Conduction properties of DNA molecular wires
description
Transcript of Conduction properties of DNA molecular wires
Sicily, May 03-04 (2008)
Conduction properties of DNA molecular wires
Rafael Gutierrez
Giovanni Cuniberti
Rodrigo Caetano
Bo Song
Institute for Materials Science and Max Bergmann Centre for Biomaterials
Collins Nganou
DNA: a complex systemDNA: a complex systemWhich physical factors are important for transport?Which physical factors are important for transport?
environmentenvironment
internal vibrationsinternal vibrations base-pair sequence(electronic structure)
metal-molecule contactmetal-molecule contact
accuracy
SizeNr. atoms
model Hamiltonians
Dynamical Effects
Static Deformations30
200
∞
DFTBModels MolecularDynamics
accuracy
SizeNr. atoms
Static Deformations30
200
∞
DFTBModels MolecularDynamics
I. Bridging first-principle and model Hamiltonian approaches: Parameterization
Benchmark: twisting of Poly(GC)
• Motivation: R. Di Felice et al. work on G-stacks• tHOMO-HOMO=f() for a GC-dimer
Ef
GC1 GC2
t
I. Bridging first-principle and model Hamiltonian approaches: Parameterization
φ (degrees)
2t
(eV)
d = 3.4 Å
(a)
(b)
(a) DFTB(b) Y. Berlin et al. CPC 3, 536 (2002)
I. Bridging first-principle and model Hamiltonian approaches: Parameterization
Twisting-stretching in Poly(GC)
• Electrical current during the stretching-twisting processΓ >> |t|
Γ ~ |t|
Γ < |t|
Molecular Computing Group http://www-MCG.uni-r.de
I. Bridging first-principle and model Hamiltonian approaches
?
d φ
l l
HOMO(GC)1-HOMO(GC)2 coupling
accuracy
SizeNr. atoms
Dynamical Effects
30
200
∞
DFTBModels MolecularDynamics
Idea: map DFTB-based electronic structure onto TB-Hamiltonian along MD trajectory
)(),( 1, tVtE jjj
.....
})({}),({ 1, jjj VPEP ( ) ( )j lE t E t Probability distributions Correlation functions
, 1 1( ) ( )( . .)j j j j j j j tunnelingj j
H E d d V d d ht Ht c
II. Model Hamiltonian and dynamical effects:short poly(GC) wires in a solvent
DFTB
DFTB
DFTB
DFTB
II. Model Hamiltonian and dynamical effects: short poly(GC) wires and time series
Parameters variation time scale ~ fs
The electron will “feel” the average of the parameters over the coarse graining time (related to tunneling time)
The rate of electrons going through the DNA for a current inorder of 1 nA is 10 e/ns
II. Model Hamiltonian and dynamical effects:adiabatic approximation and time scales
II. Model Hamiltonian and dynamical effects:short poly(GC) wires in a solvent
Average current through a G-pathway
Current strongly depends on charge „tunneling time“ tun
...1(t) (t) (t)
V1(t) V7(t)
tun
Lower bound
( ) 1j tV t fs
( )
1
1( ) ( , , )( )
N t
l lt t tj
I V I V VN t
II. Model Hamiltonian and dynamical effects:short poly(GC) wires in a solvent
...1(t) (t) (t)
V1(t) V7(t)
Probability distributions P for j(t)
Gaussian distribution(for reference)
DNA frozen
II. Model Hamiltonian and dynamical effects:short poly(GC) wires in a solvent
...1(t) (t) (t)
V1(t) V7(t)
Probability distributions P for Vj(t)
Gaussian distribution(for reference)
DNA frozen
n.n. electronic coupling mainly depends on internal DNA dynamics
II. Model Hamiltonian and dynamical effects:Linear chain coupled to bosonic bath
' ' ' ' '
( )
2( ) ( ) ( ) ( ) (1 ( )) ( ) ( )
( ) iEt t
ieI V dE dE Tr f E G E E E f E G E E Eh
E dt e e
Electrical current on lead =L,R
, 1,j j jV
Time average quantities
II. Model Hamiltonian and dynamical effects:Fluctuation-Dissipation relation
...1(t) (t) (t)
V1(t) V7(t)
0
( ) ( )
2( ) tanh cos( ) ( )2 B
C t J
J dt t C tk T
Relation between correlation functions C(t) and spectral density of the bosonic bath J() is given by FD theorem
II. Model Hamiltonian and dynamical effects:Influence of correlation times for a generic C(t)
2
1( ) , 5, 11 ( )
C tt
Gap reduction
II. Model Hamiltonian and dynamical effects:Gap reduced with
=100 fs
=1 fs
0( , ) j jJd
reorganization energy
II. Model Hamiltonian and dynamical effects:Strength of dynamical disorder
2
1( ) , 10 , 11 ( )
C t fst
II. Model Hamiltonian and dynamical effects:MD-derived correlation function
Fit to algebraic functions4
21
1( ) ( ) (0)1 ( )j
j
j
C t t t
II. Model Hamiltonian and dynamical effects:Fourier transforms of ACF for the onsite energies
( ) cos( ) ( )C dt t C t
DNA base dynamics:C=N and C=C stretch vibrations?
see e.g. Z. Dhaouadi et al., Eur. Biophys. J. 22, 225 (1993)
water modes
II. Model Hamiltonian and dynamical effects:MD-derived correlation function
...1(t) (t) (t)
V1(t) V7(t)0
( ) ( )
2( ) tanh cos( ) ( )2 B
C t J
J dt t C tk T
II. Model Hamiltonian and dynamical effects:Stochastic model Hamiltonians
How to formulate and solve a model Hamiltonian which directly uses MD informations
(t) is a random variable describing dynamical disorder (time series drawn from MD simulations)
II. Model Hamiltonian and dynamical effects:Stochastic model Hamiltonians
Formal solution for the disorder-averaged Green function, assuming Gaussian fluctuations:
Only the two-times correlation function (second order cumulant) is required !A simple case:
correlation function 2
1( )
1
C tt
Toy model: single site with dynamical disorder
II. Model Hamiltonian and dynamical effects:Stochastic model Hamiltonians
Disorder-averaged transmission T(E)
II. Model Hamiltonian and dynamical effects:Stochastic model Hamiltonians
Limits: 0
White noise
Adiabatic limit
Scaling of the transmission at the Fermi level with thecorrelation time (single site model)
I.Bridging first-principle and model Hamiltonian approaches: “static“ parameterization of minimal models
II.Bridging molecular dynamics and model Hamiltonians:„dynamical“ parameterization of minimal models
III.In progress: length and base sequence dependencies solution of random Hamiltonians contact effects
Current (and prospective) research lines