Why PET Scanners Detect Two 511 keV Gamma Photons

A PET scanner registers two gamma-ray signals on opposite sides of a patient, almost at the same time. Each photon has an energy close to 511 keV. Before reading on, predict: why doesn’t the annihilation produce one photon with an energy of 1.022 MeV instead? Energy conservation alone cannot rule that out, because both possibilities have the same total energy.

The missing constraint is momentum conservation. An electron and a positron annihilating approximately at rest have almost zero total momentum. A single photon cannot carry away their energy while leaving the total momentum unchanged. Two photons travelling in opposite directions can: their momenta cancel, while their energies add.

01Quick decision rule

Use this table when deciding which physics principle explains a PET scanner’s measurements.

02From radioisotope to gamma photons

PET stands for positron emission tomography. It uses a radioactive tracer that emits positrons, the antimatter counterparts of electrons.

Understanding the sequence matters because the scanner does not detect the positron itself.

Step 1: The nucleus emits a positron.

In positron emission, a proton in the nucleus changes into a neutron, releasing a positron and an electron neutrino:

\[ p \rightarrow n + e^+ + \nu_e \]

Here, \(p\) represents a proton, \(n\) a neutron, \(e^+\) a positron, and \(\nu_e\) an electron neutrino.

The positron is emitted with kinetic energy. It is not generally stationary when it first appears.

Step 2: The positron slows down.

As it travels through tissue, the positron transfers kinetic energy to surrounding atoms through interactions such as ionisation and excitation. It gradually loses energy.

This means the positron’s original emission point is not necessarily the point where it eventually disappears.

Step 3: The positron annihilates with an electron.

Once the positron has slowed sufficiently, it can meet an electron. The two particles annihilate, converting their mass energy, and any remaining kinetic energy, into other forms of energy. In the standard PET model, the dominant outcome is two gamma photons.

The distinction between emission and annihilation is important. PET uses the gamma photons to infer where annihilation occurred. It does not follow the positron along its path from the nucleus.

03Why each photon carries about 511 keV

The key equation is Einstein’s mass-energy relationship:

\[ E=mc^2 \]

Here:

• \(E\) is energy, measured in joules (J).
• \(m\) is mass, measured in kilograms (kg).
• \(c\) is the speed of light in a vacuum, approximately \(3.00\times10^8\ \text{m s}^{-1}\).

An electron and a positron have the same rest mass. The rest-mass energy of either particle is approximately 511 keV, where keV means kiloelectronvolt.

When both particles annihilate approximately at rest, their combined rest-mass energy is

\[ \begin{aligned} E_{\text{total}} &= 2m_ec^2\\ &=2(511\ \text{keV})\\ &=1022\ \text{keV}\\ &=1.022\ \text{MeV} \end{aligned} \]

Here, \(m_e\) is the electron rest mass, and \(1\ \text{MeV}=1000\ \text{keV}\).

So where does the factor of two go?

The total energy becomes two photons. But their energies are not simply split equally by chance. Momentum conservation explains why.

A photon with energy \(E_\gamma\) has momentum magnitude

\[ p_\gamma=\frac{E_\gamma}{c} \]

If the electron-positron pair starts with zero total momentum, the two photons must have equal momentum magnitudes in opposite directions. Since photon momentum is proportional to photon energy, their energies must also be equal.

Therefore,

\[ E_\gamma=\frac{1.022\ \text{MeV}}{2} =0.511\ \text{MeV} \]

Each photon carries approximately 511 keV.

Why the photons are only approximately opposite

The ideal model assumes that the annihilating electron and positron have negligible total momentum. Real particles in tissue do not always meet in this exact condition.

The electron has momentum associated with its motion in the surrounding material, and the positron may retain some kinetic energy. As a result, the photons can deviate slightly from exactly \(180^\circ\) separation, and their energies may differ slightly from 511 keV.

The 511 keV value is therefore the standard approximation for PET annihilation photons. It is an excellent starting point for understanding the scanner, but it does not mean every measured photon must have exactly that energy.

04How coincidence detection locates an annihilation

Knowing why the photons are produced is only half the problem. A PET scanner must also work out where the annihilation happened.

A PET scanner has detector crystals arranged around the patient. When a gamma photon interacts with a suitable crystal, it deposits energy and can produce a flash of light. A photodetector converts that light into an electrical signal.

The scanner’s electronics record the detector that registered each signal and when it occurred.

When two signals arrive within the scanner’s calibrated coincidence time window, the system can identify them as a candidate pair from the same annihilation.

The line of response

The line connecting the two detector locations is called the line of response (LOR).

Under the ideal assumption that the photons travel in exactly opposite directions without scattering, the annihilation lies on this line.

However, the detector pair alone does not tell the scanner exactly where along the line the event occurred. The annihilation could have happened near the first detector, near the second, or somewhere between them.

This is the important limitation of ordinary coincidence detection: it identifies a line, not a unique point.

A PET scanner builds an image by collecting many such events from different detector pairs and directions. Image reconstruction uses the combined information to estimate the distribution of the radioactive tracer.

05Time-of-flight PET: using the arrival-time difference

Some PET scanners also measure the difference between the two photons’ arrival times. This is called time-of-flight (TOF) information.

The idea is straightforward. Both photons travel at approximately the speed of light, \(c\). If one photon reaches its detector first, it travelled a shorter distance. The annihilation must therefore have occurred closer to that detector.

Suppose the annihilation occurs a distance \(x\) from the midpoint between detectors A and B.

Moving the annihilation point towards A shortens the path to A by \(x\) and lengthens the path to B by \(x\). The difference between the two path lengths is therefore \(2x\), not just \(x\).

If \(\Delta t\) is the magnitude of the arrival-time difference, then

\[ c\Delta t=2x \]

Rearranging gives the useful reference equation:

\[ \boxed{x=\frac{c\Delta t}{2}} \]

The symbols mean:

• \(x\): distance from the detector-pair midpoint to the annihilation point, in metres.
• \(c\): speed of light, approximately \(3.00\times10^8\ \text{m s}^{-1}\).
• \(\Delta t\): difference between the photon arrival times, in seconds.

The direction is determined by which detector registers its photon first. The annihilation point is closer to that detector.

TOF information narrows the possible location along the LOR. Real scanners have finite timing resolution, so they estimate a region of likely positions rather than identifying a mathematically exact point.

06What PET locates-and what it does not

It is tempting to assume that locating the annihilation also locates the original nuclear decay. That is not quite right.

The positron travels through tissue before annihilating. Its emission point may therefore be displaced from the annihilation point. The gamma photons carry information about the annihilation, but they do not reveal the positron’s path.

This means that even perfect measurement of the gamma-photon arrival times would not, by itself, identify the exact nuclear decay point.

In practice, several effects limit PET’s spatial precision:

• Positron range: the positron travels before annihilation, separating the decay point from the annihilation point.
• Photon non-collinearity: the photons are not always emitted at exactly \(180^\circ\) to one another.
• Detector resolution: the scanner measures interactions within detector crystals, not ideal mathematical points.
• Scattering: a photon can change direction and lose energy before reaching a detector.
• Finite timing resolution: TOF measurements narrow the possible location but retain uncertainty.

The scanner uses many events to reconstruct an estimate of the tracer’s distribution. It does not directly photograph individual positrons or their trajectories.

07Common mistakes

– Giving each photon an energy of 1.022 MeV. This confuses the total rest-mass energy of the electron-positron pair with the energy of each photon. In the ideal two-photon model, the total is 1.022 MeV, and each photon carries 511 keV.

– Adding the positron’s original kinetic energy to the annihilation photons. The positron usually loses most of that kinetic energy to the surrounding tissue before annihilation. The standard 511 keV calculation applies when annihilation occurs approximately at rest.

– Assuming coincidence detection gives an exact position. The detector pair initially identifies a line of response. TOF information narrows the location along that line, while image reconstruction combines many events.

– Treating the nuclear decay point and annihilation point as identical. The positron travels before annihilation. PET infers the annihilation location rather than tracking the original positron.

– Assuming every coincident pair gives a valid line of response. A scattered photon can arrive within the coincidence window but follow a changed path. Two unrelated photons can also arrive within the same window by chance, producing a random coincidence. Energy measurements, timing information, and statistical corrections help the scanner manage these events.

08Questions and solutions

Question 1

An electron and a positron annihilate approximately at rest. Each resulting gamma photon has an energy of about 511 keV.

Calculate the total energy carried by the two photons in MeV, and explain how this relates to the rest-mass energy of the original particles.

Solution 1

The two photons carry a total energy of approximately \(1.022\ \text{MeV}\), equal to the combined rest-mass energy of the electron and positron.

The total photon energy is

\[ \begin{aligned} E_{\text{total}}&=2E_\gamma\\ &=2(511\ \text{keV})\\ &=1022\ \text{keV}\\ &=1.022\ \text{MeV} \end{aligned} \]

Each particle contributes approximately 511 keV of rest-mass energy. When they annihilate approximately at rest, this energy becomes the energy of the two photons.

The photons travel in opposite directions so that their momenta cancel. This satisfies momentum conservation as well as energy conservation.

Question 2

A radioisotope emits a positron with a kinetic energy of \(0.75\ \text{MeV}\). The positron loses essentially all this kinetic energy to the surrounding tissue before annihilating with an electron.

Calculate the expected total energy of the two annihilation photons and the energy of each photon. Explain why the total is not \(1.772\ \text{MeV}\).

Solution 2

The expected total annihilation-photon energy is approximately \(1.022\ \text{MeV}\), with each photon carrying about \(0.511\ \text{MeV}\).

The positron’s initial kinetic energy is transferred to the tissue as it slows. It is therefore not available to be added to the rest-mass energy in the usual approximation for annihilation at rest.

The energy available from the electron-positron rest masses is

\[ E_{\text{total}}=2m_ec^2 \]

Substituting the standard rest energy of each particle,

\[ \begin{aligned} E_{\text{total}}&=2(0.511\ \text{MeV})\\ &=1.022\ \text{MeV} \end{aligned} \]

The two photons have equal energies in the ideal model:

\[ E_\gamma=\frac{1.022}{2} =0.511\ \text{MeV} \]

The incorrect result of \(1.772\ \text{MeV}\) comes from adding the positron’s original kinetic energy to the annihilation energy, even though that kinetic energy has already been transferred to the tissue.

Question 3

A student makes two claims about PET annihilation:

• A single photon with energy \(1.022\ \text{MeV}\) could be produced, provided energy is conserved.
• Any real pair of annihilation photons whose directions differ slightly from \(180^\circ\) must be rejected as physically impossible.

Evaluate both claims using conservation laws and the limitations of the ideal model.

Solution 3

The first claim is incorrect for an isolated electron-positron pair initially at rest. The second claim is too strict because real annihilations do not always satisfy the ideal model perfectly.

First claim: one photon.

Energy conservation alone is insufficient. A single photon with energy \(1.022\ \text{MeV}\) would carry momentum of magnitude

\[ p_\gamma=\frac{E_\gamma}{c} \]

This is non-zero, but the initial pair has zero total momentum. With no other object to take the recoil momentum, a single photon cannot conserve both energy and momentum.

Two photons travelling in opposite directions solve this problem. Their momenta cancel, and their combined energy is \(1.022\ \text{MeV}\).

Second claim: photons not exactly opposite.

Exactly opposite directions follow from the ideal assumption that the initial pair has zero total momentum. In real tissue, the electron’s momentum and the positron’s remaining motion can affect the photon directions and energies.

Consequently, the photons may not be separated by exactly \(180^\circ\), and their energies may not be exactly equal. A small departure from the ideal geometry is not automatically evidence that the event is impossible.

Question 4

Two detector signals arrive within the coincidence window. The photon reaching detector B has scattered in the patient’s tissue before detection, changing direction and losing some energy.

Does the coincidence automatically define a valid line of response through the annihilation point? Explain how the scanner may identify the problem, and distinguish this event from a random coincidence.

Solution 4

No. A coincidence establishes that two signals arrived within the required time window, but it does not guarantee that both photons travelled directly from the annihilation point to their detectors.

The scattered photon has changed direction. Therefore, the straight line connecting detectors A and B may not pass through the actual annihilation point. The detector pair can produce a misleading line of response even though both signals came from the same annihilation.

The photon may also arrive with less energy because of the scattering interaction. An energy window centred around the expected 511 keV value can reject many scattered events, although not every scattered photon will be identified successfully.

A random coincidence is different. It occurs when photons from two separate annihilations happen to reach the detectors within the same coincidence window. The scanner mistakenly pairs them even though they did not originate from the same event.

Coincidence timing is therefore an important selection tool, not absolute proof that a detected pair provides a correct line of response.

Question 5

An idealised PET scanner has detectors A and B separated by 60 cm. In one coincidence event, the photon at B arrives \(0.60\ \text{ns}\) before the photon at A.

Assume that:

• Both photons travel at speed \(c=3.00\times10^8\ \text{m s}^{-1}\).
• The photons travel in exactly opposite directions.
• Detector timing is perfect, and the annihilation point lies on the line AB.
• The positron travels no more than 4 mm from its nuclear decay site before annihilation, and it may travel in any direction.

A student concludes that the nuclear decay occurred exactly 9.0 cm towards B from the midpoint, with any uncertainty restricted to a 4 mm segment along AB.

Evaluate both claims. Calculate the annihilation point’s position, describe the possible region of nuclear decay sites, and explain whether two different decay sites could produce identical detector data.

Solution 5

The annihilation point is exactly 9.0 cm towards B from the midpoint under the stated ideal assumptions. However, the original nuclear decay point is not uniquely determined and need not lie on AB.

Step 1: Calculate the annihilation point.

The photon at B arrives first, so the annihilation point is closer to B. Use the TOF relationship:

\[ x=\frac{c\Delta t}{2} \]

Substitute the measured time difference:

\[ \begin{aligned} x&=\frac{(3.00\times10^8) (0.60\times10^{-9})}{2}\ \text{m}\\ &=0.090\ \text{m}\\ &=9.0\ \text{cm} \end{aligned} \]

The factor of two accounts for both path changes: the path towards B is shorter by \(x\), while the path towards A is longer by \(x\).

Step 2: Identify what the timing measurement actually locates.

The timing measurement locates the annihilation point, which we can call Q. It measures the difference in the gamma photons’ travel times from Q to the detectors.

It does not measure how long the positron travelled before reaching Q, nor does it reveal the direction of that travel.

Step 3: Determine the possible nuclear decay sites.

The positron can travel up to 4 mm in any direction before annihilation. Therefore, assuming the surrounding tissue is accessible, the original decay site could lie anywhere within a region of radius 4 mm centred on Q.

This region is three-dimensional. It is not restricted to a 4 mm segment along AB.

Step 4: Test whether the decay site can be uniquely recovered.

Consider two possible decay sites, P and P’, both within 4 mm of Q. A positron emitted at either site could travel to the same annihilation point Q. If both annihilations produce the same photon geometry and timing, the detector data would be identical.

The scanner cannot distinguish those original decay sites from this single event. Even perfect timing does not reveal the positron’s path; it only determines the gamma photons’ travel-time difference from the annihilation point.

In a real scanner, finite timing resolution, photon non-collinearity, and detector resolution introduce additional uncertainty.

The crucial distinction is that time-of-flight information improves the location of the annihilation, not the direct observation of the positron’s original emission point.

Understanding that distinction is useful whenever you interpret PET images or explain how a measured signal can reveal a source’s location without directly observing the particle that produced it.