Lecture 11 transitions and dissipation
- Qubit: single state in two level well.
- Hamiltonian diagonal is the energy shift and Delta is the tunneling.
- Wave functions correspond to the two energy states.
- Qubit has stationary states, reversibility, delocalization, which are not
compatible with dissipation.
- Spin-boson model Hamiltonian: qubit + environment + coupling.
- H_coupling = -q_0*sigma_z*F_hat. Sigma_z*q_0 can be seen as the
displacement of the particle. When moving it. We will excite the
environment to interact with this friction force.
- Response function is response to displacing particle with q(omega),
and there is a response to this which is the friction force.
- Analogy of friction force is voltage over a capacitor: V(omega) =
Z(omega)I(omega).
- Response function is i*omega*Z(omega). So the real part of the
impedance is the response function.
- Shifted oscillators: disregard tunneling. We would like to get a
Hamiltonian which is quadratic in bosonic operators, but now we have
a quadratic term with a linear term. You can shift the bosonic operators
to get the quadratic form back. Now you get shifted bosonic operators
with: , plus a constant term.
-
- Vacuum minus is a coherent state with respect to bosons in plus
position and vice versa. Left vacuum has different fractions of all
possible states of right vacuum.
- Probability to get energy E from shake up by hand given by
delta_peaks. But if we have many modes (infinitely many) we have to
sum the probabilities up and take continuous limit. ,
probability to shake up no oscillator.
- Orthogonality catastrophe: for any finite number of boson modes this
probability stays finite. However, when we go to the continuous limit,
the integral corresponding to the sum over modes could diverge,
resulting in zero probability of a quiet shake up. P(0) is the square of
two vacua. If it is zero, the overlap vanishes, and the two vacua are
orthogonal. So, they do not know each other or mix with each other.
Divergence of low frequencies if Im(X(omega)) approaches zero at
omega going to zero slower than omega.
- P_1(E) is probabilty to emit one boson in the limit of small q_0
(coupling):
- Switch tunneling -> energy difference epsilon goes to environment.
- Becomes P_1(E) if one photon process dominates.
- Two types of P(E) function, so probability of energy loss: orthogonal
vacuum, P(0) = 0 and P(E) is small at low energies (classical), elastic
tunneling: coupling is small (exponentially supressed), so big
probability of delta peak to channel without energy loss (quantum),
due to single boson emissions, and third type is borderline (power law)
(?).
- How P_1 goes at low energies: ,
if s < -1 -> subohmic, s > -1 -> superohmic, s = -1 -> ohmic.
Probability for ohmic is constant over E.
- Subohmic: tranisitions are rare and require energy, localized in one of
the wells.
- Ohmic: P(E) goes with E^{2*alpha - 1}. Rate goes with
epsilon^{2*alpha - 1}. It is a many boson process.
- Rest do via book.
Lecture 12 relativistic quantum mechanics
- We live in a Minkowski space with one point in space being an event
(x,t).
- , with
delta_x_4 being ic*delta_T.
- Transformation from one reference system moving with certain
velocity to another is given by matrix. It is a linear transformation,
rotation.
- 4 vector of momentum is invariant: (p,iE/c). Two times the 4 gradient
gives the scalar d’Alembertian.
-
