BREAKING NEWS

## Summary

In physics, the Rabi cycle (or Rabi flop) is the cyclic behaviour of a two-level quantum system in the presence of an oscillatory driving field. A great variety of physical processes belonging to the areas of quantum computing, condensed matter, atomic and molecular physics, and nuclear and particle physics can be conveniently studied in terms of two-level quantum mechanical systems, and exhibit Rabi flopping when coupled to an oscillatory driving field. The effect is important in quantum optics, magnetic resonance and quantum computing, and is named after Isidor Isaac Rabi. Rabi oscillations, showing the probability of a two-level system initially in $|1\rangle$ to end up in $|2\rangle$ at different detunings Δ.

A two-level system is one that has two possible energy levels. These two levels are a ground state with lower energy and an excited state with higher energy. If the energy levels are not degenerate (i.e. not having equal energies), the system can absorb a quantum of energy and transition from the ground state to the "excited" state. When an atom (or some other two-level system) is illuminated by a coherent beam of photons, it will cyclically absorb photons and re-emit them by stimulated emission. One such cycle is called a Rabi cycle, and the inverse of its duration is the Rabi frequency of the photon beam. The effect can be modeled using the Jaynes–Cummings model and the Bloch vector formalism.

## Mathematical treatment

A detailed mathematical description of the effect can be found on the page for the Rabi problem. For example, for a two-state atom (an atom in which an electron can either be in the excited or ground state) in an electromagnetic field with frequency tuned to the excitation energy, the probability of finding the atom in the excited state is found from the Bloch equations to be

$|c_{b}(t)|^{2}\propto \sin ^{2}(\omega t/2),$

where $\omega$  is the Rabi frequency.

More generally, one can consider a system where the two levels under consideration are not energy eigenstates. Therefore, if the system is initialized in one of these levels, time evolution will make the population of each of the levels oscillate with some characteristic frequency, whose angular frequency is also known as the Rabi frequency. The state of a two-state quantum system can be represented as vectors of a two-dimensional complex Hilbert space, which means that every state vector $|\psi \rangle$  is represented by good complex coordinates:

$|\psi \rangle ={\begin{pmatrix}c_{1}\\c_{2}\end{pmatrix}}=c_{1}{\begin{pmatrix}1\\0\end{pmatrix}}+c_{2}{\begin{pmatrix}0\\1\end{pmatrix}},$

where $c_{1}$  and $c_{2}$  are the coordinates.

If the vectors are normalized, $c_{1}$  and $c_{2}$  are related by $|c_{1}|^{2}+|c_{2}|^{2}=1$ . The basis vectors will be represented as $|0\rangle ={\begin{pmatrix}1\\0\end{pmatrix}}$  and $|1\rangle ={\begin{pmatrix}0\\1\end{pmatrix}}$ .

All observable physical quantities associated with this systems are 2 × 2 Hermitian matrices, which means that the Hamiltonian of the system is also a similar matrix.

## Procedure

One can construct an oscillation experiment through the following steps:

1. Prepare the system in a fixed state; for example, $|1\rangle$
2. Let the state evolve freely, under a Hamiltonian H for time t
3. Find the probability $P(t)$ , that the state is in $|1\rangle$

If $|1\rangle$  is an eigenstate of H, $P(t)=1$  and there will be no oscillations. Also if the two states $|0\rangle$  and $|1\rangle$  are degenerate, every state including $|1\rangle$  is an eigenstate of H. As a result, there will be no oscillations.

On the other hand, if H has no degenerate eigenstates, and the initial state is not an eigenstate, then there will be oscillations. The most general form of the Hamiltonian of a two-state system is given

$\mathbf {H} ={\begin{pmatrix}a_{0}+a_{3}&a_{1}-ia_{2}\\a_{1}+ia_{2}&a_{0}-a_{3}\end{pmatrix}}$

here, $a_{0},a_{1},a_{2}$  and $a_{3}$  are real numbers. This matrix can be decomposed as,

$\mathbf {H} =a_{0}\cdot \sigma _{0}+a_{1}\cdot \sigma _{1}+a_{2}\cdot \sigma _{2}+a_{3}\cdot \sigma _{3};$

The matrix $\sigma _{0}$  is the 2 $\times$  2 identity matrix and the matrices $\sigma _{k}\;(k=1,2,3)$  are the Pauli matrices. This decomposition simplifies the analysis of the system especially in the time-independent case where the values of $a_{0},a_{1},a_{2}$  and $a_{3}$ are constants. Consider the case of a spin-1/2 particle in a magnetic field $\mathbf {B} =B\mathbf {\hat {z}}$ . The interaction Hamiltonian for this system is

$\mathbf {H} =-{\boldsymbol {\mu }}\cdot \mathbf {B} =-\gamma \mathbf {S} \cdot \mathbf {B} =-\gamma \ B\ S_{z}$ , $S_{z}={\frac {\hbar }{2}}\,\sigma _{3}={\frac {\hbar }{2}}{\begin{pmatrix}1&0\\0&-1\end{pmatrix}},$

where $\mu$  is the magnitude of the particle's magnetic moment, $\gamma$  is the Gyromagnetic ratio and ${\boldsymbol {\sigma }}$  is the vector of Pauli matrices. Here the eigenstates of Hamiltonian are eigenstates of $\sigma _{3}$ , that is $|0\rangle$  and $|1\rangle$ , with corresponding eigenvalues of $E_{+}={\frac {\hbar }{2}}\gamma B\ ,\ E_{-}=-{\frac {\hbar }{2}}\gamma B$ . The probability that a system in the state $|\psi \rangle$  can be found in the arbitrary state $|\phi \rangle$  is given by ${|\langle \phi |\psi \rangle |}^{2}$ .

Let the system be prepared in state $\left|+X\right\rangle$  at time $t=0$ . Note that $\left|+X\right\rangle$  is an eigenstate of $\sigma _{1}$ :

$|\psi (0)\rangle ={\frac {1}{\sqrt {2}}}{\begin{pmatrix}1\\1\end{pmatrix}}={\frac {1}{\sqrt {2}}}{\begin{pmatrix}1\\0\end{pmatrix}}+{\frac {1}{\sqrt {2}}}{\begin{pmatrix}0\\1\end{pmatrix}}.$

Here the Hamiltonian is time independent. Thus by solving the stationary Schrödinger equation, the state after time t is given by

$\left|\psi (t)\right\rangle =\exp \left[{\frac {-i\mathbf {H} t}{\hbar }}\right]\left|\psi (0)\right\rangle ={\begin{pmatrix}\exp \left[{\tfrac {-iE_{+}t}{\hbar }}\right]&0\\0&\exp \left[{\tfrac {-iE_{-}t}{\hbar }}\right]\end{pmatrix}}|\psi (0)\rangle ,$

with total energy of the system $E$ . So the state after time t is given by:
$|\psi (t)\rangle =e^{\frac {-iE_{+}t}{\hbar }}{\frac {1}{\sqrt {2}}}|0\rangle +e^{\frac {-iE_{-}t}{\hbar }}{\frac {1}{\sqrt {2}}}|1\rangle$ .

Now suppose the spin is measured in x-direction at time t. The probability of finding spin-up is given by:

${\left|\langle +X|\psi (t)\rangle \right|}^{2}={\left|{\frac {\left\langle 0\right|+\left\langle 1\right|}{\sqrt {2}}}\left({{\frac {1}{\sqrt {2}}}\exp \left[{\frac {-iE_{+}t}{\hbar }}\right]\left|0\right\rangle +{\frac {1}{\sqrt {2}}}\exp \left[{\frac {-iE_{-}t}{\hbar }}\right]\left|1\right\rangle }\right)\right|}^{2}=\cos ^{2}\left({\frac {\omega t}{2}}\right),$

where $\omega$  is a characteristic angular frequency given by $\omega ={\frac {E_{+}-E_{-}}{\hbar }}=\gamma B$ , where it has been assumed that $E_{-}\leq E_{+}$ . So in this case the probability of finding spin-up in x-direction is oscillatory in time $t$  when the system's spin is initially in the $\left|+X\right\rangle$  direction. Similarly, if we measure the spin in the $\left|+Z\right\rangle$ -direction, the probability of measuring spin as ${\tfrac {\hbar }{2}}$  of the system is ${\tfrac {1}{2}}$ . In the degenerate case where $E_{+}=E_{-}$ , the characteristic frequency is 0 and there is no oscillation.

Notice that if a system is in an eigenstate of a given Hamiltonian, the system remains in that state.

This is true even for time dependent Hamiltonians. Taking for example ${\textstyle {\hat {H}}=-\gamma \ S_{z}B\sin(\omega t)}$ ; if the system's initial spin state is $\left|+Y\right\rangle$ , then the probability that a measurement of the spin in the y-direction results in $+{\tfrac {\hbar }{2}}$  at time $t$  is ${\textstyle {\left|\left\langle \,+Y|\psi (t)\right\rangle \right|}^{2}\,=\cos ^{2}\left({\frac {\gamma B}{2\omega }}\cos \left({\omega t}\right)\right)}$ .

## Derivation using nonperturbative procedure by means of the Pauli matrices

Consider a Hamiltonian of the form

${\hat {H}}=E_{0}\cdot \sigma _{0}+W_{1}\cdot \sigma _{1}+W_{2}\cdot \sigma _{2}+\Delta \cdot \sigma _{3}={\begin{pmatrix}E_{0}+\Delta &W_{1}-iW_{2}\\W_{1}+iW_{2}&E_{0}-\Delta \end{pmatrix}}.$

The eigenvalues of this matrix are given by
$\lambda _{+}=E_{+}=E_{0}+{\sqrt {{\Delta }^{2}+{W_{1}}^{2}+{W_{2}}^{2}}}=E_{0}+{\sqrt {{\Delta }^{2}+{\left\vert W\right\vert }^{2}}}$

and
$\lambda _{-}=E_{-}=E_{0}-{\sqrt {{\Delta }^{2}+{W_{1}}^{2}+{W_{2}}^{2}}}=E_{0}-{\sqrt {{\Delta }^{2}+{\left\vert W\right\vert }^{2}}},$

where $\mathbf {W} =W_{1}+\imath W_{2}$  and ${\left\vert W\right\vert }^{2}={W_{1}}^{2}+{W_{2}}^{2}=WW^{*}$ , so we can take $\mathbf {W} ={\left\vert W\right\vert }e^{i\phi }$ .

Now, eigenvectors for $E_{+}$  can be found from equation

${\begin{pmatrix}E_{0}+\Delta &W_{1}-iW_{2}\\W_{1}+iW_{2}&E_{0}-\Delta \end{pmatrix}}{\begin{pmatrix}a\\b\end{pmatrix}}=E_{+}{\begin{pmatrix}a\\b\end{pmatrix}}.$

So

$b=-{\frac {a\left(E_{0}+\Delta -E_{+}\right)}{W_{1}-iW_{2}}}.$

Applying the normalization condition on the eigenvectors, ${\left\vert a\right\vert }^{2}+{\left\vert b\right\vert }^{2}=1$ . So

${\left\vert a\right\vert }^{2}+{\left\vert a\right\vert }^{2}\left({\frac {\Delta }{\left\vert W\right\vert }}-{\frac {\sqrt {{\Delta }^{2}+{\left\vert W\right\vert }^{2}}}{\left\vert W\right\vert }}\right)^{2}=1.$

Let $\sin \theta ={\frac {\left\vert W\right\vert }{\sqrt {{\Delta }^{2}+{\left\vert W\right\vert }^{2}}}}$  and $\cos \theta ={\frac {\Delta }{\sqrt {{\Delta }^{2}+{\left\vert W\right\vert }^{2}}}}$ . So $\tan \theta ={\frac {\left\vert W\right\vert }{\Delta }}$ .

So we get ${\textstyle {\left\vert a\right\vert }^{2}+{\left\vert a\right\vert }^{2}{\frac {({1-\cos \theta })^{2}}{\sin ^{2}\theta }}=1}$ . That is ${\left\vert a\right\vert }^{2}=\cos ^{2}\left({\tfrac {\theta }{2}}\right)$ , using the identity ${\textstyle \tan({\tfrac {\theta }{2}})={\tfrac {1-\cos(\theta )}{\sin(\theta )}}}$ . Taking arbitrary phase angle $\phi$ , we can write $a=\exp(i\phi /2)\cos \left({\tfrac {\theta }{2}}\right)$ . Similarly $b=\exp(-i\phi /2)\sin \left({\tfrac {\theta }{2}}\right)$ .

So the eigenvector for the eigenvalue $E_{+}$  is given by

$\left|E_{+}\right\rangle ={\begin{pmatrix}\cos \left(\theta /2\right)\\e^{\imath \phi }\sin \left(\theta /2\right)\end{pmatrix}}.$

As overall phase is immaterial, we can write

$\left|E_{+}\right\rangle ={\begin{pmatrix}\cos \left({\tfrac {\theta }{2}}\right)\\e^{i\phi }\sin \left({\tfrac {\theta }{2}}\right)\end{pmatrix}}=\cos \left({\tfrac {\theta }{2}}\right)\left|0\right\rangle +e^{i\phi }\sin \left({\tfrac {\theta }{2}}\right)\left|1\right\rangle .$

Similarly, the eigenvector for eigenenergy

$E_{-}$

is
$\left|E_{-}\right\rangle =\sin \left({\tfrac {\theta }{2}}\right)\left|0\right\rangle -e^{i\phi }\cos \left({\tfrac {\theta }{2}}\right)\left|1\right\rangle .$

From these two equations, we can write

$\left|0\right\rangle =\cos \left({\tfrac {\theta }{2}}\right)\left|E_{+}\right\rangle +\sin \left({\tfrac {\theta }{2}}\right)\left|E_{-}\right\rangle$

and
$\left|1\right\rangle =e^{-\imath \phi }\sin \left({\tfrac {\theta }{2}}\right)\left|E_{+}\right\rangle -e^{-\imath \phi }\cos \left({\tfrac {\theta }{2}}\right)\left|E_{-}\right\rangle .$

Suppose the system starts in state $|0\rangle$  at time ${\textstyle t=0}$ ; that is,

$\left|\psi \left(0\right)\right\rangle =\left|0\right\rangle =\cos \left({\tfrac {\theta }{2}}\right)\left|E_{+}\right\rangle +\sin \left({\tfrac {\theta }{2}}\right)\left|E_{-}\right\rangle .$

For a time-independent Hamiltonian, after time t, the state evolves as

$\left|\psi \left(t\right)\right\rangle =e^{\frac {-i{\hat {H}}t}{\hbar }}\left|\psi \left(0\right)\right\rangle =\cos \left({\tfrac {\theta }{2}}\right)e^{\frac {-iE_{+}t}{\hbar }}\left|E_{+}\right\rangle +\sin \left({\tfrac {\theta }{2}}\right)e^{\frac {-iE_{-}t}{\hbar }}\left|E_{-}\right\rangle .$

If the system is in one of the eigenstates $|E_{+}\rangle$  or $|E_{-}\rangle$ , it will remain the same state. However, for a time-dependent Hamiltonian and a general initial state as shown above, the time evolution is non trivial. The resulting formula for the Rabi oscillation is valid because the state of the spin may be viewed in a reference frame that rotates along with the field.

The probability amplitude of finding the system at time t in the state $|1\rangle$  is given by ${\textstyle \left\langle \ 1|\psi (t)\right\rangle =e^{i\phi }\sin \left({\tfrac {\theta }{2}}\right)\cos \left({\tfrac {\theta }{2}}\right)\left(e^{\frac {-iE_{+}t}{\hbar }}-e^{\frac {-iE_{-}t}{\hbar }}\right)}$ .

Now the probability that a system in the state $|\psi (t)\rangle$  will be found to be in the arbitrary state

$|1\rangle$

is given by
{\begin{aligned}P_{0\to 1}(t)&={|\langle \ 1|\psi (t)\rangle |}^{2}\\&=e^{-\imath \phi }\sin \left({\frac {\theta }{2}}\right)\cos \left({\frac {\theta }{2}}\right)\left(e^{\frac {+\imath E_{+}t}{\hbar }}-e^{\frac {+\imath E_{-}t}{\hbar }}\right)e^{+\imath \phi }\sin \left({\frac {\theta }{2}}\right)\cos \left({\frac {\theta }{2}}\right)\left(e^{\frac {-\imath E_{+}t}{\hbar }}-e^{\frac {-\imath E_{-}t}{\hbar }}\right)\\&={\frac {\sin ^{2}{\theta }}{4}}\left(2-2\cos \left({\frac {\left(E_{+}-E_{-}\right)t}{\hbar }}\right)\right)\end{aligned}}

This can be simplified to

$P_{0\to 1}(t)=\sin ^{2}(\theta )\sin ^{2}\left({\frac {(E_{+}-E_{-})t}{2\hbar }}\right)={\frac {{\left\vert W\right\vert }^{2}}{\Delta ^{2}+{\left\vert W\right\vert }^{2}}}\sin ^{2}\left({\frac {(E_{+}-E_{-})t}{2\hbar }}\right)$

(1)

This shows that there is a finite probability of finding the system in state $|1\rangle$  when the system is originally in the state $|0\rangle$ . The probability is oscillatory with angular frequency $\omega ={\frac {E_{+}-E_{-}}{2\hbar }}={\frac {\sqrt {{\Delta }^{2}+{\left\vert W\right\vert }^{2}}}{\hbar }}$ , which is simply unique Bohr frequency of the system and also called Rabi frequency. The formula (1) is known as Rabi formula. Now after time t the probability that the system in state $|0\rangle$  is given by ${|\langle \ 0|\psi (t)\rangle |}^{2}=1-\sin ^{2}(\theta )\sin ^{2}\left({\frac {(E_{+}-E_{-})t}{2\hbar }}\right)$ , which is also oscillatory.

These types of oscillations of two-level systems are called Rabi oscillations, which arise in many problems such as Neutrino oscillation, the ionized Hydrogen molecule, Quantum computing, Ammonia maser etc.

## In quantum computing

Any two-state quantum system can be used to model a qubit. Consider a spin-${\tfrac {1}{2}}$  system with magnetic moment ${\boldsymbol {\mu }}$  placed in a classical magnetic field ${\boldsymbol {B}}=B_{0}\ {\hat {z}}+B_{1}\left(\cos {(\omega t)}\ {\hat {x}}-\sin {(\omega t)}\ {\hat {y}}\right)$ . Let $\gamma$  be the gyromagnetic ratio for the system. The magnetic moment is thus ${\boldsymbol {\mu }}={\frac {\hbar }{2}}\gamma {\boldsymbol {\sigma }}$ . The Hamiltonian of this system is then given by $\mathbf {H} =-{\boldsymbol {\mu }}\cdot \mathbf {B} =-{\frac {\hbar }{2}}\omega _{0}\sigma _{z}-{\frac {\hbar }{2}}\omega _{1}(\sigma _{x}\cos \omega t-\sigma _{y}\sin \omega t)$  where $\omega _{0}=\gamma B_{0}$  and $\omega _{1}=\gamma B_{1}$ . One can find the eigenvalues and eigenvectors of this Hamiltonian by the above-mentioned procedure. Now, let the qubit be in state $|0\rangle$  at time $t=0$ . Then, at time $t$ , the probability of it being found in state $|1\rangle$  is given by $P_{0\to 1}(t)=\left({\frac {\omega _{1}}{\Omega }}\right)^{2}\sin ^{2}\left({\frac {\Omega t}{2}}\right)$  where $\Omega ={\sqrt {(\omega -\omega _{0})^{2}+\omega _{1}^{2}}}$ . This phenomenon is called Rabi oscillation. Thus, the qubit oscillates between the $|0\rangle$  and $|1\rangle$  states. The maximum amplitude for oscillation is achieved at $\omega =\omega _{0}$ , which is the condition for resonance. At resonance, the transition probability is given by $P_{0\to 1}(t)=\sin ^{2}\left({\frac {\omega _{1}t}{2}}\right)$ . To go from state $|0\rangle$  to state $|1\rangle$  it is sufficient to adjust the time $t$  during which the rotating field acts such that ${\frac {\omega _{1}t}{2}}={\frac {\pi }{2}}$  or $t={\frac {\pi }{\omega _{1}}}$ . This is called a $\pi$  pulse. If a time intermediate between 0 and ${\frac {\pi }{\omega _{1}}}$  is chosen, we obtain a superposition of $|0\rangle$  and $|1\rangle$ . In particular for $t={\frac {\pi }{2\omega _{1}}}$ , we have a ${\frac {\pi }{2}}$  pulse, which acts as: $|0\rangle \to {\frac {|0\rangle +i|1\rangle }{\sqrt {2}}}$ . This operation has crucial importance in quantum computing. The equations are essentially identical in the case of a two level atom in the field of a laser when the generally well satisfied rotating wave approximation is made. Then $\hbar \omega _{0}$  is the energy difference between the two atomic levels, $\omega$  is the frequency of laser wave and Rabi frequency $\omega _{1}$  is proportional to the product of the transition electric dipole moment of atom ${\vec {d}}$  and electric field ${\vec {E}}$  of the laser wave that is $\omega _{1}\propto \hbar \ {\vec {d}}\cdot {\vec {E}}$ . In summary, Rabi oscillations are the basic process used to manipulate qubits. These oscillations are obtained by exposing qubits to periodic electric or magnetic fields during suitably adjusted time intervals.