phononic crystals NATOASI2002 - 2002... · 2012-04-19 · Phononic Crystals J.H. Page University of...
Transcript of phononic crystals NATOASI2002 - 2002... · 2012-04-19 · Phononic Crystals J.H. Page University of...
Phononic CrystalsJ.H. Page
University of ManitobawithSuxia Yang and M.L. Cowan at U of M, Ping Sheng and C.T. Chan at HKUST, &Zhengyou Liu at Wuhan University.
We study ultrasonic waves in complex (strongly scattering) media.
Ordered media (phononic crystals):- ultrasound tunneling- focusing effects
Random media: - ballistic and diffusive wave transport- new ultrasound scattering techniques (DSS & DAWS)
For more info and papers, see www.physics.umanitoba.ca/~jhpage
Outline: Phononic Crystals
Motivation: Why study phononic crystals? (acoustic and elastic analoguesof photonic crystals):
Possible applications- lenses and filters?- sound insulation?
Conclusions
Wave phenomena in 3D phononic crystals (experiments and theory) : - our crystals and the experimental setup- spectral gaps - ultrasound tunneling- near field imaging- focusing effects
-4
-2
0
2
4-4
-2
0
2
4
y (cm
)
x (cm)
PHONONIC CRYSTALSe.g. ordered arrays of mm-sized beads in a liquid or solid matrix �acoustic and elastic analogues of photonic crystals.
Why study phononic crystals?• richer physics?
•scattering contrast � � may be easier to achievecomplete spectral gaps•longitudinal + transverse modes � novel wave phenomena?
• ultrasonic techniques have some advantages•measure the field not the intensity•pulsed techniques are easy
• good theory is available - e.g. Multiple Scattering Theory for elastic andacoustic waves. [Kafesaki and Economou, PRB 60, 11993 (1999); Liu et al. PRB 62,2446 (2000); Psarobas et al., Phys. Rev. B 62, 278 (2000)]
• relatively few experiments have been performed on 3D systems (unlike2D).•new applications?
�� �v,
Questions:
• Can complete spectral gaps be readily achieved?
• How do waves travel through phononic band gaps? No propagating mode � Tunneling?
How long does it take?
• Can phononic crystals be used to focus ultrasound?
Our 3D Phononic Crystals:
Close-packed periodic arrays ofspherical beads surrounded by a liquidor solid matrix:
• hcp arrays of stainless steel balls inwater
• fcc arrays of tungsten carbide balls inwater
• fcc arrays of tungsten carbide beadsin epoxy
For all these crystals:
• very high contrast scatterers Zball/Zmatrix � 30 to 60
• very monodisperse spheres: diameter d = 0.800 � 0.0006 mm
• very high quality crystals: produced by a manual assembly techniqueusing a hexagonal template.
for hcp: ABAB… layer sequence. (slabs with c-axis � layers)
for fcc: ABCABC… layer sequence. ( [111]-axis � layers)
Schematic graph shows how the beads are packed in a FCC crystal along the [111] direction.Beads A are on the bottom layer, B the second and C the third. The sequence is ABCABC...
A
B B B B
B B BC
C
C C
C C
C
Matrix:
water
Lv = 1.49 km/s
� = 1.0 kg/m3
Scatterers:0.800-mm-diametertungsten carbide beads
LvTv
= 6.65 km/s= 3.23 km/s
� = 13.8 kg/m3
Material parameters forour phononic crystalsmade from tungstencarbide beads in water
Sample
SampleHolder
Ultrasound GeneratingTransducer
Ultrasound DetectingTransducer
Close-up of sample
y1
y2L
phase velocity:vp(�) = � L/(� + 2�n)(� = phase delay)
group velocity: vg(�) = L/tpeak
(tpeak = peak delay time)
transmissioncoefficient: Atrans(�) / Aref(FFT ratio)
Experimental setup
Planar input pulse(far field; large y1)
Compare transmittedpulses with referencepulses to measure:
Multiple Scattering TheoryMultiple scattering theory (MST) for acoustic and elastic waves -- ideally suitedto spherical scatterers. (cf. KKR theory) [Liu et al. PRB 62, 2446 (2000) ]
Band-structure: calculate elastic Mie scattering of waves from all the scatterers inthe crystal, and solve the resulting secular equation for the eigenfrequencies.
Transmission:calculate thetransmission througha multi-layer samplewith thickness Lusing a layer MST.
2(2)IQ
(2)IVQ
(2)IIIQ
(2)IIQ
1(1)IQ
(1)IVQ
(1)IIIQ
(1)IIQ
12(12)IQ
(12)IVQ
(12)IIIQ
(12)IIQ
( , ) ( , ) exp[ ( , )]T L A L i L� � � ��
A normalized amplitude (transmission coefficient) � cumulative phase relative to the incident wave.
Transmitted wave
( ) /
( )
p
g
L
d dLdk d
� � �
� ��
�
�
� �
v
v
Phase velocity
Group velocity
�
0.0
0.2
0.4
0.6
0.8
1.0
KHLA�MK
� a
/ 2�
vw
hcp bandstructure: stainless steel spheres in water
a lattice constantsound velocity in water.wv
WAVEVECTOR
Very small complete gap
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
WAVE VECTORKWX�LUX
� a
/2�
vw
Fcc bandstructure: tunsgsten carbide beads in water
a lattice constantsound velocity in water.wv Reciprocal
latticeComplete gap: centre 19 %� �� �
Fcc bandstructure and transmission: tunsgsten carbide beads in epoxy
An even biggercomplete gap!
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
WAVE VECTORUXX U L �
FREQ
UEN
CY (M
Hz)
0.1 0.011E-31E-41E-51E-61E-71E-80.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.57-Layer Sample
K
TRANSMISSION COEFFICIENT
centre 90 %� �� �
0-1.0
-0.5
0.0
0.5
1.0
NO
RMA
LIZE
D F
IELD
0 10 20 30 40 50
-0.08
-0.04
0.00
0.04
0.086-layer sample
TRANSMITTED PULSEINPUT PULSE
TIME (�s)
FCC tungsten carbide beads in water:Input and transmitted pulses near the lowest band gap, k [111]�
0.8 1.20.0
0.2
0.4
0.6
0.8
1.0 INPUT
NO
RMA
LIZE
D A
MPL
ITU
DE
FREQUENCY (MHz)0.8 1.2
0.00
0.04
0.08
0.12
0.16TRANSMITTED
FCC tungsten carbide beads in water:Fourier Spectrum of these input and transmitted pulses near the
lowest band gap0
-1.0
-0.5
0.0
0.5
1.0
NO
RMA
LIZE
D F
IELD
0 10 20 30 40 50
-0.08
-0.04
0.00
0.04
0.086-layer sample
TRANSMITTED PULSEINPUT PULSE
TIME (�s)
0.6 0.8 1.0 1.2 1.4
1E-3
0.01
0.1
1
6 LAYER: DATA THEORY
12 LAYER: DATA THEORY
AM
PLIT
UD
E TR
AN
SMIS
SIO
N
COEF
FICI
ENT
FREQUENCY (MHz)
Transmission coefficient
0.2 0.4 0.6 0.8 1.0 1.2 1.41.0
1.2
1.4
1.6 EXPERIMENT THEORY
PHA
SE V
ELO
CIT
Y (k
m/s
)
FREQUENCY (MHz)
0
20
40
60 EXPERIMENT THEORY
CU
MU
LATI
VE
PHA
SE (r
ad)
Cumulative Phaseand the
Phase Velocity
Use these data tocompute thedispersion curve,� versus k = � /vp,and compare withthe bandstructurecalculation…
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
WAVE VECTORKWX�LUX
� a
/2�
vw
Fcc bandstructure: tunsgsten carbide beads in waterThe black symbols show experimental data from phase velocity measurements
a lattice constantsound velocity in water.wv Reciprocal
latticeComplete gap: centre 19 %� �� �
-30 -20 -10 0 10 20 30-1
0
1
x1796
THROUGH SAMPLE
TIME (�s)
-1
0
1
x1
THROUGH WATER
NO
RM
ALIZ
ED F
IELD -1
0
1 INPUT
Measuring the group velocity: Digitally filtered input andtransmitted pulses (bandwidth 0.05 MHz) for a 12-layer phononic crystalin the middle of the gap, compared with the pulse transmitted through thesame thickness of water.
Frequency dependence of the group velocity for a 5-layer sample.
0.6 0.8 1.0 1.2 1.40
2
4
6
8 THEORY FIT EXPERIMENT
GR
OU
P VE
LOC
ITY
(km
/s)
FREQUENCY (MHz)
0 2 4 6 80
4
8
12
16
1D Theory
vwater = 1.49
Group Velocity Theory Data
vbead = 6.655
VEL
OCI
TY (k
m/s)
SAMPLE THICKNESS (mm)
Group Velocity versus sample thickness in the middle of the gap
The group velocity increases linearly with thickness! � tunneling.Yang et al., Phys. Rev. Lett. 88, 104301 (2002)
Effect of absorption - the two-modes modelSimple physical picture:
Absorption cuts off the long scattering paths � destructive interference isincomplete.
Effectively, absorption introduces a small additional component having a realwave vector inside the gap.
Two-modes model: (Dominant) Tunneling mode (constant tunneling time) + Propagation mode (constant velocity)
� Group velocity becomes
222
2 2 2 22 2
1
1 1
LLll ba
L L L Ll l l la b a b
ga eae
aae a e ae a e b
LLt
��
� � � �
�
� � � �
�
�
v
vwhere L is the thickness, a is the coupling coefficient, ta and la are thetunneling time and tunneling length, vb is the group velocity of the propagatingmode and lb is its decay length.
0 2 4 6 80
4
8
12
16
1D Theory
vwater
Group Velocity Theory Data
vbead=6.655
VEL
OCI
TY (k
m/s)
SAMPLE THICKNESS (mm)
The effect of absorption: Fits of the two-modes model to 1 D theory and toour data for 3 D phononic crystals (dashed curves).
0 2 4 6 80
4
8
12
16
Phase Velocity Theory Data
1D Theory
vwater
Group Velocity Theory Data
vbead=6.655
VEL
OCI
TY (k
m/s)
SAMPLE THICKNESS (mm)
Add: the phase velocity in the middle of the gap for the same sample thicknesses.
Transmitted waveIncident
wave
V0
Electron tunneling through a barrier
0
0 00
0
xV V x L
x L
���
� � ��� ��
2 2
2 ( ) ( )2
d V E xm dx
��� � � �
�
From the transmitted phase at large large L, find
� �
� �0
0
2arctan
2p
VL
V�
�
� �
� ��� ��
�� �� �
��
� �
v 0g
VL � �� �
� �� �� �
��
v
Barrier: Both phase and group velocities increase with L.
Band gap: Group velocity increases while the phase velocity is constant.
Schrodinger equ.
Boundary conditions
(0 ) (0 ); (0 ) (0 )
( ) ( ); ( ) ( )L L L L
� � � �
� � � �
� � � �
� � � �
� �� �
� �� �
10
1
86421.5
Experiment
Theory
Phononic crystals
Tungsten carbide
water
TRA
NSI
T TI
ME
(�s)
SAMPLE THICKNESS (mm)
Magnitude of the tunneling time: ttunnel ~ 1 / ��gap
Tunneling Time
(also true for lightand electrons!)
Question: Is ultrasound tunneling dispersionless?
Some general points about pulse propagation in dispersive media:
� � � � � �o0, expx t e t i t� �� �
� � � � � �1o, exp exp
2 e
xx t E ik xl
� � � ��
� �� �� �� � � � �� �� � �
� �
� � � � � �2
21o o o2! 2
o o
dk d kk kd d
� � � � �
� �
� � � � � ��
Incident pulse at x = 0 � Fourier transform:
pulse envelope E(�) is F.T. of e(t)
Pulse after traveling a distance x
extinction length, 1 1 1e s al l l� � �
� �
scattering absorption
Expand the wave vector k(�) in a Taylor series about the central frequency�o:
group velocity dispersion (GVD)
� � � � � �ot E� � � �� �� � � � �� �
inverse group velocity
� � � �
� � � � � �� � � � � �� �
� � � � � � � �� � � �� �
o
21 2 21o o o2o o
1 2 2 21o o 2o o
o
, exp exp2
exp exp
exp exp exp exp exp2
exp exp2
e
e
e p g
xx t ik xl
E i dk d x i d k d x
x ik x i t E i dk d x i d k d xl
x x xi t tl
�
� � � � � � � �
� � � � � �
� �
�
�
� �� � � �� �
�
�� � � � � � �� �
� � �� � � � � �� � �� ��
�� � �� � � �� � � � �� � � �
� �� �� �� �
�
v v� � � �� �1 2 2 21
2 oexpe t i d k d x� ��
� ��� � �� �� �� �� �� � �� �� ���
Pulse at (x,t) becomes
attenuation
carrier frequency oscillationstravel at vp = �o/ko
peak travels at the group velocityof the central frequency
GVD changes the width symmetrically
convolution
Good approximation, so long as the pulse bandwidth is sufficiently narrow that:• higher order terms in k(�) expansion are negligible• frequency dependence of le is not important.
0.7 0.8 0.9 1.0 1.1 1.2 1.30
5
10
f (MHz)
0
2
Data Theory
-20
0
20
d2 k/d�
2dk
/d�
k
Frequency dependence of the wave vector, the inverse group velocity and the groupvelocity dispersion near the band gap in a 5 layer tungsten carbide/water crystal.
[c.f. measurements near the band edge in photonic crystals by Imhof et al. PRL 83, 2942 (1999)]
-20 -10 0 10 20 300
1 Shifted Reference Sample6.14
f = 0.744 MHz
TIME (�s)
0
1 Shifted Reference Sample
f = 0.972 MHz1.0
NO
RMA
LIZE
D A
MPL
ITU
DE
0
1 Shifted Reference Sample
f = 1.2 MHz4.9
Pulse broadening (a) just above the band gap, (b) in the middle of the gap and(c) just below the band gap (bandwidth 0.02 MHz). � Very weak dispersionin the gap, despite very strong scattering and a large variation in vp with �.