Scintillation Detectors, Photomultipliers & Gamma Spectroscopy

Scintillation Physics & Crystal Properties

Scintillation detectors operate through a two-stage detection process: an ionizing photon or charged particle deposits energy in a scintillating material to produce low-energy fluorescence photons (optical or ultraviolet), which are subsequently converted into an amplified electrical signal by a photosensor (such as a photomultiplier tube or silicon photomultiplier).

Inorganic Scintillator Bandgap Mechanism

In pure inorganic ionic crystals (e.g., pure sodium iodide), the electronic energy structure consists of a lower valence band (filled electron states) and an upper conduction band (empty electron states) separated by a wide forbidden energy bandgap ($E_g \approx 6 - 9\text{ eV}$). When ionizing radiation traverses the lattice, it excites electrons into the conduction band, leaving holes in the valence band, or forms bound electron-hole pairs called excitons.

Direct radiative recombination across the wide bandgap is inefficient and produces energetic UV photons that match the bandgap energy. Consequently, pure crystals reabsorb their own emission (self-absorption). To overcome this, inorganic crystals are doped with trace activator impurities (such as thallium, $\text{Tl}^+$, or cerium, $\text{Ce}^{3+}$) at concentrations of ~0.1 to 0.2 mole percent:

1. The activator ions introduce localized, discrete electronic energy levels within the forbidden bandgap.
2. Ionization energy and mobile excitons migrate through the crystal lattice and transfer their energy to these activator centers.
3. The excited activator states relax back to their ground state via prompt radiative electronic dipole transitions, emitting visible photons (e.g., $\lambda \approx 415\text{ nm}$ for $\text{NaI(Tl)}$).
4. Because the transition energy between activator levels is significantly less than the bulk bandgap $E_g$, the emitted optical photons cannot excite electrons across the bandgap. The crystal is optically transparent to its own scintillation light.

Conduction Band  ==============================================
                                |              ^
                                | Energy       | Ionization
                                v Transfer     | Excitation
Activator States                +----+         |
(within forbidden bandgap)      |    |         |
                                |    | hv (scintillation light ~415 nm)
                                +----+         |
Valence Band     ==============================================

Scintillation Crystal Comparison

---

Photomultiplier Tube Architecture & Signal Processing Chain

To translate faint scintillation light into a measurable voltage pulse, the crystal is optically coupled (via silicone optical grease) to a Photomultiplier Tube (PMT).

Scintillation Photons
        |     +--------------------------------------------------------------+
        v     |                                                              |
  [ Photocathode ] --> [ Focusing ] --> [ D1 ] --> [ D2 ] ... [ D10 ] --> [ Anode ]
    (Bi-alkali,         Electrode        Gain = delta^N ~ 10^6 - 10^7        Pulse Q
     eta ~ 25%)                                                                |
                                                                               v
                                                                         [ Preamplifier ]
                                                                               |
                                                                               v
                                                                        [ Linear Amp ]
                                                                               |
                                                                               v
                                                                           [ MCA/ADC ]

The PMT Dynode Chain

1. Photocathode: A thin bi-alkali layer (typically antimony-potassium-cesium, $\text{Sb-K-Cs}$) deposited on the inside of the PMT glass window converts arriving scintillation photons into photoelectrons via the photoelectric effect. The quantum efficiency $\eta$ represents the probability of photoelectron emission per incident photon: $\eta = \frac{N_{\text{photoelectrons}}}{N_{\text{scintillation photons}}} \approx 20\% - 30\%$
2. Focusing Electrodes: Electrostatic fields direct the diverging photoelectrons onto the first dynode ($D_1$).
3. Dynode Electron Multiplier: A cascade of 10 to 14 curved dynodes, each held at an increasingly positive electrical potential ($\,\Delta V \approx 100\text{ V}\,$ per stage) via a resistive voltage divider. When an accelerated electron strikes a dynode coated with a secondary emission material (such as $\text{BeO}$ or $\text{GaP}$), it ejects $\delta$ secondary electrons (typically $\delta \approx 4 - 6$). The overall PMT current gain $G$ across $N$ dynodes is: $G = \alpha \cdot \delta^N \approx 4^{10} \approx 1.05 \times 10^6$
4. Anode: The final anode collects the dense cloud of multiplied electrons, producing a total charge packet $Q = N_{\text{pe}} \cdot e \cdot G$.

The Pulse-Height Electronics Chain

• Charge-Sensitive Preamplifier: Mounted adjacent to the PMT anode to minimize cable capacitance and electronic noise. It integrates the charge packet $Q$ on a feedback capacitor $C_f$, generating a step-voltage pulse with amplitude $V_{\text{peak}} = Q / C_f$ and a long exponential tail (~50–100 $\mu\text{s}$).
• Linear Shaping Amplifier: Performs CR-RC differentiation and integration to reshape the long-tailed preamplifier signal into a narrow, semi-Gaussian pulse (shaping time $\tau_{\text{shape}} \approx 0.5 - 2.0\ \mu\text{s}$). It incorporates pole-zero cancellation to prevent pulse undershoot and baseline restorer circuits to eliminate spectral distortion at high count rates.
• Multichannel Analyzer (MCA): Converts analog peak voltages into discrete digital channels through an Analog-to-Digital Converter (ADC):
  • Wilkinson ADC: Charges a capacitor to peak pulse height and discharges it with a linear constant-current source while a high-frequency clock counts clock pulses. It provides exceptional channel linearity, but the conversion dead time is proportional to pulse amplitude ($t_{\text{conv}} \propto V$).
  • Successive Approximation / Flash ADC: Employs binary comparator ladders to digitize pulses in a fixed, fast conversion window ($<1\ \mu\text{s}$), making it the modern standard for high-throughput nuclear medicine spectroscopy.

---

Gamma Spectrometry Features & Interaction Kinematics

When monoenergetic gamma photons interact within a scintillator, three primary interaction mechanisms shape the resulting pulse-height distribution:

Counts
  ^
  |        Compton Continuum           Full-Energy Photopeak
  |        +-----------------+               +---+
  |        |                 |               |   |
  |        |    Compton Edge |               |   |
  | Back-  |          |      v               |   |
  | scatter|          |     ---              |   |
  |  Peak  |          |    /   \             |   |
  |   +-+  |          |   /     \            |   |
  |   | |  |          v  /       \           |   |
  +---+-----------------+---------+----------+---+-------------> Energy (keV)
      E_BS             E_C                   E_0

1. The Full-Energy Photopeak

Occurs when the incident gamma ray deposits its entire energy $E_0$ inside the crystal through a single photoelectric absorption event, or through successive Compton scattering events terminating in photoelectric capture. The resulting peak is centered at energy $E_0$.

2. Compton Continuum and the Compton Edge ($E_C$)

Compton scattering occurs when a photon with energy $E_\gamma$ scatters off an atomic electron at an angle $\theta$. The energy of the scattered photon $E_\gamma'$ is governed by relativistic kinematics:

$E_\gamma' = \frac{E_\gamma}{1 + \frac{E_\gamma}{m_e c^2}(1 - \cos\theta)}$

The recoil electron kinetic energy $T_e = E_\gamma - E_\gamma'$ is absorbed in the crystal. Maximum energy transfer to the electron occurs during a direct 180° backscatter collision ($\cos 180^\circ = -1$):

$E_C = T_{e,\max} = E_\gamma - \frac{E_\gamma}{1 + \frac{2 E_\gamma}{m_e c^2}} = \frac{E_\gamma}{1 + \frac{m_e c^2}{2 E_\gamma}}$

This maximum electron energy forms the sharp upper boundary of the Compton continuum, termed the Compton edge ($E_C$).

Worked Example: Compton Edge and Backscatter Calculation for Cesium-137

Calculate the Compton edge ($E_C$) and backscatter peak ($E_{\text{BS}}$) energies for the 661.7 keV gamma emission of cesium-137 ($m_e c^2 = 511.0\text{ keV}$).

1. Compute the Compton edge energy $E_C$: $E_C = \frac{661.7}{1 + \frac{511.0}{2(661.7)}} = \frac{661.7}{1 + \frac{511.0}{1323.4}} = \frac{661.7}{1 + 0.38612} = \frac{661.7}{1.38612} \approx 477.37\text{ keV}$
2. Compute the 180° backscatter peak energy $E_{\text{BS}}$ (photons scattering at 180° in the surrounding lead shield and re-entering the crystal): $E_{\text{BS}} = \frac{E_\gamma}{1 + \frac{2 E_\gamma}{m_e c^2}} = E_\gamma - E_C = 661.7 - 477.37 \approx 184.33\text{ keV}$

Note the exact kinematic symmetry: $E_C + E_{\text{BS}} = E_\gamma$.

3. Pair Production Escape Peaks

For high-energy gamma rays ($E_\gamma > 2 m_e c^2 = 1.022\text{ MeV}$), pair production occurs in the nuclear Coulomb field, creating an electron-positron pair. When the positron thermalizes, it annihilates with an atomic electron, producing two co-linear 511 keV annihilation photons:

• Single Escape Peak (SEP): One 511 keV photon escapes the crystal without interacting, while the other is completely absorbed: $E_{\text{SEP}} = E_0 - 511\text{ keV}$.
• Double Escape Peak (DEP): Both 511 keV annihilation photons escape the crystal boundaries: $E_{\text{DEP}} = E_0 - 1022\text{ keV}$.

---

Energy Resolution Calculation

The energy resolution of a spectrometer reflects its ability to resolve closely spaced gamma-ray energies. Resolution is defined as the percentage ratio of the Full Width at Half Maximum (FWHM) of the photopeak to the centroid energy $E_0$:

$\% R = \frac{\text{FWHM}}{E_0} \times 100\%$

Counts
   ^
   |               /\  <-- Peak Centroid (E_0 = 661.7 keV)
   |              /  \
   |   Half Max  +----+ <-- FWHM = 46.3 keV
   |            /      \
   |           /        \
---+----------+----------+--------------------> Energy

Worked Example: NaI(Tl) Energy Resolution Calculation

A 3" × 3" $\text{NaI(Tl)}$ detector is calibrated using a certified $^{137}\text{Cs}$ standard source. The 661.7 keV photopeak is centered at MCA channel 662. The peak width at half the maximum net peak height spans from channel 639 to channel 685 (FWHM = 46.0 channels = 46.0 keV).

$\% R = \frac{46.0\text{ keV}}{661.7\text{ keV}} \times 100\% = 6.95\%$

Standard clinical acceptance criteria require a 3" × 3" $\text{NaI(Tl)}$ crystal to achieve an energy resolution $\% R \le 7.5\%$ at 662 keV. Resolutions $>8.5\%$ indicate crystal hydration, yellowing, optical decoupling, or PMT dynode degradation.

Statistical Foundation of Resolution

Unlike HPGe semiconductor detectors which require only $\sim 2.96\text{ eV}$ to create an electron-hole pair (yielding $\% R < 0.2\%$), an inorganic scintillator requires $\approx 26\text{ eV}$ to produce one optical photon, of which only ~25% liberate photoelectrons at the photocathode (~100 eV per collected photoelectron). For a 662 keV gamma ray, this produces only:

$N_{\text{pe}} \approx \frac{661{,}700\text{ eV}}{100\text{ eV/pe}} \approx 6{,}600\text{ photoelectrons}$

Poisson counting statistics dictate an irreducible theoretical standard deviation of $\sigma_N = \sqrt{N_{\text{pe}}}$. Because $\text{FWHM} = 2.355 \cdot \sigma$, Poisson statistics set an absolute physical resolution limit:

$\% R_{\text{statistical}} = \frac{2.355 \sqrt{N_{\text{pe}}}}{N_{\text{pe}}} \times 100\% = \frac{2.355}{\sqrt{6{,}600}} \times 100\% \approx 2.9\%$

Optical light collection non-uniformities and dynode multiplication variance increase actual $\text{NaI(Tl)}$ resolution to the observed 6.5%–7.5% range.