[hamdani yusuf (OP)]

Can we demagnetize a permanent magnet by applying electromagnetic field in reverse polarity?

[chiralSPO]

Yes. You can even completely reverse the polarity, if the electromagnet is strong enough, and the permanent magnet "flexible" enough.

[PmbPhy]

Can we demagnetize a permanent magnet by applying electromagnetic field in reverse polarity?

It can only be done with a special type of EM field. Magnetic material has what's called a hysteresis curve. It's a curve which describes the applied magnetic field to a magnetic material. To demagnetize such material the applied field must be altered and reduced in a continuous manner.
There's no such thing as a

[RD]

https://en.wikipedia.org/wiki/Degaussing

[PmbPhy]

https://en.wikipedia.org/wiki/Degaussing

Note in particular the link in that page to the following
https://en.wikipedia.org/wiki/Magnetic_hysteresis

[hamdani yusuf (OP)]

when a magnet is reversed, what is actually changed? is it electron spin of the atoms in it? or electrons motion around atoms' nuclei? or electrons motion around the crystal lattice? or spin of the nuclei themselves?

[PmbPhy]

when a magnet is reversed, what is actually changed? is it electron spin of the atoms in it? or electrons motion around atoms' nuclei? or electrons motion around the crystal lattice? or spin of the nuclei themselves?

The magnetic domains rearrange.

[hamdani yusuf (OP)]

when a magnet is reversed, what is actually changed? is it electron spin of the atoms in it? or electrons motion around atoms' nuclei? or electrons motion around the crystal lattice? or spin of the nuclei themselves?

The magnetic domains rearrange.

Let's say I have a microscopic permanent magnet cube containing 3^3 domains (imagine Rubik' cube). At first they all directing upward.
After electromagnetically reversed, what happens to the atoms in the interface between adjacent domains? Are they sliding to each other?

[PmbPhy]

when a magnet is reversed, what is actually changed? is it electron spin of the atoms in it? or electrons motion around atoms' nuclei? or electrons motion around the crystal lattice? or spin of the nuclei themselves?

The magnetic domains rearrange.

Let's say I have a microscopic permanent magnet cube containing 3^3 domains (imagine Rubik' cube). At first they all directing upward.
After electromagnetically reversed, what happens to the atoms in the interface between adjacent domains? Are they sliding to each other?

To be exact requires an understanding of quantum mechanics. I'll try to simplify. But a greater understanding is much more involved so if you want than then see - https://en.wikipedia.org/wiki/Magnetic_domain

Ferromagnetism involves the magnetic dipoles of unpaired associated with unpaired electrons in atoms. In a magnetic domain magnetic moments of the atoms are aligned with one another and they point in the same direction. When the polarization of that domain changes it occurs because the magnetic moments of the atoms are changed.  No sliding.

In simplistic terms think of this as the orbits of the electrons in the atoms undergoing a change in the plane of their orbits. While not literally correct it does give you a general picture. If you want literally correct then you'll have to learn quantum mechanics.

[evan_au]

The key to degaussing is to use an oscillating magnetic field in the electromagnet, which starts off strong, and then decays to zero.

This moves the magnetic domains back and forth until they end being balanced with no overall magnetic field.

If you just applied a strong magnetic field with opposite orientation, you would reverse the existing magnetic field, but not eliminate it.

Note that normal mechanical vibration will jiggle the domain boundaries so the material starts to line up with Earth's weak magnetic field. So degaussing is never entirely complete.

Experiments which must be run with no ambient magnetic field are often enclosed in a Maxwell coil, to cancel Earth's magnetic field.

[PmbPhy]

The key to degaussing is to use an oscillating magnetic field in the electromagnet, which starts off strong, and then decays to zero.

The term oscillate means to vary in a regular fashion. However the magnetic field need to be changed in regular fashion. That's what was meant when I wrote

To demagnetize such material the applied field must be altered and reduced in a continuous manner.

and later posted the link to magnetic hysteresis