Will thermal disturbance change an atom's spin?

[univector]

TL;DR

For quantum computers, we often talk about cooling the system to almost absolute zero degree. Is it because thermal disturbance rotate the atoms and change the direction of the spin?

If thermal motion (collision of atoms) changes the direction of an atom, will that change the direction of spin? If so, how much time does it take from the change in the atom orientation to the change in the spin?

 

[Twigg]

Yes, thermal disturbances (blackbody photons scattering off atoms, atom-atom collisions, etc.) result in loss of coherence. Each collision imparts a little bit of random phase to the spin, which washes out entanglement.

The reason collisions result in spin changes has less to do with the collision itself and more to do with short-range interactions between the spins. For example, each spin has its own magnetic moment and creates a magnetic field. When two atoms collide, one atom's spin feels the magnetic field of the other atom, and this causes both spins to precess until the atoms separate again. It's a highly time-dependent precession. The spin dynamics during a collision depend largely on the interaction Hamiltonian, and how it scales with distance. However, I don't see any reason why there would be a lag between momentum transfer and spin precession.

 

[univector]

Thanks for shedding the light on the cause to the change in the spin. So cooling the system is necessary, but not sufficient to maintain a 'long-time' coherence. For example, if the temperature is close to absolute zero, if two atoms are close, they may interact according to the joint Hamiltonian and lose coherence. Is that understanding correct?

Also, I try to visualize motion of an atom and a spin. If an atom is rotating from west to east, and the spin is pointing to west at time $t_1$. After the atom finishes rotating half circle at time $t_2$, will the spin point to east at time $t_2$?

 

[256bits]

Also, I try to visualize motion of an atom and a spin. If an atom is rotating from west to east, and the spin is pointing to west at time t1t1t_1. After the atom finishes rotating half circle at time t2t2t_2, will the spin point to east at time t2t2t_2?

From that statement it would suggest that you have you not researched what spin means for elementary particles?
Even for an everyday object, its rotation is a vector than does point along the axis of rotation, and does not change every half rotation. Spin of an elementary particle is analogous, and called that simply because it acts as if the particle is rotating.
Look up magnetic moment, electron spin, orbital spin, or nuclear spin to perhaps acquire a better understanding.
Or even angular momentum for an everyday object as a comparison.

 

[PeterDonis]

After the atom finishes rotating half circle at time t2t2t_2, will the spin point to east at time t2t2t_2?

Does the direction of the Earth's spin change as it goes around the Sun?

 

[Twigg]

Answering the first paragraph of univector's latest post, not quite. Atom based quantum computers require the interaction Hamiltonian to entangle adjacent spins. For atoms, no interaction means you can't create entanglement. The difference is whether the interaction is fluctuating or stable. Thermal motion results in interactions whose intensity fluctuates randomly in time, and that means the spin will precess in random directions. You can think of it as the spin becoming entangled with the instantaneous momentum of the other atom, which is subsequently scrambled by further collisions. Interactions between cold atoms is acceptable, desirable even, because there is no loss of information to the random bouncing of atoms. Each atom sits in a ground state of motion in a potential well, so its motion and position are in a steady state.

I think the above replies have covered the second question. A little further reading and thinking should clear that up for you.