Given that the state of the particle at time $t=-\infty$ is $\left|S_z^+\right>$, a magnetic field of the form $\mathbf{B}=B \tanh(t/\tau) \hat{\mathbf{z}}$, the Hamiltonian is $H=-\vec{\mu}\cdot\mathbf{B}$, where $\vec{\mu}=-\gamma \mathbf{S}$, how do you find the probability that the particle is in the state $\left|S_z^-\right>$ at time $t$?

I've approached the problem so far by setting up the Schödinger equation using the above definition of the Hamiltonian to find, by representing the state of the particle $\left|\Psi,t\right>$ as the column vector:

$$ \left|\Psi,t\right>=\begin{pmatrix}\psi_+\\\psi_-\\\end{pmatrix}, $$

and solving:

$$ i\hbar\frac{\partial}{\partial t}\left|\Psi,t\right>=\hat{H}\left|\Psi,t\right>. $$

In doing so, with $H$ being, more precisely:

$$ \hat{H}=\gamma B\frac {\hbar}{2}\tanh{\Big(\frac{t}{\tau}}\Big)\begin{pmatrix}1&0\\0 &-1\\\end{pmatrix},$$

I've found the general solution to be, taking into account the initial condition:

$$ \left|\Psi,t\right>=\begin{pmatrix}C_1 e^{\frac{-i\gamma B\tau}{2}}\cosh{\Big(\frac{t}{\tau}\Big)}\\0\\\end{pmatrix}. $$

The thing is that I don't know if I've answered my own question correctly, as intuitively it seems to me that upon the flipping of the direction of the magnetic field at $t=0$, the state could or would change to $\left|S_z^-\right>$, meaning that the probability of finding the particle at time $t$ in state $\left|S_z^+\right>$ would not always be one.

Further, by operating the Hamiltonian on the original state, one finds that original state multiplied by a constant (if the Hamiltonian is evaluated at $t=-\infty$), meaning (I think) that it exists in a stationary state.

I would appreciate either affirmation of my result, or a pointing out that it is incorrect, and suggestions as to how I should re-tackle the problem if indeed my current solution is incorrect.


1 Answer 1


You can certainly flip a spin using a magnetic field. It does not even have to be time dependent; a time-independent field will work, provided it is pointing in the right direction. But the direction is key; you cannot change the $z$ projection of the spin with a magnetic field that points entirely in the $z$-direction.

A single spin-$\frac{1}{2}$ particle behaves essentially like a classical magnetic moment. If it is exposed to a magnetic field in the $z$-direction, the spin will precess around the $z$-axis. (Classically, this is a consequence of the torque being $\vec{N}=\vec{\mu}\times\vec{B}$.) If the spin does not point precisely along the $\pm z$-axis, the $x$- and $y$-components of $\vec{S}$ will change; however, the $z$-component will remain constant. This is what you have found with your calculation. If the magnetic field pointed had a component in the $xy$-plane, your initial $\vec{S}$ would precess around a different direction, and the $z$-projections of the spin would change in the way you were hoping to see.

  • $\begingroup$ Hmmm. So, based on your comment, it seems to me then that my calculation in the way that I carried out is correct, due to the initial state of the particle being $\left|S_z^+\right>$ and the magnetic field being aligned with it. Classically, however, would not the direction of the moment change direction? The potential energy of the particle classically is: $U=-\vec{\mu}\cdot\mathbf{B}$, and to minimize this one would have $\mu$ align with $\mathbf{B}$. $\endgroup$ Commented Nov 1, 2018 at 6:03
  • $\begingroup$ @T.Zaborniak Yes, it's the fact that the field and the initial spin direction are aligned that leads to the spin not changing. (There is something a odd about the expression for the time-dependent wave function in the question though; the absolute value of the upper component of the spinor should be 1 for all time.) $\endgroup$
    – Buzz
    Commented Nov 1, 2018 at 6:08
  • $\begingroup$ @T.Zaborniak The classical torque is what I wrote in my answer, and if you take the dot product of $\vec{N}=d\vec{S}/dt$ with the direction $\hat{z}$ of $\vec{B}$, you can see that $S_{z}$ is independent of time. $\endgroup$
    – Buzz
    Commented Nov 1, 2018 at 6:12
  • $\begingroup$ I still don't get your classical explanation exactly. I'm thinking that in writing $S_z$ you mean $L_z$, where $L_z$ is the classical angular momentum in the $z$-direction, and by $\vec{S}$ you mean the classical angular momentum. That being the case, isn't it necessary that $L_z$ change when $t$ goes from $t<0$ to $t>0$ to have $|d\vec{L}/dt|$ remain a constant? The magnitude of $\vec{L}$ obviously would have to stay the same, but because it is $L_z\hat{\mathbf{z}}$, its direction would need to change... $\endgroup$ Commented Nov 1, 2018 at 6:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.