trapped-ion qubit
A trapped-ion qubit stores quantum information inside a single charged atom (an ion). Picture one atom floating almost perfectly still in empty space, pinned there by carefully shaped electric and magnetic fields the way a marble settles into the bottom of an invisible bowl. The atom has internal energy levels, and you pick two of them to stand for |0> and |1>. Because the atom is held away from any surface and is naturally identical to every other atom of its kind, it is extremely well isolated from the noisy outside world, and it stays in a usable quantum state for a long time. You write to it and read from it by shining precisely tuned lasers (or microwaves) on it: the laser nudges the atom between its two chosen levels, and a different laser makes the atom glow brightly for one state and stay dark for the other so a camera can tell which it is.
The big practical strengths are long coherence times (the qubit holds its state for a relatively long while before noise scrambles it) and very high gate fidelity (operations come out very close to what you intended, with few errors). Trapped ions can also talk to many partners in the same trap because they share a common vibrational motion, which makes two-qubit gates flexible. The trade-offs are real, though: each operation is driven by lasers and that mechanical motion, so gates tend to be slower than on superconducting chips, and scaling up is hard. Keeping a hundred or a thousand ions in one trap, all addressed by their own steady laser beams, gets unwieldy fast, which is why much of the engineering effort goes into linking many smaller traps together rather than building one giant one.
The two states |0> and |1> are just two chosen internal energy levels of the ion; before you measure, the qubit can be any combination with |alpha|^2 + |beta|^2 = 1, and the measurement laser returns 0 or 1 with probability |alpha|^2 or |beta|^2.
Trapped ions are one of the leading qubit technologies today, but like every platform we are still in the NISQ era: these are small, error-prone machines, not large fault-tolerant quantum computers, and choosing ions over superconducting qubits is a trade of speed and easy scaling for longer coherence and higher per-gate fidelity.