Finite Size Scaling and

Quantum Criticality

Sabre Kais
Department of Chemistry
Birck Nanotechnology Center
Purdue University
West Lafayette, IN 47907
 Kais
Research Group
Department of Chemistry
Purdue University

Finite Size Scaling Quantum Computing

in and
Quantum Mechanics Quantum Information
Theory

Qi Wei Zhen Huang

Winton Moy Hefeng Wang
Dr. Pablo Serra Xu Qing
 Introduction

Finite Size Scaling In Quantum Mechanics

Applications
(A) Stability of atoms, molecules and quantum dots

(B) Atomic and molecular stabilization in superintense

laser fields
 Statistical Mechanics
Classical Quantum

Phase Transitions Phase Transitions

Free Energy T 0
F(Ki)=-KBT log(Z) E 0 (λi )
Ground State ……:

Critical Phenomena Critical Phenomena

Correlation Length Mass Gap of H
ξ ~ (T − TC ) −ν ξ → Δ1E ~ (λ − λC ) −ν

Finite Size Scaling Finite Size Scaling

Thermodynamic Limit Number of Basis Functions

N →∞ M →∞

Applications Applications
 Phase Transitions
™ Classical: Classical phase transitions are driven by thermal energy
fluctuations P CP
Liquid
Solid
Like the melting of an ice cube:

Solid → Liquid → Gas

Gas
T

™ Quantum: Quantum phase transitions, at T=0, are driven by the

Heisenberg uncertainty principle

Like the melting of a Wigner crystal: Two dimensional electron

layer trapped in a quantum well

Wigner crystal Fermi liquid

 Quantum Phase Transitions
Transitions that take place at the absolute zero of
temperature, T=0, where crossing the phase
boundary means that the quantum ground state
energy E 0 ( λ ) , of the system changes in some
fundamental way.

This is accomplished by changing some parameter

in the Hamiltonian of the system H 0 ( λ ) .

We shall identify any point of non-analyticity in the

ground state energy λ = λ C , as a quantum phase
transition.
 Phase Transition

Partition
Function
Free energy: f[K]=-kBT log[Z]

Coupling Constants: {K1, K2; KD}

As a function of [K], f[K] is analytic almost everywhere

Possible non-analyticities of f[K] are points (DS=0),

lines (DS=1), planes (DS=2), …

Regions of analyticity of f[K] are called phases

 Phase Transitions

First-order:

∂f ∂K i is discontinuous across a phase boundary

Continuous Phase Transition:

All ∂f ∂K i are continuous across the phase boundary.

But, second derivatives or higher derivatives are
discontinuous or divergent
 Phase Transitions

P T
One Phase
CP
PC TC
Liquid
Solid Gas Liquid

Two Phases
Gas

TC T ρ

β Critical
Order
Parameter
ρ + − ρ − ∝ T − TC Exponent

β = 0.327
 Critical Exponents
The critical exponents describe the nature of the singularities in various measurable
quantities at the critical point [α , β , γ , δ ,η ,ν ]

In the limit T → TC
−α
™ Heat Capacity: C ~ T − TC
β
™ Order Parameter: M ~ T − TC
™ Susceptibility: χ ~ T − TC
−γ

™ Equation of State: M ~ H1δ

™ Correlation Length: ξ ~ T − TC
−ν

Scaling Laws:
™ Fisher: γ = ν (2 − η )
™ Rushbrooke: α + 2β + γ = 2
™ Widom: γ = β (δ − 1)
™ Josephson: νd = 2 − α
 Critical Exponents
[α , β , γ , δ ,η ,ν ]

Exponent TH EXPT MFT ISING2 ISING3 HEIS3

α 0-0.14 0 0 0.12 -0.14
β 0.32-0.39 ½ 1/8 0.31 0.3
γ 1.3-1.4 1 7/4 1.25 1.4
δ 4-5 3 15 5
ν 0.6-0.7 ½ 1 0.64 0.7
η 0.05 0 ¼ 0.05 0.04
α + 2β + γ 2 2.00 ±0.01 2 2 2 2
( βδ − γ ) β 1 0.93 ±0.08 1 1 1
(2 − η )ν γ 1 1.02 ±0.05 1 1 1 1
( 2 − α ) νd 1 4/d 1 1 1
TH. Theoretical values (from scaling laws); EXPT. Experimental values (from a variety of
systems); MFT. Mean field theory; ISINGd. Ising model in d dimension; HEIS3. classical
Heisenberg model. D=3
Kenneth G. Wilson (1982)
The Nobel Prize in Physics 1982

Kenneth G. Wilson
Cornell University

Renormalization Group

"for his theory for critical phenomena in

connection with phase transitions"
 Quantum Partition Function Z
Z=Tr exp[-βH]
Operator density matrix : e − βH → e − iHT (Time evolution operator)
T → −i β

e− βH
= e [ ]
− δτH N
Feynman
δτ is time interval and Nδτ = β Path-Integral

Z(β) = ∑ ∑ n e−δτH m1 m1 e−δτH m2 m2 e−δτH m3 ⋅⋅⋅ mN e−δτH n

n m1,m2 ,m3 ,...,mN

Has the form of a classical partition function, i.e. a sum over

configurations expressed in terms of a transfer matrix, if we think
about the imaging time as an additional spatial dimension

Quantum Classical
⇒
d − dimensions + it (d + 1) − dimensions
 Analogies

Classical (KT) Quantum (T 0)

(D+1) space dimension D space, 1 time
Transfer Matrix Quantum H
Partition Function Generating Functional
Equalibrium State Ground State
Ensemble Average Expectation Values
Free Energy Ground State of H
Temperature Coupling Constant
Correlation Functions Propagators
Correlation length (ξ )) Mass Gap of H ξ ……
→ Δ1E
 Finite Size Scaling
In statistical mechanics, the finite size scaling method provides a
systematic way to extrapolate information obtained from a finite
system to the thermodynamic limit

Importance
The existence of phase transitions is associated with singularities
of the free energy. These singularities occur only in the
thermodynamic limit.

Yang and Lee Phys. Rev. 87, 404 (1952)

The Nobel Prize in

Physics 1957
 Finite-size effects in
Statistical Mechanics
∞
Cv

N'''

N''

N'

TC T
In the present approach, the finite size corresponds not to the
spatial dimension, as in statistics, but to the number of elements
in a complete basis set used to expand the exact eigenfunction of a
given Hamiltonian.
Quantum Mechanics
∞ M
ψ = ∑ anφn ≅ ∑ anφn
n =0 n =0

(Variational Calculations)

Classical Quantum
⇔ ( M → ∞)
( N → ∞)

Kais et. al Phys. Rev. Letters 79, 3142 (1997)

Finite Size Scaling: Quantum Mechanics
H = H 0 + Vλ
M (N )
ψλ (N )
= ∑ an( N ) (λ )φn
n

The FSS ansatz

O
(N )
λ
~ O λ FO N λ − λC ( ν
)
Δ O (λ ; N , N ′) =
(
ln O
(N )
λ
O
( N ′)
λ
)
ln( N ′ N )

The curves intersect at the critical point

Δ O (λC ; N , N ′) = Δ O (λC ; N ′′, N )

 Short Range Potentials
Yukawa Potential

Hamiltonian

1 2 e−r
H (λ ) = − ∇ − λ
2 r
Basis Set
1
φn ( r , Ω) = e −r / 2 L(n2−)l (r )Yl ,m (Ω)
(n − l + 1)(n − l + 2)

( 2)
Where Ln−l (r ) is the Laguerre polynomial of degree
n and order 2 and Yl ,m (Ω) are the spherical harmonic
functions of the solid angle
 Yukawa

λC = 0.839903

ΔH λexact = 0.839908
Γα = C

Δ H − Δ ∂∂Vλ
Phys. Rev. A 57, R1481 (1998)
Gaussians Slater
Finite Size Scaling : Data
Collapse
O ~ + (λ − λC ) μ
∞ λ →λ
C

O N
~ O ∞ O
(
F N λ − λC
ν
)
F ( x ) ~ x − μ /ν
x →0

O ~ N − μ /ν
N N →∞

O N
(
~ N − μ /ν G N 1 /ν (λ − λC ) )
(
E0 N α /ν ~ G N 1 /ν (λ − λC ) )
Data Collapse E0 N α /ν ~ G ( (λ − λC )N1/ν )

α = 2, ν = 1

(λ − λC )N 1 /ν
Chem. Phys. Letters 319, 273 (2000)
Im(λ )
0.30

5 N =4
0.20
6
11
0.10 λc = 0.840

0.00
0.82 0.86 0.90 0.94 0.98
Re(λ )
-0.10

-0.20

Yukawa N =4
-0.30

Kais et al. Chem. Phys. Letters 423, 45 (2006)

 Quantum Mechanics

Complex
Complex Complex Complex
Angular
Time Energy Charge
Momentum

Regge Instanton Rotation Variation

Theory Method Complex Method
Coordinates

Stability
Scattering
Tunneling Resonances of
Amplitude
Ions
 Two Electrons Atoms
1 2 1 2 1 1 λ 1
H (λ ) = − ∇1 − ∇ 2 − − + λ=
2 2 r1 r2 r12 Z
e-
r12 = r1 − r2 r12

e-
r2
ZC=0.911 r1

Z
ET=-0.5

He like system H like +e-

E gs (λ ) H- He Z

ZC=0.911
 Finite Size Scaling procedure
™ Hamiltonian:
1 2 1 2 1 1 λ 1
H = − ∇1 − ∇ 2 − − + ; λ =
2 2 r1 r2 r12 Z

™ Basis Set:

Ψ= ∑ Ci, j ,k 2
(
1 i j −α r1 − β r2
r1 r2 e )
+ r1 j r2 i e − β r1 −α r2 r12 k
i , j ,k

i+ j+k ≤ N

™ Hamiltonian Matrix:
Leading Eigenvalues, E 0 , E1
™ Renormalization Equation:
N N +1
⎛ E1 ( N ) ⎞ ⎛ E ( N + 1) ⎞
⎜⎜ ⎟⎟ = ⎜⎜ 1 ⎟⎟
⎝ E0 ( N ) ⎠ ⎝ E0 ( N + 1) ⎠
 λC = 1.097

Kais et. al, Phys. Rev. Letters 80, 5293 (1998)

Conditional Probability

λ > λC

λ = λC

λ < λC
 D=3

Fluids
Liquid Gas

Electronic Structure
Bound Continuum
 Critical Charges and Stable Atoms and Ions

H- He

N=2

0.91 Z

H-- He- Li

N=3

2.08 Z

He-- Li- Be

N=4

2.85 Z
 D=3

Fluids
Liquid Gas

Electronic Structure
Bound Continuum
Surcharge
Se=N-ZC

Int. J. Quantum Chem. 75, 533 (1999)

¾ Do doubly charged negative atomic ions exist
in the gas phase?

NO
¾ What is the smallest object that can bind two
extra electrons?

This is a challenge for

experiment and theory!
The two electrons must be
separated by at least 5.6 Å
 Stability of Dianions of
Fullerenes
_
EA1 = + 2.65 eV C 60
EA2 = - 0.3 eV

__
R c = 5.6 A C 79

EA1 = + 3.14 eV
EA2 = + 0.44 eV __
C 84

Advances in Chemical Physics, Volume 125, 1 (2003).

Stability of Linear Dianions
e- e-

RC=5.6 Å

e- e-

- -
C ≡ C - Be - C ≡ C 1
∑g
+

BeC4--

BeC4-- BeC2- + C2-

 Phase Transitions and Stability of
Three Body Coulomb Systems

(m,q) He ↔ He + + e -
(m,q) (m,q)
+
H2 ↔ H + p

(M,Q) (M,Q) (M,Q) Q Q=−q q

r → fr
1 2 1 2 1 qQ qQ Q 2 H → uH f 2
H =− ∇1 − ∇ 2 − ∇1 ⋅ ∇ 2 + + +
2μ 2μ m r1 r2 r12 f = u Qq
mM
u=
m+M
∇12 ∇ 22 1 1 1
H =− − − − − κ∇1 ⋅ ∇ 2 + λ
2 2 r1 r2 r12 0≤λ = Q q ≤∞
0 ≤ κ = (1 + m M )−1 ≤ 1
 Three Body Coulomb Systems (ABA)

Particles Mass
e: electron 1
p: proton 1836.15
u: muon 206.76
d: deuteron 3670.5
t: tritium 5476.92
D=∞

(e- μ + e - ) Stable (μ -p+ μ − ) Stable

Phys. Rev. A 62, 06050 (2000)
 Phase Transitions and Stability of
Three Body Coulomb Systems
Charge density probability

H2+

λ = 1.24 < λCN = 20 = 1.2402

Bound States

λ = 1.241 > λCN = 20 = 1.2402

Coulomb Explosion
 FSS for Critical Conditions
for Stable Dipole Bound
Anions
The Hamiltonian
- e

R
-Z +Z z

Slater Basis Set:

Chem. Phys. Lett. 372,205 (2003)

µc=0.655
a.u

µc=0.655 a.u =1.625D without B.O =2.5D

Electron will be trapped with µc>2.5D
H3C-CN µ= 4.3 D Ea=108 cm-
H2CCC µ= 4.34 D Ea=173 cm-
C3H4O3 µ= 5.5 D Ea=40 meV
C3H2O3 µ = 4.5 D Ea=20 meV
 FSS for Critical Conditions for
Stable Quadrupole Bound Anions
Hamiltonian: consists of a charge q at the origin and two charges –q/2 at
z = +1 & -1

Slater Basis Set:

Where β is the variational parameter used to optimize the numerical results

and P2l (θ ) is the Legendre polynomial of order l.

q q=QR
Journal of Chemical Physics, 120, 8412 (2004)
qc=1.46 a.u ⎛⎜ − ,+ q,− ⎞⎟
q q
⎝ 2 2⎠

qc=3.9 a.u ⎛ q q⎞
⎜ + , − q , + ⎟
⎝ 2 2⎠

KCl2- Qzz=10 a.u

K2Cl- Qyy=27 a.u
(BeO)2- (13,13,0.5)
CS2- Qzz=4 a.u

Journal of Chemical Physics, 120, 8412 (2004)

Hexagonal Lattice of Quantum Dots
 Hubbard Model

H = −t ∑ i c jσ + U ∑ ni↑ ni↓ − μ ∑ (ni↑ + ni↓ )

c +

<i , j > ,σ i i

Where t is the nearest-neighbor hopping

term, U is the local repulsive interaction
+
and µ is the chemical potential. ciσ (ciσ )
creates (annihilates) and electron with
spin σin a Wannier orbital located at
site i; the corresponding number
operator is niσ = ci+σ ciσ and < > denotes
the nearest-neighbor pairs.
 Experimentally, one can prepare quantum dots with:
¾ Different metal and this primarily determines the chemical
potential µ
¾ Different sizes and this primarily determines U
¾ Different packing distance and this primarily determines t

So the “experimental” Hamiltonian has 3 different parameters,

µ, U and t, each of which can be independently varied. The
essential physics are embodied in the Hubbard Hamiltonian
 Why do
Why do we
we need
need to
to use
use
RG method?
RG method?

¾N=7, 784 states in total

14 States 210 States 420 States 140 States

¾N=19, 2 billion states

 RG Procedures
RG Procedures
1. Divide the N-site lattice into appropriate ns -site blocks
labled by p (p=1,2,…,N/ns) and separate the Hamiltonian H
into intrablock part HIB and interblock HB ,

H = H B + H IB = ∑H
p
p + ∑V
< p , p'>
p , p'

2. Solve HB exactly to get the eigenvalues, the

eigenfunctions of E pi ( i = 1,2,....4 ns )

3. Treat each block as one site on a new lattice and

the correlations between blocks as hopping
interactions
Triangular lattice with
Hexagonal Block
Calculation Results for Metal-
Insulator Transition
Charge gap:
Δ g = E ( N + 1) + E ( N − 1) − 2 E ( N )
10

6 (U/t)c=12.5
Δp/t

2
Metal Insulator
0
0 5 10 15 20
U/t

Phys. Rev. B66, 081101 (2002)

Finite-size Scaling for Hubbard
model on triangular lattice
4
10
3
10
2
10

/t
0.405
1
10

ΔgN
0 2 3 4 5 6 7
10 N =7,7 ,7 ,7 ,7 ,7 ,7
-1
10
-2
10 (U /t) c =12.52
-3
10
6 8 10 12 14 16 18 20 22
U /t

4
10
3
10
/t
0.405

2
10
ΔgN

1
10
0
10
-1
10
-2
10
-3
10
-6000 -3000 0 3000 6000 9000
0.5
(U-U c )N /t

J.X.Wang, S. Kais, Phys.Rev.B66,081101(2002)

Atomic & Molecular Stabilization
by
Superintense Laser Fields
with
Prof. Dudley Herschbach
Harvard University

The Nobel Prize in Chemistry 1986

 Multiply charged negative
atomic ions
in superintense laser fields
 The peak-power of
Superintense Laser Fields (I > a.u.) pulsed lasers has
increased by 12 orders
of magnitude during
the past 4 decades.

QUESTION: What is the

highest-level intensity
presently possible?

ANSWER: The highest

possible focused laser
intensity =1019 W/cm2

This level of intensity can

be achieved with
femtosecond laser based
on Chirped Pulse
Amplification (CPA).
 Laser Atom Interaction

The electric field of a monochromatic plane wave

can be written
E (t ) = E0 (eˆx cos ωt + tan δ eˆ y sin ωt )
With ex and ey orthogonal to each other and to the
propagation direction
δ = 0 corresponds to linear polarization
δ = ± π/4 corresponds to circular polarization.

N ⎧⎪ 1 2 1 i −1
1 ⎫⎪ ∂Ψ
∑ ⎨ pi + r r + ∑ r r
i =1 ⎪ 2 ri + α (t ) j =1 ri − rj
⎬Ψ = i
∂t
⎩ ⎪⎭

Moving frame of reference which follows the quiver motion of the classical electron
r r
where α (t ) = (α 0 E0 ) E (t )

α0=E0/ω2, where E0 and ω are the amplitude and frequency of

the laser field.
 High Frequency Floquet Theory

Hamiltonian of atomic system in super-intense laser fields

H = kinetic term + coulomb term + laser term + e-e term

N ⎧⎪ 1 2 i −1
1 ⎫⎪
H = ∑ ⎨ pi + Vo (ri , a0 ) + ∑ ⎬
i =1 ⎪⎩
2 j =1 ri − r j ⎪⎭
Vo is the “dressed” Coulomb potential, is the time average of the shifted
Coulomb over one period of the laser
2π
1 dφ
V0 ( a0 , r ) = −
2π ∫
0
v
r + a0 (eˆx cos φ + tan δ eˆ y sin φ )

α o = I ω 2 , where I is the time - averaged beam intensity and ω is the frequency.

Structure Equation: H Ψ =E( α0 ) Ψ

Hydrogen atom in super-intense linear laser fields

M. Pont, et al. Phys. Rev. Lett. 61, 939 (1988) J. H. Eberly, et al Science 262, 1229 (1993)
The electronic orbitals for the ground states. The presentation is given in a plane
passing through the axis of the field taken as polar axis z
Do atoms in superintense laser fields behave
like diatomic molecules ?
 Summary
™ Symmetry breaking of electronic structure
configurations resemble classical phase transitions.
Phys. Rev. Letters 77, 466 (1996)
J. Chem. Phys. 120, 2199 (2004)
™ Finite Size Scaling method can be used to obtain
critical parameters for the quantum Hamiltonian:
Critical charges, critical dipole and quadrupole-bound
anions,…
Phys. Rev. Letters 79, 3142 (1997)
J. Chem. Phys. 120, 8412 (2004)
™ Quantum phase transitions can be used to explain
and predict the stability of atoms, molecules and
quantum dots.
Phys. Rev. Letters 80, 5293 (1998)
Adv. Chem. Phys. 125, 1-100 (2003)
Atomic dianions are unstable in the gas phase

Multiply charged negative ions are stable in

superintense laser fields
(Intensity > 1 a.u. ~ 1016 W/cm2)

J. Chem. Phys. 124, 201108 (2006)

Thank You!
Question 1
Question 2