRANSDUCER PEAK OUTPUT FREQUENCY

RADIATION FORCE-TARGET DEFLECTION

53

F.

MEASUREMENT
1.

56 56
59

Theory
Procedure

2.

G.

DETERMINATION OF TRANSIENT CAVITATION

THRESHOLDS
V.

62

RESULTS, DISCUSSION, RECOMMENDATIONS

64 64 74

A
a

RESULTS AND DISCUSSION

RECOMMENDATIONS

APPENDIX A
APPENDIX B

COMPUTER CONTROL PROGRAM

PVDF HYDROPHONE CONSTRUCTION AND
GENERAL INFORMATION

76

81

LIST OF REFERENCES
INITIAL

90
93

DISTRIBUTION LIST

ACKNOWLEDGMENTS
The following people helped us greatly by providing their
knowledge,
skills,

forebearance, and/or personal dedication:

Anthony

A. Atchley

George Jaksha
Steve Blankschein

Carol Crooker

Maria Zanger

Amy

and Aaron Bruce

in

Our classmates

UX

61

I.

INTRODUCTION
was
first

The term

"cavitation"

used

in the late nineteenth cen-

tury by Sir Charles Parsons to describe the

mechanism

that caused

reduced thrust in marine propellers.

fied cavitation

Lord Rayleigh, in 1917, identi-

as the cause of erosion in

pump components that were

to

sub-

ject to

reduced liquid pressures. Cavitation has come

be defined as

the rupture of a liquid or a liquid /solid interface due to the reduction

of liquid pressure [Ref.
1].

Since the theoretical tensile strength of

pure liquids
(=

is

exceptionally high, extremely large negative pressures

to

-1000 bars) should be required

cause the liquid bonds

to break.

Since rupture in real liquids can occur at relatively low negative

pressures

(= -1 bar), it

has been postulated that inhomogeneities exist

Further-

within the fluid which cause this reduced tensile strength.

more, experimental results point to the necessity of these microscopic inhomogeneities containing gas [Ref.
2].

Thus, cavitation

is

actually the formation of macroscopic bubbles from microscopic bubbles (cavitation nuclei).

Acoustic cavitation can therefore be defined as

is

the activity of bubbles within a liquid that

subject to an acoustic

pressure

field.

Microcavitation, or the formation of micron-sized bubbles from

cavitation nuclei

due

to the influence of ultrasonic acoustic

energy

(ultrasound), has benefits

when used

in

such applications as ultrasonic

cleaning and etching, acceleration of certain chemical reactions, and,

medically, to destroy

unwanted

cell

formations.

Microcavitation may,

however, produce adverse biological effects as well.

In particular, the

improper use of diagnostic or therapeutic ultrasound could possibly

result in the unintentional destruction of or

damage

to

healthy

cells.

Investigation of microcavitation

phenomena and

their possibly

adverse biological effects has produced a significant amount of infor-

mation on cavitation thresholds.

ultrasound cavitation effects
biological
effects
[Ref.

Robert Apfel's studies of pulsed

3]

have summarized the possible

to

in single

cells

due

excessive sound pressure

liq-

amplitudes.

His work has accounted for the effects of frequency,

uid inertia, and fluid viscosity in the cavitation process.

H. G. Flynn

has done extensive numerical modelling of cavitation phenomena, also

taking into account viscosity and surface tension as well as compressibility

and heat conduction

al.

[Ref. 4].

R. A. Roy, et at [Ref.

5],

and
for

A. A.

Atchley. et

[Ref. 6],

have produced precision techniques

mea-

suring transient acoustic cavitation thresholds by detecting the light

and sound emissions produced during violent bubble

their research that
this thesis.

collapse.

It is

has become the basis

for the

research reported in

Although some aspects of cavitation thresholds have been

little

previously investigated, there has been

work done

to

determine
of this

the time evolution of the cavitation event.

The objectives

research are to

(1)

develop an apparatus in which cavitation can be

induced and detected simultaneously by two methods acoustic

scattering

and sonoluminescence;
(3)

(2)

determine transient cavitation

thresholds; and
cavitation events.

examine the time evolution of the transient

II.

THEORY

A.

BACKGROUND
The behavior
of

bubbles under the influence of an acoustic presIn this study,

sure

field is

indeed complex.

we

will

examine the pres-

sure amplitudes of pulsed ultrasound which will induce the rapid

growth and collapse of cavitation nuclei.

the acoustic signal, the
rest time

We

will

vary the frequency of

number

of acoustic cycles per pulse,

and the
effects

between pulses

in order to

more

fully

understand the
to

of each

parameter on the pressure amplitude required

In addition,

cause the

nuclei to behave in this manner.

we

will

develop an appa-

ratus that provides for both acoustic and optical observation of the
cavitation events.
of the bubbles

We must

therefore be aware not only of the behavior

field

when

influenced by the sound

but also of their

light.

propensity for scattering sound and producing visible

In this chapter,

we

will

review the above-mentioned aspects of

the three regimes of acoustic

bubble behavior.
cavitation:
tion.

We

will first look at

stable cavitation, transient cavitation,

and gas body

activa-

We

will describe

the theory that describes bubble activity under

static equilibrium conditions,

under the influence of a continuous

sound pressure

field in

which controlled bubble oscillations occur,

result in controlled growth of the bubbles,
to

under conditions which

and

under conditions which lead

unstable growth and violent collapse.

theory will be described that allows one to predict the peak acoustic
Finally,

pressure which produces the collapse of a gas bubble.

we

will

10

examine the theories which explain both

cence)

light

emission (sonolumines-

upon bubble

collapse and acoustic scattering from multiple

bubbles.

B.

CAVITATION REGIMES
It is

generally acknowledged that two, and possibly three, cavita-

tion regimes exist.

The focus

of our research will

be transient cavitathis

tion

produced by pulsed ultrasound.

is

As the name indicates,

regime

transient in nature; bubbles are likely to exist for less than a

few acoustic cycles.

The

results of the

measurements

to

be described

later indicate that effects of the

second regime, stable cavitation,

may

also be important.

Stable cavitation occurs

its

when a bubble

subjected to
is

a sound field oscillates about

equilibrium radius.

This regime

characterized by bubbles that are relatively permanent; noticeable

changes in the equilibrium radius occur only

third regime (considered
is

after

many

cycles.

The

by many not

to

be a true cavitation regime)

known

as gas body activation, which involves the oscillations of gas

pockets permanently embedded in cracks, crevices, or within biological systems.

As gas body activation does not include actual formation

will

and growth of gas bubbles, only stable and transient cavitation

discused further in this chapter.

be

C.

BUBBLES UNDER QUASI-STATIC EQUILIBRIUM CONDITIONS

The
first

stage of any cavitation problem

is

nucleation, or the for-

mation of a macroscopically observable bubble from a cavitation

nucleus.

This formation usually involves the loss of mechanical

11

stability of the nucleus.

Following the developments of Walton and

[Ref. 8],

Reynolds
brief
ical

[Ref. 7],

Noltingk and Neppiras

will

and Blake

[Ref. 9],

examination

be made of the conditions required

a
liquid.

for

a spher-

bubble

to lose its static equilibrium in

Given a bubble of radius Ro in

infinite extent at

static equilibrium

with a liquid of

a hydrostatic pressure Po, the pressure inside the

bubble

is

given by

P = P0 +

(jg)
is

(2.D

The term ht

known

as the Laplace or

Young pressure and

is

due

to

the surface tension of the liquid, a (the usually minor contribution of

the vapor pressure has been neglected).

Assuming that the

to

liquid

undergoes an incremental change in hydrostatic pressure

value
Poo,

new

the bubble radius will undergo an incremental change to

R.

some new radius

Assuming isothermal conditions and an

ideal gas,

we note

p 1 Vi = P2V2 = constant

where

2cA4 ^3 P1V1- (^ Po+pTTrcRo Ro 3

and

12

P 2 V 2 = |Poo+-p- 3-7TR3.
1

The assumption
of evaporation

of isothermal conditions is justified

because the rate

and condensation

of vapor at the bubble surface is suffi-

cient to maintain this condition.

After combining these two expres-

sions and cancelling

common

terms,

we

find that

Figure 2.1, which

is

a plot of

Poo

as a function of R, offers insight

is

into the stability of bubbles.

Assuming there

no gas diffusion

into or

out of a bubble,

Poo is

the value of the liquid pressure necessary to

balance the

sum

of the internal pressure (which tends to

make

the

bubble expand) and the Laplace pressure (which makes the bubble
collapse).
Poo is

minimum

for

a bubble of radius R*. Bubbles of radius

less

than R* are in stable equilibrium, whereas bubbles of radius

greater than R* are in unstable equilibrium.

To see

this

more
P<

clearly,

consider a bubble of radius R<R*.

If

is

increased,

decreases

because the decrease in the Laplace pressure

decrease in internal pressure.

(2&,

-p-| is less

than the

Consequently, the external liquid

is

pressure needed to maintain equilibrium

external pressure.

now

less

than the actual

to R.

As a

result, the

bubble shrinks back

In other

words,

the

collapsing

pressure

dominates

over

the

13

Figure 2.1

Graph of Pressure vs. Bubble Radius Showing Stable and Unstable Equilibrium States of Bubbles
Within a Liquid (from
expanding pressure.
If

[Ref. 7]

the radius had been

made

smaller, P,

would

increase corresponding to a dominance of the expanding pressure.

Again, the bubble returns to R.
Therefore, the bubble
is

in a state of

stable equilibrium perturbations in the radius cause the bubble to

return to

its initial

radius.

By invoking a
it is

similar

argument

for
is

bubbles

of radius greater

than R*,

seen that the equilibrium

unstable.

An
the

increase in radius causes

Poo to

increase due to the dominance of

in

expanding pressure.

The bubble would increase

size

14

indefinitely.

decrease in radius causes the bubble to shrink

to the

corresponding stable value of R.

stability of bubbles,

Because of

its role

in determining the

R*

is

called the critical radius. in equilibrium at

If

Now, consider a bubble of radius Ro < R*

pressure Po-

According

to

Figure 2.1,

it is

stable.

the pressure
it

is

lowered to some value Po - Pa. the bubble

will

grow

until
If,

reaches a

new

stable equilibrium radius R' (since R' < R*).

is

however, the

pressure

reduced

to

Po - Pb. where Pb

is

the Blake threshold

pressure, the bubble reaches the critical radius R*.

Any

further

reduction in pressure provides no equilibrium value for the bubble

radius,

and the bubble grows uncontrollably. To

Poo

solve for R*,

we must

minimize

as a function of radius.

Rewriting equation

(2.2) slightly,

Poo

P +

2o^

Ro~ POIR*

2o R'

(2.3)

then differentiating with respect

zero,

to

R and

setting the result equal to

we

find

SIPo+IQRo'
R* =

1/2

~2a

(2.4)
!

Previously,
initial

we considered

the Blake Pressure to be a function of

is

bubble radius.

The converse view

to

consider the

initial

radius to be a function of Blake pressure.

Accordingly, the Blake

15

radius

Rb

is

the

initial

radius of a bubble that,

to a

when

subjected to the

(acoustic)

Blake pressure Pb, grows

radius of R*. becomes

unstable, and grows uncontrollably.

pressure,

To find the Blake threshold

we

first

rewrite R* (equation 2.4) in terms of Rb:

3|Po +

H)4
.

1/2

R* =

~2a

(2.5)

Letting

P +

RB"fB

(2.6)

allows R* to be expressed more simply as

R* =

'3KV/2

2c

(2.7)

or

2o
(2.8)

Setting

Ro = Rb,

= R*. and P^ = P

Pb

in equation (2.3),

we

find

Poo

PO

PB 3R*

2q

2c R*

(2.9)

or

Po - Pb = - i_2_ 3 R*

(2.10)

16

Rewriting this expression using equations

(2.7)

and

(2.6),

we

find

Po - Pb = - q a

nfn 2(P

3a 2a
+

1/2

r^Rb

3,
)

(2.11)

which can be written

in

normalized form as

Pb =

Pb
1
i-

pjJ-=

4 B qXB [4(1 -X B 9

3X

1/2

(2.12)
)J

where

XB

^.
this discussion, a
its stability if

Summarizing
librium will lose

bubble of radius Rb in stable equi-

subjected to an acoustic pressure of

amplitude Pbtimes
its

Upon
size;

loss of stability, the

i.e.,

bubble

will

expand

to

many

initial

it

will

form a macroscopically observable

bubble.

D.

BUBBLES UNDER THE INFLUENCE OF A SOUND FIELD

We
a sound

begin the examination of bubble motion under the influence of

field

with a description of bubble motion by Rayleigh

[Ref.

10].

He

initially

considered the collapse of an empty spherical cavity

infinite

located in

an

mass
is

of incompressible fluid.

The hydrostatic
the

pressure in the liquid

assumed constant. Rayleigh observed that

starting from rest

work done by the collapsing bubble

to the kinetic

had

to

be equal

energy of the

fluid.

Thus, the following holds true:

P^

(R5 - R3) =

J V 2(r) 4*r2d]

(2.13)

17

where

r is

the distance from the center of the bubble to a point in the

liquid that is greater

velocity of the fluid.
f

than the radius of the bubble

R and

v(r) is

the
471

Poo is

the hydrostatic pressure of the liquid and -&-

Rq

R3

is

the decrease in volume of the spherical cavity.

is

The

rightis

hand term

the kinetic energy of the liquid,

^mv2

Since the fluid

is

incompressible, the change in volume of the liquid

equal to the
is true:

change

in

volume

of the

bubble and the following relationship

4rcr 2 v(r)

= constant.
= 4rcR2 v(R) or v 2

Therefore, 47tr 2
it

v(r)

(r)

R4 2 prv (R).

Letting v(R) = R,

can be shown that

v2 =

R4 R2

^p-.

(2.14)

Substituting this result into equation (2.13) and performing the integration on the right-hand side yields
4tt
*}

Poo

-j1 (Rq -

R3

2tc

R3 R2

(2.15)

Solving for

R2

r2 K

- n " 3p

(2.16)

18

Since

is

equal to -xr
is

dR

it is

apparent that dt =
infinity.

dR R

As

decreases,

however,

compelled to approach

At this point, Rayleigh

initial

strives to avoid this obvious flaw

by admitting that the

assumpthat the

tion of zero internal pressure is invalid.

He then assumes

cavity
sion.

is filled

with a perfect gas undergoing an isothermal compresis true,

If

this

assumption

Boyle's law holds

and the pressure

If

within the bubble will increase until the bubble begins to rebound.
fluid viscosity is zero

and

all

other losses ignored, the bubble will con-

tinue to oscillate between this

minimum

radius and the

initial

initial

radius

Ro-

Again invoking the assumption that

is

pressure inside the

fol-

bubble

zero or at least

some

constant. Rayleigh has derived the

p, at

lowing equation governing pressure within the liquid,

any

point:

p
Poo "

R f*S 1= 3? [w-\
.

R4 fsS 3r* [w-

I 1

(2.17)
)

The work

of Plesset, Poritsky,

and Noltingk and Neppiras con-

tributed to the final formulation of an equation that describes the

motion of a bubble under the influence of an acoustic

[Ref.

field.

Plesset
12]

11]

extended Rayleigh's analysis, as did Poritsky

[Ref.

[Ref.

and
comfinal

Noltingk and Neppiras

13].

In 1976, Lauterborn [Ref. 14]

bined these results with previously described results to form the

characteristic equation governing bubble motion, the

RPNNP

equation.

19

Lauterborn so named this equation in honor of the pioneering

work done by Rayleigh,

review, he set the

Plesset, Noltingk, Neppiras,

and

Poritsky.

In

work done by the

liquid at infinity

minus the work

bubble equal
to

done by the

liquid layer immediately adjacent to the

the kinetic energy of the liquid surrounding the bubble:

(P L - Poo) 47tr2

dr = 2rcpR 3 R2

(2.18)

where Pl

is

the pressure in the liquid layer immediately surrounding

the bubble.

The reader should note that

Poo in this

development

is

varying with time:

Poo

= P

P(t)

where Po

is

a static pressure about which

P(t)

varies.

Therefore, from

Reference 14, Pl can be expressed as

^ = ^0 + f-Pv)^K -2|-4^|
where k
is

,2.19,

the polytropic exponent of the gas inside the bubble,

|i

is

the shear viscosity of the liquid,

Differentiating (2.18)

and Pv

is

its

vapor pressure.
P(t),

and substituting

(2.19) for

PL and P TO = Po +

we

find that

RR

|R

2=I

P + Rn--Pv iV RO

R-f URj -R-- R

"

P 0-

P(t)

(2.20)

20

The following assumptions are

equation
1.

implicit in this

comprehensive

[Ref. 7]:

The bubble remains spherical

Uniform conditions

at all times;

2. 3. 4.
5.

exist within the bubble;

X, is

The acoustic wavelength,

much

greater than R;

No body

forces are present;

is

Bulk viscosity

ignored;

6.
7.

The

liquid is incompressible;
of the bubble is

Gas content

constant no diffusion occurs;

8.

Vapor pressure,

Pv,

is

constant during the bubble's motion.

in deriv-

Although many simplifying assumptions have been made

ing equation (2.20),
it

is

a better-than -reasonable description of the

to

dynamics

of a

bubble subjected

an ultrasonic sound

field [Ref. 7].

is

In

addition, the

RPNNP,

or simply the Rayleigh-Plesset, equation

the

foundation of the theories discussed in the next two sections.

E.

STABLE BUBBLE GROWTH DUE TO AN ACOUSTIC PRESSURE FIELD

Stable cavitation and rectified diffusion, although neglected in the

RPNNP

equation, are two related and interesting processes which

may
The

be important in the time evolution of transient cavitation events.

stable growth of cavitation nuclei to sizes
is

which allow transient growth

Rather than

a likely

mechanism

for

producing transient cavitation.

undertake a rigorous mathematical development of

rectified diffusion

21

as in Reference 15, a qualitative description of this

phenomenon

will

be presented here.

The importance
ment, was
first

of rectified diffusion,
al.,

or incremental enlarge[Ref. 16].

noted by Harvey, et

in

1944

Blake

[Ref. 9]

was the
1949.

first to

attempt the derivation of a mathematical model in

the area effect and the shell
effect,

Two

effects,

appear

to

work

in concert to
1.

produce bubble growth.

Effect

The Area

When

a bubble contracts, the concentration of gas within the

Since gas diffuses from regions of high concentra-

bubble increases.
tion to those of

low concentration, the gas diffuses out of the bubble.

When

the bubble expands, the concentration of gas decreases, and the

concentration gradient reverses, the result being a diffusion of gas into

the bubble.

The

diffusion rate

is

proportional to the area through

it is

which gas may

flow.

Therefore,
it is

more gas enters the bubble when

contracted.

expanded than leaves when

of gas during
2.

The net

result is

an influx

one complete oscillation of the bubble.

The Shell Effect

The
rate of diffusion, as noted above, is proportional to the

concentration gradient.
liquid

Crum

[Ref.

15] considers

a spherical shell of

surrounding a bubble. He argues

that,

when

the bubble contracts,

the shell expands, with a consequent reduction in gas concentration near

the bubble surface.

This increases the concentration gradient in the

of gas out

outward direction over that discussed above, so the diffusion

of the bubble is even greater.

Conversely,

when

the bubble expands.

22

the shell volume decreases, the gas concentration increases,

and the
is

concentration gradient into the bubble increases.

the

The

overall result

enhancement
3.

of the area effect.

Threshold Equation
There exist both high- and low-frequency thresholds
for

rectified diffusion.

Crum has

formulated expressions for both.

Of

interest here is the high-frequency approximation [Ref. 15]:

2

~2 2 ^ pRoC0
hi

12 \2 + b2
CO
co

2a
RoPoc
>

Q
Co.

pi

(3+4K)

2a RoPo

(2.21)

where
Pa
= pressure threshold for rectified diffusion
17.

3 -4 VT
2
J

so
3
Poo

RoJ
^

( 1

4a

[3 Poo ROJ.
2

= p

Q)2

Rq/3

Po

b
2
co

4co|j./3 Poo

(Damping term)
+

\(
)

4 a\

Poo

n~

(small amplitude resonance frequency

pR^

squared)

Co

= equilibrium concentration of gas in the liquid

23

Ci

= concentration of dissolved gas in the liquid far from the

bubble
= viscosity of the liquid'^

\i

co

= angular frequency of the acoustic field

Ro

= equilibrium bubble radius

= liquid density

Poo

= ambient pressure

= surface tension of the liquid

F.

BUBBLES DRIVEN TO UNSTABLE GROWTH AND VIOLENT COLLAPSE

The
final step in

development of the theory which defines bubble

taken from research by R. A. Apfel and H. G.

behavior in a sound
Flynn.

field is

In Reference 3, Apfel indicates that the possibly adverse bio-

logical effects of acoustics are linked to the

growth and collapse of

submicron cavitation
acoustic cycle
logical cell
[Ref. 3].
will,

nuclei.

nucleus "that grows to 5 |im in one

upon

collapse, deposit in the size scale of

one bio-

approximately 300 million electron volts (MeV)" of energy

This energy, in the form of heat, results in internal gas tem-

peratures in the range of 1000-5000 K.

effects

At such high temperatures,

such as
effects

free radical formation in the

and recombination can occur.

These

can result

production of visible light that can be

detected optically (this will be discussed in detail later in this

24

chapter).

Apfel's

work culminates

in a formula that predicts the

peak

acoustic pressure which produces a cavitation event (collapse) given

the initial bubble radius Ro

and maximum internal gas temperature

density, viscosity,

5000 K with acoustic frequency,

of the liquid as parameters.

and surface tension

He develops the

relationship between the

maximum
=

collapse temperature

and normalized acoustic pressure p

is

Pa p

(Pa

is

peak acoustic pressure; Po

ambient pressure) by relating

p to Rmax (maximum bubble radius) and

Rm ax
will

to

T. Although a

rigor-

ous development of these two relations

Figure 2.2 shows readily
acoustic pressure.

not be conducted here,

how bubble
is

size

corresponds

to

applied

The reader

referred to Reference 2 for a

vs. p.
It is

more

rigorous development of

vs.

Rm ax and Rmax

necessary,

however, to examine Figure 2.2 carefully to determine the character of

the

Rmax

vs.

p relationship.

The maximum bubble radius can be

Ri, which occurs
is

thought of as being composed of two parts:

when

AP, the pressure difference across the bubble,

force);

negative (outward
yet the bubble

and Rmax. which occurs when AP

is positive,

continues to expand due to outward liquid

momentum.
is

The time over which maximum bubble radius

noted in Figure 2.2.
algebraic

computed
negative,

is
is

also

The time
and
At.

x.

during which AP

is

the

sum

of t2,

ti,

t2

-ti -At,

(2.22)

25

MINIMUM BUBBLE

SIZE;

MAXIMUM
TIME
Figure 2.2

TEMPERATURE

Schematic of Bubble Radius as a Function of

Time When Subjected

where
.

to a Single Acoustic Cycle (from

[Ref. 3])

7c

j 2tp-ir
Zip - Pb)

(2.23)

'*

iW

(2.24)

26

and

^f^V^+fp'
Pa where p = -5 and pb
by equation
(2.12).
is

(2.25)

the normalized Blake threshold pressure given

is

The value oft

2(p-p b
P

therefore

-27cfLV

J_ J 2(p-1)
p

"

|Ro

6u_

Vr

(p-i:

Po (P-D
(2.26)

The maximum bubble radius can now be estimated by

2
/Polp "
1)

1/3

Rmax
From

= Ro +

fV

l+|(p-

1)

(2.27)

the perfect gas law,

it

PR3 -~

= constant, and

its

adiabatic form

PR3 ^

= constant,

can be shown that

fR-max
1)

= To

(Y

(2.28)

From equations

(2.26)

and

(2.27),

above,

we

find

To

(y

1)

i^W

Pq

(p -

l+o(p-

1)

(2.29)

It is

now

convenient to find frequency (embedded in

x),

resulting

in:

27

/Pox

f=

3tcRo

Z 3.46
o
.

LV
:

./(P-1) + Jlp-Pb)
p
,

J
,

,,

(2.30)

^+1
.3

1/3

+iN?-pox- a4C

where

% = pb +

p-2

+ V(p -

1)

(p

-Pb)-

In Table 2.1 from Reference 3

and Figure 2.3 from Referand frequency

in transient
fre-

ence

4,

the effects of initial bubble radius

It

cavitation threshold can be seen.

is

important to note that

little

quency and internal temperature have

effect

on the threshold

pressures of smaller bubbles, but become increasingly important as

initial

bubble radius gets

to

larger.

This approximate formulation

is

not

meant

be a substitute

for

more exact numerical calculations such as

it

are found in Reference 4, but

does give a reasonable idea of the

of frequency, viscosity,

manner

in

which the parameters

and

initial

bubble radius
G.

affect the transient cavitation threshold of

liquid.

SONOLUMINESCENCE
Since one of the goals of this thesis
is to
it

develop both optical and

is

acoustical cavitation detection schemes,

important

to

present

some

theories concerning sonoluminescence.

[Ref. 71,

Following the review of

to the origin of

Walton and Reynolds

two primary theories as

sonoluminescence have been advanced.

28

TABLE

2.1

PRESSURE THRESHOLDS PREDICTED BY EQUATION

(2.30)

T K.

f(MHz)
1.0

Pressure (bar) Pressure (bar) R = 0.1 nm R = 0.5 um

2500 5000

6.4 6.4

2.0
2.1

2500 5000

2.5

6.4 6.4 6.4 6.4

2.8 3.1'
4.2
5.1

2500 5000

5.0

Puise

wiatr*

l0" 6 i

Biote mreshoio

-5
10

IQ-'

Imtiol

Radius, R n 'cm)

Figure 2.3

Threshold Pressure Pt (In Bars) for Transient Cavitation as a Function of Initial Radius Ro Over a Range of Frequency fX at a Pulse Width of ljis [Ref. 4]

29

Noltingk and Neppiras

is

[Ref.

13]

proposed that sonoluminescence

the result of black-body radiation from the gas contained within the

bubble, which theoretically rises to extreme temperatures as the bubble undergoes a rapid adiabatic compression during its collapse.

Griffing [Ref.

17]

proposed that the formation of highly reactive

during bubble collapse were responsible
for

free radials in the liquid

rapid chemical changes in the fluid, resulting in the emission of

light.

For the most recent examinations of sonoluminescence, the

interested reader
is

referred to References 18 through 22.

to the

The

pre-

ponderance of evidence points

subsequently play a
role in

formation of free radicals that

sonoluminescence.

These

free radicals, or

"reactive molecules (or atoms) with unpaired electrons" [Ref. 23]

may

be formed as a result of the dissociation of molecules subjected

high temperatures produced by a rapid adiabatic compression.
if

to the

Often,

a free radical (with

its full

its

unpaired electron) reacts with a molecule

of paired electrons, the result is another
all

having

complement

free radical.

This chain of events will be self-sustaining until

of the

suitable material has been reacted with, the free radicals have com-

bined with each other to form stable molecules, or

radicals have collided with the wall of the container,

all

of the free

where they are

removed from the process.

light

According to Griffing, the emission of

free radical reactions that

may be due

to

one of these

happens

to

be of a chemiluminescent nature. This hypothesis has been confirmed

by a number of subsequent experiments

[Refs. 18-22].

30

H.

ACOUSTIC SCATTERING FROM BUBBLES

The other detection method that we hope
to

use

to investigate

transient cavitation

is

the scattering of an acoustic signal (either the

source signal or an alternate lower-frequency active signal) from

bubbles generated by such events. This technique

is

expected to take

advantage of the exceptionally large scattering cross-section of

resonant bubbles.
Since our measurements do not require numerical

solution of scattering parameters, however, a qualitative explanation

rather than a rigorous mathematical development will be given here.

This theory

is

taken from Clay and Medwin

[Ref. 24].

small, rigid sphere

whose density and

elasticity are greater

than

that of the

medium

backscatters a very small

amount

of the acoustic

energy
is

when

ensonified by acoustic frequencies where the wavelength

much

greater than the sphere's radius.

is,

The acoustical cross-

section of the sphere

in fact,

much

smaller than the geometrical

field

cross-section because of diffraction of the incident sound

the small, rigid body.

around

Bubbles, however,
specific acoustic

differ

from
(pc)

rigid

spheres in two ways:

(1)

the

impedance

of the gas within the bubble is

(2)

much

less

than that of the water, and

bubbles resonate.

Bubbles are

extremely effective absorbers and scatterers of sound at frequencies

near resonance frequencies.

are on the order of

Scattering cross-sections at resonance

their geometrical cross-sections.

400 times

(See

Figure 2.4.)

31

1000

4C0

100

ka-

Figure 2.4

Ratio of Scattering Cross-Section to Geometrical Cross-Section for a Bubble and a Rigid Sphere (from [Ref.
Transient cavitation events studied in this investigation
in the generation of

24])

will result

many bubbles

of varying sizes.

The

effects of

multiple scattering

must be considered.

When

scatterers are widely

is

spaced, the scattering cross-section per unit volume

simply the

sum

of the individual scattering cross-sections of the

bubbles within

32

that volume.

The bubble "cloud" expected as a

result of the transient

cavitation event will be

will

much more

densely packed, however, which

cause the individual scattered

fields of the oscillating

bubbles

to

interact.

The

result

is

that the resonance curve of a single bubble

shown

in Figure 2.4 is
is

broadened and the scattering cross-section of

less

an array of bubbles
sections.

than the

sum

of the

individual cross-

cloud of bubbles bubbles packed randomly in a three-

dimensional array acts as a pressure release surface at the face of the

cluster,

thus reflecting acoustic energy at this interface.

It

is

this

effect that is

expected to reveal the presence of acoustic cavitation.

33

III.

EXPERIMENTAL APPARATUS

A.

BACKGROUND
The apparatus with which these
cavitation

measurements
L.

will

be

made
and

is

based upon a system developed by A. A. Atchley,

A. Frizzell,

R. E. Apfel [Ref. 6].

(1)

Our apparatus

differs

from theirs in three

major areas:
methods:

cavitation detection can be accomplished

by two

optically (via sonoluminescence)

(2)

and acoustically, either

will rely

individually or simultaneously;

this

apparatus

on random
artificially
(3)

bubble nuclei attaining threshold radii vice providing an

introduced uniform cavitation nucleus to the sound
field;

and

this

apparatus

is

computer-controlled.
the fluid

The system

is

made up

of two

major subsystems:
system.

management system and

the electronics

The

fluid

management system

will distill, filter,

and degas
electronics

ordinary tap water for use in the cavitation enclosure.

The

system

will provide the capability to induce, detect,

and record tran-

sient cavitation events, all

under computer

control.

B.

CAVITATION ENCLOSURE AND LIGHT-TIGHT ENCLOSURE

The
cavitation enclosure (Figure 3.1)
of 1.2
is

a 25

cm by

15

cm by

15
the

cm box made

cm

thick plexiglass.
it is

Plexiglass

was chosen as

primary structural material because

installed transducers

sufficiently strong to

support
it

and

fluid inlet

and outlet connections, yet

allows the optical detection apparatus to "see"

insonified region of the system.

what occurs

in the

One end

of the

chamber holds a

34

Figure

3.

Cavitation Enclosure

Figure 3.2

Light-Tight Enclosure and Photomultiplier Tube Assembly

35

Harisonic Labs focused transducer that supplies the cavitation generation signal.

A wedge

and two

vertical walls (also plexiglass)

have been

installed at the opposite

end of the enclosure

reflections that

to scatter the incident

sound and thus minimize

region.

would

interfere at the focal

The source transducers

cm.

(rated at 1.0

MHz and

2.25 MHz) have

focal lengths of 7.6

Two

additional walls with holes cut to allow

passage of the focused sound beam are installed on either side of the
focal region,

again to minimize interference from reflections.

source transducer allows for rapid changing of

The
ele-

mounting

for the

ments so that the other source frequency can be examined.

Two

Panametrics V301 transducers are permanently installed in the bot-

tom and one

side of the enclosure in

such a manner that

their

beam

pattern intersects the focal region of the source transducer.

The top

of the enclosure is removable to allow access to the interior. cavitation enclosure

for fluid inlet
is

The

placed in a light-tight box that has connections

and

outlet as well as

BNC

connections for source and

receive transducer signals (Figure 3.2).

The photomultiplier tube

(PMT) assembly
that
it

is

mounted on one

side of the light-tight enclosure so

also

can "see" the

focal region for

sonoluminescent indications
sensitive circuitry
is

of cavitation.

The photomultiplier tube's

pro-

tected

by a mechanical shutter that can be lowered

to block light

and

by an interlock switch

in the lid of the enclosure that interrupts the

high -voltage power supply to the

PMT when

the

lid is

removed.

36

C.

FLUID MANAGEMENT SYSTEM

The
fluid

management system

(Figure 3.3) consists of a four-elefilter,

ment

deionization/filtration system, a bag-type 0.2 |jm

a small

variable-speed electric centrifugal pump, an ultraviolet

sterilizer,

and

an Erlenmeyer
infusion.

flask with connections for flask is also

system evacuation and argon

The Erlenmeyer

used as a

fluid reservoir

and

expansion tank. The filtration/deionization system produces the equivalent of quintuply distilled water
silicon,

and removes

all

ionized minerals,

and

free

carbon dioxide down

to four parts

per

billion.

The

bag

filter

ensures that no particles larger than 0.2 jim are introduced

into the system.

is

These two elements are used only when the system

being

filled.

When

the system

is fully

charged, the water

is

circu-

lated through the cavitation enclosure via a closed loop consisting of

the
voir.

pump, the U/V

Argon
is

sterilizer,

the enclosure, and the expansion reserfluid to help

bubbled into the

enhance the sonolumines-

cent reaction.

The system

is

simultaneously evacuated (15 inches Hg

is

vacuum)
ued

in order to

remove dissolved gases. The evacuation

contin-

until visible air

bubbles are removed (usually about one hour). The

circulation system is then secured

and the sample

for

is

returned to

atmospheric pressure, where

it

remains quiescent
After about

one-half hour

before data accumulation begins.

lation (usually

an hour of data accumurefilled,

20

to

40 data

points), the

system must be drained,

is

and reprocessed.
becomes

Refilling the

system

necessary because the

fluid

"tired" after multiple cavitation events,

to fewer

causing higher-than-

normal thresholds (possibly due

remaining cavitation nuclei).

37

Q.

CO

L 3

H-

M O c
UJ

CO

<I

c o

^<ff

> id >
>

u u

id

s
(A

CO CO CO
<V

ed

id

u,

>
\
ID

a
B 3
0_

C
id

C o
VI

c
a
a.

a
f=

x
UJ

^3
a.

a.

3 3
o

>

id

CZD

a.

m
c o L
a:

CO

38

D.

CAVITATION GENERATION AND DETECTION SYSTEM

The electronics system consists
of the cavitation generation sys3.4).

tem and the cavitation detection system (Figure

cavitation generation system
is

The heart

of the

the Hewlett-Packard

9000

Series

300

personal computer, which controls the output frequency, pulse width

(duty cycle), pulse repetition rate, and acoustic pressure amplitude of
the source transducer via the Hewlett-Packard
ator.

3314A function

gener-

The output

of the function generator is boosted

is

by an Amplifier

Research 100-A15 Power Amplifier and

transducer.

then applied to the source

is

printout of the control program

contained in

Appendix A.

The

cavitation detection system consists of two subsystems:

the

acoustic detection system and the photomultiplier tube (PMT) assembly.

The acoustic detection system

is

made up

of the

two Panametrics

V301 transducers

installed in the cavitation enclosure

and the associa

ated elements for either passive or active

detection of cavitation events.

(via scattering of

cw

signal)

In the passive

mode, the output

of

one
is

of the detection transducers is boosted

by a small RF amplifier and

displayed on the Tektronix

signal

2245A

oscilloscope.
of a

The trace

of this

on the oscilloscope consists primarily

bundle of acoustic

pulses that have either been diffracted or scattered to the detection

transducer face.

The bundle

is

delayed from the source pulse by

approximately 100 microseconds, which corresponds to the time the

pulse would take to
focal

make

the 15

cm

journey from the source

to the

region

and then

to

the

detection

transducer face.

The

39

<n

a o

M u Q
CO
<u

u 3

a
Q.

3
U)

a
a o

4)

L O 3 o

>

o
o
+*

* m o n c
-*

a.

o
*-

3
e

a
o
(J

4*

o 3
u.

a O
CT7

a
o
4* J*

0.

40

system operator must monitor the signal in the

vicinity of this time

delay for a transient that would indicate an increase in the scattered

source signal, thus signifying the occurrence of a cavitation event.

operator
for the

The

must then manually

to calculate

interrupt the computer

ramp

function

computer

and display the threshold pressure.

is

In the active
at a

mode, the second Panametrics transducer

driven
result-

continuous frequency in the 100 kHz-500 kHz range.

is

The

ing active signal

received

by the detection transducer along with the

of the detection transducer is

source signal.

The output
filter to

passed

through a low pass

remove the cavitation generation signal

while passing the active detection signal.

sharp increase in the

amplitude of the received signal indicates scattering from the bubble

cloud that
is

formed by a cavitation event.

The photomultiplier

assembly and associated circuitry provide optical detection of the

sonoluminescence generated by a transient cavitation event.
pulse from the

The

PMT

is

pre-amplified and sent to an

ORTEC 9302

pulse height discriminator, which determines the sensitivity of the

mechanism. The discriminated signal

2071 A counter.

is

then sent through a Canberra

light noise within the light-

Because of background

tight enclosure, there is

always some pulsed output from the PMT.

Since transient events nearly always occur in groups, events will likely

produce numbers of pulses significantly higher than background.

the time of this writing, attempts to detect cavitation via this

At

method

have proved unsuccessful.

41

E.

TRANSDUCER CALIBRATION APPARATUS

In addition to the apparatus used for the generation

and detection

of cavitation, the apparatus

is

used

for the

source transducer calibration

Calibration

sufficiently involved to

warrant further description.

involves measuring the transverse

and

axial

beam

pattern of the source

transducer and determining the transducer's total acoustic power

output by force target deflection.

The

axial

and transverse beam patterns

found

of both the source

and

detection

transducers were

using a small
for

PVDF probe
use.

hydrophone designed and

built

specifically

this

The

construction and calibration of this hydrophone were accomplished to

fulfill

requirements

for

a final project in a transducer theory

and

design course.

Pertinent portions of the report submitted on the

project are contained

on Appendix B.
jig that

The PVDF hydrophone was

its

mounted

in a

micrometer

allowed precise measurement of

position relative to the transducer face (Figure 3.5).

The

force target deflection

measurement was used

to

determine

peak output frequencies and

transducers (Figure
3.6).

total acoustic

power output of the source

description of the theory behind the force

In this procedure, the source

target procedure follows in Chapter IV.

transducer
at the

is

driven by the function generator via the power amplifier

setting as will be

same gain

used

in the threshold experiment.

The

target is a small plexiglass plate,

of

suspended

at a 45-degree angle
line of equal

by four pieces

two-pound

test

monofilament fishing

length (approximately .04

mm

diameter, 225 jxgrams/cm

linear

42

Figure 3.5

PVDF Hydrophone Determination

if

of Source Transducer

Beam

Pattern

Figure 3.6

Equipment Arrangement for Force Target Deflection Measurements

43

density).

The

plexiglass target
It

is

rectangular in shape, of dimensions

to

3.8 x 5.1

x 0.7 cm.
1

has been carefully weighted using lead strips

air).

achieve a weight of
0.3
stiff

g in water (14.7 g in
into the target

cavity approximately
clear, thin,

cm

deep

is

machined

and covered by a

plastic to

form a pressure-release surface.

is

This ensures that the

radiation force

almost entirely transformed into a deflecting force.

is

Essentially no

acoustic energy

reflected

back

to

the

driving

transducer, but rather toward the bottom of the test tank.

deflection

The

was measured by a graduated telescope mounted

jig

to the

micrometer

described above.

44

IV.

EXPERIMENTAL PROCEDURE

A.

BACKGROUND
Prior to data taking,
it is

necessary to determine the performance

This involves reciprocity
for the

characteristics of the transducers involved.

calibration of the two Panametrics

V301 transducers used

Also,

acoustic detection of cavitation events.

once the source

it

transducers' true peak output frequencies are determined,

is

necessary to determine their axial and transverse

beam

patterns.
field

These transducers are used

induces cavitation.

to

generate the ultrasonic

that

The

axial

beam

patterns are used to verify the

transducers' focal lengths while the transverse

integrated to yield the effective area,
Aeff,

beam

patterns are

which

relates

peak intensity

and

total

transducer output power.

it is

Having determined the beam patterns,

a radiation force-target deflection

necessary to perform
of the

measurement

Harisonic

transducers in order to calculate transducer output power.

Having

experimentally determined Aeff and total power, the output peak

intensity

and consequently the output pressure characteristics can be

accurately determined.

Upon completion
tation

of transducer source strength calibration, caviin

samples are prepared and threshold measurements are made

order to establish a data base.

45

B.

RECIPROCITY CALIBRATION
The technique employed
to find the sensitivities of the

Panamet-

rics

V301 transducers

is

identical to that described in Reference 25.

is

See Figure 4.1.

distance
is

The distance between transducers

20.2 cm.

This

well into the far field of the transmitter,

which

is

found

using the following approximation:

r min
1

L
(4 1]
-

~1T~4X
where

= acoustic wavelength 1.5

fresh water

mm

(assuming

c =

1481

m/s

for

and

986 kHz),

L
r

= diameter of V301 transducer = 3.2 cm, n =

minimum

distance to the acoustic far

field (in

cm).

By employing

the appropriate equations established in Reference 25,

sensitivities
serial

graphs of the transducers'

4.2.

can be constructed.

See Figure

The V301 transducers with

numbers 596 and 598 were used.

C.

AXIAL BEAM PATTERN TRANSDUCERS

MEASUREMENT OF SOURCE

Using the polyvinylidene fluoride (PVDF) hydrophone discussed

above (see Appendix B) and a precision measuring
the axial
jig (see Figure 3.5),

beam

patterns of the two Harisonic source transducers are

determined.

A CW

sinusoid at the transducer's rated operating

46

DC
(a)

T
(b)

\>

v'x

03*
-e
tile)

*D

Figure 4.1

Reciprocity Calibration of Transducers (from

[Ref. 25])

-80
to

*- V301
00
"O
1

00

5ER.

-**-

V301 SER. V301 5ER.

NO 596 NO. 598 NO. 600

>

a
c
if)

-1

10

120

350

550
Frequency (kHz)

750

950

Figure 4.2

Sensitivity Calibration Results of 3

Paname tries V3 01 Transducers

47

frequency

is

generated by the

HP 3314A

function generator and

passed
amplifier

through the Amplifier Research model

and
finally to the test transducer.

100A15 power
is

The amplifier gain

set at

200

for

a constant
is

15

VAC

(peak-to-peak) output.

The PVDF

hydrophone
measuring

inserted into the mounting bracket attached to the

jig.

The

active element of the

hydrophone

is

oriented such

that the hydrophone face

axis as the micrometer

moves along the transducer's longitudinal

is

knob

rotated.

By recording

the received
jig's

voltage from the hydrophone

and the position from the measuring

micrometer knob, an axial beam pattern can be determined and the

transducer focal length
verified.

The Harisonic transducers used

to

in

this series of experiments

were found

have focal lengths of 7.29 and

7.37 cm.

There

is

not a large enough difference between these values

and the rated

focal length of 7.62

cm

(3 in.) to

warrant redesign of the

cavitation tank or relocation of the transducers used for acoustic

detection of cavitation.

D.

TRANSVERSE BEAM PATTERN MEASUREMENT OF SOURCE TRANSDUCERS

1.

Theory
It

is

necessary to relate peak intensity and total acoustic

power
voltage

in order to establish a relationship

between transducer input

In order to establish this

and transducer output pressure.

,

relationship

we

define

an

effective

beam

area Aeff by

48

where Ipeak = peak intensity (W/m 2 and Ptot =

)

total

output power
is

(W).

Since

a p2 a

V2

= K'V2 where K'

is

a constant; p

acoustic

pressure and

is

the corresponding hydrophone ouput voltage.

This
face,

voltage, a function of distance

is

from the center of the transducer

determined by a transverse scan of the focal region.

At the focal

point.

Ipeak =

k v peak'

Substituting this expression into equation

(4.2) gives

A e ff=

Ptot

KV2 PEAK
that

(4.3)

It is

known

Ptot = The

IdS;

dS = incremental area normal

to the

beam.

integral is taken over the entire cross-sectional area of the

(in

beam

and can be approximated

polar coordinates) by

Ptot =

Jf
I
I

K^rdrdG

2kK

\^rdr

(4.4)

where

= radial distance from transducer axis

is

and a

is

some

radial

distance within which

In practice, a is

contained practically

all

of the

beam

energy.

chosen equal

to the radius of the

transducer face.

Therefore,

2kK' F \^rdr

2n
=

A e ff

"2

Kry

PEAK

J 2 VPEAK

\^rdr

4 5
-

49

Finally, since Aeff

it

can be determined from the transverse beam

(as

profile,

is

necessary only to calculate Pjot

the

described in a subsequent

section) in order to calculate

PTOT/Aeff.

peak intensity since Ipeak =

Further,

since peak pressure,

ppeak.

is

related to peak

intensity by the following relationship (assuming either plane or

spherical waves),

Ppeak = (pocIpeak)

1/2
.

(4.6)

Ppeak can be rewritten

in

terms of Aeff and Ptot:

Ppeak =

^IH

1/2
-

(4 7)
-

2.

Procedure

The procedure used

is

for

measuring transverse beam patterns

very similar to that described in the previous section.

The

jig is

positioned in a

manner such

that the hydrophone travels in a line

perpendicular to the longitudinal axis at a distance from the

transducer face equal to the focal length.

Accurate position measure-

ments (10
readings.

jim)

can be made by recording the micrometer knob

Concurrently, hydrophone output voltages are recorded

from the oscilloscope display, resulting in the transverse beam patterns of Figures 4.3

and

4.4.

Plots of

V^r

vs. r are included as Figures to

4.5

and

4.6.

Only the right half of each beam pattern was used

50

>
E c o
-C CL
<u

O
1_

"O

X u_ Q
>
CL

>^

Q.

o
-2.5-2.0-1.5-1.0-0.5 0.0 0.5
1.0
1.5

2.0

2.5

Distance from Center of Transducer Face (cm)

Figure 4.3

Transverse

Beam

Pattern of 1.0

MHz

Transducer

>
E a o
.c Q.

O
(_

"O

X
u. Q.

>.

Q >

13

Q.

-3-2-1

Distance from Center of Transducer Face (cm)

Figure 4.4

Transverse

Beam

Pattern of 2.25

MHz

Source Transducer

51

4.0
1 s 3.0
CM
1

J I

\
\
J

>
2.5

b
x
CM

2.0
i

,5

>

1.0
J

0.5

0.0

0.0

0.5

1.0

1.5

2.0

2.5

Distance from Center of Transducer Face (cm)

Figure 4.5

V2r

vs. r (Integrand of

1.0

MHz

Equation (4.4)); Source Transducer

>

CM

>

2 -~

0H>
0.0
1.0

2.0

Distance from Center of Transducer Face (cm)

Figure 4.6

V2 r

vs. r (Integrand of

2.25

MHz

Equation (4.4)); Source Transducer

52

be symmetrical.

This

is

not exactly true, but no appreciable error

is

introduced by making this assumption.

Integrations of these plots are

performed by using a Macintosh Plus personal computer and

software.

DBGPLOT

The

results are plotted in Figures 4.7

and

4.8.

E.

DETERMINATION OF TRANSDUCER PEAK OUTPUT FREQUENCY

In order to optimize the output from the source transducers,
it is

necessary to

know

their

peak output frequency.

Using the deflection

apparatus illustrated in Figure

in the focal region of

3.6, the rectangular target is

suspended

a source transducer.
is

sinusoidal signal at the

rated operating frequency

amplifier.

applied to the transducer via the power

is

constant input voltage of 15 volts (peak to peak)

is

main-

tained (monitored on the oscilloscope) while frequency

varied above

and below the manufacturer's quoted peak output values.

deflection
is

Target

measured

for the various test

input frequencies by noting

the change in target position with a graduated telescope

mounted on

the measuring jig used previously.

Plots of target deflection versus

transducer frequency are reproduced as Figures 4.9 and 4.10.

transducers specified as
1

The

MHz and

2.25

MHz peak

output frequencies

were found to have true peak output frequencies of 986 kHz and 2.17

MHz,

respectively.

53


CM

E
<N

03

3
'

<>

> b

U.O

CO

U.D H

A 4/l H U.
CM

>
0.0<
0.0
1

'

0.5

1.0

1.5

2.0

2.5

Distance from Center of Transducer Face (cm)

Figure 4.7

Integral of

V^dr

for 1.0

MHz

Transducer

CM

E
CM

>

b
X

CM

>

Distance from Center of Transducer Face (cm)

Figure 4.8 Integral of

V 2 rdr

for 2.25

MHz

Transducer

54

E E

ra

0.97

0.98

0.99

1.00

1.01

1.02

Transducer Operating Frequency (MHz)

Figure 4.9

Target Deflection versus Frequency for Determination of Peak Operating Frequency of 1.0 MHz Source Transducer

u.

E E c o

U.D

"

en
C3

0.41.90

n
1.95

2.00 2.05

2.

10

2.15 2.20

2.25 2.30

Transducer Operating Frequency (MHz)

Figure 4.10

Target Deflection versus Frequency for Determination of Peak Operating Frequency of 2.25 MHz Source Transducer

55

F.

RADIATION FORCE-TARGET DEFLECTION MEASUREMENT

1.

Theory
The method used
to
is

determine the peak output pressure of

a variation of the method described by

the Harisonic transducers

Dunn,

et al. [Ref. 26].

Rather than using a sphere, as in Dunn's work,

is

the rectangular plexiglass target

employed.

rectangular target
it

is

used rather than a sphere because, as a

field,

sphere deflects,

tends to move out of the sound

while a rela-

tively large plate is less susceptible to this effect.

In addition, there is

no need

to calculate the

form function.

Referring to Figure 4.11 (reproduced from Reference 26 with

some

clarifying remarks),

it is

a relatively simple matter to construct

the appropriate free-body diagram and the equations governing the

static equilibrium:

IFX IFy

= -F r + Tcos9 =

0,

-mg + Tsin0

0.

The above equations imply that

T-^--^&" cos0 " sine*

1

* 8) (4 lOJ
l

Therefore,

Fr

= r-4r.

However, tan0 = -g

If

Lh,

tanG L/d.

Therefore,

_jng_m_mgd

f4 gl

56

Transducer

Figure 4.11

Free Body Diagram of Force Target Deflection

From References 25 and

27,

it

can be shown that the excess pressure

2 p e = pc s, where p = density of the medium, s = condensation, and c =

phase speed of the acoustic waves.

For a high-intensity sound

s).

field,

variations in density are significant, so p = po(l +

By

substitution,

57

2 2 Pe = pc s = po(l + s)c 2 s = poc s + poc 2 s 2

If

s = socoscot,

the time-

averaged excess pressure, <p e >t. can be rewritten as follows:

2 <Pe>t = Po c2s o <coscot> + poc s

<cos 2 cot>.

The
term

first

term on the right-hand side reduces

to zero,

but the second

(or

radiation pressure) can be expanded by employing the

appropriate trigonometric identity.

2
2

poc 2 s

<COS 2 C:> = poc 2 s

q" [1

+ <COS 2 G)t>]

poc 2 s

poc 2 s

<cos2cot>.
1 Poc^s
tz

This reduces finally to poc 2 Sq <cos 2 cot> =

However, p e =

where
it

= intensity.

Recalling that power, Ptot =

where
is

= area,

can be seen that the relationship Ptot = <Pe>tcA

also true.

Setting <p e >t

A = Fr

Ptot =

"T"

A = Fr

where

c =

1481

m/s

since

these tests are conducted in fresh water.

By

substitution,

PTOT= I2^ d
where
mfic

(4.10)

g
By

is

a constant and d

is

the

measured

target deflection.

substituting the appropriate values for m,

g, c,

and

L,

58

Ptot =

(0.99 x lO- 3 kg) (9 .8

m/ s 2

(1481 m/s)

14 05 x 10"^
(

d:

1.02 x lO 2

N^

d;

d in meters.

This equation

is

used

in the radiation force-target deflection

mea-

surements.

Upon completion

of

such measurements,

all

ingredients

necessary to characterize transducer output pressure

will

have been

determined.
2.

Procedure

A CW

sinusoid

(at

the transducer's peak output frequency)

generated by the

HP 3314A

function generator

is

sent through the

Amplifier Research model 100A15 power amplifier and finally to the

Harisonic transducer undergoing

target
is

test.

The

deflection of the plexiglass

(as

measured as the transducer input voltage

monitored on

the Tektronix
vs.

2445A

oscilloscope) is varied.

Plots of target deflection

input voltage squared are included as Figures 4.12 and 4.13.

The

slopes of these plots are proportional to output power for a given input
voltage.

Note that the "mirror" images of the recorded data points are

plotted as well.

This ensures that a line determined

by linear

regression will pass through the origin, resulting in zero power output
for zero

input voltage, as

is

dictated
all

by common sense.

Having determined

of the terms necessary to characterize

output pressure, the pressure as a function of input voltage can be

plotted.

See Figures 4.14 and 4.15.

Again, the "mirror" images are

plotted as well in order to drive the linear regression through the

59

c o
o
<^-

Q
4->
<1>

"

'

-3000 -2000
(-/-)

-1000

1000

2000
(V 2
)

3000

Input Voltage Squared

Figure 4.12

Target Deflection vs. V2 for Source Transducer Calibration 1.0

MHz

^
b E c o 4-> o
<L>

<

Q
4-1

<D
1

<U

u>
t_ TO

-2

3000 -2000 -1000

1000

2000

3000

(+/-) Input Voltage Squared (V 2 )

Figure 4.13

Target Deflection vs. V2 for Source Transducer Calibration 2.25

MHz

60

TO

If)

in

Q.

o
u
c
TO
(_ a>

Transducer Input Voltage (Volts)

Figure 4.14

Output Pressure Amplitude

for 1.0

MHz

Input Voltage Source Transducer

vs.

TO
n*

-O
<D (_ Z3

If)

in
<z>

t_

Q.

O
t_ <u

3 o
If)

c
TO

-
20 30 40

50

-40

-30-20-10

10

50

Transducer Input Voltage (Volts)

Figure 4.15

Output Pressure Amplitude

for 2.25

MHz

vs. Input Voltage Source Transducer

61

origin.

Note that, in Figure 4.15, the relatively large deviations from

the linear regression at lower voltages are of no concern since the

transducer

is

not operated in this region.

The relationships between

the peak-to-peak transducer input voltage and the peak output

pressure are included below.

For the

MHz

transducer:

Ppeak

(bar) =

0.0522 Vi N (mV).
transducer:

For the 2.25

MHz

Ppeak

(bar) =

0.0403 Vi N (mV).

G.

DETERMINATION OF TRANSIENT CAVITATION THRESHOLDS

All threshold

data were taken using the passive acoustic (scatter-

ing of the source signal from the bubble "cloud" produced

sient event) detection scheme.

by a tran-

This method was the

first to

show

acceptable
tionally

reliability.

As the data required was substantial and excepto take,

time-consuming

we concentrated on data accumulasacrifice

tion via this

method rather than

more time

refining the other

two detection methods.

The HP 9000

series

300 computer was

pro-

grammed

to allow

user input of desired frequency, pulse repetition

frequency (PRF), pulse duration, and pressure amplitude sweep para-

meters

(via

input of function generator starting amplitude, voltage

per step, and total

amplitude increase

number

of

amplitude steps).

Data was obtained at 986 kHz and 2.17

MHz

using 200 Hz, 500 Hz,

62

and 1000 Hz PRF. Pulse durations of and 40

jisec

1.5, 2, 5,

10, 15, 20, 25, 30, 35,

were examined.

All voltage

ramps were

started at 300in

400

mV

peak input voltage

to the

power

amplifier,

which results

an

initial

peak output pressure from the source transducers

of approxi.

mately 3-4 bar.

Incrementally increasing pressure steps of

bar

were held
tion event
at

for 1.5 seconds.

The ramps were continued

until a cavita-

was

detected.

Twenty threshold measurements were taken

each pulse duration, PRF, and frequency combination.

63

V.

RESULTS. DISCUSSION. RECOMMENDATIONS

A.

RESULTS AND DISCUSSION

Figures 5.1 and 5.2 are graphs of threshold pressure versus pulse

duration taken at 986 kHz with PRFs of 200 Hz and 1000 Hz, respectively.

The

error bars in these

and subsequent

figures represent one

standard deviation of the data above and one standard deviation below
the

mean

threshold pressure.
jisec

It

should be noted that thresholds were

undetectable at the 2

pulse width for the 200 Hz PRF, most likely

due

to lack of scattered signal

from a two-cycle burst at such a low

PRF. At 1000 Hz PRF, the thresholds are essentially constant down to

the 10 (isec pulse duration.

At 5 and 2

jisec

pulse durations, the

thresholds

show a marked

increase.

The data
6.

in Figures 5.1

and 5.2

show good agreement with Reference

20

CO

3
to CO o>
i_

Qo

o
-C
co

'

'

'

10

20

30

40

50

Pulse Duration (microseconds)

Figure 5.1

Threshold Pressure

vs. Pulse

Duration 986 kHz, 200 Hz

PRF

64

'

'

p, o

1ft-<u _
1

*4

A -

en 01
<l>
'

c_

O Z

JL.
( 1

Q.

o
o
.c
to
fl o

1 l

J-

} 1

ft D

--

'

'

'

40
Pulse Duration (microseconds)

50

Figure 5.2

Threshold Pressure

vs. Pulse

Duration 986 kHz, 1000 Hz

PRF

Figures 5.3 and 5.4 are graphs of threshold pressure versus pulse

duration at 2.17

MHz

for

PRFs

of

200 and 1000 Hz,

at

respectively.
for pulse

There

is

an essentially constant threshold

|j.sec

200 Hz PRF

durations of 15
a

and

larger, while shorter pulse durations indicate

marked increase

in threshold pressure.

In contrast, however, at
for all pulse

1000 Hz PRF, the thresholds are essentially constant

durations

down

to 2 |isec,

with a marked increase evident at a pulse

duration of 1.5
Figure 5.5

|isec. is

a comparison of the thresholds obtained for the two

is

source frequencies at 200 Hz PRF; Figure 5.6

the

same

at

1000 Hz
[Ref. 4]

PRF.

These results support theoretical predictions by Flynn

[Ref. 3] that

and Apfel

the threshold

is

independent of frequency

for

sufficiently small cavitation nuclei.

For parameters used in this

65

16
TO

nrwwwnwWKWVWW

.o
<D

14
12 10

__

in

A-

it
to

Mj

10

20

30

40

50

Pulse Duration (microseconds)

Figure 5.3

Threshold Pressure vs. Pulse Duration 2.17 MHz, 200 Hz

PRF

20
<*-v

18
16

L. TO

.a
a> L. 05 en

14
12

Q.

o
o
sz

10

8
6

i
-I
10

n
P"
20

r
30

r
40

50

Pulse Duration (microseconds)

Figure 5.4

Threshold Pressure

vs. Pulse

Duration 2.17 MHz, 1000 Hz

PRF

66

CO

Z3

0)

a
o
.c
to

30

40

50

Pulse Duration (microseconds)

Figure 5.5

Threshold Pressure vs. Pulse Duration 986 kHz and 2.17 MHz at 200 Hz PRF

20

-O

16

-o ^^^ 0"KrtZ

O 3
<n u) <D

14
fr 12

7MHZ

Q. "O

10 8 6

O
.c

U5 0> l_

1
1

<r

4
10

'

20

30

40

50

Pulse Duration (microseconds)

Figure 5.6

Threshold Pressure vs. Pulse Duration 986 kHz and 2.17 MHz at 1000 Hz PRF

67

experiment, threshold pressure values are essentially equal

to

Blake

pressure thresholds, which have no frequency dependence (see

equation 2.11). This conclusion
is

valud due to the close relationship

for the frequencies

between transient and Blake thresholds

consideration (see Figure 2.3).

under

Figure 5.7

is

a histogram of pressure thresholds (rounded to the

nearest bar) measured for each individual transient cavitation event.

Again from equation (2.11) above, there

is

a direct relationship Figures 5.8

for individual

between bubble

size

and Blake threshold pressure.

through 5.11 are histograms of threshold measurements

frequency/ PRF combinations.
It

should be noted that the graphs

may

be slightly skewed due to the starting pressure amplitudes of most

ramps.

Since our experience indicated that expected threshold

pressures were in the 6-9 bar range, source pressure ramps were
started in the 3-4 bar range.

An

occasional low threshold event due to

the presence of an unusually large nucleus

may have been

missed.

Keeping this in mind,

exhibit a roughly

it

appears that

all

histograms except Figure 5.8

in the 7-8 bar

Gaussian distribution with mean values

range.
(j.m.

This corresponds to bubble radii in the range of 0.075-0.09

As stated above, when

compared
in Figures 5.5

larger pulse width

measurements

are

and

5.6, the thresholds are

is

independent of

pulse duration.
threshold.

As pulse durations decrease, there

an increase

in

This observation would indicate that the larger pulse

68

200
en

o c

o
o>

.Q

4 5 6

10111213141516171819 20 21 22232425

Threshold Pressure (bar)

Figure 5.7

Histogram of

Threshold Pressure Measurements (rounded to nearest bar)

All

to a>

c o o

<D

4 5 6 7

9 1011

12131415161718192021 22232425

Threshold Pressure (bar)

Figure 5.8

Histogram of

All

Threshold Pressure Measurements 986 kHz, 200 Hz PRF

69

en a>
<_>

tz <D (_ t_

o
<D

O o

E
Z3

4 5 6

101112131415161718192021 22 23 2425
Threshold Pressure (bar)

Figure 5.9

Histogram of All Threshold Pressure Measurements 986 kHz, 1000 Hz PRF

V)

O c

o
<1>

4 5 6 7 8 9 10111213141516171819 20 21 22 23 2425 Threshold Pressure (bar)

Figure 5.10

Histogram of

All

Threshold Pressure Measurements 2.17 MHz, 200 Hz PRF

70

en <D

c
<D Z3

o
.a

Z
2 3

4 5 6 7

10111213 14151617181920:

Threshold Pressure (bar)

Figure 5.11

Histogram of

All

Threshold Pressure Measurements 2.17 MHz, 1000 Hz PRF

durations allow time for a growth mechanism, such as rectified

diffusion, to

cause nuclei

to evolve to a

Blake or transient threshold

for

radius.

There appears

to

be insufficient time

the

growth

mechanism
that,

to take effect at shorter pulse durations.

It is

postulated

during the rest time between pulses, nuclei dissolve to their

original equilibrium sizes.

Further examination of Figures 5.1 through 5.4 reveals that there

is

a distinct relationship between the

Specifically, Figure 5.1

flat

and sloped portions

of the

graphs.

(low frequency/low PRF) reveals a

continuously decreasing threshold pressure for increasing pulse

duration.

Although no additional data was taken at pulse durations

greater than

40 usee, there were essentially no thresholds recorded

at

71

levels lower

than the threshold determined at that point, so

it

is

anticipated that the graph flattens out beyond this point.

An attempt

was made

to evaluate this

hypothesis experimentally, but problems

were encountered with the 986 kHz source transducer that prevented

any further
5.2

reliable

data from being obtained at that frequency.

Figure

(low

frequency/high
for

PRF)

shows

relatively

flat

threshold

measurements

pulse durations of 10 |isec and greater, with

significantly higher thresholds occurring at 5

and 2

jisec

pulse

durations.

Figure 5.3 (high frequency/low PRF) shows

for pulse

flat

threshold

measurements

durations of 15-20
5,

jisec

and

greater,

and

increasing thresholds for 10,

and 2

|j.sec

pulse durations.
flat

Finally,

Figure 5.4 (high frequency/ high PRF) shows essentially

threshold

measurements down

to

|isec

pulse duration.

This observation
at 2.17

prompted an additional MHz, 1000 Hz PRF, and

ascertain whether there

set of 10 threshold

measurements

1.5 |isec pulse duration (3 cycles per pulse) to

was a

rise in threshold

below 2 usee.

The

mean
2.90,

value of the ten readings was 12.9 bar, with standard deviation of

which indicates a "knee"

in this

graph as well in the vicinity of

the 2 (isec pulse duration.

There appears to be a relationship between

frequency, pulse duration, and

PRF which

allows a certain range of

pulses per unit time to

mark

the break point where shorter pulse

In Reference

durations begin to affect cavitation thresholds.

it

was

postulated that threshold became independent of pulse duration

when
the

pulses contained more than 10 cycles.

Our measurements show

"knee" to be located as indicated in Table 5.1:

72

TABLE

5.1

CYCLES PER UNIT TIME FOR VARIOUS FREQUENCY. PRF. AND PULSE DURATION COMBINATIONS

PRF

Pulse length
(-KNEE")

Cycles/pulse

Cycles/sec
(PRFx
Cycles/pulse)

986 kHz 986 kHz

2.17 2.17

200 Hz
1000 Hz

40
10 15 2

n.sec fisec jisec jisec

40
10

8,000
10,000

MHz MHz

200 Hz
1000 Hz

33
4

6,600 4,000

Although the value of PRF x (cycles/pulse) product varies from

4,000
to 10,000, this analysis indicates that the

knee of the graph

is

predictable within a reasonable range of pulse widths.

In addition, our

choice of pulse widths

of the graph.

was not geared toward

finding this break point

There

may

be other pulse durations that more accurately

indicate the knee between those examined which yield a narrower

range of cycles per unit time values. To test the hypothesis, a series of
threshold measurements were taken at 2.17

MHz and 500 Hz

PRF.

It

was anticipated that the knee

of the

of the curve should occur in the vicinity

10 usee pulse duration.

of these

Figure 5.12, which illustrates the

results

measurements, shows that the knee occurs as

it

predicted.

Although

is

not prudent to generalize

in

when discussing
fields, it is felt

mechanisms as complex as bubble dynamics

sound

that this aspect of these findings bears further investigation.

73

18 17 16
TO

-Q
CD (_ Z> in CO CD

15

14
13 12
1 1
1

"1

10 9 8 7

~T~
_

Q
tr

ur

6 5

4
10

'

'
1

'

20

30

40

Pulse Duration (microseconds)

Figure 5-12

Threshold Pressure vs. Pulse Duration 2.17 MHz, 500 Hz PRF

B.

RECOMMENDATIONS
All experimental

measurements were taken using the passive

acoustic scattering method, which proved to be an excellent detection

method when pulse widths were

large.

The resulting

signal scattered

from the bubble "cloud" generated by the cavitation event resulted in a

clear, distinct indication of

a transient event. As pulse durations were

decreased, however, the scattered signal

lacking definition.
It

was sometimes weak and

is

essential that sonoluminescence detection

and /or

active acoustic detection

schemes be

fully incorporated into

the procedures to minimize any systematic errors that

may

exist in

the passive detection system.

limited

The

optical detection
(a

system was

only by unavailability of luminol

final

sonoluminescence

enhancer) until the very

weeks of the

project.

By

then,

it

was

too

74

late to bring the

system

to full operation.
will

It

is

anticipated that this

system,

when

fully operational,

allow the replacement of the

human
system.

element of the current detection scheme with the electronics

As gathering

of cavitation data via

manual observation

of the

transient signals on an oscilloscope requires exceptional vigilance and

perseverance on the part of the operators, fully automating the

detection
apparatus.
It

mechanism

is

essential to the continued success of this

is

recommended, when the system

is

fully operational,

that

additional threshold

measurements be taken

at specific pulse dura-

tion /PRF/ frequency combinations in order to confirm the cycles per

unit time effect noted in the hypothesis above.

Extremely low or high

PRFs (10 Hz, 8 kHz) should

merit.

indicate whether the hypothesis

has
of

Localization of a specific

number

of pulses or period

insonification
to

which marks the knee

of the threshold curve

may

lead

a better determination of the

mechanisms which allow the

cavitation events to fully evolve.

75

APPENDIX A

APPARATUS CONTROL PROGRAM "CAVPAT"

I

****##***** + *#*****#****#*****#
>*

23

33 40 50 63 70
p3

!*

i*
!*

!
i

PROGRAMMERS: R.L. BRUCE AND R.D. MIDOLETQN 15 NOVEMBER 9S7 DATE: PROFESSOR A. A. ATCHLEY THESIS ADVISOR: PROGRAM NAME: CAVDAT HP BASIC 5.3 PROGRAMMING LANGUAGE: HP 30O0 SERIES 300 COMPUTER:
t

#* + #####*###***#*#***#####***##*******#****#**#*#*

30 03
;

##*** + ##* + ****** + *##p jRPOSE*******************************

i

110 120 130 140 150 ISO !70 130 130

I

300 210 330 230 340 250 350 270 2S0 230 300 310 320 330 340 350 350 370 330 390 400 410 420 430 440 450 4 63 470 330 490 530

'

"CAVDAT" IS TO CREATE AN ULTRASONIC OUNO FIELD PROGRAMMING THE HP 3314A FUNCTION GENERATOR. USER-DEFINED PULSE BURSTS ARE SENT TO A FOCUSSED TRANSDUCER WHICH I? THEN THE USERS CAN THEN SWEPT THROUGH A USER-DEFINED AMPLITUDE RANGE. DEVOTE T^EIR ATTENTION TO DETECTION O c CAVITATION EVENTS DISPLAYED ON AN OSCILLOSCOPE. FURTHER, WHEN A CAVITATION EVENT IS DETECTED, AND THE COMPUTER IS MANUALLY CUED, THE COMPUTER WILL DETERMINE TRANSDUCER SOURCE LEVEL USING CONVERSION FACTORS PRE-DETERMINED EXPERIMENTALLY. THE CRITICAL USER INPUTS ARE PULSE REPETITION THIS ALGORITHM WILL CALCULATE FREQUENCY PRF AND PULSE DURATION. THE REQUIRED NUMBER OF CYCLES PEP BURST, N, TO ACHIEVE T HE DESIRED WAVEFORM SUBJECT TO THESE PARAMETERS.
By'
t
.

THE D URPOSE OF

'

*****#+*#************** *vari able defini tions* ****************** *****

Aeipcut = HP 3314A OUTPUT VOLTAGE AMPLITUDE IN MILLIVOLTS. Arcpstart - HP 331 4A STARTING VOLTAGE AMPLITUDE IN MILLIVOLTS. Gain = ROWER AMPLIFIER GAIN. Incr = HP 331 4A VOLTAGE INCREMENT IN MILLIVOLTS. Inivl - SW/TR INTERVAL IN MILLISECONDS. N = NUMBER OF CYCLES IN PULSE BURST; HP 33 UA IS IN N-CYCLE MODE. K - COUNTER INDEX.
- COUNTER INDEX. Nunbetepv = NUMBER OF STEPS IN VOLTAGE AMPLITUDE' SWEEP. Opfreq = TRANSDUCER OPERATING FREQUENCY. PeaMK) - PEAK PRESSURE AT CAVITATION EVENT IN EAR. Prf - PULSE REPETITION FREQUENCY IN HZ. Pulse - PULSE DURATION IN MICROSECONDS. Ren = REMAINDER OF THE INDICATED DIVISION.

CLEAR SCREEN CLEAR 7 CLEAR HP- IB INTERFACE ON KBD ALL GOTO 1470 (ALLOW FOR USER INTERRUPT 0> PROGRAM.
I i
i

*#*** **###* + #**##***^AP, I ABLE DECLARATIONS************************

INTEGER K ,L ,N .Numbstepa REAL Amp out Anpe tart ,Ga in Incr Intvl Op free ,Prf .Pulse .Rem
, ,

76

" '

"

i*** * ************** +u rHENSIGN LINEAR

fiRR A

v '~

**********************

E30 340

DIM APinc< ieea

,PeakMS00)

~3

i*****#* + ####********'AD DRESSES ******************************

i i

5E0 G~Z 580

5 90
ot?V3

'

7! 7 IS THE ADDRESS OF TnE HF 33 4* FUNCTION GENERATOR. 732 IS THE ADDRESS OF THE HP ZZ.25A THINKJET PRINTER.
1

'

'

e:o ^z^
5-18

prior to any test run, verify power amplifier bain.

353 S53 572 5S3 590 703

71

*********************** i^f \jj piJLSE PARAMETERS***********************

i

7Z0 730 ?4B "S3 750 770 780 "SO 300 310 ECO 330 640
S50-

SS0 572 530 S30 300

Si

SCO S30
54.3

CLEAR SCREEN DIS^ "PLEASE ENTER TRANSDUCER OPERATING FREQUENCY IN KHZ. INPUT Opfreq DI3P "FLEAEE ENTER POWER AMPLIFIER GAIN '.ZOO OR LESS)."! INPUT Gain (OPERATOR INPUTS THE GAIN FOR FUTURE REFERENCE. IF Sain>20B THEN GOTO "42 GOTO 770 CIS- "THIS GAIN MAY RESULT IN TRANSDUCER DAMAGE." WAIT 2 GCTO 700 OISP "PLEASE ENTER DESIRED PULSE REPETITION FREQUENCY (IN HZ " INPUT r f 'CALCULATE SW.< TP INTERVAL IN MILLISECONDS. Ir,lvl( \/Prf >*J000 CIS- "PLE.nSE ENTER DESIRED PULSE DURATION IN MICROSECONDS " INPUT Pulse N=< <Pulse*Gpfreq.VI000) 'CALCULATE NUMBER OF CYCLES. IECHO THE INPUT PARAMETERS AND AllOW USER TO EDIT AS NECESSARY. PRINT 'TRANSDUCER OPERATING FREQUENCY 'IS: " lOpfreq "KHZ. PRINT "PULSE REPETITION FREQUENCY IS " PrfV'HZ " PRINT "PULSE DURATION IS :" -.Pulse; "MICROSECONDS. PRINT "NUMBER OF CYCLES PER BURST IS:",iN PRINT "POWER AMPLIFIER GAIN IS: # -Gain PRINT "SW/TR INTERVAL IS Intv ;" MILLISECONDS PRINT "ARE THESE PARAMETERS CORRECT? <Y/N>" INPUT DeclS IF UPCS'OecU >="N" THEN GOTO 570
"
, .

!*<?##*#**#**###* INPUT SWEEP PARAMETERS******* ****** ******* **#

!

950 550 570 330 330

'1000

CLEAR DISP INPUT CI3P

SCREEN -PlEASE INPUT SWEEP STARTING AMPLITUDE IN MILLIVOLTS."; Ampstart "PLEASE INPUT AMPLITUDE INCREMENT IN MILLIVOLTS. " INPUT Incr

11

"

'

!G!0 :?20 030

'.040

0T5F
IF

'

P LSASE

INPUT NUMBER OF S7EFS FER SWEEP";

1

INPUT Nurvo steps

1053 1050 1370 10S0 1090 1100

Anpstart; 303-2 T\->E1. SOTO 2?0" v?" 3 IF Nunb3teps*Incr 3000 THEN 5CTC c Anpstart-KNuraosteps + Incr I 3006 THEN GOTO 1070 BOTC it 30 DISF "ENTRY OP THESE PnRAME T ERS MAY RESULT IN TRANSDUCER DAMAGE." WAIT 2 SOTC 350
7
1
,'

MIC
!!20 :!30

-ECHO THE SWEEP PARAMETERS.

\

M40
1150 !160
]

170

M30
!!S0 Z00 1218 1230 1230 1240 ;2S0 1Z60 1Z70
!

PRINT "STARTING AMPLITUDE IS: " Ampstart TMILLIVOLTS RINT "AMPLITUDE INCREMENT 15: " Incr? "MILLIVOLTS. PRINT -NUMBER OF STEPS FER SWEEP IS " *Nuibsteps PRINT "ARE THESE PARAMETERS CG c EECT r (Y/N)."i INPUT Dec 2% IF UPC$<Dac2S>-"N" THEN SOTO SE0 'INITIALIZE 5T0RA6E FIELD COUNTER-. K-1
;

'

*************** *AMPL TUCE SWEEP INFORMATION*********************

I

260

CLEAR SCREEN OUTPUT 717; FR" Op f ^eq "KZ SET TRANSDUCER OPERATING FREQUENCY. OUTPUT 717; "M03" 'SET N-CYCLE MODE. OUTPUT 717iNH" ;N;"EN" 'SET NUMBER OF CYCLES. OUTPUT 717i"TI"Intvl;"MS" 'SET SW/TR INTERVAL. WAIT ALLOW TRANSIENTS TO SETTLE OUT.
;,

'

"

1Z90 1300 :3I0 3Z0 1330 1340 1350 13S0 1370 1330 1390 1400 1410 1420 1430 1440 1450 1460 1470 1460 1490 1530
i

Apnpoui=Ar"ip5tart

FOP 1=1 TO Nunbstepa STEP 'COMMENCE AMPLITUDE SWEEP. Ampinc< I )=Ampout 'SET FUNCTION GENERATOR OUTPUT 717; " AP " Amp inc< I ) " NUQ " Ampout =Arnpi.nc( I HIncr 'INCREMENT COUNTER. WAIT 1.5 'ALLOW TRANSIENTS TO SETTLE OUT. 'CONTINUE WITH SWEEP UNTIL COMPLETE. NEXT I PRINT "AMPLITUDE SWEEF COMPLETE." IP Opfraq=SSB THEN GOTO 1390 IF QpfreQ2l70 THEN GOTO 1410 PeakU: = .0522*( Awpout-Incr )*( .2) >) GOTO 1420 Peak(K >-< .0403*( Anpotu-Incr )', .2 PRINT "SINCE NO EUENT DETECTED, THRESHOLD IS GREATER THAN: GOTO 1580 'QUERY USER FOR FURTHER INSTRUCTIONS.
!

>

"

tPeak< K

>?

"BAR.

!###**##*###*# ##***CAUI TAT ION INFORM AT I ON ******************

DI8P "WAS A CAVITATION EUENT DETECTED?"; INPUT Dec3$ IF UPCS(0ec3$ >="N" THEN GOTO 1530 IF UPC$<Dec3S)*"Y" THEN GOTO 1510

78

"

"

"

1510 5520 ETC 1540

i

IF
IF

GpfreqS8S THEN SOTO 530 0pfreq=2i70 THEN GOTO 550

1 i (

SEC 1550 5^0 '550 '530 1500 !5!0 '520 '530 '540 !SE0 560 1B70
!
'

Feat SOTO Fear

'

.0522*' Amp
.
.

ir.ci

>*<

>

<SE0
i

=
'.

PRINT
KK+I

'^EMr.

0403* Awe nc CI * 2 PRESSURE AT CAVITATION EVENT NUMBER"

i
>
.

?h

"IS:

"

;Peak< K) "BAR.

DTSF 'DO t'GO WISH TO RUN AGAIN *I7H THE SAME PARAMETERS?"; INPU1 Dec4J r HEN SOTO iZ0 IF UFC$<Dec4$>="Y = UFC*Dec4S = "N" THEN GOTO 1740 l OUTPUT 717; "PF" 'RESET Hf3 3314A FUNCTION GENERATOR. wFIT 50 ESTABLISH QUIESCENT PERIOD BETWEEN TRIALS. 5EEF 375.56". iCUE uSER THAT ANOTHER TRIAL IS nED\JT TO START.
,!

30

302. 06,. EETlF 54S .22 .5 BEEP 1790. 35,. 75 EF-E^ 'Z'ZZ.^c .25
!
i

SEE C

'.630
!

700

EEEP 1790.36 GOTO 1240

1710!
1720 1730 ]~40 1750
i

.*******-***** + ****#* *pj< INT-GuT OPTION* ** + ***#****+*** + #*#*

-50
1770 1730 1790 1600 :5!0 1520 :S30 1540 !550 1553 1570 1550 1630 1500 1910 1323 1330 1340 1350 1350 '570 1530 1330

CIGF "WOULD YOU LIkE A HAROCQPV uF PARAMETERS AND RESULTS""; INPUT Dec5$ IF UPC$<Dee5$>-"Y" FriEJii GOTG 1730" IF 'JPC${Dsc5$>="N" THEN. GOTO 520. IDENTTFv ADDRESS OF TH I NX JET PRINTER. PRINTER 15 ~0Z Upfreq "KHZ PRINT "TRANSDUCER OFERATING FREQUENCY IS PRINT "NUMBER OF CYCLES FEF BURST 15: "iN PRINT "POWER AMPLIFIER GAIN IS: "ream PRINT "SUi/TR INTERVAL IS :" Intvl "MILLISECONDS. PFIN T "PULSE DURATION I 5 " :Ful se "MICROSECONDS PRINT " d 0l5E REPETITION FREQUENCY IS Prf "HZ STEP 'CREATE DC L00 D FOR PRINTING OF RECORDED EVENTS FOR L=l TC (K-1 "BAR PRINT "PEAK PRESSURE AT CAVITATION EVENT NUMBER" L " I 5" PaaM L NEXT L UNTIL COMPLETE. 'CONTINUE PRINTER IS CRT (RETURN TO CRT
1
'

'

;
;

"

"

>

*##***#*##*##*CONTINUE TE5TIN6
.

IF

NECESSARY********************

DISP "DO YOU WISH TO ALTER PARAMETERS AND RUN AGAIN?"; INPUT Dec5$ IF UPC$'DecaS )"Y" THEN GOTO i960 IF UPC$(Dec5$ >*N" THEN GOTO 2050 OUTPUT 717; PR" RESET HP'3314A FUNCTION GENERATOR. 'WAIT 50 'ESTABLISH QUIESCENT PERIOD BETWEEN TRIALS. BEEF 975.55,1 CUE USER THAT ANOTHER TRIAL IS ABOUT TO START. BEEP 1302. 06 ,.5 2000 EEEP 154S.ZZ ..5
r
!

79

2030
2 "40

2050 2050
2370"

BEEP I7SG.3G..75 BEEP '302. 05, .25 SEEP 790 36 SOTO "0 PRINT PROGRAM COMPLETE, riJITPiJT 7 ftt PE5ET HP 3314A FUNCTION SENERA70R rUI END
!

'

'"

1
i

.'

80

APPENDIX B

PVDF HYDROPHONE CONSTRUCTION AND GENERAL INFORMATION

A.

CONSTRUCTION DETAILS
Our
desire for small size led us to choose a large plastic syringe

and a

large -diameter

hypodermic needle as the hydrophone housing.

This gave us a convenient method for insulating the "hot" lead from
the grounded elements, as well as providing ample storage area close
to the active

elements

for the preamplifier circuitry.

Our primary

dif-

ficulty in construction

was working with small

sizes of

PVDF
to a

film.

After

numerous unsuccessful attempts

for

to cut the

PVDF

shape

which allowed
element size
(1

good
),

electrical

connection while minimizing active

mm 2

we moved
in length.

to

PVDF

strip

approximately 2

mm

wide and about 4

of the film

cm

We

reduced the bottom

(positive) side

by etching away the PVDF using a Q-tip soaked

in ferric

chloride,

which gave us a marginally accurate method

size while

of reducing

element

keeping a capability to have a reliable electrical

connection.

The top (ground) was not etched.

The PVDF

strip

was

then centered on a 3.6

lead.

mm x 6.4 mm brass

plate attached to the "hot"

The

barrel of the needle

was insulated with

several layers of heat

shrink tubing to prevent any of the "hot" side from short-circuiting to

ground.

The remaining lengths

of the

PVDF

strip

were attached

to the

insulation along the needle barrel and connected to the surface of the

needle with nickel paint.

Gaps between the PVDF

strip

and needle

81

were

filled

with epoxy adhesive to add

to

stability.

Resistance measure-

ments were taken

ensure continuity of connections and that no

short circuits existed.

The unity gain preamplifier

installed in the syringe

(voltage follower)
electrical

was constructed and

tions

and

connec-

made

to the

hot lead and external ground connections.

"Plasti-Dip."

The
See

entire apparatus

was insulated by being coated with

Figures B.l and B.2.

B.

SENSITIVITY CALCULATION

[Ref. 28]

= Vo/t where

= Electric Field developed

Vo
t

= Voltage Induced

= Thickness of Piezoelectric film

is

Also,

= g33X3 where g33

a constant indicating field/applied stress

X3

= Applied mechanical stress

* *, o = r^. Pressure Sensitivity =

Mq
tv/t

v/D V/P

(Et)

=-tt

= g33X3t = g33t y

Numerically,

Mo

= g33t = 339 x 10-3

f^j^

(9

x 10-6 m)

= 3.051

x 10-6 Yjj^P = 3051 x 10_6

Pa"

Mo

= 20 logio (3.051 x 10-6

dB

re

lV/Pa = -110.31 dB

re

lV/Pa

82

o
ji

O
u

E
fa

83

4>
fl

O
43 fa

a
fa

CQ

fa

ea

t-.

H-

Q
a
J

E
i

KJ

o
a-

*5

te

a.

O
-x

^~

6
U-

a.
t-,x

a
>

1 >
CN
>

a-

84

This calculation indicates that transducer pressure sensitivity

is

frequency independent.
sensitivity is contingent

It

should be mentioned that the calculated

nickel paint

upon the assumption that the

to

used

to

mount

the

PVDF

the brass mounting block effectively

clamps the active face in the

(1)

and

(2)

directions, thus confining the

face to operation solely in the thickness-expander

B.3.

mode.

See Figure

0 T3
-4-"

>

2.0

2.5

3.0

Frequency (MHz)

Figure B.3

PVDF Hydrophone
C.

Sensitivity Calibration Curves

DESCRIPTION OF TEST AND CALIBRATION TECHNIQUE

We

chose the reciprocity technique for calibration of our

sensitivity

hydrophone

because calibrated hydrophones

for

in

the

megahertz range were not available

comparison.

Our frequency

band

of interest

was 0.5 MHz

to

4.0

MHz.

For the reciprocity

the trans-

calibration,

we used two Panametric V323 transducers as

85

mitting and reversible elements of the experimental setup.

These

transducers are designed

for

operation at 2.25

MHz and

provided

smooth response through the frequency range. The

calibration set
4.1.

up

was as noted

in the

equipment diagram included as Figure

The

distance between the transducers

was 30 cm, which

is

well into the

far field for all frequencies of interest.

Far-field calculation for the

maximum
r

frequency

is

as follows:

[Ref. 25]

where L = 5 x 10~ 3
wavelength at 4.0

= largest extent of source active element, X in fresh water, r =

MHz

minimum

extent of far

field,

and
x 10-3m) 2 = 4(3.70 x l<Hm) =
fS

01689

1689 cm

We
3478A

took Voltage Level

(in

dBV) and input current

(in

milli-

amperes) readings from the

Digital Multimeter in

HP 3585A Spectrum

Analyzer and

HP

accordance with standard reciprocity

techniques.

Open-circuit voltage sensitivity of our

PVDF hydrophone was

determined using the following equation:

M 0x =
When

[pocj VrIr J

[Ref. 25]

converted to sensitivity

level, this

calculation became:

86

ML

VLR + V^ - V Lr -

Lr -

20

log f +

20 log^ x

D.

DIRECTIVITY EFFECTS/EXCITATION OF FIRST THICKNESS

MODE
1.

Analysis of Directivity Effects

One
dicted

possible cause of the difference between actual and pre-

sensitivities

may be

the change in the directivity of the

hydrophone as frequency increases.

frequency
is

As the transmitting transducer

increased, both the hydrophone

and transducer become

more

directional, implying

more

efficient

transmission by the transof the acoustic

ducer and more

efficient reception

by the hydrophone
is

power

in a given direction.

The trade-off

that the

main lobes

of

both the transducer and the hydrophone become narrower.

Using the

approximation for a piston-like source

[Ref. 25],

= 1/4 k 2 LiL2 where Li = 2

mm and L2 = 5 mm

k =

co/c = 27nVc

At 500 kHz, DI = 10 log

= 10.5

dB

These "corrected" calculated values are plotted as such in Figure B.3.

2.

Calculation of First Thickness

Mode

of Brass

Mounting
would

It

was
if it

originally believed that the brass

mounting

plate

behave as

were perfectly
led

rigid in the

frequency range of interest.

not be the case.

Measurements

us

to believe that this

may

For this

87

reason,

it

was decided

to

determine

if

the

first

thickness

mode

of the

brass mounting plate was being excited somewhere in the range

between 0.5
excited with
t

MHz and
= X/2 or
fk, f

4.0 MHz.

The

first

thickness

mode

will

be
X =

when
= c/X:

X = 2t.

For a thickness of 1.3

mm,

2.6

mm.

Since c =

= c/X = (4700 m/s)/2.6 x 10~ 3

m=
it

1.81

MHz
that the
first

It

becomes evident that

is likely

thickness

mode
E.

is

being excited to some extent.

DISCUSSION
As noted above,
theoretical sensitivity

was calculated assuming

that the longitudinal

and transverse modes were clamped and that

excited.

only the thickness

mode was

Data showed

little

agreement

with these original theoretical calculations, however.

The PVDF

hydrophone's active face was considerably larger than the wavelengths

of the frequencies of interest,

which makes the receiving beam patThis directivity also

is

tern

more narrow as the frequency increases.

to the sensitivity

adds considerably
Directivity

when sound

received "on axis."

and

Directivity Index as calculated previously in this

paper

contributed greatly to the sensitivity of the hydrophone.

As can be

seen from the graph of Pressure Sensitivity

B.3), there is

vs.

Frequency (See Figure

good agreement between theoretical and experimental

At

values at the lower end of the frequency range (up to 1.5 MHz).

higher frequencies, additional variables are introduced which limit

88

agreement.

likely contributor to the

higher frequency disparity

might be slight misalignment of the hydrophone.

up. the

As frequency goes
Slight misalign-

beam

pattern becomes increasingly narrow.

ment could
energy.

result in poorer reception of the transmitted acoustic

it

Lastly, as alluded to previously,

is

possible that the

is

first

thickness

mode

of the brass

mounting plate

being excited in the

frequency range of interest.

89

LIST
Ross, D., Mechanics Press. Inc.. 1976.

OF REFERENCES
Underwater Noise
,

1.

of

p.

202, Pergamon

2.

Strasberg, M., "Onset of Ultrasonic Cavitation Tap Water," Journal of the Acoustical Society of America v. 31, pp. 163-176, February
,

1959.
3.

Apfel, R. E., "Possibility of Microcavitation

,

from Diagnostic Ultrasound," IEEE Transactions on Ultrasonics. Ferroelectrics. and Frequency Control v. UFFC-33, pp. 140-146, March 1986.
Flynn, H. G., "Generation of Transient Cavities in Liquids by Microsecond Pulses of Ultrasound," Journal of the Acoustical Society of America v. 72, pp. 1926-1932, December 1982.
,

4.

5.

Roy, R. A., Atchley, A. A., Crum, L. A., Fowlkes, J. B., and Reidy, "A Precise Technique for the Measurement of Acoustic Cavitation Thresholds and Some Preliminary Results," Journal of the Acoustical Society of America v. 78, pp. 1799-1805, November 1985.
J. J.,
,

6.

Atchley, A. A., Frizzell, L. A., and Apfel, R. E., 'Thresholds for Cavitation Produced in Water by Pulsed Ultrasound," submitted for publication in Ultrasonics
.

7.

Walton, A. in Physics

J.,
,

v.

T., "Sonoluminescence," Advances 595-660, 1984. 33, pp.

and Reynolds, G.

8.

Neppiras, E. A., and Noltingk, B. E., "The Onset of Cavitation Produced by Ultrasonics," Proceedings of the Physical Society of London Section B v. 64, pp. 1032-1038, 1951.
.

9.

Harvard University Acoustics Research Laboratory Technical Memorandum No. 12, The Onset of Cavitation in Liquids: I by F. G. Blake, Jr., p. 28, 2 September 1949.
.

10.

Rayleigh, Lord, "On the Pressure Developed in a Liquid During the Collapse of a Spherical Cavity," Philosophy Magazine v. 34, pp.
,

94-98, August 1917.

11.

M. S., "The Dynamics of Cavitation Bubbles," Journal Applied Mechanics v. 16, pp. 277-282, September 1949.
Plesset,
,

of

90

12.

Poritsky, H., 'The Collapse or Growth of a Spherical Bubble or Cavity in a Viscous Fluid," Proceedings of the First U.S. National Congress on Applied Mechanics ed. E. Sternberg (New York), pp.
,

813-821, 1952.
13.

Noltingk, B. E., and Neppiras, E. A., "Cavitation Produced by Ultrasonics," Proceedings of the Physical Society of London Section B v. 63, pp. 674-685, 1950.
.

14.

Lauterborn, W., "Numerical Investigation of Nonlinear Oscillations of Gas Bubbles in Liquids," Journal of the Acoustical Society of America v. 59, pp. 283-293, February 1976.
,

15.

Crum, L. A., "Rectified September 1984.

Diffusion," Ultrasonics

v.

22, pp. 215-223.

16.

Harvey, E. N., Barnes, D. K., McElroy, W. D., Whitely, A. H.. Pease, D. C, and Cooper, K. W., "Bubble Formation in Animals," Journal of Cellular and Comparative Physiology, v. 24, pp. 1-40, 1944.

17.

Theoretical Explanation of the Chemical Effects of Ultrasonics," Journal of Chemical Physics v. 18, pp. 997-998,
Griffing, V.,
,

1950.
18.

Sehgal,

Steer, R. P., Sutherland, R. G., and Verrall, R. E., "Sonoluminescence of Aqueous Solutions," Journal of Physical Chemistry v. 81, pp. 2518-2620, December 1977.
,

C,

19.

C, Sutherland, R. G., and Verrall, R. E., "Optical Spectra Sonoluminescence from Transient and Stable Cavitation in Water Saturated with Various Gases," Journal of Physical Chemistry v. 84, pp. 388-395, February 1980.
Sehgal,
of
,

20. Sehgal,

C, and Verrall, R. E., "A Review of the Electrical Hypoof Sonoluminescence," Ultrasonics v. 20, pp. 37-39, January 1982.
thesis
,

21.

Vaughan,

P., and Leeman, S., "Some Comments on Mechanisms of Sonoluminescence," Acustica v. 59, pp. 279-281, 1986.
.

22.

Crum,

L. A.,

and Reynolds, G.

T.,

Stable Cavitation," Journal 78, pp. 137-139, July 1985.

23.

of the Acoustical Society of

"Sonoluminescence Produced by America v.

,

Daniels, F.,

and Alberty,
Inc.,

R. A., Physical

Chemistry 4th
.

ed., p.

328,

John Wiley & Sons,

1975.

91

24.

Clay, C. S., and Medwin, H., Acoustical Oceanography: Principles and Applications pp. 184-204, John Wiley & Sons, Inc., 1977.
,

25. Kinsler, L. E.,

and others, Fundamentals 389-391, John Wiley & Sons, Inc., 1982.

of Acoustics

3rd

ed., pp.

26.

for the

Averbuch, A. J., and O'Brien, W. D., "A Primary Method Determination of Ultrasonic Intensity With the Elastic Sphere Radiometer," Acustica v. 38, pp. 58-61, 1977.

Dunn,

F.,

27.

Hill

Lindsay, R. B., Mechanical Radiation pp. 217, 292-293, Book Co., 1960.
,

McGrawManual

28.

Pennwalt Corporation,
1983.

KYNAR

Piezo Film Technical

92

INITIAL DISTRIBUTION LIST

No.
1.

Copies
2

Defense Technical Information Center

Cameron
2.

Station

Alexandria,

VA

22304-6145
2

Library Code 0142 Naval Postgraduate School Monterey, CA 93943-5000

Dr. L. E. Hargrove

3.

Office of Naval Research Physics Division - Code 1112 800 N. Quincy Street Arlington, 22217

VA

4.

Commander, Naval Coastal Systems Center Dr. Elan Moritz - Code 4120 Panama City, FL 32407-5000
Professor A. A. Atchley,

5.

Code 61 Ay

Naval Postgraduate School 93943-5000 Monterey, CA

6.

Professor S. R. Baker,

Code 61Ba

Naval Postgraduate School 93943-5000 Monterey, CA

7.

Professor S. L. Garrett,

Code 61Gx

Naval Postgraduate School 93943-5000 Monterey, CA

8.

Chairman, Department of Physics, Code 61 Naval Postgraduate School Monterey, CA 93943-5000

9.

LCDR

R. L. Bruce

6002 Crocus Court Alexandria, VA 22310

10.

LT
141

R. D. Middleton, John Street

Jr.

Hauppauge,

NY

11787

93

1 1

and Aaron Bruce 213 Brookdale Drive League City, TX 77573

Amy

12.

CAPT

Robert Perron

Department of Physics Naval Postgraduate School Monterey, CA 93943-5000

13.

Mr. and Mrs. R. D. Middleton, 141 John Street

Sr.

Hauppauge,
14.

NY

11787

Maris R. Zanger

300 Glenwood Circle #164 Monterey, CA 93940

94

Thesis B82475
c.l

Bruce An investigation of acoustic cavitation produced by pulsed ultrasound.

Thesis B82475
c.l

Bruce An investigation of acoustic cavitation produced by pulsed ultrasound.