Mixed-Field Radiation Monitoring and Beam Characterization Through Silicon Diode Detectors

We present a calibration of a commercial silicon diode with proton and alpha beams and gamma rays. The diode together with a fast acquisition chain can be exploited for both direct and indirect (through the secondary radiation field) beam characterization. Within this work, we demonstrate the detector capabilities of resolving single-energy-deposition events and independently measuring dose rate and beam flux. Profiting from the mixed radiation field in CERN’s high-energy accelerator mixed field facility (CHARM) we show how the silicon detector can be exploited to characterize the mixed radiation field present in it, which in turn is used for validating the radiation tolerance of components and systems to be installed in the European Organization for Nuclear Research (CERN) accelerator complex.

its immediate surroundings.Depending on the beam energy, the secondary radiation field consists of multiple particle species at various energies (mixed-field radiation).In the largest European Organization for Nuclear Research (CERN) accelerators, such as large hadron collider [1], [2], [3] or super proton synchrotron (PS) [4], the induced radiation poses a threat to the reliability of the installed electronic systems.Therefore, CERN's radiation hardness assurance (RHA) approach requires: 1) knowledge about the expected radiation environment that is provided by the beam monitoring and simulations and 2) radiation-tolerant design of electronic systems.The CERN in-house designs are based often on the commercial-off-the-shelf (COTS) components, which require radiation tests, both at the component (before the system design) and system levels.In the case of CERN, the latter is performed in the CERN high-energy accelerator mixed field facility (CHARM) [5].The radiation field, produced via the spallation reaction with a target, mimics the mixed radiation field that can be encountered along the CERN accelerator infrastructure.
In addition to system-level irradiation tests, the CHARM facility is used as a validation for dosimeters and radiation monitors intended for the accelerator's mixed-field monitoring, for example, for distributed optical fiber radiation sensor (DOFRS) [6] or CERN radiation monitors (RadMons) [7], [8].The facility provides multiple instruments for radiation monitoring, together with the validated simulations in the FLUKA Monte Carlo code [9], [10], [11], [12].This work extends the previous studies that demonstrated the use of a 300-µm passivated implanted planar silicon (PIPS) detector for heavy-ion beam characterization [13], [14], [15], to a characterization of the CHARM's mixed-field.When compared to a previous study [16], the measurements involve an additional radiation-exposed location (m5) and provide the intercomparison with the standard beam instrumentation available in the facility.This is exploited for the indirect beam characterization through the secondary mixed field, for example, for the spill duration optimization.The presented silicon detector is intended to be used as a complementary instrument for CHARM monitoring, and in the future, for the accelerator's mixed-field radiation environment [17].Schematic top view drawing of the CHARM facility with the highlighted detector testing positions: m5, m6, and G0.Adapted from: [12].
The article is structured as follows: Section II introduces the CHARM facility, where mixed-field measurements were performed, and the silicon diode detectors that were used in this study.Section III presents the calibration of the detectors with proton and alpha beams performed at Centro Nacional de Aceleradores (CNA; Seville, Spain).The section shows as well dose-rate calibration with the Co-60 source provided at the same facility.Section IV describes the proton measurements from Paul Scherrer Institut (PSI) and TRIUMF facilities, which are used to demonstrate the capabilities and limitations of the detector in the case of direct beam characterization (based on the interaction of the primary beam with the silicon).Section V demonstrates the use of the detector in mixed-field radiation environments, similar to ones present near the accelerators.Section VI exploits the use of the detector for a beam characterization through the secondary mixed field.Section VII contains the summary of the work with the related outlook.

II. EXPERIMENTAL SETUP A. CHARM Facility
The CHARM facility [5] is part of CERN's PS East Area [18].It provides a mixed field through a spallation reaction between 24 GeV/c protons from a PS and a target (the most frequently Cu).In addition to the target selection, the facility allows the modification of the particle spectral composition by the shielding arrangement.CHARM contains several standard test locations, varying in terms of particle spectra and intensity.Within the scope of this work, measurements at two positions were performed: G0 and m5.During tests, the copper target was used and no shielding was in place.The schematic drawing with the highlighted testing positions is depicted in Fig. 1.

B. Canberra Silicon Diode Detectors
The experimental setup, depicted in Fig. 2, is based on the Canberra FD 50-14-300RM commercial silicon detectors.Within this work, measurements from three detectors (the same model) are presented.The detectors were operated in the reverse-bias mode, under a voltage listed in Table I, that according to the manufacturer is sufficient to obtain a full depletion.Whereas the exposed surface is equal to 0.5 cm 2 , this model of detectors is known to have some silicon active volume under the steel case [13].Based on the ratio of direct ionization events from: 1) heavy ions directly punching silicon and 2) heavy ions degraded by the case before the interaction [14], we estimate the actual sensitive region to be in the 0.75-0.85cm 2 , that is, up to 70% larger.This was reflected in the FLUKA Monte Carlo model of the detector depicted in Fig. 3.
Both detectors were operated with either of two CIVIDEC preamplifiers: C1-HV or C2-HV.The certified gain was equal to 21.7 and 43.9 dB, respectively.The amplified output signal was digitized with the CAEN DT5751 1 GS/s digitizer.

A. Energy Calibration at CNA
Each energy deposition event is recorded as a voltage trace over time.After the conversion (R = 50 impedance) to current, the signal is integrated to retrieve the number of the collected charges.This step assumes that the setup does not encounter signal losses and the amplification is equal to the one specified by CIVIDEC.In the following step, the total deposited energy is calculated as the measured number of charges multiplied by the energy that is needed to create an e-hole pair in silicon (W ≈ 3.6 eV).On top of that, the obtained deposited energy value is further multiplied by a k correction factor, arising from a calibration.
For both detectors, the calibrations were performed in the 3-MV Tandem accelerator at CNA under vacuum conditions, as depicted in Fig. 4 [19], [20].In the case of detector I, the calibration involved a proton beam with ten energies between 0.5 and 2.97 MeV, also exploitable for electronics testing [21].Detector II was calibrated with an alpha beam of four different energies between 2.97 and 8.95 MeV.The obtained k factors are equal to 0.95 in the case of detector I and 0.92 for detector II.The factors are similar and the difference is likely due to the detectors themselves (e.g.due to the irradiation history and/or doping difference).As it has been reported [22], the value might depend on the depositing particle type.Within the accuracy of the presented setup, there is no evidence that the difference in k factors comes from the different particle types used during the calibration (proton, alpha), mainly thanks to a very thin dead layer (<50 nm).

B. Dose-Rate Calibration With Co-60 at CNA
A current that is measured at the preamplifier's bias output is directly proportional to the dose rate.Profiting from this fact, a calibration of the setup (detector II) against Co-60 was performed.During the test, the beam was not collimated and the detector was covered with an aluminum foil.The reference  Fig. 4. Total deposited energy (equal to the beam energy) as a function of the measured deposited energy that assumes on average 3.6 eV to create a single e-hole pair and setup parameters according to the specification.The calibration was performed for both detectors (I and II) listed in Table I.The slope of each fit corresponds to a correction factor k that should be applied to the measurements in a similar energy regime.
dose-rate measurements in the facility were obtained using PTW 30010 Farmer ionization chamber (volume 0.6 cm 3 , with a 4.55-mm-thick polymethyl methacrylate (PMMA) build-up cap).The measurements with the linear fit are depicted in Fig. 5.The retrieved current to dose-rate conversion is equal to 6.2 × 10 4 Gy/s/A.
It has to be noted that the presented detector contains a case that partially covers the active silicon region, as described in Section II-B.During the calibration with gammas from Co-60, the case helps in reaching the electron equilibrium, which would not be achieved if pure 300-µm-thick silicon was used, as there would not be enough material for electron build-up.The electron equilibrium was achieved in the reference condition, in the used ionization chamber.According to performed FLUKA Monte Carlo simulations under irradiation with photons of 1 MeV, the measured dose would be within 30% lower when compared with the dose in thicker silicon volume (in electron equilibrium).While considering only the active region itself and without surroundings, the respective dose would be factor 3.3 underestimated.
The calibration is useful for dose rate estimations in high-radiation environments, where the use of event-by-event acquisition mode is not feasible due to pile-ups.With the presented limitations due to potential case impact, the uncertainty of the dose rate measurements in a mixed-field is within 41%, as the calibration could introduce 30% overestimation due to electron-equilibrium condition and the case collimation (assuming the worst case of 0.35 cm 2 shielded surface) could reduce the active silicon volume by 41%, leading to the dose rate underestimation.However, in the accelerator's mixed field, for example, in CHARM, the expected deviation would be lower due to the highly penetrating nature of the radiation.

IV. DIRECT BEAM CHARACTERIZATION
This section profits from the previously described calibration and applies it to the measurements with proton Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.Fig. 6.Energy deposition spectra under proton irradiations as measured by detector I (PSI-PIF) and detector III (similar to detector II, TRIUMF-PIF).Additionally, at 6.07 MeV, a vertical dashed line is displayed, which indicates the maximum energy that a proton can deposit in 300 µm of silicon through electromagnetic interactions.Fig. 7. Flux over time as measured by detector I, via the event-by-event energy depositions, in PSI-PIF facility for 17-MeV protons, together with the flux profile provided by the facility.The calculation for the diode assumes that the exposed Si surface (0.5 cm 2 ) is the only active surface.However, some protons punchthrough the diode case.beams provided by PSI-proton irradiation facility (PIF) and TRIUMF-PIF facilities [23], [24], [25].The presented measurements are further used to demonstrate the capabilities of direct beam characterization, that is, by placing the detector in a beam.The tests in TRIUMF involved 230-and 480-MeV proton beams.In the case of PSI, two primary energies were used: 200 and 70 MeV.Both PSI beams were also further degraded to 151 and 101 MeV (from 200 MeV) and 17 MeV (from 70 MeV).The last one was also further degraded by the Al slabs of 1, 1.2, and 1.5 mm.

A. Energy Deposition Spectra
The setup allows retrieving each single energy deposition event (created by a single hitting particle).Depending on the beam type, the resulting energy deposition spectra provide information concerning the beam energy spread and the interaction of the radiation field with silicon, the latter being essential also for Monte Carlo simulations benchmarking.Examples of energy deposition histograms collected with the aforementioned proton beams are depicted in Fig. 6.
There is a good agreement between the ∼200-MeV spectra measured in PSI and TRIUMF.With the decreasing energy, there are fewer open nuclear reaction channels, and therefore fewer high-energy deposition events.In the case of significantly degraded beams, the majority of the measured events are due to proton direct ionization and were not measured by the high-energy proton beam, due to their lower linear energy transfer (LET).Moreover, due to the energy degradation and associated beam energy spread, some protons do and some do not deposit their entire energy in the silicon, leading to the step at ∼6 MeV, which corresponds to the maximum energy that can be deposited via d E/d x by protons in the thickness of the diode.The step can be used as an ad hoc calibration in the radiation environment with a significant presence of protons, for example, when the presented calibration is not valid due to additional cable attenuation.

B. Diode as a Flux Monitor
For particle beams with a high probability of interaction with silicon, that is, excluding neutrons, the setup can be used as a flux monitor, as events can be resolved in time.Fig. 7 demonstrates an example of the diode application as a flux monitor, for direct beam characterization.However, the proposed detector has one drawback, namely that some active silicon region is shielded by the metal casing.If no prior assumptions concerning a beam can be made (capability of Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.punching through the casing), the flux measurement is with 70% uncertainty, because of the active surface uncertainty (between 0.75 and 0.85 cm 2 ).However, in case these events can be identified in energy deposition spectra (e.g., for heavy ions with high LET), the accuracy of the flux measurement is significantly improved.

C. Dose-Rate Estimations
Profiting from the dose rate calibration, described in Section III-B, through the leakage current measurements, the setup can be exploited for a beam characterization, in terms of absolute delivered dose and the time variability.Fig. 8 depicts the dose rate during 230-MeV-proton irradiation at TRIUMF-PIF.Over the irradiation with the total fluence of 8.8 × 10 7 cm −2 , the measured total ionizing dose (TID) was 58.0 mGy.Assuming the LET of the beam is equal to 3.37 MeV cm 2 /g [26], the expected dose is 47.4 mGy, within 19% agreement with diode the measurements.According to the FLUKA Monte Carlo simulations for 230-MeV protons, the case could lead up to 3.4% higher dose when compared with uncovered silicon volume.The higher measured value could arise from the calibration, as explained in Section III-B, due to not fully achieved electron equilibrium.
In general, the aforementioned calibration is a good approximation for particles that are capable of penetrating the diode case without significant energy loss.The estimated uncertainty for such mixed fields is within 41%, arising from the difference in the exposed/covered silicon volumes and electron equilibrium during calibration.In the mixed field where there is a significant presence of thermal neutrons, the case would act as a converted and the related deposited dose would be significantly larger as compared with pure silicon.

D. Event Classification via Pulse Shape Discrimination
The setup allows collecting not only deposited energy values but entire current profiles with 1-ns resolution.This feature can be exploited for pulse shape discrimination (PSD), for example, in terms of amplitude or falling time [27], [28].Fig. 9 represents the density histogram of amplitudes versus deposited energies for irradiation with 17-MeV proton beam (degraded from 70 MeV), further degraded by 1.5 mm of Al.The corresponding energy deposition spectra are depicted in Fig. 11.Simulated particle spectra, for the dominating particle species, in G0 (left side) and m5 (right side), for CHARM configuration with copper target, and no shielding in place.Fig. 6.As the beam encountered multiple interactions before reaching the diode, the beam energy spread was expected to be large.For 300-µm silicon thickness, the highest energy that can be deposited through the direct ionization by protons is 6.07 MeV [29], and therefore the particles above this energy mostly punchthrough the silicon volume.A similar conclusion could be drawn based on the analysis of the distribution of the fall time versus deposited energy.The PSD is especially useful during heavy-ion beam characterization for the detection of contaminants [30].

V. CHARM MIXED-FIELD MEASUREMENTS
The silicon diode setup during CHARM tests in the G0 position is shown in Fig. 10, together with other radiation sensors, covered in Section VI-A.The experimental measurements were collected with detector I during operation with a Cu target and no shielding in place.The related FLUKA Monte Carlo simulations were performed in a two-step approach.In the first one, the entire geometry of the CHARM facility was simulated, and particles were scored in 20 × 20 × 20 cm 3 air volumes located in G0 and m5 positions.The resulting particle spectra in G0 and m5 locations, for the most relevant particle species are depicted in Fig. 11.The particle spectrum in m5 spans toward higher energies and contains more pions and protons when compared with neutrons.The contribution of EM particles is visible mostly at lower deposited energies, and therefore they were neglected.In the second step, simplified simulations were performed, involving the particle spectra from the first step and the silicon diode model.This approximation neglects the directionality of the radiation in the primary field and particles come as a diode-orthogonal beam.The measurements from G0 and m5 positions are depicted in Fig. 12, together with FLUKA Monte Carlo simulations.There is a very good agreement at the G0 position.In the m5 testing position, the agreement between simulation and measurements is worse which could be due to: 1) the use of a two-step simulation approach that neglects the directionality of the radiation field, which is known to introduce discrepancies [31]; 2) other issues with the simulations, for example, related to modeling of the detector's dimensions/materials; 3) radiation degradation or malfunctioning of the silicon detector; and 4) the misplacement with respect to the m5 position, as the alignment was performed remotely.The exact reason remains unknown and the tests should be repeated to confirm that the discrepancy arises from the simulation side.

VI. INDIRECT BEAM CHARACTERIZATION
The detectors can be successfully used for beam characterization.Within this work, we investigate the detector used for indirect beam diagnostics, that is, measuring the secondary field produced through the target, as opposed to the direct exposition of the detector to the primary beam (covered in Section IV).For this purpose, the detectors can be used in two acquisition modes, also simultaneously.The first one measures each energy deposition event.While counting a number of events over time, as illustrated in Section VI-C, the information about the relative beam intensity can be retrieved.Whenever the information about single-energy depositions is not necessary or not available (e.g., due to the digitizer saturation), the second mode can be exploited, as presented in Section VI-A.It makes use of the current that is measured at the output of the preamplifier.With the calibration presented in Section III, the measured current can be converted to the dose rate (Co-60 equivalent).

A. Beam Intensity Through TID Measurements
For constant settings, the TID in each CHARM location will be proportional to the number of protons on the target (POT).In the case of the silicon diode, the leakage current Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.Fig. 13.Dose rate (retrieved through the calibration presented in Section III) over time as measured by detector I in the G0 position (CHARM configuration: no shielding in place with the Cu target).Due to a limitation of the source measure unit (no measurement while accessing buffered data), only selected spills were captured entirely and thus used for TID estimation.Fig. 14.Example of a spill recorded with the digitizer by means of the event count rate (1 ms binning), together with the raw total event count, and event count that was corrected to compensate for buffer saturation.
was measured only during several spills, not throughout the entire CHARM run (one week).Within the selected 11 spills, selected those depicted in Fig. 13, corresponding to 6.49 × 10 POT, the detector 10.96 mGy (1.69 × 10 −15 Gy/POT).As the CHARM configuration constant, extrapolating the following result to the entire one-week run yields (36.04 ± 0.82) Gy (std.error).
Over the CHARM run (one week), a radiophotoluminescence dosimeter [32] was installed very close to the diode, as shown in Fig. 10.During the one week of acquisition, the RPL dosimeter measured 32.6 Gy, that the POT = 2.13 × 10 16 p, corresponds to a dose rate of 1.53 × 10 −15 Gy/POT (copper target no shielding).At the same time, the nearby RadMON measured 26.9 (1.26 10 −15 Gy/POT).The agreement of the diode TID measurement is within 25% with RadMON and within 10% once compared with the RPL dosimeter.
Repeating the procedure for all possible shielding/target configurations would allow for indirect beam intensity measurements, through the TID deposited in each spill.

B. Mitigations for High-Intensity Beams
The proposed setup has a limitation concerning the maximum event rate (30 kHz for 500-ns events) due to the bandwidth of the USB 2.0 link between the digitizer and the PC.With the dedicated firmware settings, the number of trigger events can be preserved in the data, providing Fig. 15.Averaged spill time profile over 2853 spills as measured via 1) event rate from silicon diode (trigger threshold at 13 ADC counts) in G0 position calibrated to secondary emission chamber 70; 2) SEC70; and 3) differentiated signal from the beam current transformer (BCT) in the PS accelerator (before extraction).Whereas the diode measurement was calibrated to the SEC 70 one (total number of counts versus total number of measured protons), the BCT measurement is independent of the other two.information about how many events were lost due to buffer saturation.This allows retrieving the valid normalization of energy deposition spectra, even in the case of events loss due to the buffer saturation, under the assumption that the composition of the particle spectra is constant in time, enabling the exploitation of the detector as a flux monitor for highintensity beams, as long as pile-ups are negligible.An example of a spill, collected in CHARM's G0 position, when the event rate led to dead time, is depicted in Fig. 14, together with a related dead-time correction.

C. Beam Time Profile From the Event Rate
Assuming the constant test settings (target, shielding, testing position, amplification, and acquisition threshold), the diode detector can be used to resolve the primary beam time profile, through the measured count rate, which is corrected for a dead time, as described in Section VI-B.
As a first step, for the selected test settings, the count rate has to be calibrated against the total beam intensity.Within this study, we used a secondary emission chamber [33] (position 70) that provides an absolute number of protons per spill in the transfer line supplying CHARM with a proton beam.Over the selected time period, 2853 CHARM spills were considered.To cross-calibrate the diode, the total number of counts was divided by the total number of protons delivered to CHARM, leading to the 4.1 × 10 7 POT/count.This factor was retrieved for a G0 position with no shielding and copper target, the analog-to-digital converter (ADC) threshold of 13 ADC counts while running with a CIVIDEC C1-HV amplifier and an additional 6-dB attenuator.The resulting beam time profile is depicted in Fig. 15 and it is in good agreement with the decrease rate of the intensity in the PS accelerator (primarily due to extraction), which is independent of the profile measured by diode through the SEC70 cross-calibration.
The described procedure could be used to assess the beam time profile with a better time resolution than through the SEC70 instrument (20 ms resolution).The direct time profile measurement of the CHARM primary beam, with the diode installed directly in the beam, would lead to quick detector damage.

VII. CONCLUSION
This article introduced a silicon diode, manufactured by Canberra, with an energy deposition calibration in the few MeV energy ranges and dose rate calibration up to 8 mGy(Si)/s.The detector is commercially available and is widely used in the radiation effects community for heavy-ion beam characterization.Despite the partial coverage of the active silicon (up to 0.35 cm 2 ) by the detector case, we highlighted several use cases for both beam and radiation characterization, with the related limitations.
As a novelty, based on the CERN CHARM's mixed field, we have demonstrated how the diode can be used indirectly to retrieve key beam characteristics (intensity, time profile) from the secondary radiation field when direct in-beam measurement is not possible.
The benchmark of the experimental measurements in the CHARM G0 position yields very good agreement with the FLUKA Monte Carlo simulations.In this position, the measured TID is within 25% when compared to RadMON and within 10% when compared to the RPL dosimeter.

Fig. 2 .
Fig.2.Scheme of the experimental setup used in this study, consisting of a silicon diode detector, power supplies, a preamplifier, and a digitizer.

Fig. 3 .
Fig. 3. Model of the detector implemented in the FLUKA Monte Carlo particle transport code.

Fig. 5 .
Fig. 5. Dose-rate (in Si) as a function of the leakage current amplitude under the Co-60 irradiation (detector II).

Fig. 8 .
Fig. 8.Dose-rate (in Si) during the irradiation with 230-MeV protons in TRIUMF-PIF, as measured by detector III, via the leakage current.The integrated dose rate was 58 mGy, corresponding to the proton fluence of 8.8 × 10 7 cm 2 .The variations in the dose rate were due to the intended flux changes.

Fig. 9 .
Fig. 9. Amplitude of the energy deposition events as a function of the event deposited energy, as measured with detector I under the 17-MeV proton irradiation, further degraded by 1.5 mm of Al slab.In the bottom right side of the figure, there are averaged pulse shapes, both with 4-MeV deposited energy, one by protons that punched through the case, and the second one that deposited entire energy in silicon.

Fig. 10 .
Fig. 10.Silicon diode during experiments in CHARM, in G0 position.Additionally, in its direct proximity, RadMON and RPL dosimeters were installed.

Fig. 12 .
Fig.12.Measured energy deposition spectra in testing positions G0 and m5, as measured by detector I at CHARM for the Cu target and no shielding in use.The traces were normalized to the delivered POT.Additionally, simulated energy deposition distribution was depicted, as retrieved via the FLUKA Monte Carlo simulations, considering circular beam composed of particle spectra shown in Fig.11.