24
Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

  • View
    214

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Computational PhysicsQuantum Monte Carlo methods

Dr. Guy Tel-Zur

Page 2: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

References

• MHJ Chapter 11• Giordano & Nakanishi Chapter 10• R. Landau et al Chapter 15

Page 3: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Many Body System

Schrödinger’s equation

for a system of A nucleons (A = N + Z), N being the number of neutrons and Z the number of protons). There are:

coupled second-order differential equations in 3A dimensions. For a nucleus like 10Be this number is 215040. This is a truly challenging many-body problem.

Page 4: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Eq. (11.1) is a multidimensional integral. As such, Monte Carlo methods are ideal for obtaining expectation values of quantum mechanical operators.

Our problem is that we do not know the exact Wave function Psi(r1, .., rA, α1, .., αN).

Our goal in this chapter is to solve the problem using Variational Monte Carlo approach to quantum mechanics.

We limit the attention to the simple Metropolis algorithm, without the inclusion of importance sampling. Importance sampling and diffusion Monte Carlo methods are discussed in chapters18 and 16.

Before that we give a short review about Quantum Mechanics in the next slides

Page 5: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Postulates of Quantum Mechanics

Schrödinger’s equation for a one-dimensional one body problem:

if we perform a rotation of time into the complex plane, using τ = it/hbar, the time-dependent Schrödinger equation becomes

With V = 0 we have a diffusion equation in complex time!

Page 6: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

The diffusion constant:

The wave function have to satisfy:

Page 7: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 8: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 9: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 10: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 11: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

As a result of Postulate III:

Page 12: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

We could briefly summarize the above variational procedure in the following three steps.

Page 13: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Variational Monte Carlo for quantum mechanical systems

The quantum PDF:

Page 14: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 15: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 16: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

We proceed with the Metropilis algorithm in a similar way as we did with the Ising Model:

Page 17: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 18: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

An illustration of variational Monte Carlo methods

Page 19: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Recall Eq. 11.16:

Page 20: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur
Page 21: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Variational Monte Carlo for atoms

Page 22: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

The Born-Oppenheimer Approximation

The separation of the electronic and nuclear degrees of freedom is known as the Born-Oppenheimer approximation

The first term of Eq. (11.41) represents the kinetic energy operator of the center of mass. The second term represents the sum of the kinetic energy operators of the N electrons, each of them having their mass m replaced by the reduced mass μ = mM/(m + M) because of the motion of the nucleus. The nuclear motion is also responsible for the third term, or the mass polarization term.

Page 23: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

Taking the limit M → ∞ in Eq. (11.41), the kinetic energy operator reduces to

The inter-electronic potentials are the main problem in atomic physics. Because ofthese terms, the Hamiltonian cannot be separated into one-particle parts, and the problem must be solved as a whole. A common approximation is to regard the effects of the electron-electron interactions either as averaged over the domain or by means of introducing a density functional, such as by Hartree-Fock (HF) or Density Functional Theory (DFT).

Page 24: Computational Physics Quantum Monte Carlo methods Dr. Guy Tel-Zur

The next topics :

• The Hydrogen atom• The Helium atom