QHO Coherent States

Let all the variables in Quantum Harmonic Oscillator

From definition, define zz to be the eigenvalue of annihilation operator a^\hat{a} and they are the most classical QHO states. They have the least uncertainty and follow classical trajectories under time evolution

a^z=zz,zC\hat{a}\ket{z} = z\ket{z}, \quad z \in \mathbb{C}

Note that a^\hat{a}^\dagger has no eigenstates. It is only possible because the Hilbert Space is infinite dimension.

Expanding in the numbers basis where cnc_n is a Probability amplitude

z=n=0cnn\ket{z} = \sum_{n=0}^{\infty} c_n \ket{n}

Apply a^\hat{a}. Note in QHO Quantum Dynamics we proved a^n=nn1\hat{a}\ket{n}=\sqrt{n}\ket{n-1}

a^z=ncnnn1=zz\hat{a}\ket{z} = \sum_n c_n \sqrt{n}\ket{n-1} = z\ket{z}

Let m=n1m=n-1

cm+1m+1=zcmcnn=zcn1c_{m+1}\sqrt{m+1} = z\, c_m \Longrightarrow c_n\sqrt{n} = z\, c_{n-1}

So

cn=zncn1=znzn1cn2==znn!c0c_n = \frac{z}{\sqrt{n}}c_{n-1} = \frac{z}{\sqrt{n}}\frac{z}{\sqrt{n-1}}c_{n-2} = \cdots = \frac{z^n}{\sqrt{n!}}c_0     z=c0n=0znn!n\implies\boxed{\ket{z} = c_0 \sum_{n=0}^{\infty} \frac{z^n}{\sqrt{n!}}\ket{n}}

Normalize to get c0c_0

zz=c02nz2nn!=c02ez2=!1\braket{z|z} = |c_0|^2 \sum_n \frac{|z|^{2n}}{n!} = |c_0|^2 e^{|z|^2} \stackrel{!}{=} 1 c0=ez2/2\Longrightarrow \boxed{c_0 = e^{-|z|^2/2}}

Coherent State Observable

The Observable for a coherent state becomes Where N^\hat{N} is Number Operator

zN^z=za^a^z=z2\langle z|\hat{N}|z\rangle = \langle z|\hat{a}^\dagger \hat{a}|z\rangle = |z|^2 P(n)=nz2=cn2=ez2z2nn!P(n) = |\braket{n|z}|^2 = |c_n|^2 = e^{-|z|^2}\frac{|z|^{2n}}{n!}

this is a Poisson Distribution with mean E[n]=z2\mathbb{E}[n]=|z|^2

Note this is the energy expectation

E=H^=ω ⁣(N^+12)=ω ⁣(z2+12)\langle E \rangle = \langle \hat{H} \rangle = \hbar\omega\!\left(\langle \hat{N}\rangle + \tfrac{1}{2}\right) = \hbar\omega\!\left(|z|^2 + \tfrac{1}{2}\right)

Coherent State Wavefunction

The Wave function for a coherent state becomes

a^z=zz\hat{a}\ket{z}=z\ket{z}

Let variables from Quantum Harmonic Oscillator

    12(xα+αx)ϕz(x)=zϕz(x)\implies\frac{1}{\sqrt{2}}\left(\frac{x}{\alpha} + \alpha\frac{\partial}{\partial x}\right)\phi_z(x) = z\,\phi_z(x)

rearranging gives

ϕzx=(2zαxα2)ϕz\frac{\partial \phi_z}{\partial x} = \left(\frac{\sqrt{2}\,z}{\alpha} - \frac{x}{\alpha^2}\right)\phi_z lnϕz=2zαxx22α2+const\ln\phi_z = \frac{\sqrt{2}\,z}{\alpha}x - \frac{x^2}{2\alpha^2} + \text{const} ϕz(x)=Nze(xb)2/2α2eikx\phi_z(x) = N_z\, e^{-(x-b)^2/2\alpha^2}\, e^{ikx} b=2αRe[z],k=2αIm[z]b = \sqrt{2}\,\alpha\,\text{Re}[z], \qquad k = \frac{\sqrt{2}}{\alpha}\text{Im}[z]

Note Expected Value of an Observable are

zx^z=b,zp^z=k\braket{z|\hat{x}|z} = b, \qquad \braket{z|\hat{p}|z} = \hbar k

Note Gaussian property

Δx=α2,Δp=α2ΔxΔp=2\Delta x = \frac{\alpha}{\sqrt{2}}, \qquad \Delta p = \frac{\hbar}{\alpha\sqrt{2}} \Longrightarrow \Delta x\,\Delta p = \frac{\hbar}{2}

Note

z=12(x^α+iαp^)\boxed{z = \frac{1}{\sqrt{2}}\left(\frac{\braket{\hat{x}}}{\alpha} + i\frac{\alpha}{\hbar}\braket{\hat{p}}\right)}

This can be proven but I don't think the professor proved it.