One weak transition switches many photons on and off
In 1986, three groups reported the quantum jumps of single trapped ions. The ion stayed in place, but its fluorescence switched between bright stretches and dark intervals. A change inside one atom had become visible through many detected photons. The groups used different ion species and level schemes; Bergquist and colleagues at the National Bureau of Standards, now NIST, worked with a mercury ion.
The method grew from Dehmelt’s idea of electron shelving. A weak, narrow transition is hard to see by collecting its own scarce photons. Instead, use it to move the ion into a state that interrupts a much brighter cycle. The interruption amplifies the effect of the weak transition into a sustained change in the light.
Read the evidence: W. M. Itano, J. C. Bergquist and D. J. Wineland (2015), Early Observations of Macroscopic Quantum Jumps in Single Atoms — participants’ historical accountW. Nagourney, J. Sandberg and H. Dehmelt (1986), Shelved Optical Electron Amplifier: Observation of Quantum Jumps, Phys. Rev. Lett. 56, 2797 — author abstract
The recorded photons come from the bright cycle
The mercury experiment used a strong ultraviolet transition for the bright cycle. Picture an ion with a lower level and an upper level connected by a strong optical transition. The laser lifts the ion up; it drops back and emits a photon; the laser lifts it again. Repeating this cycle produces a stream of fluorescence photons. A detector collects a fraction of them, enough to identify the bright condition.
Now add a third level. A separate, narrow transition can transfer the ion into it, but the strong detection laser has no suitable transition out of that level. The bright cycle stops. This long-lived level is called metastable: the ion eventually leaves it, but slowly compared with the bright cycle. When it returns, fluorescence resumes.
The detector is watching the strong transition throughout. It can report that the weak transition occurred even if it never collects the photon associated with that weak transition. The absence of many expected photons becomes the amplified signal.
Read the evidence: J. C. Bergquist et al. (1986), Observation of Quantum Jumps in a Single Atom, Phys. Rev. Lett. 57, 1699–1702 — NIST collection, TN-46–49
A bright ion can still give an empty bin
The detector collects only a fraction of the scattered light, so a bright ion can occasionally yield zero registered photons. A simple Poisson counting model illustrates the problem. In the formula, N counts the detected photons, eta is collection and detection efficiency, the two R terms are the scattering and background rates per second, and Delta t is the time-bin duration in seconds.
This illustration assumes steady rates, no switching within a bin and bins much longer than the fluorescence’s short-time correlations; it ignores detector dead time. Within those assumptions, a longer bin reduces the chance of a false dark interval. The full emission statistics require a quantum-optical description, and the historical readout fidelity requires the experiment’s own calibration. This coarse counting model isolates how missed photons can create a false dark bin.
Read the evidence: J. C. Bergquist et al. (1986), Observation of Quantum Jumps in a Single Atom, Phys. Rev. Lett. 57, 1699–1702 — NIST collection, TN-46–49Leibfried et al. (2003), Quantum dynamics of single trapped ions, Rev. Mod. Phys. 75, 281
Bin size changes what looks like a jump
Longer bins have a cost: averaging can blur a real switch. Bins that are too short make ordinary missed photons look like interruptions. The researchers therefore compared count distributions, background, the time spent bright or dark, and the response to controlled changes of the weak drive. Together these checks connected a sustained dark stretch to the internal-level scheme.
A switch detected within one bin is timed only to that observation window. Quantum-trajectory methods update the internal-state estimate from those outcomes. Any classical path between orbitals remains unspecified; the observations follow changes in the detected light.
The trace also has a different shape from the smooth curve obtained by averaging many repetitions. One record contains individual bright and dark stretches; the average gives the fraction of repetitions that are bright at each time. Both describe the same experimental procedure at different levels of detail.
Read the evidence: J. C. Bergquist et al. (1986), Observation of Quantum Jumps in a Single Atom, Phys. Rev. Lett. 57, 1699–1702 — NIST collection, TN-46–49W. M. Itano, J. C. Bergquist and D. J. Wineland (2015), Early Observations of Macroscopic Quantum Jumps in Single Atoms — participants’ historical account
Telling states apart and keeping coherence are separate tests
Fluorescence distinguishes the bright and dark alternatives at the cost of their coherence, the phase relationship that would allow a superposition to interfere later. Over a suitable interval the same state label can remain stable through repeated detection, even though an arbitrary initial superposition has been disturbed.
For spectroscopy, experimenters can separate the narrow probe pulse from the bright readout interval. They first turn off the detection light, probe the transition, then turn the light on to read the result. This avoids letting the strong readout light shift or broaden the transition during the probe. Trapping makes it possible to repeat the sequence on the same ion and compare individual switching histories with ensemble averages.
Read the evidence: D. J. Wineland (2013), Nobel Lecture: Superposition, Entanglement, and Raising Schrödinger’s Cat, Rev. Mod. Phys. 85, 1103–1114 — published version of the 2012 lectureJ. C. Bergquist et al. (1986), Observation of Quantum Jumps in a Single Atom, Phys. Rev. Lett. 57, 1699–1702 — NIST collection, TN-46–49W. M. Itano, J. C. Bergquist and D. J. Wineland (2015), Early Observations of Macroscopic Quantum Jumps in Single Atoms — participants’ historical account