The Ionization-Chamber-Beammonitor I-BM for thermal and cold neutron beams First Differential Mobility Analysis (DMA) Measurements of Air Ions Produced by Radioactive Source and Corona

The Ionization-Chamber-Beammonitor I-BM is a beam monitor for thermal neutrons thought as a extremely robust beam monitoring solution for extreme environments of radiation exposure and thus very limited accessibility. The device, the front-end chamber itself, needs no servicing over its lifetime. It is composed as an entirely passive device without exposed gain stage or requirements concerning operation gas or other consumables. The beam monitor is designed as an ionization chamber with ambient air as gas filling and a natural Boron coating to give it sensitivity to thermal neutrons. In the environment of application the device is, apart from neutrons, typically exposed to very high gamma dose rates. To increase the devices specificity, it is designed as an ionization chamber pair with a measurement cell and a reference cell, the former being coated with a neutron converter, the latter being uncoated and serving as a gamma dose reference in a difference measurement. Thermal beam neutrons entering the ionization chamber will be captured by the neutron converter with a probability that depends upon the capture cross section as well as the overall thickness of the converter layer that is coated


CDT CASCADE Detector Technologies GmbH
Hans-Bunte-Straße 8-10 D -69123 Heidelberg chamber, operated sandwiching the common high voltage electrode. The neutron signal is fed to the inverting input of the TIA whereas the signal on the reference cell is fed to the non-inverting input as depicted in Figure 1. Both current signals are fed through the same impedance so that any signal generated identical (common mode) in both chambers is nulled. Only differential signals, which in this case are the neutron signals, generate a signal at the output of the TIA. The scheme simultaneously gives complete insensitivity to noise on the HV-supply, as this by symmetry appears as common mode on the amplifier´s input as well. Figure 1: Schematic presentation of the ionization beam monitor I-BM. The symmetric parallel plate ionization chamber pair on the left is fed with a biasing voltage on its central electrode. One chamber has an internal coating with neutron sensitive 10 B converter material, the other one is left uncoated. Neutron signals appear as differential signals, whereas Gamma signals together with noise on the HV supply are common mode and thus annihilated. The device is read-out by a noise optimized TIA, limited in bandwidth to about 125 kHz and thus optimally adapted to the fastest signals of ~ 10µsec duration that are projected for ESS. Between the ionization chamber and the electronic readout a shielded cable of up to 10m length may be inserted.

Lifetime of the converter coating at ESS
For the I-BM, 10 B as part of natural Boron is employed as neutron converter. Upon neutron capture, Boron undergoes the following nuclear reaction: The neutron capture cross section of σ = 3838 Barn implies that the probability for a Boron nucleus to capture a neutron and disintegrate in the neutron beam of capture flux Φ C is P = σ Φ C . For ESS with Φ C_average up to 5 x10 9 n/(cm 2 s) for the unchopped macro pulses with typical beam divergence (see below), this probability amounts to 2 x 10 -11 s -1 . In other words, a Boron atom and thus the monitor´s coating will have a lifetime in such a beam of 5 x 10 10 sec ~1600 years. But even for an allocation of the ionization chamber at a position covering much larger divergence the lifetime is still well acceptable.

Signal formation
Upon neutron capture the boron nucleus disintegrates into the two fragments stated above. The energy released is carried by the alpha and Li ions which are released back to back. As the converter is attached to a substrate and since only the fragments released into the detector gas can be detected, the energy deposit in the detector is a line spectrum of 4 lines with an energy of up to 1,78 MeV. These energetic ions generate electron ion pairs in the detector gas where they lose about 34eV/ion-pair generated in air. The range of these ions in air is several mm. With the ionization chambers gap of 2mm, most of the ions will deposit their kinetic energy only partially into the counting gas before they hit the inactive material of the opposing electrode. For any geometry of ionization chamber depth and coating thickness, an average energy deposit and thus charge deposit per neutron may be determined. For the example configuration of 1,2 µm coating thickness and (2,2 +/-0,1) mm chamber depth, this average Energy deposit was computed to amount to (521 +/-11) keV which is equivalent to a total charge deposit (positive and negative together) per converted neutron of Q o = 30.600 e. The corresponding signal energy spectra are depicted in Figure 2. Figure 2 Computed pulse height spectra of neutrons detected in a gas detector with a 1,2 µm thick 10 B 4 C converter. For very thin converter layers the energy spectrum is a line spectrum. For thick layers, it washes out to a continuous spectrum as the secondary particles lose energy in the passive converter layer on their way from point of conversion into the counting gas (left). This spectrum is further distorted if the depth of counting gas is limited to below the range of secondary ions in the counting gas. For a chamber depth of 2,2 mm the resulting pulse height spectrum is depicted on the right.
For a beam monitor, most of the neutrons will traverse the device undetected. Only a small fraction is being converted with a detection efficiency ε n . With the prototype device presented here, the 10 B 4 C converter coating was chosen comparatively thick with 1,2 µm, additionally both electrode surfaces of the chamber were coated, giving an overall detection efficiency of 3,8% for thermal neutrons. Detection efficiency for this configuration was computed and is depicted in Figure 3. A monitor device for operation at a facility like ESS would be adapted to the available flux and be equipped with a detection efficiency of about 10 -4 to 10 -3 , the goal being a broad dynamic range for beam intensities that can be covered with an adequate signal height.  Figure 3 Computed detection efficiency of the beam monitor configuration with 1,2µm thick 10 B 4 C converter coatings on both electrodes.

CDT CASCADE Detector Technologies GmbH
With these numbers at hand, the neutron capture flux Φ C through the ionization chamber (at vertical incidence) can be computed to: Φ 5,4 * 10 1 9 * 10 1 _ ! This holds true for dry air as detector gas. With humidity present, the effective signal height is somewhat compromised. The following plot gives a calibration factor for operation in humid 2 air. Figure 4 The effect of humidity on detector response for detectors where electrons and ions are drifted through air. Here a Honeywell PID was studied, which however relies upon a very similar detection and charge collection mechanism as the I-BM. For typical environmental conditions of relative humidity between 40 and 60 per cent, the signal, that is the output current, is degraded to about β humidity ~ 2/3 as an effect of humidity. For this typical case of operation in humid air with a correction factor β hum = 2/3 to the measured current, the neutron capture flux is: The uncertainty in calibration introduced through the uncertainty in changing environmental conditions amounts to about +/-7,5% over typical environmental conditions in relative humidity, atmospheric pressure and temperature, where the former clearly predominates. It is, apart from the signal rise time which is for operation in air radically slowed down by three orders of magnitude, the price that is payed for device robustness. However, even the shortest pulse durations envisioned for installations like ESS can be resolved with this device. Further, the calibration factors related to environmental conditions show time constants in the order of hours or days, which are not in the realm of what these devices are intended to monitor and serve for.

CDT CASCADE Detector Technologies GmbH
Conversely, assuming the peak capture flux projected for ESS as described below at a value of Φ C_peak ~10 11 n/(cm 2 s), the ionization chamber realized and described herein would generate a current of 15µA or an output signal of 90V with a 1 cm 2 beam, about an order of magnitude higher than the dynamic range. Consequently, an ESS dedicated device should be equipped with one boron layer only of about 100nm thickness and of natural isotopic mixture. Further adaptation may be realized by modification of the feedback resistance R.

Noise
The noise performance of the transimpedance amplifier TIA is optimized through the choice of input amplifier, feedback resistor, input capacitance (sensor and cable) and compensation capacitance. In effect, the amplifier output noise is constant at constant bandwidth, where bandwidth is determined by the RC of the feedback resistor and the compensation capacitor which again is directly proportional to the input capacitance. If a long operating cable is needed to allocate the active electronics in a safe place, thus introducing a large input capacitance to the electronics, then a large compensation capacitance is needed to secure stability. In order to maintain a certain bandwidth that may be needed to locally detect a pulsed neutron beam, the feedback resistance R f needs to be adapted. In final conclusion, the signal to noise ratio at constant bandwidth is proportional to the feedback resistance R f .

Realization
A first realization of the Ionization-Chamber-Beam-Monitor I-BM is depicted in Figure 6. The device was built with an aperture of 70mm for evaluation purposes. The ionization chamber itself can be adapted in size to the particular application. It consists of Aluminum electrodes and ceramic or FR-4 based insulators. Each device contains two ionization chambers formed by a central HV electrode sandwiched between two current collecting electrodes and 2,2mm wide detector gaps. One of the ionization chambers is coated with isotopically enriched 10 B 4 C. The other ionization chamber is insensitive to neutrons as it has no coating. It serves as reference chamber to gain complete insensitivity to gamma-irradiation induced current signals. Insensitivity is achieved through complete symmetry. The ionization chambers are mounted in a housing that has shielding functions as well. It may be adapted to the particular application. . The device is read-out through an adopted DAQ-Box that contains the transimpedance amplifier circuitry, ADC and DAQ electronics with USB-link.

Performance Tests
The device was tested at the TRIGA reactor in Mainz. With the similar configuration of two converter coatings of identical thickness, the conversion rate was determined in a multi-wire proportional chamber prior to the experiments with the I-BM. The beam on the thermal column of the TRIGA reactor showed a signal rate of 15,4kHz with the wire chamber at the maximum CW operating power of 100kW. With each converted neutron generating 30600 e charges, 2/3 of which are detected in humid atmosphere , this amounts to a CW current of about 50 pA, just about the targeted noise related detection limit. The TRIGA reactor may however be pulsed as well. A single pulse, where the reactor´s control rods are being shot out of the core, has a peak power of 250MW, 2500 times higher than the highest CW power. Such reactor pulses were detected with the I-BM, one of which is depicted in Figure. 7.
The overall measured amplitude of the pulse was 766mV. The full integrated energy of the reactor pulse was determined by the operators to 9,86 MWs, corresponding to the area of the pulse. This calibration then results in a measured peak reactor power in the pulse of 246 MW, a factor of 2460 higher than ordinary CW operation. At a CW power of 100kW, a CW detection rate of 15,4kHz was determined as stated before. Consequently, a current of 2460 x 50pA = 123 nA must have been detected in the peak, corresponding to 743mV, giving an impedance U out /I in = 6,0 MΩ, very well in accordance with the TIA feedback resistance of 3,0 MΩ together with the additional voltage gain of 2 of the subsequent buffer to the ADC of the electronics employed.

Figure 6:
One of two pulses of the TRIGA reactor measured with the prototype I-BM. The reactor operators determined a full pulse power of 9,86 MWs. Taking the full pulse energy for calibration, the pulse peak power as measured with the I-BM device amounted to 246 MW. The width of the pulse was measured to be 34,1ms FWHM. Both results are consistent with the reactor´s specifications. The current was sampled at 702kHz sampling rate, the signal is offset at -4,75V to fully exploit the dynamic range of the electronics for the unipolar signal. Some pick-up oscillations were revealed in this measurement and allowed to improve the shielding as well as the internal powering scheme of the electronics.
The TIA output voltage signal was recorded with an 18bit ADC at 702kHz sampling rate. Such high sampling rate was chosen to avoid spurious high frequency noise to fold back into the bandwidth of the signal and the monitor itself, targeted at around 125kHz.