Energy transfer at the nanoscale: diodes and pumps Dvira Segal - - PowerPoint PPT Presentation

energy transfer at the nanoscale diodes and pumps
SMART_READER_LITE
LIVE PREVIEW

Energy transfer at the nanoscale: diodes and pumps Dvira Segal - - PowerPoint PPT Presentation

Energy transfer at the nanoscale: diodes and pumps Dvira Segal Chemical Physics Theory Group University of Toronto J Motivation T Quantum open systems out of equilibrium: Transport and dissipation. Quantum energy flow: Heat


slide-1
SLIDE 1

Energy transfer at the nanoscale: diodes and pumps

Dvira Segal Chemical Physics Theory Group University of Toronto

slide-2
SLIDE 2

Motivation

  • Quantum open systems out of equilibrium: Transport and

dissipation.

  • Quantum energy flow: Heat conduction in bosonic/fermionic systems.
  • Nonlinear transport: diode, NDC
  • Control: Pumping of heat
  • Nanodevices: Understand and manipulate heat transfer in molecular

systems and nanoscale objects.

J ΔT J

ΔT

slide-3
SLIDE 3

Outline

  • I. Motivation
  • II. Models for studying the fundamentals of quantum heat flow.
  • III. Static case: Nonlinear effects

Thermal rectification-diode

  • 1. Experiment
  • 2. Formalism
  • 3. Sufficient conditions for thermal rectification
  • IV. Dynamic case: Active control

Stochastic heat pumps

  • 1. Mechanism
  • 2. Formalism
  • 3. Examples: Control of the noise properties/ solid characteristics.
  • 4. Efficiency: Approaching the Carnot limit
  • V. Summary and Outlook
slide-4
SLIDE 4

Quantum energy flow Vibrational heat flow Photonic heat conduction Electronic energy transfer

Introduction/Motivation

slide-5
SLIDE 5

Vibrational energy flow in molecules

S S

Molecular electronics

Heating in nanojunctions.

Fourier law in 1 D.

  • C. Van den Broeck, PRL (2006).

Nanomachines IVR

carbon nanotubes

slide-6
SLIDE 6

Phonon mediated energy transfer

  • G. Schultze et al. PRL 100,

136801 (2008)

STM tip Adsorbed molecules Metal

  • Z. Wang, et al., Science 317, 787 (2007)

Strong laser pulse gives rise to strong increase of the electronic temperature at the bottom metal surface. Energy transfers from the hot electrons to adsorbed molecule.

J TR TL

slide-7
SLIDE 7

Single mode heat conduction by photons

  • D. R. Schmidt et al., PRL 93, 045901

(2004). Experiment: M. Meschke et al., Nature 444, 187 (2006).

2

The electromagnetic power (blackbody radiation) flowing in the device is given by: ( ) ( ) coupling coefficient 4 ( )

e B B e e

P r n n d R R r R R

γ γ γ γ

ω ω ω ω

⎡ ⎤ = − ⎣ ⎦ = +

slide-8
SLIDE 8

Exchange of information

  • K. Schwab Nature 444, 161 (2006)

Radiation of thermal voltage noise

The quantum thermal conductance is universal, independent of the nature

  • f the material and the

particles that carry the heat (electrons, phonons, photons) .

2 2

3

B Q

k T G h π =

slide-9
SLIDE 9

Electronic energy transfer

Coherence EET in poly- conjugated polymers

(Collini and Scholes Science 323, 369 (2009) ).

The lines show the characteristic anticorrelation theoretically predicted for

  • scillations caused by

electronic coherences.

slide-10
SLIDE 10
  • II. Models: Energy flow in hybrid systems

S L R L R

H H H H V V = + + + +

S n n

H E n n =∑

µR

collection of phonons; electron-hole excitations; spins. Hν

, , n m n m

V F S n m

ν ν

=

slide-11
SLIDE 11
  • 1. Harmonic system

J TR TL

( )( )

† 0 0 † , , † † , , , S L R L R S k k k k k k k k

H H H H V V H b b H b b V b b b b

ν ν ν ν ν ν ν

ω ω λ = + + + + = = = + +

∑ ∑

.

slide-12
SLIDE 12
  • 2. Two Level System

J TR TL

( )

† , , † , , ,

2 ; =

S L R L R S z k k k k x k k k k

H H H H V V B H H b b V F F b b

ν ν ν ν ν ν ν ν ν

σ ω σ λ = + + + + = = = +

∑ ∑

slide-13
SLIDE 13

J TR ; µR TL ; µL

  • 3. Energy transfer between metals

S L R L R

H H H H V V = + + + +

† 0 0 † , , † , ; , ' , , ' , ' S k k k k k k k k k k

H b b H c c V c c S

ν ν ν ν ν ν ν ν

ω ε λ = = =

∑ ∑

No charge transfer

slide-14
SLIDE 14
  • III. Static Case: Nonlinear effects

( )

n n a n

J T T α = Δ

1 2 3

lim / Conductance ( ) ( ) Thermal rectification ( ) / 0 Negative differential

T

J T J T J T J T T α α α

Δ →

= Δ → Δ ≠ −Δ ≠ < → ∂ Δ ∂Δ < thermal conductance

TL TR

;

a L R L R

T T T T T T = + Δ = −

slide-15
SLIDE 15

( ) ( ) ( )

L R B B

J n n d ω ω ω ω ω ⎡ ⎤ = − ⎣ ⎦

∫T

  • D. Segal, A. Nitzan, P.

Hanggi, JCP (2003).

Harmonic model

slide-16
SLIDE 16

2

( ) ( ) ( ) Thermal rectification

n n a n

J T T J T J T α α = Δ → Δ ≠ −Δ ≠

Thermal rectification

Reed 1997 Electrical rectifier

Asymmetry + Anharmonicity Thermal Rectification

  • M. Terraneo, M. Peyrard, G. Casati, PRL (2002);
  • B. W. Li, L. Wang, G. Casati, PRL (2004);
  • D. Segal and A. Nitzan, PRL (2005),JCP (2005).
  • B. B. Hu, L. Yang, Y. Zhang, PRL (2006)
  • G. Casati, C. Mejia-Monasterio, and T. Prosen, PRL (2007)
  • N. Yang, N. Li, L. Wang, and B. Li, PRB (2007)
  • N. Zeng and J.-S. Wang, PRB (2008)
slide-17
SLIDE 17
  • C. W. Chang, D. Okawa, A. Majumdar,
  • A. Zettl, Science 314, 1121 (2006).

C9H16Pt Non uniform axial mass distribution

sensor heater

Experiment: thermal rectifier

slide-18
SLIDE 18

S L R L R

H H H H V V = + + + +

( )

2 2

cos(2 ) 2 2

S

p V H x m π π = −

( )

2 , 2 , , 1 , 2

1 ( ) cos(2 ) 2 2 2

L i L L L L i L i L i i i

p V H k x x x m π π

+

= + − −

2 int ,

( ) 2

L L L N

k V x x = −

Simulations B. W. Li, L. Wang, G. Casati, PRL (2004)

slide-19
SLIDE 19

Formalism: Master Equation

[ ]

( )

Heat current: Tr , 2

S

i J H H V

ν ν ν ρ

= −

, ,

Model: ; ; collection of phonons; electron-hole excitations; spins.

S L R S n n L R n m n m

H H H H V H E n n V V V V F S n m F B H

ν ν ν ν ν ν

λ = + + + = = + = =

∑ ∑

[ ]

, , ,

Dynamics: Liouville equation in the interacation picture [ ( ), (0)] ( ), ( ), ( )

t m n m n m n

d i V t d V t V dt ρ ρ τ τ ρ τ ⎡ ⎤ = − − ⎣ ⎦

slide-20
SLIDE 20

Formalism: Master Equation

Liouville Equation Pauli Master equation

2 2 , , , ,

( ) ( ) ( ) ( ) ( )

n n m m m n n n m n m m m

P t S P t k T P t S k T

ν ν ν ν ν ν → →

= −

∑ ∑

&

,

2

( ) ( ); ( ) ( ) (0)

n m

iE n m T

k T f T f T d e B B

ν

τ ν ν ν ν ν ν ν ν ν

λ τ τ

∞ → −∞

= = ∫

2 , , ,

1 ( ) ( ) ( ) 2

L R m n n m n n m L n m R n m

J E S P t k T k T

→ →

⎡ ⎤ = − ⎣ ⎦

Weak system-bath coupling limit; <Bρ(0)>=0; Factorization of the density matrix of the whole system; Markovian limit.

slide-21
SLIDE 21

Sufficient conditions for thermal rectification

(1) ( ) ( ) The reservoirs have different mean energy

L L R R

T H T H ρ ρ ≠

2 2 2 2

( ) 1 1 ( ) 1 1 (2) ( ) ( ) The relaxation rates' temperature dependence should differ from the central unit occupation function, combined with some spatial asymmetry.

H C H L R C L R

n n f T f T ω ω λ λ λ λ ⎛ ⎞ ⎛ ⎞ − − − ≠ − ⎜ ⎟ ⎜ ⎟ ⎝ ⎠ ⎝ ⎠

L.A. Wu and D. Segal, PRL (2009). L.A. Wu, C.X. Yu, and D. Segal arXiv: 0905.4015 Harmonic force field Anharmonic force field

,

2

( ) ( ); ( ) ( ) (0)

n m

iE n m T

k T f T f T d e B B

ν

τ ν ν ν ν ν ν ν ν ν

λ τ τ

∞ → −∞

= = ∫

g(TH) g(TC)

TH TC

slide-22
SLIDE 22

Spin-boson thermal rectifier

  • D. Segal, A Nitzan PRL (2005).

( )

L R L R B B B B L R B B

J n n ω Γ Γ = − Γ + Γ ( ) ( ) ( )

1 2 1 2

L R L R B B B B L L R R B B B B

n n J n n ω Γ Γ − = Γ + + Γ +

slide-23
SLIDE 23
  • III. Dynamic Case: Active control

Until now: Heat was flowing from hot objects to cold objects. Question 1: Can we direct heat against a temperature gradient? Answer 1: Add (i) external forces (ii) asymmetry Heat pump moves heat from a cold bath to a high temperature bath. Question 2: Do we need to shape the external force in order to achieve the pumping operation? Answer 2: Random noise can lead to pumping. J W J

slide-24
SLIDE 24

Simple model: Stochastic heat pump

( )

† , , † , , ,

( ) 2 ; =

S L R L R S z k k k k x k k k k

H H H H V V B t H H b b V F F b b

ν ν ν ν ν ν ν ν ν

ε σ ω σ λ = + + + + + = = = +

∑ ∑

JR>0 JR<0 JL<0 JL>0

ωL ωR

TL TR

( )

2 ,

Spectral function of the reservoirs ( ) 2

k k k

ν

ω π λ δ ω ω = −

slide-25
SLIDE 25

Mechanism: Random fluctuations catalyze heat flow

W

L

ω

R

ω

R

ω W

L

ω

  • D. Segal, A. Nitzan, PRE (2006).

D Segal PRL (2008); JCP (2009).

The subsystem is coupled to both ends The subsystem is coupled to the left side only. TLS temperature is effectively high TTLS>TL>TR

slide-26
SLIDE 26

Formalism: Population

Liouville equation Pauli Master equation

( ) ( )

. 1 1 1 1 1 1

( ) ( ) ( )

L R L R

P t k k P t k k P t

ε ε ε → → → →

= − + + +

( )

1 1

Transition rates: ( ) 1 ( ) ( ); ( ) ( ) ( ) k d g n I B k d g n I B

ν ν ν ν ν ν

ω ω ω ω ω ω ω ω

∞ ∞ → → −∞ −∞

= + − = −

∫ ∫

Spectral lineshape of the Kubo oscillator: 1 ( ) exp ( ') ' 2

t i t

I e i t dt d

ω ε

ω ε ω π

∞ −∞

=

∫ ∫

slide-27
SLIDE 27

Spectral lineshape of the Kubo oscillator: 1 ( ) exp ( ') ' 2

t i t

I e i t dt d

ω ε

ω ε ω π

∞ −∞

=

∫ ∫

2 1 1 1 2 1 2 1 2

exp ( ) ( ) ( ) ( ) ( ) ( ) ( )

t t t

K t K t i dt t i dt dt t t t t

ε ε ε ε

ε ε ε ε ε = ⎡ ⎤ = + − + ⎣ ⎦

∫ ∫ ∫

K

Formalism: Random Frequency modulations

(Kubo Oscillator)

2 2

For a Gaussian process in the fast modulation limit / Define ' ( ) ( ') Obtain: ( ) dt t t t I γ π γ ε ε ω ω γ

≡ + = +

slide-28
SLIDE 28

[ ] [ ]

1 1 1 1

( ) 1 ( ) ( ) ( ) 1 ( ) ( ) ( ) ( ) ( ) ( ) k d g n I B k g B n B k d g n I B k g B n B

γ ν ν ν ν ν ν γ ν ν ν ν ν ν

ω ω ω ω ω ω ω ω

∞ → → → −∞ ∞ → → → −∞

= + − → = + = − → =

∫ ∫

Formalism: Transition rates

2 2

/ For a Gaussian process in the fast modulation limit ( ) I γ π ω ω γ = +

Field-free vibrational relaxation rates Kubo oscillator transition rates

slide-29
SLIDE 29

Formalism: Current

[ ]

ˆ Current operator: ( ) , 2

R S R R

i J H t H V = −

1 1 1

Master equation description:

R R R

J P f P f

ε ε ε → →

= −

[ ]

1 1

( ) ( ) ( ) ( ) ( ) ( ) 1 f d g I B n f d g I B n

ν ν ν ν ν ν

ωω ω ω ω ωω ω ω ω

∞ → −∞ ∞ → −∞

= − = − +

∫ ∫

1 0 1

B k B k

γ ν γ ν → → → →

→ →

Kubo oscillator transition rates Field-free vibrational relaxation rates

slide-30
SLIDE 30

( ) ( )

. 1 1 1 1 1 1

Population: ( ) ( ) ( )

L R L R

P t k k P t k k P t

ε ε ε → → → →

= − + + + ( )

1 1

Transition rates: ( ) 1 ( ) ( ); ( ) ( ) ( ) k d g n I B k d g n I B

ν ν ν ν ν ν

ω ω ω ω ω ω ω ω

∞ ∞ → → −∞ −∞

= + − = −

∫ ∫

1 1 1

Heat current:

R R R

J P f P f

ε ε ε → →

= −

[ ]

1 1

( ) ( ) ( ) ( ) ( ) ( ) 1 f d g I B n f d g I B n

ν ν ν ν ν ν

ωω ω ω ω ωω ω ω ω

∞ → −∞ ∞ → −∞

= − = − +

∫ ∫

Formalism: Summary

slide-31
SLIDE 31

Numerical Results

γ B0 2 4 6 12 14 16 18 20 22

0.1 0.2 Colormap of the heat flux at the Right contact JR>0 JR<0 JL<0 JL>0

ωL ωR

TL TR RL ωL =200, ωR=3, TL=TR=25.

( )

2 2

/ ( ) ( ) exp / I g A

ν ν ν

γ π ω ω γ ω ω ω ω = + = −

slide-32
SLIDE 32

γ TL-TR 2 4 6 8 2 4 6 8 10 12

0.1 0.2

Colormap of the heat flux at the Right contact ωL =200, ωR=3, B0=15 TR=25.

Numerical Results

Pumping JR>0 JR<0 JL<0 JL>0

ωL ωR

TL TR

slide-33
SLIDE 33

Proof of principle for a dichotomous noise

L

ω

R

ω

R

ω

L

ω

E1-E0=B0-Ω

( ) ( )

Dichotomous noise 1 ( ) ~ 2 I ω δ ω δ ω − Ω + + Ω ⎡ ⎤ ⎣ ⎦

( )

2 ,

( ) 2 Assumption: The R Reservoir spectral denisty strongly varies within the noise spectral window ( ) ( )

k k k R R

g g B g B

ν ν

ω π λ δ ω ω = − + Ω << − Ω

E1-E0=B0+Ω

slide-34
SLIDE 34

If TL=TR, it can be shown that current is catalyzed from the R side into the L side when the following condition is satisfied

( ) ( )[

] [ ]

( ) ( ) ( ) ( ) 1 ( ) 1 ( ) 1 Or ( ) ( )

L R L L R L

g g n B g n B n B n B g g n B g n B n B n B

− − + − − +

+ − Ω + + Ω − Ω < − Ω + + − Ω + + + Ω + + Ω < − Ω

( ) g g B

ν ν ± =

± Ω

slide-35
SLIDE 35

Efficiency: Approaching the Carnot limit

2

Assume Einstein solids with g ( ) 2 ( )

ν ν ν

ω πλ δ ω ω = −

max

[ ( ) ( )] Pumping condition: Work: ( ) [ ( ) ( )] Cooling efficiency =

L L R R L R L R R R R R L L L R R R R R L R L R

J n n T T J T W n n J T W T T

ν ν ε ε ε ε

ω ω ω ω ω ω ω ω ω ω ω η η ω ω = − − − < → < = − − − = < = − − T฀ T฀ ωL ωR

TL TR

slide-36
SLIDE 36
  • Noise processes in nanomechanical resonators: Adsorption-desorption

noise, temperature fluctuations.

(Clealand and Roukes, J. App. Phys., 92 2758 (2002), Y.T. Yang et al. Nano Lett (2006). ).

  • The resonance frequency can be tuned by the gate- voltage fluctuations

change the characteristic modes.

(Sazanova et al. Nature 431, 285 (2004).

Physical Realizations

slide-37
SLIDE 37

Exciton stochastic pump

(a) The Metals have different band structure. Delta-like DOS will lead to the Carnot efficiency. Related ideas, showing reversible particle transfer, were considered in Humphrey and Linke PRL 2005.

slide-38
SLIDE 38

Pumping of heat by modulating the reservoirs temperatures

Computer simulation of a heat pump where the temperature of one reservoirs is modulated periodically.

slide-39
SLIDE 39

N.Li, B. Li and P. Hanggi,EPL (2008)

slide-40
SLIDE 40

We studied quantum heat transfer in minimal models, seeking to connect the transport characteristics with the microscopic description (H J ?)

  • Static transport: We discussed sufficient conditions for

thermal rectification.

  • Dynamical control: We studied pumping of heat due to

the shaping of the reservoirs properties, given that the system is suffering a stochastic noise.

Summary

slide-41
SLIDE 41

Formal issues:

  • Going beyond the perturbative treatment.
  • Add coherent effects
  • Consider more realistic models.

Basic challenges:

  • Static case: Understand other nonlinear phenomena,

e.g. negative differential thermal conductance, from first principles. Understand the ballistic- diffusive heat- flow crossover (Fourier law).

  • Dynamic case: Control heat flow by modulating the

reservoirs temperatures.

Outlook

slide-42
SLIDE 42

Rectifiers: Abraham Nitzan Tel Aviv University Lianao Wu University of Toronto & The Basque Country University at Bilbao Claire Yu University of Toronto Pumps: Abraham Nitzan Tel Aviv University

Thermal rectification:

  • D. Segal and A. Nitzan, PRL 94, 034301 (2005); JCP 122, 194704 (2005).
  • D. Segal, PRL 100 105901 (2008).

L.A. Wu and D. Segal, PRL 102, 095503 (2009); L.A. Wu, C. X. Yu, and D. Segal arXiv 0905:4015. Molecular heat pumps:

  • D. Segal and A. Nitzan, PRE 73, 02609 (2005).
  • D. Segal PRL 101 260601 (2008); JCP 130, 134510 (2009).

Thanks-

slide-43
SLIDE 43

,

2

( ) ( ); ( ) ( ) (0)

n m

iE n m T

k T f T f T d e B B

ν

τ ν ν ν ν ν ν ν ν ν

λ τ τ

∞ → −∞

= = ∫

2

Harmonic bath, bilinear coupling ( ) ( ) 2 ( )

B j j

k T n

ν ν ν ν

ω πλ δ ω ω ⎡ ⎤ = − − − ⎢ ⎥ ⎣ ⎦

Noninteracting spin bath ( ) ( ) ( )

S S

k T n

ν ν ν ν

ω ω = − Γ

The rate constant is reflecting the bath properties

2 ,

Noninteracting spinless electrons ( ) ( ) 2 ( )[ ( ) ( )]

B i j F i F i i j

k T n n n

ν ν ν ν ν ν

ω πλ δ ε ε ω ε ε ω ⎡ ⎤ = − − − + − + ⎢ ⎥ ⎣ ⎦

/ 1

( ) [ 1]

T B

n e

ν

ω ν ω −

= −

/ 1

( ) [ 1]

T S

n e

ν

ω ν ω −

= +

( )/ 1

( ) [ 1]

T F

n e

ν ν

ω µ ν ω − −

= +