How a PET Scanner Turns Antimatter Into an Image

The scanner sees a two-photon signature

During a PET scan, radioactive decay creates tiny, short-lived bits of antimatter - positrons - inside the body.

Not a tank. Not a beam. One positron at a time. A PET radiopharmaceutical contains unstable atoms that can create positrons as they decay. A positron is the antimatter counterpart of an electron: it has the same mass and the opposite electric charge.

The positron does not travel all the way out to the scanner. It moves a short distance through tissue, loses most of its kinetic energy, and meets an electron. The positron and electron annihilate.

The principal usable PET signal is a pair of approximately 511-keV photons.

That is the first big idea. A PET scanner does not see the positron. It detects the two-photon signature left behind.

So here is the board question: How does one two-photon event become a line, and how do many lines become a map?

What the finished map represents

PET stands for positron emission tomography. Its job is to estimate where a radiotracer is concentrated inside the body.

A radiotracer combines a molecule with a radioactive atom. The molecule is chosen because of how it moves through or binds within the body. The radioactive atom supplies a signal the scanner can detect.

One common example is fluorine-18 fluorodeoxyglucose, or FDG, a molecule similar to glucose. Its distribution can provide information related to glucose metabolism. FDG is only one tracer, however. Different PET tracers can target different molecular processes.

A PET image is not simply a photograph of anatomy. It is a reconstructed map of tracer activity. In a PET/CT scanner, the PET part provides functional or molecular information while the CT part helps place that information in anatomical context.

The radioactive decay makes a positron

In beta-plus decay, a proton in an unstable nucleus changes into a neutron. The decay emits a positron and a neutrino.

The neutrino escapes without becoming part of the PET image. The positron interacts strongly enough with matter to slow down nearby. It does not annihilate at exactly the spot where the nucleus decayed, so positron travel is one physical limit on how precisely PET can locate the original tracer atom.

This also clears up a useful misconception: the annihilation is not the same event as the nuclear decay. First the nucleus emits the positron. Then the positron travels, slows, and annihilates with an electron.

Why each photon has about 511 keV

Put the familiar equation at the top of the board: E = mc^2.

The rest-mass energy of an electron is approximately m_e c^2 = 0.511 MeV = 511 keV.

A positron has the same mass. When a slow positron and an electron annihilate, their combined rest-mass energy is approximately E_pair = 2m_e c^2 = 1.022 MeV. In the simplest and most useful PET model, that energy leaves as two photons, each with approximately 511 keV.

Momentum must be conserved too. If the electron-positron pair has very little net momentum before annihilation, the two photons leave in nearly opposite directions. Nearly matters. Residual net momentum of the pair produces a small departure from a perfect 180-degree angle, which is another limit on spatial resolution.

One accepted coincidence gives a line, not a point

Picture a ring of detectors around the patient. Two detector elements on opposite sides record photons with energies near 511 keV within a short timing window.

The scanner first treats that matched pair as a candidate coincidence. If the event passes its energy and timing tests, the system provisionally assigns it to the straight line connecting those two detectors. That line is called a line of response.

One accepted coincidence does not identify one exact point. It gives a whole line of possible locations. Scattered and random coincidences can still pass the initial tests, so real systems also estimate and correct for that contamination.

That is the geometry behind PET. The coincidence supplies the direction information electronically.

Many lines become an image

One line is not much of an image. A scan records a large number of coincidence events at many positions and angles around the detector ring.

The computer then solves an inverse problem. It asks: what tracer distribution would be most consistent with all of these measured lines? The answer is reconstructed as a three-dimensional activity map.

Regions with more recorded tracer-related activity can appear more intense, but intensity does not diagnose a condition by itself. The meaning depends on the tracer, protocol, corrections, anatomy, and interpretation by trained medical professionals.

The simple crossing-lines sketch is therefore a starting model, not the whole scanner. Real reconstruction must account for detector response, count statistics, motion, photon attenuation in the body, scattered photons, and random coincidences between unrelated detections.

Time of flight is an extra refinement

At this point, the core PET chain is complete: decay -> annihilation -> photon pair -> line of response -> reconstructed map.

If the two photons arrive at the same time, the annihilation was likely near the midpoint between the two detectors. If one arrives first, the event was likely closer to that detector.

For an arrival-time difference Delta t, the displacement from the midpoint along the line is Delta x = c Delta t / 2.

The factor of two is easy to miss. If the event moves by Delta x, one photon path becomes shorter by that amount while the other becomes longer by the same amount. The path-length difference changes by 2 Delta x.

Suppose the measured time difference were 100 picoseconds: Delta x = (3.00 x 10^8 m/s)(100 x 10^-12 s) / 2 = 1.5 cm.

That calculation does not mean the final image resolution is 1.5 cm. It shows how timing can narrow an event from an entire line to a more likely segment. The reconstructed image still depends on many events and the rest of the system.

The scanner must reject convincing impostors

Not every apparent pair is a true coincidence.

A photon can scatter in the body before detection. That deflection can cause the detected pair to define a line of response that misses the true annihilation location. Two unrelated photons can also arrive close enough in time to look like a pair. Other photons may be absorbed or leave the detector's field of view.

PET systems use energy and timing information, calibration, and mathematical corrections to estimate these effects.

Attenuation, scatter, and random coincidences are not small vocabulary details. They affect where events appear and how accurately image intensity represents tracer concentration.

That is why the finished image is the result of particle physics, detector physics, statistics, and computation working together.

A precise way to say what PET sees

PET does not see cancer. PET does not see the positron. PET does not directly see the original nuclear decay.

It detects annihilation photons and reconstructs a radiotracer activity map. Trained medical professionals interpret that map together with the tracer's biology, the imaging protocol, anatomical images, and the patient's clinical information.

PET uses ionizing radiation. The tracer, amount, protocol, patient, and any CT portion of the examination all affect exposure. Medical professionals weigh the expected benefit against the radiation risk for a specific examination. Patients should follow their imaging team's instructions and tell the team about pregnancy or breastfeeding.

That is the careful conclusion: a PET scanner turns a short-lived piece of antimatter into an image by never trying to photograph the antimatter at all. It looks for candidate pairs near 511 keV, uses accepted coincidences to define lines of response, and combines many corrected measurements to estimate the tracer distribution.

Learn the physics behind this

PET connects radioactive decay, antimatter, mass-energy equivalence, photon energy, momentum conservation, radiation detection, and modern imaging. The related Mousseau Physics course is Modern Physics.