Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

Lecture 2.6: Time Evolution and Rabi Oscillations

Purdue University

Review: The Bloch Sphere and Rotations

Qubit States on the Bloch Sphere

Any qubit state can be written as:

ψ=cosθ20+eiϕsinθ21|\psi\rangle = \cos\frac{\theta}{2}|0\rangle + e^{i\phi}\sin\frac{\theta}{2}|1\rangle

where θ\theta and ϕ\phi are the polar and azimuthal angles on the Bloch sphere.

The six cardinal states are:

Rotations: SU(2)

Last lecture we learned that SU(2) is the group of rotations on the Bloch sphere.

Rotation about the z-axis:

Rz(θ)=eiσzθ/2=cosθ2Iisinθ2σzR_z(\theta) = e^{-i\sigma_z\theta/2} = \cos\frac{\theta}{2}\,I - i\sin\frac{\theta}{2}\,\sigma_z

where σz=(1001)\sigma_z = \begin{pmatrix} 1 & 0 \\ 0 & -1 \end{pmatrix}.

Rotation about an arbitrary axis n^\hat{n}:

Rn^(θ)=eiθ(n^σ)/2=cosθ2Iisinθ2(n^σ)R_{\hat{n}}(\theta) = e^{-i\theta(\hat{n}\cdot\vec{\sigma})/2} = \cos\frac{\theta}{2}\,I - i\sin\frac{\theta}{2}\,(\hat{n}\cdot\vec{\sigma})

where n^σ=nxσx+nyσy+nzσz\hat{n}\cdot\vec{\sigma} = n_x\sigma_x + n_y\sigma_y + n_z\sigma_z.


iClicker Question 1

Consider the rotation Rz(θ)=eiσzθ/2R_z(\theta) = e^{-i\sigma_z\theta/2}.

Which pair of states are unchanged by this rotation (up to a global phase)?

Answer: (C). The states 0|0\rangle and 1|1\rangle are eigenstates of σz\sigma_z, so they only pick up a phase under Rz(θ)R_z(\theta):

Rz(θ)0=eiθ/20,Rz(θ)1=e+iθ/21R_z(\theta)|0\rangle = e^{-i\theta/2}|0\rangle, \qquad R_z(\theta)|1\rangle = e^{+i\theta/2}|1\rangle

States on the equator (+|+\rangle, |-\rangle, +i|+i\rangle, i|-i\rangle) are not eigenstates of σz\sigma_z — they precess around the z-axis.

General principle: Eigenstates of a generator are invariant under the transformation it generates.


The Stern-Gerlach Experiment: Discovering Spin

The 1922 Experiment

Before we write down the spin Hamiltonian, let’s see where it comes from experimentally.

In 1922, Otto Stern and Walther Gerlach performed one of the most important experiments in quantum mechanics. They sent a beam of silver atoms through an inhomogeneous magnetic field and observed where the atoms landed on a detector screen.

Classical expectation: A magnetic dipole in a non-uniform field feels a force proportional to its component along the field gradient. If the atomic magnetic moments were oriented randomly (like tiny compass needles pointing in all directions), the beam should spread into a continuous smear.

What they observed: The beam split into exactly two discrete spots.

Why Deflection Measures Spin

The key physics is straightforward:

  1. Energy of a dipole in a field: A magnetic moment μ\vec{\mu} in a magnetic field has energy

    U=μBU = -\vec{\mu}\cdot\vec{B}
  2. Force from a gradient: Force is the negative gradient of energy. For a field pointing in the z-direction with a gradient:

    Fz=Uz=μzBzzF_z = -\frac{\partial U}{\partial z} = \mu_z\frac{\partial B_z}{\partial z}
  3. Magnetic moment comes from spin: For an electron, μz=γSz\mu_z = \gamma S_z where γ\gamma is the gyromagnetic ratio.

  4. Putting it together:

    Fz=γSzBzzF_z = \gamma S_z \frac{\partial B_z}{\partial z}

The force is directly proportional to SzS_z. Classically, SzS_z could be anything, giving a continuous range of forces and a smeared spot. But quantum mechanics says SzS_z can only be +/2+\hbar/2 or /2-\hbar/2 — nothing in between.

Two allowed values of spin → two allowed forces → two spots.

This was shocking. The atoms behaved as if their magnetic moment could only point in two directions — “up” or “down” relative to the field — with nothing in between.

The Interpretation: Spin-1/2

We now understand that:

  1. Silver has one unpaired electron in its outer shell (configuration: [Kr] 4d¹⁰ 5s¹)

  2. The electron has an intrinsic angular momentum called spin, with quantum number s=1/2s = 1/2

  3. When measured along any axis, spin-1/2 can only yield two outcomes: +/2+\hbar/2 (“spin up”) or /2-\hbar/2 (“spin down”)

  4. The two spots correspond to these two eigenstates

The Stern-Gerlach apparatus is a qubit measurement device. It projects the electron spin onto the 0|0\rangle or 1|1\rangle state (relative to the field direction) and physically separates them.

Why This Matters for Us

The Stern-Gerlach experiment shows that:

When we write ψ=c00+c11|\psi\rangle = c_0|0\rangle + c_1|1\rangle and talk about the Bloch sphere, we’re describing something real: the quantum state of an electron spin.


Back to Physics: Spin in a Magnetic Field

The Physical Setup

An electron (or any spin-1/2 particle) in a magnetic field B\vec{B} has Hamiltonian:

H=γSB=γ2σBH = -\gamma\vec{S}\cdot\vec{B} = -\frac{\gamma\hbar}{2}\vec{\sigma}\cdot\vec{B}

where γ\gamma is the gyromagnetic ratio.

The spin precesses around the magnetic field direction — this is the quantum analog of a gyroscope.

The Question

Given an initial state ψ(0)|\psi(0)\rangle, how do we find ψ(t)|\psi(t)\rangle?


The Schrödinger Equation

The Fundamental Law

Quantum states evolve according to:

Hψ(t)=itψ(t)H|\psi(t)\rangle = i\hbar\frac{\partial}{\partial t}|\psi(t)\rangle

Solving with the Time Evolution Operator

Define the time evolution operator U(t)U(t) by:

U(t)ψ(0)=ψ(t)U(t)|\psi(0)\rangle = |\psi(t)\rangle

Substituting into the Schrödinger equation:

HU(t)ψ(0)=itU(t)ψ(0)H\,U(t)|\psi(0)\rangle = i\hbar\frac{\partial}{\partial t}U(t)|\psi(0)\rangle

Since this must hold for any initial state:

HU(t)=iUtH\,U(t) = i\hbar\frac{\partial U}{\partial t}

The solution is:

U(t)=eiHt/\boxed{U(t) = e^{-iHt/\hbar}}

You can verify this solves the equation: teiHt/=iHeiHt/\frac{\partial}{\partial t}e^{-iHt/\hbar} = \frac{-iH}{\hbar}e^{-iHt/\hbar}.

The Hamiltonian as a Generator

In the infinitesimal limit (Δt\Delta t small):

U(Δt)IiHΔtU(\Delta t) \approx I - \frac{iH}{\hbar}\Delta t

Compare to our rotation formula: R(δθ)I+δθGR(\delta\theta) \approx I + \delta\theta\, G.

The Hamiltonian is the generator of time evolution.

Just as the Pauli matrices generate rotations in space, the Hamiltonian generates translations in time.


U(t)U(t) is Unitary

Why Unitarity Matters

We start with a normalized state: ψ(0)ψ(0)=1\langle\psi(0)|\psi(0)\rangle = 1.

We need it to stay normalized:

ψ(t)ψ(t)=ψ(0)UUψ(0)\langle\psi(t)|\psi(t)\rangle = \langle\psi(0)|U^\dagger U|\psi(0)\rangle

For this to equal 1 for any initial state, we need:

UU=I\boxed{U^\dagger U = I}

This is the definition of a unitary operator.

Verification

For U(t)=eiHt/U(t) = e^{-iHt/\hbar} with Hermitian HH (i.e., H=HH^\dagger = H):

U=(eiHt/)=e+iHt/=e+iHt/U^\dagger = \left(e^{-iHt/\hbar}\right)^\dagger = e^{+iH^\dagger t/\hbar} = e^{+iHt/\hbar}
UU=e+iHt/eiHt/=e0=IU^\dagger U = e^{+iHt/\hbar}e^{-iHt/\hbar} = e^0 = I \quad \checkmark

Unitarity guarantees probability conservation.


iClicker Questions

iClicker A: The Generator Pattern

The Hamiltonian generates time evolution: eiHt/ψ(0)=ψ(t)e^{-iHt/\hbar}|\psi(0)\rangle = |\psi(t)\rangle.

What does the operator eip^x0/e^{-i\hat{p}x_0/\hbar} do to a wavefunction ψ(x)\psi(x)?

Answer: (B). Just as the Hamiltonian HH generates time translations, the momentum operator p^\hat{p} generates spatial translations:

eip^x0/ψ(x)=ψ(x+x0)e^{-i\hat{p}x_0/\hbar}\psi(x) = \psi(x + x_0)

This is the same pattern: Hermitian operator in the exponent -> unitary transformation that “moves” the state.


The Generator Pattern

We’ve now seen the same mathematical structure three times:

TransformationUnitary OperatorHermitian Generator
Spin rotationRz(θ)=eiσzθ/2R_z(\theta) = e^{-i\sigma_z\theta/2}σz\sigma_z (and σx,σy\sigma_x, \sigma_y)
Time evolutionU(t)=eiHt/U(t) = e^{-iHt/\hbar}HH (Hamiltonian)
Space translationT(x)=eip^x/T(x) = e^{-i\hat{p}x/\hbar}p^=ix\hat{p} = -i\hbar\frac{\partial}{\partial x}

The pattern: Hermitian operators generate unitary transformations through exponentiation.

This is one of the deepest ideas in physics: symmetries (unitary transformations) are generated by observables (Hermitian operators).


Example 1: Magnetic Field Along z — The Phase Gate

Setup

B=Bzz^\vec{B} = B_z\hat{z}
H=γBz2σz=ω02σzH = -\gamma B_z \frac{\hbar}{2}\sigma_z = \frac{\hbar\omega_0}{2}\sigma_z

where ω0=γBz\omega_0 = -\gamma B_z is the Larmor frequency.

Time Evolution

U(t)=eiHt/=eiω0tσz/2=Rz(ω0t)U(t) = e^{-iHt/\hbar} = e^{-i\omega_0 t\,\sigma_z/2} = R_z(\omega_0 t)

This is rotation around the z-axis at angular frequency ω0\omega_0.

What Happens to Different States?

Starting in 0|0\rangle:

U(t)0=eiω0t/20U(t)|0\rangle = e^{-i\omega_0 t/2}|0\rangle

Just a global phase — the state doesn’t move on the Bloch sphere.

Starting in +|+\rangle:

U(t)+=12(eiω0t/20+e+iω0t/21)U(t)|+\rangle = \frac{1}{\sqrt{2}}\left(e^{-i\omega_0 t/2}|0\rangle + e^{+i\omega_0 t/2}|1\rangle\right)

The relative phase changes — the state precesses around the equator.

This is precession: the quantum version of a spinning top in a gravitational field.


Example 2: Magnetic Field Along x — Rabi Oscillations

Setup

B=Bxx^\vec{B} = B_x\hat{x}
H=Ω2σxH = \frac{\hbar\Omega}{2}\sigma_x

where Ω\Omega is the Rabi frequency (proportional to BxB_x).

Time Evolution

U(t)=eiΩtσx/2=Rx(Ωt)U(t) = e^{-i\Omega t\,\sigma_x/2} = R_x(\Omega t)

Using the rotation formula:

Rx(θ)=cosθ2Iisinθ2σx=(cosθ2isinθ2isinθ2cosθ2)R_x(\theta) = \cos\frac{\theta}{2}\,I - i\sin\frac{\theta}{2}\,\sigma_x = \begin{pmatrix} \cos\frac{\theta}{2} & -i\sin\frac{\theta}{2} \\ -i\sin\frac{\theta}{2} & \cos\frac{\theta}{2} \end{pmatrix}

Evolution of 0|0\rangle

Starting in the ground state 0|0\rangle:

ψ(t)=U(t)0=(cosΩt2isinΩt2)=cosΩt20isinΩt21|\psi(t)\rangle = U(t)|0\rangle = \begin{pmatrix} \cos\frac{\Omega t}{2} \\ -i\sin\frac{\Omega t}{2} \end{pmatrix} = \cos\frac{\Omega t}{2}|0\rangle - i\sin\frac{\Omega t}{2}|1\rangle

The Probabilities Oscillate!

P0(t)=cos2Ωt2P_0(t) = \cos^2\frac{\Omega t}{2}
P1(t)=sin2Ωt2P_1(t) = \sin^2\frac{\Omega t}{2}

The population oscillates between 0|0\rangle and 1|1\rangle. This is fundamentally different from precession (Example 1), where the populations stayed constant.

Special Pulses

π-pulse (Ωt=π\Omega t = \pi):

0π-pulsei11|0\rangle \xrightarrow{\pi\text{-pulse}} -i|1\rangle \equiv |1\rangle

Complete population inversion — the qubit flips from ground to excited state.

π/2-pulse (Ωt=π/2\Omega t = \pi/2):

0π/2-pulse12(0i1)|0\rangle \xrightarrow{\pi/2\text{-pulse}} \frac{1}{\sqrt{2}}(|0\rangle - i|1\rangle)

Creates an equal superposition — the state moves from the pole to the equator.


Eigenstates: States That Don’t Change

Definition

An eigenstate of the Hamiltonian is a state that evolves only by a global phase:

Hψn=EnψnH|\psi_n\rangle = E_n|\psi_n\rangle

Under time evolution:

eiHt/ψn=eiEnt/ψne^{-iHt/\hbar}|\psi_n\rangle = e^{-iE_n t/\hbar}|\psi_n\rangle

The state just picks up a phase factor — it doesn’t move on the Bloch sphere.

Examples

For H=ω02σzH = \frac{\hbar\omega_0}{2}\sigma_z:

For H=Ω2σxH = \frac{\hbar\Omega}{2}\sigma_x:

General rule: Eigenstates of the Hamiltonian are stationary. Superpositions of different energy eigenstates oscillate.


Qiskit Demonstration: Rabi Oscillations

import numpy as np
import matplotlib.pyplot as plt
from qiskit import QuantumCircuit
from qiskit.quantum_info import Statevector

# Rabi oscillations: Rx(θ) applied to |0⟩
thetas = np.linspace(0, 4*np.pi, 100)
P0 = []
P1 = []

for theta in thetas:
    qc = QuantumCircuit(1)
    qc.rx(theta, 0)  # Rotation around x-axis by angle theta
    state = Statevector(qc)
    probs = state.probabilities()
    P0.append(probs[0])
    P1.append(probs[1])

# Plot
plt.figure(figsize=(10, 5))
plt.plot(thetas/np.pi, P0, 'b-', linewidth=2, label='$P_0(t) = \\cos^2(\\Omega t/2)$')
plt.plot(thetas/np.pi, P1, 'r-', linewidth=2, label='$P_1(t) = \\sin^2(\\Omega t/2)$')
plt.xlabel('$\\Omega t / \\pi$', fontsize=14)
plt.ylabel('Probability', fontsize=14)
plt.title('Rabi Oscillations: Population Transfer Between $|0\\rangle$ and $|1\\rangle$', fontsize=14)
plt.legend(fontsize=12)
plt.grid(True, alpha=0.3)
plt.axhline(y=0.5, color='gray', linestyle='--', alpha=0.5)
plt.axvline(x=1, color='green', linestyle=':', alpha=0.7, label='π-pulse')
plt.axvline(x=0.5, color='orange', linestyle=':', alpha=0.7, label='π/2-pulse')
plt.show()

Summary

  1. Time evolution is governed by the Schrödinger equation: ψ(t)=eiHt/ψ(0)|\psi(t)\rangle = e^{-iHt/\hbar}|\psi(0)\rangle

  2. The Hamiltonian generates time evolution, just as Pauli matrices generate spatial rotations.

  3. Unitarity (UU=IU^\dagger U = I) guarantees probability conservation.

  4. Precession (HσzH \propto \sigma_z): States rotate around the z-axis. Poles are stationary; equator precesses.

  5. Rabi oscillations (HσxH \propto \sigma_x): Population oscillates between 0|0\rangle and 1|1\rangle. π-pulse inverts; π/2-pulse creates superposition.

  6. Eigenstates of the Hamiltonian are stationary — they evolve only by a global phase.

iClicker E: Energy Eigenstates

A qubit has Hamiltonian H=ω2σzH = \frac{\hbar\omega}{2}\sigma_z, so the energy eigenstates are 0|0\rangle (energy +ω/2+\hbar\omega/2) and 1|1\rangle (energy ω/2-\hbar\omega/2).

The state +=12(0+1)|+\rangle = \frac{1}{\sqrt{2}}(|0\rangle + |1\rangle) is prepared at t=0t=0. At time tt, the state is:

Answer: (B). Energy eigenstates each pick up their own phase factor eiEnt/e^{-iE_n t/\hbar}. Since +|+\rangle is a superposition of two energy eigenstates with different energies, the relative phase changes with time. The state precesses around the z-axis.


iClicker F: What Doesn’t Change?

Under the Hamiltonian H=Ω2σxH = \frac{\hbar\Omega}{2}\sigma_x, which quantity is CONSTANT in time?

Answer: (B). The Hamiltonian HσxH \propto \sigma_x generates rotations around the x-axis. Under such rotations, the x-component of any vector is unchanged — only the y and z components rotate into each other.

More formally: [H,σx]=[σx,σx]=0[H, \sigma_x] = [\sigma_x, \sigma_x] = 0, so σx\langle\sigma_x\rangle is a conserved quantity.


Homework 2.6


Problem: Time Evolution Practice

A qubit has Hamiltonian H=ω2σzH = \frac{\hbar\omega}{2}\sigma_z.

(a) Write the time evolution operator U(t)=eiHt/U(t) = e^{-iHt/\hbar} as an explicit 2×22\times 2 matrix.

(b) If the initial state is 0|0\rangle, find ψ(t)|\psi(t)\rangle. Does the state change physically, or just by a global phase?

(c) If the initial state is +=12(0+1)|+\rangle = \frac{1}{\sqrt{2}}(|0\rangle + |1\rangle), find ψ(t)|\psi(t)\rangle.

(d) For part (c), compute P0(t)=0ψ(t)2P_0(t) = |\langle 0|\psi(t)\rangle|^2 and P1(t)=1ψ(t)2P_1(t) = |\langle 1|\psi(t)\rangle|^2. Do the populations change with time?

(e) For part (c), compute P+(t)=+ψ(t)2P_+(t) = |\langle +|\psi(t)\rangle|^2. At what time tt does the state first become orthogonal to +|+\rangle? What state is it at that moment?


Problem: Rabi Oscillations

A qubit has Hamiltonian H=Ω2σxH = \frac{\hbar\Omega}{2}\sigma_x and starts in state 0|0\rangle.

(a) Using Rx(θ)=cosθ2Iisinθ2σxR_x(\theta) = \cos\frac{\theta}{2}I - i\sin\frac{\theta}{2}\sigma_x, show that:

ψ(t)=cosΩt20isinΩt21|\psi(t)\rangle = \cos\frac{\Omega t}{2}|0\rangle - i\sin\frac{\Omega t}{2}|1\rangle

(b) Compute P1(t)=1ψ(t)2P_1(t) = |\langle 1|\psi(t)\rangle|^2. Sketch this function from t=0t=0 to t=4π/Ωt = 4\pi/\Omega.

(c) At what time does the qubit first have a 50% probability of being in 1|1\rangle? What is this pulse called?

(d) At what time does the qubit first reach 1|1\rangle with certainty? What is this pulse called?

(e) If Ω=2π×10\Omega = 2\pi \times 10 MHz, calculate the duration of a π-pulse in nanoseconds.

Problem: The Hadamard Gate

The Hadamard gate is one of the most important single-qubit gates:

H=12(1111)H = \frac{1}{\sqrt{2}}\begin{pmatrix} 1 & 1 \\ 1 & -1 \end{pmatrix}

(a) Verify that HH is unitary: compute HHH^\dagger H and show it equals II.

(b) Verify that H2=IH^2 = I. What does this tell you geometrically about the rotation?

(c) The general rotation formula is:

Rn^(θ)=cosθ2Iisinθ2(n^σ)R_{\hat{n}}(\theta) = \cos\frac{\theta}{2}I - i\sin\frac{\theta}{2}(\hat{n}\cdot\vec{\sigma})

For H2=IH^2 = I, we need θ=π\theta = \pi (a 180° rotation). So HH should equal i(n^σ)-i(\hat{n}\cdot\vec{\sigma}) for some axis n^\hat{n}, up to a global phase.

Show that:

H=12(σx+σz)H = \frac{1}{\sqrt{2}}(\sigma_x + \sigma_z)

(d) From part (c), what is the rotation axis n^\hat{n}? Express it as a unit vector.

(e) Describe in words: where is this axis on the Bloch sphere? (Hint: it’s in the xz-plane, halfway between...)

(f) The Hadamard swaps the z-basis and x-basis:

Verify one of these by explicit matrix multiplication.

(g) Why do you think the Hadamard gate is preferred over, say, Ry(π/2)R_y(\pi/2) for creating superpositions? (Hint: look at the matrix elements of each.)


Problem: Unitarity Implies Hermitian Generator

This problem proves the deep connection between unitary operators and Hermitian generators.

(a) Let U=eiGU = e^{iG} for some operator GG. Write an expression for UU^\dagger in terms of GG^\dagger.

(b) The condition for UU to be unitary is UU=IU^\dagger U = I. Substitute your expression from (a) and show that this requires:

eiGeiG=Ie^{-iG^\dagger}e^{iG} = I

(c) For the equation in (b) to hold, we need G=GG^\dagger = G. Explain why.

Hint: If AA and BB are matrices, eAeB=eA+Be^A e^B = e^{A+B} only when [A,B]=0[A,B] = 0. What does eiGeiG=I=e0e^{-iG^\dagger}e^{iG} = I = e^0 tell you?

(d) Conclude: If U=eiGU = e^{iG} is unitary, then GG must be Hermitian.

(e) We’ve been writing time evolution as U(t)=eiHt/U(t) = e^{-iHt/\hbar}. Using your result, explain why the Hamiltonian HH must be Hermitian for time evolution to be unitary.

(f) The Pauli matrices σx\sigma_x, σy\sigma_y, σz\sigma_z are Hermitian. Verify that σx=σx\sigma_x^\dagger = \sigma_x by writing out the matrix and its conjugate transpose.

(g) Using your result from (d), explain why Rz(θ)=eiσzθ/2R_z(\theta) = e^{-i\sigma_z\theta/2} is guaranteed to be unitary without doing any explicit calculation.


Problem: Space Translation Operator

The momentum operator in one dimension is p^=iddx\hat{p} = -i\hbar\frac{d}{dx}.

(a) Consider the translation operator T(a)=eip^a/T(a) = e^{-i\hat{p}a/\hbar}. Using p^=iddx\hat{p} = -i\hbar\frac{d}{dx}, show that:

T(a)=eaddxT(a) = e^{-a\frac{d}{dx}}

(b) Taylor expand eaddxe^{-a\frac{d}{dx}} acting on a function ψ(x)\psi(x):

eaddxψ(x)=(1addx+a22!d2dx2)ψ(x)e^{-a\frac{d}{dx}}\psi(x) = \left(1 - a\frac{d}{dx} + \frac{a^2}{2!}\frac{d^2}{dx^2} - \cdots\right)\psi(x)

(c) Recognize this Taylor series as the expansion of ψ(xa)\psi(x-a) around xx. Conclude:

T(a)ψ(x)=ψ(xa)T(a)\psi(x) = \psi(x - a)

(Note: The sign convention varies. With T(a)=e+ip^a/T(a) = e^{+i\hat{p}a/\hbar}, you get ψ(x+a)\psi(x+a).)

(d) Verify that T(a)T(a) is unitary using the result from the previous problem.