torsional oscillator experiments on helium films on ...science-visits.mccme.ru/doc/nyeki.pdf · 2d...
TRANSCRIPT
-
Ján Nyéki, Anastasia Phillis,
Jeevak Parpia*, Brian Cowan
and John Saunders
29/09/10, IPP Moscow
Torsional Oscillator Experiments on
Helium Films on Graphite;
The Search for a Two Dimensional
Supersolid
Department of Physics, Royal Holloway University of London, Egham, UK
*LASSP, Department of Physics, Clark Hall, Cornell University, Ithaca, USA
-
Why Graphite as Substrate ?
Joly et al, Surf. Sci. 264 (1992) 419
•Flat - minimalised disorder
•Helium film grows in layers.
•We can control the growth.
•Two solid He layers solidify
•Triangular lattice symmetry
•Cooling to submillikelvinrange
Adsorption potential
+ Periodic potentialarising from atomic structure of surface
Ground state
-
2D Helium on Graphite
• Strongly correlated fermions or bosons• Subject to an external triangular lattice potential•Tuning parameters: interatomic distance
binding potential
• No impurities
Experimental techniques
3He: NMR, heat capacity, neutron diffraction,
vapour pressure
4He: heat capacity, torsional oscillator,
neutron diffraction, third sound, vapour pressure
Temperature range ~ 15 µK to 15 K
QuestsSuperfluidity of a 3He monolayer
2D superfluid-insulator transition
2D supersolid
-
Exfoliated Graphite
Taub et al., PRB 16 (1977) 4551
Fukuyama, JPSJ 77 (2008) 111013
Grafoil ~ 20 m2/g
Grafoil – atomically flat platelets 10-20 nm long, rolled, angle spread ±15°- surface area ~ 20 m2/g - density ~ 50% of bulk- “deep sides”, edges 3-5%- possible to cool below 1mK
-
Layered Helium Films
Registered solid
(commensurate phasesuperlattice)
√3 ×√3 6.4 nm-2
3He/Gr
4/7 phase
Fukuyama, JPSJ 77 (2008) 111013
-
Sibling system: 4/7 Phase in 3He films
Elser, PRL62 (1989) 2405Takagi, J Phys Conf Ser 150 (2009) 032102
7
4
densitylayer 1st
densitylayer 2nd
1
2 ==ρρ
Existence of 4/7 solid is clearly identified
both experimentally and theoretically
•Frustrated spin ½ magnet•Mott insulator•Gapless spin-liquid ground state
-
Motivation I
Can we study this phenomenonin 2D helium-4 ?
(alternative system of bosons in externally imposed lattice potential)
Revived theoretical interest
Heidarian & Damle, PRL 95, 127206 (2005),
Melko et al, PRL 95, 127207 (2005),
Bonisegni & Prokof‘ev, PRL 95, 237204 (2005) + ...
Wessel and Troyer, PRL 95, 127205 (2005)
Greiner et al, Nature 415, Jan 2002
Stoof, ibid, N&V
Triangular lattice symmetry supports formation of supersolid
-
Searching for Supersolid
Superfluid response Broken translational symmetry+
-
4He/Gr 2nd Layer - Phase Diagram
Pierce& Manusakis, PRL 81, 156 (1997) and PRB 59, 3802 (1999)
1st layersolid
Commensurate4/7 phase
Total coverage on PM scale [ nm-2 ]
11 12 13 14 15 16 17 18 19 20 21 22
L CL + G IC
vacancies ?
Puddling
IC phase
-
4/7 Phase in 4He films
Corboz et al., PRB78 (2008) 245414
Total 4He coverage on PM scale [ nm
-2]
15 16 17 18 19 20 21 22 23
T [
mK
]
0
200
400
600
800
1000
1200
1400
1600
1800
L CL + G IC
Total coverage [ nm-2 ]
12 13 14 15 16 17 18 19 20 21 22
C [
mJ/K
]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
5mK 10mK 15mK 20mK 30mK 40mK 50mK 60mK
L C ICL + G
Ziouzia et al.: Physica B (2003) & PhD Thesis
3He tracer
Greywall, Busch: PRL 67 (1991) 3535
IC
C
-
Motivation IICrowell & Reppy, PRB 53, 2701 (1993)
superfluid response in 2nd layer
20 mK – 600 mK
-lnT like dependence
no dissipation peaks reported
Total 4He coverage on PM scale [ nm
-2]
15 16 17 18 19 20 21 22 23
T [
mK
]
0
200
400
600
800
1000
1200
1400
1600
1800
L CL + G IC
3D “Supersolidity”
Sample annealing(disorder) and impurities playimportant role
Chan, Science 319, 1207 (2008)
-
Torsional Oscillator
E.L.Andronikashvili, Zh.Eksp.Teor.Fiz. 16 (1946) 780
Andronikashvili’s experimentBulk liquid 4He
ρn/ρ
ρs/ρBishop & Reppy, PRB 22, 2701 (1980)
-
Torsional Oscillator Signal
Kosterlitz-Thouless theory
Resonant frequency
12
22)(
h
mk
T
TB
c
cs
πρ
=−
I
k=ω
Bishop & Reppy, PRB 22, 2701 (1980)
AmplitudeDrive=const ~ Q
≈
∆
Q
1Dissipated Energy
Liquid 4He films
-
Experimental Cell: Torsional Oscillator
torsion element 1: contains exfoliated graphite sample ~ 35 m2
torsion rod
torsion element 2: coin silver cylinder, with two Mg electrodes biased at 100V
torsion rod
isolation mass + torsion rod (copper) mounted to cold-plate
Measure frequency shift and dissipationquality factor Q ~ 250000 - 550000working frequency ~ 1423 Hz; detecting changes < 10 µHz ~ 2x10-7g
2 mK - 4000 mK (weak thermal linkto the copper demagnetisation stage)
-
L+G Region
Temperature [ K ]
0.0001 0.01 0.1 1
Fre
qu
en
cy
- 1
42
3 H
z
[ m
Hz ]
-208
-206
-204
-202
-200
-198
-196
13.25 nm-2
15.42 nm-2
16.35 nm-2
17.31 nm-2
17.57 nm-2
17.91 nm-2
18.14 nm-2
18.29 nm-2
18.45 nm-2
18.57 nm-2
18.69 nm-2
Temperature [ K ]
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
10
8 x
∆∆ ∆∆(1
/Q)
0
10
20
30
40
50
60
70
80
90
Total coverage on PM scale [ nm-2 ]
10 11 12 13 14 15 16 17 18 19 20 21 22
L CL + G IC1st
layer
Dissipation maximum
at constant temperature
in L + G region;
moves to lower T
in L region
10 11 12 13 14 15 16 17 18 19
T [
mK
]
0
300
600
900
1200
L + G
1st layer
∆∆∆∆f
∆∆∆∆(1/Q)
-
Transition L+G ���� L
Temperature [ K ]
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
10
8 x
∆∆ ∆∆(1
/Q)
0
10
20
30
40
50
60
70
80
90
Total coverage on PM scale [ nm-2 ]
10 11 12 13 14 15 16 17 18 19 20 21 22
L CL + G IC1st
layer
18.14 nm-2
chosen as a reference to
evaluate additional decoupling
on top of the slip process
Temperature [ K ]
0.0001 0.01 0.1
Fre
qu
en
cy -
14
23
Hz
[ m
Hz ]
-200
-199
-198
-197
-196
13.25 nm-2
15.42 nm-2
16.35 nm-2
17.31 nm-2
17.57 nm-2
17.91 nm-2
18.14 nm-2
18.29 nm-2
18.45 nm-2
18.57 nm-2
18.69 nm-2
Extra decoupling = superfluid
For n ≥ 18.14 nm-2
0.001
∆∆∆∆(1/Q)
∆∆∆∆f
-
Homogeneous Fluid Region
Temperature [ K ]
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
10
8 x
∆∆ ∆∆(1
/Q)
0
10
20
30
40
50
60
70
80
90
Total coverage on PM scale [ nm-2 ]
10 11 12 13 14 15 16 17 18 19 20 21 22
L CL + G IC1st
layer
New peak appears
&
sharp increase of ∆∆∆∆f
for n4 > 18.85 nm-2
13 14 15 16 17 18 19 20
T [
mK
]
0
100
200
300
400
LL + G
Temperature [ K ]
0.0001 0.01 0.1 1
∆∆ ∆∆f
[ m
Hz ]
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
18.14 nm-2
18.29 nm-2
18.45 nm-2
18.57 nm-2
18.69 nm-2
18.85 nm-2
19.01 nm-2
19.09 nm-2
19.18 nm-2
0.001
∆∆∆∆(1/Q)
∆∆∆∆f
18.14 nm-2
-
Super- response around C phase
Total coverage on PM scale [ nm-2 ]
10 11 12 13 14 15 16 17 18 19 20 21 22
L CL + G IC1st
layer
Temperature [ K ]
0.0001 0.01 0.1 1
∆∆ ∆∆f
[ m
Hz ]
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
19.18 nm-2
19.26 nm-2
19.34 nm-2
19.42 nm-2
19.51 nm-2
19.60 nm-2
19.69 nm-2
19.78 nm-2
19.87 nm-2
19.96 nm-2
20.05 nm-2
20.16 nm-2
20.28 nm-2
20.40 nm-2
20.52 nm-2
20.64 nm-2
20.74 nm-2
Temperature [ K ]
0.0001 0.01 0.1 1
10
8 x
∆∆ ∆∆(1
/Q)
-10
0
10
20
30
40
50
60
70
80
90
0.001 0.001
∆∆∆∆(1/Q)∆∆∆∆f
-
Temperature [ K ]
0.00 0.05 0.10 0.15 0.20 0.25 0.30
∆∆ ∆∆f
[ m
Hz ]
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Super- response around C phase
Total coverage on PM scale [ nm-2 ]
10 11 12 13 14 15 16 17 18 19 20 21 22
L CL + G IC1st
layer
19.18 nm-2
19.26 nm-2
19.34 nm-2
19.42 nm-2
19.51 nm-2
19.60 nm-2
19.69 nm-2
19.78 nm-2
19.87 nm-2
19.96 nm-2
20.05 nm-2
20.16 nm-2
20.28 nm-2
20.40 nm-2
20.52 nm-2
20.64 nm-2
20.74 nm-2
after full composite background subtraction
∆f ~ ρs( *)( *) 1
(0) (0) 2
s
s
Tf T
f
ρρ
∆= =
∆
-
Super- Response Overwiev
Total coverage on PM scale [ nm-2
]
15 16 17 18 19 20 21 22
∆∆ ∆∆f(
0)
[ m
Hz ]
0
1
2
3
T
[ m
K ]
0
400
800
1200
1600
L CL + G IC
Not a simple structureless superfluid
-
Total coverage on PM scale [ nm-2
]
19.0 19.5 20.0 20.5 21.0
∆∆ ∆∆f(
0)
[ m
Hz ]
0
1
2
3
4
Superlattice density is a “pivotal” point in
coverage dependence of super- response
Scaling
C
T/T*
0 1 2 3 4 5 6 7 8 9 10
∆∆ ∆∆f(
T)/∆∆ ∆∆
f(0)
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
T/T*
0 1 2 3 4 5 6 7 8 9 10
∆∆ ∆∆f(
T)/∆∆ ∆∆
f(0)
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
ρρρρs(T)/ρρρρs(0)vs
T/T*
ρρρρs(T)/ρρρρs(0)vs
T/T*( *)( *) 1
(0) (0) 2
s
s
Tf T
f
ρρ
∆= =
∆
-
T/T*
0.1 1 10
∆∆ ∆∆f(
T)/∆∆ ∆∆
f(0
)
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
T/T*
0.1 1 10
∆∆ ∆∆f(
T)/∆∆ ∆∆
f(0
)
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Scaling
Jason Ho, private communication, 2010
α
ρρ
)/(1
)0()(
0TTT ss +
=
n4 on PM scale [ nm
-2 ]
19.0 19.5 20.0 20.5 21.0
T*
[ m
K ]
0
10
20
30
40
50
60
*TT
α~1.0
≤
*4TT 0.4T
α~1.0
≤≤*
-
Super- Fraction
Temperature [ K ]
0.0001 0.01 0.1 1
∆∆ ∆∆f
[ m
Hz ]
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
19.18 nm-2
19.26 nm-2
19.34 nm-2
19.42 nm-2
19.51 nm-2
19.60 nm-2
19.69 nm-2
19.78 nm-2
3rd layer
What is the value of super- fraction?
Compare with measurements of a KT superfluid monolayer in the same cell (3rd layer, n3 ~ 5.5 nm
-2)
Super- fraction at T����0:
~ 39% of 2nd layer for 19.18 nm-2 sample
~ 27% of 2nd layer for 19.78 nm-2 sample
ΧΧΧΧ-factor for this cell = 0.942
Measurements at perimeter velocity 25 ÷÷÷÷ 55 µµµµm/s
checked up to ~ 500 µµµµm/s0.001
-
Super- response from superlattice?
These number fluctuations are intrinsic to the 4/7 superlattice structure[Watanabe, Imada, JPSJ 76 2007]. Fluctuations into the 3rd layer.
Zero-point vacancies [Andreev & Lifshitz] appear in the superlattice when ∆> 2UV .
In 2D , ∆ for 3He impuriton is much larger than in the bulk [Ziouzia, PhD Thesis, London ]
The 2nd layer superlattice phase is unstable. [Corboz et al., PRB 78 2008]
Key: since ∆N ∆φ ≥ 1, number fluctuations required to define phase.[For this reason a commensurate solid cannot be a supersolid]
g(E)
E
∆
-
Monte Carlo simulations
2nd
layer coverage n2 [ nm
-2 ]
3 4 5 6 7 8
∆∆ ∆∆f(
0)
[ m
Hz ]
0
1
2
3
4
G + L SF IC
Total coverage on PM scale [ nm-2
]
16 17 18 19 20 21
∆∆ ∆∆f(
0)
[ m
Hz ]
0
1
2
3
4
L CL + G IC
Corboz et al., PRB78 (2008) 245414
Pierce& Manousakis, PRL 81, 156 (1997)
and PRB 59, 3802 (1999)
Two Monte Carlo simulations
of the 2nd layer disagree
about stability of the 4/7
superlattice solid phase
stable 4/7 phase
superfluid phase only
-
In the 4/7 phase
a ~ 4.1 Ǻ � vc~ 3.9 m/s
Momentum deficit in Quantum Glass
Reduced rotational inertia due to presence of tunelling two-level systems.
If “bulk” velocity v(t) ≠0 the TLS has nonzero mean values of momenta
12 in both stationary states. In the ground state the projection of
1 onto v is negative.
No superflow
[Andreev, JETP Lett 85 2007, ZhETF 136 2009, Korshunov Pis’ma v ZhETF 90 2009 ]
U - characteristic height of
energy barrier
J - tunneling amplitude
∆∆∆∆ ≈ 1 mK ?Temperature [ K ]
0.0001 0.01 0.1 1
∆∆ ∆∆f
[ m
Hz ]
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
-
Summary
Motivation: Does vacancy doped commensurate phase √7 x √7(4/7 phase) have a supersolid response?
•High precision TO measurements from 2mK to 4K. Second layer density is the tuning parameter.
•1st layer interaction with the substrate present all time ���� careful measurements can distinguish that from ‘super-’ response; sensitive to
2nd layer structure
•Unusual temperature dependence of momentum deficit; no sharp onset:∆f(T) ~ -lnT near onset while ∆f(T) ~ T as T ���� 0.
•∆f(0) is tuned by n2. Plateau-feature when n ≈ n0 (4/7 ) and vanishes near at incommensurate 2nd layer density nI. Possible QCP at nI.
•More than 25% of the 2nd layer contributes to super- signal when approaching n0.
-
Summary II
Motivation: Does vacancy doped commensurate phase √7 x √7(4/7 phase) have a supersolid response?
• After “pre-cursor” feature plot of ∆f(0) vs T* shows two clear regimes. Boundary correlates with qualitative change in dependence of ∆f on lnT
• For a significant coverage range there is data collapse in ∆f(T)/∆f(0) vsT/T* coordinates.
• Quantum glass model qualitatively agrees with observed ∆f(T) ~ -lnTdependence. Energy scales involved need to be addressed.
•We need experimental evidence on the structure of putative 4/7 phase in the T ���� 0 limit.