CsI(Na) physical properties

CsI(Na) is an inorganic scintillator with a complex light output function. It has excellent stopping power and can deliver energy resolutions superior to NaI(Tl). The material responds to γ-rays with a rise time of about 150 ns at room temperature. The light pulse subsides according to a complex non-exponential time law. Over the first 10 μs the time evolution of the scintillation light pulse can be parametrized quite accurately as shown below. The parameters are given in the table at the end of this section.

A(t) = (1 - e^-t/σ) × (A₀e^-t/τ + A₁e^-t/Τ)

At room temperature the rise time (10% to 90%) is about 150 . The light pulse decay can be grouped into a fast and a slow component, which carry equal integrated light or measured PMT anode charge. The exponential time constants associated with the two components are 0.5 μs and 5 μs, respectively.

This mixture of time constants makes it difficult for classical analog electronics to operate CsI(Na) detectors with precision at pulse rates above a few kcps. But the all-digital eMorpho has no difficulty operating a CsI(Na) detector core at rates of 20 kcps with very good spectrosopic performance.

A great advantage of this material is its tolerance to mechanical and thermal shock. It is malleable and does not cleave. Hence, it is ideal for outdoor and fieldable instruments where there is a risk of serious mechanical or thermal shock.

Characteristic	Value
Density	4.51 g/cm³, 2.61 oz/inch³

Atomic weight	259.81 g/mol
Molecular density	0.01736 mol/cm³, 1.046 10²² molecules/cm³
Refractive index	1.78
Rise time constant	σ = 50 ns
Fast component amplitude	A₀ = 1.0
Fast component decay time	τ = 0.50μs
Slow component amplitude	A₁ = 0.1
Slow component decay time	Τ = 5.0μs

Features

Plus:	Very high stopping power Excellent energy resolution (7.0% fwhm @ 662 keV) Best performance at E_γ>2 MeV
Minus:	Moderate cost Moderate speed <50 kcps

Go to

eMorpho page

CsI(Na) Pulse Shape

Figure 1: The graph to the left shows the CsI(Na) light pulse shape as measured using a 60 MHz, 10-bit eMorpho. The 10% to 90% rise time is about 150 ns at room temperature, but varies considerably from pulse to pulse.

Figure 2: This is a plot of the integrated charge of the electronic image of the light pulse vs. integration time. The obvious expectation is that collecting all the light, by integrating over all of the pulse, would yield the best energy resolution. In practice, integrating over the tail end of the pulse will be counter productive due to noise and ADC digitizing errors.

For CsI(Na) we find that integrating over 10 μs produces good energy resolution. Longer integration times will improve the energy resolution for large pulses, but will deteriorate the energy resolution for small signals due to noise and discretization errors.

Features

Pulse rise time:	150 ns for the light pulse
Decay time:	Complex law involving a 0.5 μs and a 5.0 μs component with equal charge (i.e. light) content
Very slow component:	Extends to more than 50 μs
Light collection:	50% in 1.0 μs 90% in 6.0 μs (compared to light collection at 10 μs)

CsI(Na) Non-proportionality

All inorganic scintillators are non-proportional. This means that the amount of scintillation light generated in response to energy deposited by a γ-ray is not strictly proportional to the deposited energy. Aside from photo-electron statistics, and especially at energy deposits beyond 1 MeV, this is a limiting factor for the energy resolution any scintillator can achieve.

One way to quantify this phenomenon is to compute the apparent brightness of the scintillator as a function of deposited energy. Here we measure brightness in units of PMT anode pulse charge vs deposited energy. We arbitrarily define the value at 662 keV as 100% and quantify the non-proportionality by how much the charge-to-energy ratio deviates from the pivot point at 662 keV.

Figure 3: Non-proportionality as measured in a 3-inch CsI(Na) crystal. The general trend is that the CsI(Na) crystal appears brighter at lower γ-ray energies and loses luster as the γ-ray energy increases. Note that this non-proportionality is a function of the γ-ray energy, not the total deposited energy. Cascade sums are not subject to scintillator non-proportionality.

⁶⁰Co is the perfect example for this phenomenon. Place a 1 μCi (37 kBq) ⁶⁰Co-source right against the front face of a 3-inch radiation sensor. The excited daughter of ⁶⁰Co emits an 1172 keV and a 1333 keV γ-ray in coincidence. When both are fully absorbed in the scintillator, the sum signal should precisely match the formula 1172 keV + 1333 keV = 2505 keV.

Note that the non-proportionality is mostly, but not exclusively a characteristic of the material. Especially at the low energy end the effect is more pronounced in small crystals due to geometry and light collection effects.

Summary

Non- prop.:	Light output per deposited energy decreases with γ-ray energy.
Cascades:	Cascade, ie coincidence, summing is not affected by this phenomenon.

Go to

R2D sensors page