1 of 37

Discretizing the sphere with an irregular� lattice: Application to the �transverse-field Ising model

Grigorios Makris

Institute for Quantum Information, RWTH Aachen University

Supervisors:

Prof. Dr. Fabian Hassler Prof. Dr. Stefan Wessel

Institute for Theoretical Solid State Physics, RWTH Aachen University

2 of 37

Outline of presentation

    • Regular and irregular lattices
    • The free particle on the sphere
    • The transverse-field Ising Hamiltonian on the sphere
    • Symmetry breaking of the Ising model
    • Scaling and conformal symmetry
    • Results on the sphere

​

​

2 / 19

3 of 37

Regular lattices in general

Objective homogeneity quantification of a periodic surface using the Gini coefficient

3 / 19

4 of 37

spectrum(100,1.3,0,5)

Lattices on the sphere

  • Regular lattices are only possible for specific number of points

Examples of Platonic solids

4 / 19

5 of 37

Irregular lattices

Nearest neighbor couplings equal.

5 / 19

6 of 37

Motivation for the sphere

The plane is conformally related to the cylinder

6 / 19

7 of 37

Fibonacci lattice

Allocate points on spiral around the z-axis of the sphere.

golden angle,

Lattice implementations

golden ratio

7 / 19

8 of 37

Free particle on the sphere

Decide on a lattice by solving the free particle problem.

Discretize on the tangent plane

of every point on the sphere

8 / 19

9 of 37

Free particle on the sphere

degeneracies

Fibonacci lattice

9 / 19

10 of 37

Transverse-field Ising Hamiltonian on the sphere

same as free particle hoppings!

Constant couplings

Free particle couplings

10 / 19

11 of 37

  • symmetry

Symmetry breaking of the Ising model

  • No order without symmetry breaking
  • Symmetry may be broken when ground states of both sectors have same energy.

11 / 19

12 of 37

Symmetry breaking of the Ising model

  • At infinite system size the first excited becomes truly degenerate with the ground state.

​

  • Phase transition happens at the critical point, more degeneracies.

Finite transverse field Ising chain

12 / 19

13 of 37

for large separation

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

Quantum to classical mapping

At the critical point there are degeneracies, so that

Implies scale invariance

Original quantum model exists along chain

13 / 19

14 of 37

Scale invariance

Two-point correlation functions of scaling fields have power behavior at the critical point

is the scaling dimension of the field

Ising model

14 / 19

15 of 37

Conformal symmetry

  • Scale invariance is characteristic in second order phase transitions.

​

  • Dilations or rescalings are part of the more general conformal transformations.

​

  • Conformal transformations preserve angles.

​

15 / 19

16 of 37

  • Conformal symmetry at the critical point:

​

​

​

​

J L Cardy 1985 J. Phys. A: Math. Gen. 18 L757

Conformal symmetry

16 / 19

17 of 37

Finite-size scaling

Scaling assumption

​

​

The thermodynamic observables scale with system size as power laws of to ratios of scaling exponents.

Consider a system of size .

17 / 19

18 of 37

Results with quantum monte carlo

Flat model:

18 / 19

19 of 37

  • Thank you for your attention!

19 / 19

20 of 37

Scaling

  • On the classical Ising model we can tune two variables, the temperature and the magnetic field.

​

  • The phase transition happens at a critical temperature at zero magnetic field.
  • Scaling of thermodynamic observables on infinite systems is a power law with universal scaling exponents.

implies scale invariance

20 / 19

21 of 37

Maximizing the smallest nearest neighbor distance

Minimizing interacting potential

  • For arbitrary number of points irregular lattices are used

Refinement of the platonic solids. �Creation of the icosahedron-seed.

Lattice implementations

maximize

  • d
  • d

Spiral points – Fibonacci lattice

21 / 19

22 of 37

0-generation

1-generation

2-generation

Icosahedron-seed

Lattice implementations for the sphere

Most points connect to 6 neighbors.

22 / 19

23 of 37

Free particle on the sphere

degeneracies

Fibonacci lattice

Icosahedron-seed

We choose the Fibonacci lattice!

23 / 19

24 of 37

Isotropy of the spherical Ising model

​

  • Our model locally approximates the flat 2D transverse field Ising model.
  • Isotropy established up to second order similarly to flat model.

24 / 19

25 of 37

Free particle on the sphere

How good is the spectrum on the Fibonacci lattice?

Minimum angle of rotation:

25 / 19

26 of 37

Free particle on the sphere

How do the behave?

Almost one-to-one correspondence with nearest neighbor distances .

26 / 19

27 of 37

Free particle on the sphere

How do the behave?

27 / 19

28 of 37

Outline of presentation

  • Constructing a discrete sphere

​

    • Lattice implementations
    • The free particle on the sphere
    • The transverse-field Ising Hamiltonian on the sphere

​

  • Transverse-field Ising model at criticality

​

    • Basics of phase transitions
    • Finite-size scaling and conformal symmetry
    • Results on the sphere

​

​

28 / 19

29 of 37

Quantum to classical mapping

Insert completeness relations

Transfer matrix:

many spin state

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

29 / 19

30 of 37

Quantum to classical mapping

due to gap in the spectrum for finite systems

for large separation

30 / 19

31 of 37

Finite-size scaling

constant for all system sizes.

Binder Cumulant is constant for all system sizes.

​

Don’t need to know scaling exponents a priori.

Suppose we want to find the critical temperature .

31 / 19

32 of 37

Conformal symmetry

The and the are conformally related,

32 / 19

33 of 37

Results on the sphere

Examine also the classical Ising model by setting the transverse field to zero.

33 / 19

34 of 37

Results on the sphere

Nevertheless, fitting with the exponents of the 3D Ising universality class may be applicable.

Collapse of Binder Cumulant

34 / 19

35 of 37

Conclusions and Outlook

  • The crossing points are very sensitive to very small rescalings of the couplings.
  • Attempts to find couplings on more uniform lattices yield more uniform distribution of couplings and better spectra.
  • The Fibonacci geometry yields better spectrum.
  • TFIM has more degenerate energy levels when Laplacian-couplings are used.

​

  • Pronounced finite-size effects on the spin system.
  • Better results on classical system.

35 / 19

36 of 37

36 / 19

37 of 37

37 / 19