Why is NaI(Tl) used for gamma cameras?

NaI(Tl) combines a high effective atomic number and density (about 3.67 g/cm³) for good 140 keV absorption with a high light output that peaks near 415 nm, well matched to photomultiplier tubes. It can also be grown as a large single crystal, which is what makes a wide field-of-view camera practical. Its main drawbacks are that it is hygroscopic and relatively slow.

Why is a photopeak energy window used?

A pulse-height analyzer accepts only events whose total signal falls within a window centered on the photopeak, typically a 20% window (about 126 to 154 keV for the 140 keV technetium-99m photopeak). This rejects most scattered photons, which have lost energy and therefore produce smaller signals, improving image contrast. A camera with better energy resolution can use a tighter window and reject more scatter.

Who should perform gamma camera acceptance and quality control testing?

A qualified or board-certified medical physicist should perform acceptance testing and oversee the routine quality control program, measuring uniformity, spatial resolution and linearity, energy resolution, and sensitivity against manufacturer and NEMA specifications. DRPS provides nuclear medicine physics testing and program support.

Gamma Camera Detector Physics and Performance

Q: What are the main components of a gamma camera detector?

A conventional Anger gamma camera has a collimator, a large single-crystal sodium iodide scintillator doped with thallium (NaI(Tl)), a light guide, an array of photomultiplier tubes (PMTs), and the positioning and pulse-height-analysis electronics. The crystal converts a gamma ray into visible light, the PMTs convert that light into electrical signals, and the electronics compute where the event happened and how much energy it deposited.

Q: What is Anger positioning logic?

Anger logic estimates an event's location as the signal-weighted average (centroid) of the photomultiplier outputs. The x and y coordinates are the sums of each PMT's position multiplied by its signal, divided by the total signal, while the total signal itself is the energy signal used for pulse-height analysis. It is why a camera with a few dozen PMTs can localize events to a few millimeters.

Q: What is the difference between intrinsic and system spatial resolution?

Intrinsic spatial resolution describes the detector alone, without a collimator, and is typically about 3.5 to 4 mm full width at half maximum. System (extrinsic) resolution includes the collimator and is what the clinical image sees; it combines the intrinsic and collimator components in quadrature and is usually dominated by the collimator at clinical distances.

A gamma camera's clinical image quality is largely decided inside the detector — in the sodium iodide crystal, the photomultiplier array, and the positioning and energy electronics — before a single photon reaches the collimator. Understanding how these components set intrinsic spatial resolution, energy resolution, and the photopeak window explains exactly what acceptance and routine quality-control tests are measuring and why they matter.¹²³

The Anger scintillation camera has been the workhorse of planar and SPECT imaging for decades. Its design is deceptively simple: a large NaI(Tl) crystal turns a gamma ray into a flash of light, an array of photomultiplier tubes (PMTs) senses that flash, and the electronics compute both where the interaction happened and how much energy it deposited. Every performance metric a physicist reports — uniformity, resolution, linearity, energy resolution, sensitivity — traces back to how well that chain performs.²³⁴

Introduction

The gamma camera's job is to form an image from individual gamma-ray interactions, one event at a time, by answering two questions for each event: where and how much energy. The "where" becomes a pixel location; the "how much" decides whether the event is kept or rejected as scatter. The precision of those two answers is the detector's intrinsic performance, and it caps what any collimator or reconstruction can achieve downstream.²⁴

This guide follows a gamma ray from the crystal to the image: the scintillator, the PMT array, Anger positioning logic, pulse-height analysis and the photopeak window, and the physics that sets intrinsic spatial and energy resolution. It then connects those principles to the acceptance and routine QC tests a medical physicist performs, the clinical consequences of detector performance, and the accreditation framework that requires it. Numbers are drawn from established instrumentation references so that expectations rest on measurement.¹²³

Topic Explanation

The detector chain

A conventional Anger camera detector consists of, in the order a gamma ray encounters them:²⁴

Collimator — defines the direction of accepted photons (omitted for intrinsic measurements).
NaI(Tl) crystal — a single thallium-doped sodium iodide scintillator, commonly 3/8 inch (about 9.5 mm) thick, that absorbs the gamma ray and emits visible light.
Light guide / optical coupling — channels scintillation light to the photodetectors.
Photomultiplier tube array — dozens of PMTs that convert light into measurable electrical pulses.
Positioning and pulse-height electronics — compute event location and total energy.

Why NaI(Tl)?

Sodium iodide doped with thallium remains the standard scintillator for gamma cameras because it balances several competing requirements.²⁴ It has a density of about 3.67 g/cm³ and a high effective atomic number, giving good photoelectric absorption of the 140 keV photons emitted by technetium-99m, the most common clinical radionuclide.⁸ Its light output is high and its emission peaks near 415 nm, closely matched to the peak sensitivity of bialkali PMT photocathodes. Crucially, NaI(Tl) can be grown as a large single crystal, which is what makes a wide field-of-view camera possible. Its disadvantages are that it is hygroscopic — it must be hermetically sealed — and comparatively slow, which limits count-rate performance.²⁴

How an event becomes a pixel

When a gamma ray undergoes photoelectric absorption in the crystal, it produces a burst of visible light roughly proportional to the deposited energy. That light spreads to several nearby PMTs, each producing a signal proportional to the light it receives. The electronics then do two things: they estimate the position of the interaction from the pattern of PMT signals, and they sum all the signals to estimate the deposited energy. Only events whose summed energy falls inside the photopeak window are recorded in the image.²⁴

Key Technical Principles

Anger positioning logic

The position of each event is estimated as the signal-weighted centroid of the PMT outputs. If PMT $i$ sits at position $(x_{i}, y_{i})$ and produces signal $S_{i}$ , the estimated coordinates are:

X = \frac{\sum _{i} x _{i} S _{i}}{\sum _{i} S _{i}}, Y = \frac{\sum _{i} y _{i} S _{i}}{\sum _{i} S _{i}}

and the total (energy) signal is:

Z = i \sum S_{i}

This elegant result — first described by Hal Anger — is why a camera with only a few dozen PMTs can localize an event to a few millimeters: the centroid interpolates between tubes rather than simply assigning the event to the nearest one. Modern cameras apply energy, linearity, and uniformity correction maps to these raw coordinates, but the underlying computation is still a weighted centroid.²⁴

Pulse-height analysis and the photopeak window

The energy signal $Z$ is sorted by a pulse-height analyzer, which accepts only events within an energy window centered on the photopeak. A 20% window centered on the 140 keV technetium-99m photopeak spans:

140 keV \pm (0.10 \times 140 keV) = 126 - 154 keV

Because a photon that scattered in the patient has lost energy, it produces a smaller signal and is more likely to fall below the window. The window therefore rejects most scatter and improves image contrast.² AAPM acceptance-testing guidance describes centering a 20% window symmetrically on the photopeak, with a 15% window as a tighter alternative when energy resolution allows.²

Energy resolution

Energy resolution quantifies how sharply the camera measures deposited energy, expressed as the full width at half maximum (FWHM) of the photopeak divided by the photopeak energy:

R_{E} (%) = \frac{FWHM}{E _{photopeak}} \times 100

A modern gamma camera achieves an energy resolution of about 9–10% at 140 keV.¹⁷ Thus a 10% resolution corresponds to a photopeak FWHM of about 14 keV, which is precisely why a 20% (±14 keV) window captures essentially all photopeak events while still excluding scatter. Energy resolution improves (the percentage gets smaller) at higher photon energies because more scintillation light is produced per event, reducing statistical fluctuation. Better energy resolution lets a camera use a narrower window and reject more scatter — a direct link between a bench measurement and clinical contrast.¹²

Intrinsic versus system spatial resolution

Intrinsic spatial resolution describes the detector alone, with no collimator. It is limited by the statistics of how scintillation light divides among the PMTs and by the physics of energy deposition, and it degrades at lower photon energy. Typical intrinsic resolution is about 3.5 to 4 mm FWHM; a thinner 3/8-inch crystal yields roughly 3.5 mm FWHM versus about 3.9 mm for a thicker 5/8-inch crystal, because a thinner crystal spreads less light but absorbs fewer photons.² This is the crystal-thickness trade-off: thinner crystals improve spatial resolution at the cost of detection efficiency.²

System (extrinsic) resolution includes the collimator and is what the clinical image actually sees. The intrinsic ( $R_{i}$ ) and collimator ( $R_{c}$ ) components add in quadrature:

R_{sys} = R_{i}^{2} + R_{c}^{2}

Worked resolution example

Suppose a camera has an intrinsic resolution $R_{i} = 3.5$ mm and is fitted with a low-energy high-resolution collimator whose resolution at a 10 cm source-to-collimator distance is $R_{c} = 7.4$ mm (a representative value). The system resolution is:

R_{sys} = (3.5)^{2} + (7.4)^{2} = 12.25 + 54.76 \approx 8.2 mm

The result is dominated by the collimator, and it makes a practical point: once the collimator is attached, modest changes in intrinsic resolution have limited effect on the clinical image, but a degraded intrinsic resolution (from a PMT drift or a tuning error) still shows up. This is why QC monitors the intrinsic performance directly rather than assuming the collimator hides it.²⁴

Detector performance at a glance

Parameter	Typical value (Tc-99m, 140 keV)	What degrades it
Crystal thickness	3/8 in (≈9.5 mm)	— (design trade-off)
Intrinsic spatial resolution	≈3.5–4 mm FWHM	PMT drift, thicker crystal, lower energy
System resolution (LEHR, 10 cm)	≈8–10 mm FWHM	Collimator, source distance
Energy resolution	≈9–10% FWHM	PMT gain drift, crystal aging
Photopeak window	20% (126–154 keV)	Poor energy resolution forces a wider window
Integral uniformity	within a few percent	PMT nonuniformity, correction-map error

Clinical Impact

Detector performance sets a ceiling on everything the clinician sees. Poor energy resolution forces a wider photopeak window, admits more scatter, and lowers lesion contrast. Degraded intrinsic resolution or linearity distorts small structures and can create artifacts that mimic or mask disease. Nonuniformity is especially dangerous in SPECT: a small, stable nonuniformity at the center of rotation produces ring artifacts in reconstructed slices that can be misread as pathology.¹³

Because these failures are often gradual — a slowly drifting PMT, an aging crystal, a stale correction map — they are invisible without routine QC. The daily uniformity flood, periodic resolution and linearity bar-phantom checks, and energy-peak verification exist precisely to catch detector drift before it reaches the clinical image.¹³ For how these QC tests are performed and interpreted, see our guides to gamma camera uniformity QC and gamma camera energy resolution QC.

Practical Optimization Tips

Peak the camera and verify the window

Before clinical use each day, confirm the camera is peaked on the correct photopeak for the radionuclide in use and that the window is set appropriately. A mis-set energy peak or window is one of the most common causes of degraded images and failed uniformity floods.¹³

Match crystal and collimator to the task

Camera intrinsic resolution is fixed by the hardware, but the collimator is the user's lever. A low-energy high-resolution collimator favors resolution; a high-sensitivity collimator favors count rate. Because system resolution is dominated by the collimator at clinical distances, collimator choice usually matters more to the image than small intrinsic differences.²⁴

Keep the source close

System resolution degrades with source-to-collimator distance, so positioning the detector as close to the patient as safely possible is the simplest resolution gain available at acquisition time.²⁴

Protect the crystal

NaI(Tl) is hygroscopic and thermally sensitive. Avoid rapid temperature swings and physical shock, both of which can crack the crystal or admit moisture, producing permanent nonuniformity. Follow manufacturer warm-up and handling procedures.²⁴

Hold a disciplined QC schedule

Follow a documented QC program: daily uniformity and energy-peak checks, periodic spatial resolution and linearity (bar phantom), and acceptance-level testing after installation or major service, all compared against NEMA-style specifications and manufacturer values.¹⁵

Regulatory Considerations

Gamma camera performance sits at the intersection of accreditation requirements and radioactive-material rules. The detector itself is a radiation-detection instrument, not a radiation source, so image-performance standards come primarily from accreditation and professional guidance rather than from the Nuclear Regulatory Commission.

Performance standards and methods. NEMA NU 1 defines the standardized methods for measuring gamma camera performance characteristics — intrinsic spatial resolution, energy resolution, uniformity, sensitivity, and count-rate performance — that manufacturers and physicists use to specify and verify a camera.⁵ AAPM Report No. 6 and IAEA Human Health Series No. 6 provide acceptance-testing and quality-control frameworks.²³
Accreditation. Clinical sites typically maintain nuclear medicine accreditation (for example, through the ACR or the Intersocietal Accreditation Commission), which requires a defined QC program and a qualified medical physicist's involvement in acceptance and annual performance evaluation.
Radioactive material. The technetium-99m and other radionuclides imaged on the camera are byproduct material governed by 10 CFR Part 35 (or the equivalent Agreement State program) and 10 CFR Part 20, which separately require QC of the dose calibrator and radiation safety controls — distinct from camera image QC.
State jurisdiction. Across the states DRPS serves — Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware — radioactive-material medical use is administered by the NRC or the Agreement State program (Washington DC and Delaware are direct-NRC). Always confirm requirements with the authority having jurisdiction.

For the formal performance-testing procedures that put these principles into practice, see our guide to gamma camera NEMA NU-1 performance testing.

Frequently Asked Questions (FAQs)

What are the main components of a gamma camera detector?

A collimator, a thallium-doped sodium iodide (NaI(Tl)) crystal, a light guide, an array of photomultiplier tubes, and the positioning and pulse-height electronics. The crystal makes light, the PMTs make signals, and the electronics compute position and energy.²⁴

Why is NaI(Tl) used?

It has high density (about 3.67 g/cm³) and atomic number for good 140 keV absorption, a high light output peaking near 415 nm matched to PMTs, and it can be grown as a large single crystal. It is hygroscopic and relatively slow.²⁴

What is Anger positioning logic?

The event location is the signal-weighted centroid of the PMT outputs: the sums of each PMT position times its signal, divided by the total signal. The total signal is the energy used for pulse-height analysis.²⁴

Why a photopeak window?

A 20% window (about 126–154 keV for the 140 keV technetium-99m photopeak) accepts photopeak events and rejects most scattered photons, which have lost energy. Better energy resolution allows a tighter window and more scatter rejection.¹²

Intrinsic versus system resolution?

Intrinsic resolution (≈3.5–4 mm FWHM) is the detector alone; system resolution adds the collimator in quadrature and is usually collimator-dominated at clinical distances.²

Key Takeaways

The detector chain — NaI(Tl) crystal, PMT array, positioning and energy electronics — sets the ceiling on gamma camera image quality.²⁴
Anger logic localizes each event as the signal-weighted centroid of PMT outputs, and the summed signal is the energy used for pulse-height analysis.²⁴
A 20% photopeak window (126–154 keV at 140 keV) rejects scatter; better energy resolution (≈9–10% FWHM) allows a tighter window.¹²
Intrinsic spatial resolution is ≈3.5–4 mm FWHM and trades off against efficiency through crystal thickness; system resolution adds the collimator in quadrature and is collimator-dominated.²
Detector drift is gradual and invisible without QC; uniformity, resolution, linearity, and energy checks catch it before it reaches the image.¹³
NEMA NU 1, AAPM Report No. 6, and IAEA guidance define the measurement methods; accreditation requires a physicist-led QC program.²³⁵

Conclusion

The Anger gamma camera turns individual gamma-ray interactions into clinical images through a well-understood physical chain, and that chain — crystal, photomultipliers, positioning logic, and pulse-height analysis — determines the resolution, energy discrimination, and uniformity the clinician ultimately relies on. Knowing how intrinsic performance arises explains why acceptance and routine QC tests are designed the way they are, and why a drifting detector must be caught at the bench long before it distorts a patient study.²³⁴

How DRPS Can Help

Diagnostic Radiation Physics Services (DRPS) supports nuclear medicine and SPECT facilities across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware with PET/CT and nuclear medicine physics, gamma camera acceptance testing, annual performance evaluation, NEMA-based QC program design, and accreditation support by board-certified medical physicists. Contact us to review your gamma camera QC program.

Related Resources

References

Zanzonico P. Routine quality control of clinical nuclear medicine instrumentation: a brief review. J Nucl Med. 2008;49(7):1114-1131. doi:10.2967/jnumed.107.050203. doi.org
American Association of Physicists in Medicine. Scintillation Camera Acceptance Testing and Performance Evaluation. AAPM Report No. 6. New York, NY: AAPM; 1980. aapm.org
International Atomic Energy Agency. Quality Assurance for SPECT Systems. IAEA Human Health Series No. 6. Vienna: IAEA; 2009. iaea.org
International Atomic Energy Agency. Nuclear Medicine Physics: A Handbook for Teachers and Students. Vienna: IAEA; 2014. iaea.org
National Electrical Manufacturers Association. Performance Measurements of Gamma Cameras. NEMA Standards Publication NU 1-2018. Rosslyn, VA: NEMA; 2018. nema.org
Sánchez F, Fernández MM, Giménez M, et al. Performance tests of two portable mini gamma cameras for medical applications. Med Phys. 2006;33(11):4210-4220. doi:10.1118/1.2358199. doi.org
Russo P, Mettivier G, Pani R, Pellegrini R, Cinti MN, Bennati P. Imaging performance comparison between a LaBr3:Ce scintillator based and a CdTe semiconductor based photon counting compact gamma camera. Med Phys. 2009;36(4):1298-1317. doi:10.1118/1.3081412. doi.org
International Atomic Energy Agency, Nuclear Data Services. Live Chart of Nuclides: technetium-99m decay data (140.5 keV gamma emission). Vienna: IAEA. iaea.org
U.S. Nuclear Regulatory Commission. 10 CFR Part 35, Medical Use of Byproduct Material. nrc.gov