HOW SCIENCE WORKS · ATOMIC PHYSICS
Fluorescence imaging: collect enough photons to locate one atom
A single atom is much smaller than the wavelength of visible light, yet it can still be detected. Resonant light repeatedly excites the atom, which emits fluorescence photons in many directions. A carefully designed objective collects a small fraction and forms a bright spot on a low-noise camera.
Photon counts above camera background.
Whether a site is occupied by an atom.
Measure loss, hopping and detection fidelity.
Learning note: This guide separates raw signal, physical model and inference. Numerical examples are illustrative unless identified as experimental results.
Chapter index
Route 1 — Build the measurement
Chapters 1–3 explain the experimental problem, the interaction and the sequence that produces the signal.
Route 2 — Turn signal into evidence
Chapters 4–6 develop a calculation, describe calibration and connect the method to a primary experiment.
Route 3 — Test the interpretation
Chapters 7–8 examine failure modes, controls and the broader lesson about scientific measurement.
CHAPTER 1 OF 8
Why one atom is hard to see
An atom does not provide a solid reflecting surface. Detection relies on a cycling optical transition that lets the same atom absorb and emit many photons before it is lost or transferred into a dark state.
Only a limited solid angle reaches the objective, and detector efficiency is below one. The experiment therefore needs many scattering cycles, low background and an optical system that resolves neighbouring sites.
Return to the chapter index · Return to the overview
CHAPTER 2 OF 8
Fluorescence creates the signal
A near-resonant laser excites the atom; spontaneous emission produces photons with random directions. The expected detected count is roughly the scattering rate multiplied by exposure time, collection fraction, optical transmission and camera quantum efficiency.
The recorded spot is the instrument’s point-spread function rather than a literal image of atomic size. Its centre localises the emitter, while its integrated counts help distinguish an occupied site from noise.
Return to the chapter index · Return to the overview
CHAPTER 3 OF 8
Cooling must run during imaging
Every scattered photon gives the atom recoil momentum. Without cooling, repeated scattering heats the atom until it leaves its trap or moves to another lattice site.
Quantum-gas microscopes therefore combine fluorescence with optical molasses or Raman sideband cooling. The same light cycle that reveals the atom must be balanced by a mechanism that keeps its motion confined.
Return to the chapter index · Return to the overview
CHAPTER 4 OF 8
Worked example: a photon budget
Suppose an atom scatters 20,000 photons during an exposure. If the objective collects 8%, downstream optics transmit 60% and the detector converts 80% of arriving photons into counts, the expected detected signal is 20,000×0.08×0.60×0.80 = 768 photoelectrons.
If the local background has a standard deviation of 20 counts, that signal is easily distinguishable in this simplified example. Real fidelity also depends on loss, hopping, overlapping point-spread functions and camera statistics.
| Quantity | Raises signal when… | Possible cost |
|---|---|---|
| Scattering rate | More photons are emitted | More recoil heating |
| Collection efficiency | Larger solid angle reaches detector | Demanding high-NA optics |
| Exposure time | Counts accumulate longer | Greater loss or hopping |
Return to the chapter index · Return to the overview
CHAPTER 5 OF 8
From bright spots to occupation maps
Researchers calibrate the point-spread function, subtract structured background and fit or deconvolve the image. A threshold or likelihood model then assigns each site as occupied or empty.
Detection fidelity is tested with repeated images. An atom that vanishes signals loss; a spot that moves signals hopping; disagreement between two successive classifications estimates error rather than assuming every bright point is correct.
Return to the chapter index · Return to the overview
CHAPTER 6 OF 8
A quantum-gas microscope result
In 2010, researchers reported fluorescence images of a strongly interacting rubidium-87 Mott insulator with single-atom and single-lattice-site resolution.
They reconstructed atom distributions and resolved density and fluctuation structures across the lattice. The experiment made many-body physics visible at the level of its individual particles rather than only through an averaged cloud.
Return to the chapter index · Return to the overview
CHAPTER 7 OF 8
What can mislead the image
Stray light and camera defects create false bright sites. Two atoms on one site may collide under illumination and both be lost, producing parity projection. Heating can cause hopping, while imperfect polarization pumps atoms into dark states.
Scientists quantify these errors with prepared patterns, repeated exposures, simulated image reconstruction and independent atom-number checks. A beautiful image is not enough without a measured fidelity matrix.
Return to the chapter index · Return to the overview
CHAPTER 8 OF 8
The scientific lesson
Fluorescence imaging is an engineered compromise among information, disturbance and time. More scattered photons improve classification but add recoil and increase the chance that the system changes while being observed.
It also demonstrates how instruments extend sight: the atom remains sub-wavelength, but repeated quantum events, calibrated optics and statistical classification make its position knowable.
Read the primary research: Single-Atom-Resolved Fluorescence Imaging of an Atomic Mott Insulator (Nature, 2010).