1 Introduction

Modern radiation protection largely relies on EPDs to achieve ALARA objectives. If compared to passive dosemeters, EPDs offer real-time reading, time-resolved dose recording, alarm threshold settings, and visible/audible alarms to prevent accidental exposures. Tens of EPDs for photons are commercially available and are largely used especially in medicine [1]. Their performances usually comply with relevant international Standards [2, 3]. For neutrons, the situation is far different. The need for neutron EPD (EPD-n) exists and is increasing. EPD-n are used to monitor the staff of nuclear plants and the personnel accompanying transports of spent fuel flasks [4]. The increased age of nuclear plants, requiring refurbishment or decommissioning works, is increasing such demand. Apart from the nuclear sector, EPD-n are needed in the industrial applications of neutron sources, such as measurements of density and water content of soil [5], asphalt control [6] and in the concrete sector [7, 8]. In the medical sector, the growing number of hadron-therapy and accelerator-based BNCT installations will certainty increase the demand for these devices. Neutrons in workplaces range over ten or more decades in energy, leading to several technical difficulties when trying to measure the associated dose equivalent [9]. Due to these difficulties, EPD-n are dramatically less developed than EPD for photons in terms of design, performance and market. In the first decade of 2000, the EVIDOS project [10] evaluated the field performance of less than ten EPD-n, half of which were commercially available while the remaining ones were research prototypes. A regional comparison was organized in 2008 [11] involving three prototype EPD-n. Only two EPD-n were included in the EURADOS 2012 comparison for the whole-body neutron dosimetry [12]. To date, it was possible to find only three EPD-n on the market. Commercial EPD-n are based on one or more silicon detectors covered by different radiators, such as polyethylene to produce a fast response, and lithium or boron to produce a thermal response. Almost invariably, this structure tends to produce an overestimation in the thermal field and an underestimation in the 0.1 \(\div \) 1 MeV region. Whilst the thermal overestimation can be somehow corrected by decreasing the concentration of \(^{6}\)Li or \(^{10}\)B in the thermal radiators, the epithermal underestimation invariably yields to a “deep valley” in the energy dependent response. The energy range of commercial EPD-n is usually from 0.025 eV to 15 MeV. The energy dependence, estimable as the ratio of the maximum to minimum response as the energy varies over the rated range, is as large as 20 to 60. Clearly, they require workplace-specific calibration factors to accurately operate. As the silicon detectors have planar structure, the variability of the response as a function of the angle is quite important and can be as high as 2. In addition to the very few commercial EPD-n, some research prototypes appeared in last two decades but are not currently on the market:

  • In the early 2000s, PTB (Germany) developed an EPD-n prototype [13] based on a single silicon diode with \(^{6}\)LiF and polyethylene radiators. The variability of the energy response is about 50 from thermal neutrons up to 14.8 MeV and its typical calibration coefficient is 271 mSv\(^{-1}\) obtained for the bare \(^{252}\)Cf source;

  • In the same years, IPSN (France) developed an EPD-n based on a segmented 4 cm\(^{2}\) silicon diode with different radiators, able to reduce the energy response variability to a factor of about 2.6 from thermal neutrons up to 14.8 MeV [14];

  • The Helmholtz Zentrum Munchen (HMGU, Germany) developed a combined neutron and photon dosemeter [15, 16], whose neutron section is formed by three independently acquired silicon diodes. Relying on the use of \(^{6}\)LiF and polyethylene radiators in an albedo configuration, the variability of the response from thermal neutrons up to 14.8 MeV is about 7.

Within the recent DOIN project (2021–2022), INFN prototyped a new EPD-n for neutrons based on a patented design. Its main characteristics are:

  • The energy variability of the response is limited to about 2 when the energy varies from thermal neutrons up to the quality of \(^{241}\)Am-Be, including radionuclide and monoenergetic reference neutron fields;

  • the calibration coefficient is about 10\(^{4}\) mSv\(^{-1}\) in terms for H\(_{p}\)(10, 0\(^{\circ }\)) for the bare \(^{252}\)Cf source;

  • The response shows very limited directional dependence;

  • The parasitic photon sensitivity is lower than 2 mSv\(^{-1}\) in the range from 48 to 205 keV.

2 The DOIN

Because of ongoing procedures to protect intellectual property, the internal structure of the DOIN dosemeter cannot be disclosed here. As seen in Fig. 1, the prototype consists of a 3D printed ABS plastic case with size 7 \(\times \) 4.5 \(\times \) 17 cm\(^{3}\) plus a hemispherical dome 2.7 cm in radius. At this stage, the box has only the function of containing the components and was not designed to be compact. In view of industrializing the prototype, the box can be easily reduced by a factor of 2 in volume.

Fig. 1
figure 1

The DOIN dosemeter

Full size image

The dosemeter includes the following internal parts:

  • The detector module, including the hemispherical moderator as well;

  • The analogue board to control the detector;

  • The digital board including signal processing, storage on SD card and display control;

  • The display board;

  • The power supply board, generating all supplies for analogue and digital components based on a 1100 mAh rechargeable 3.8 V Li-P battery.

If H is the dosemeter estimate for the personal dose equivalent H\(_{p}\)(10), the following equation can be written:

$$\begin{aligned} H =\frac{L}{k(q)} \end{aligned}$$
(1)

where

  • \(L \) is the number of electrical pulses (counts) provided by the detector module;

  • \(q \) is the energy index, determined by the detector module in addition to the number of counts;

  • \(k(q) \) is the \(q \)-dependent calibration coefficient.

\(k(q) \) was experimentally derived by testing the prototype in a variety of reference neutron fields, including radionuclide and accelerator-based monoenergetic fields [17]. These tests are described in Sect. 4.

3 Computational

The whole structure of DOIN was simulated using the FLUKA-INFN code [18]. The response was simulated in the energy range from 1 meV to 20 MeV at angles of incidence from 0\(^{\circ }\) to ± 60\(^{\circ }\). The model includes a 30 \(\times \) 30 \(\times \) 15 cm\(^{3}\) ICRU tissue slab phantom to simulate the effect of the human body. Figure 2 shows the energy dependence of the \(q \)-index at various angles of incidence. Its numerical value is nearly constant from the thermal to 1 eV neutrons. At higher energies, it monotonically decreases with the energy up to about 10 MeV. This monotonic dependence allows, in principle, to infer information on the neutron spectrum starting from a measured parameter, \(q \). This is the basis for the dose estimation in DOIN. Interestingly, the energy dependence of \(q \) shows very little variation as the angle changes, being +25 % the largest deviation from the 0\(^{\circ }\) (frontal incidence) curve, occurring at 60\(^{\circ }\) and 100 eV.

Fig. 2
figure 2

Simulated energy index \(q \) of the DOIN dosemeter as a function of the energy and the incidence angle

Full size image

4 Experimental

Calibration experiments were conducted in the following reference neutron fields:

  • the thermal neutron field HOTNES at INFN / ENEA Frascati [19,20,21];

  • \(^{252}\)Cf and \(^{241}\)Am-Be at ENEA Bologna and Politecnico of Milan [22], respectively;

  • Accelerator-based monoenergetic reference neutron fields at NPL Teddington (UK) [23] with energy 0.0715, 0.144, 0.565 and 1.2 MeV.

The first three fields are obtained with the reaction \(^{7}\)Li(p,n)\(^{7}\)Be at different angles and projectile energies. 1.2 MeV was obtained with the reaction \(^{3}\)H(p, n)\(^{3}\)He at an angle of 0\(^{\circ }\). The energy domain above the quality of the \(^{241}\)Am-Be will be studied in near future. These fields were selected to follow as close as possible the IEC 61526 Standard [3], ideally recommending one thermal field, one monoenergetic neutron field between 10 and 100 keV, three monoenergetic fields from 100 keV and 1 MeV, three monoenergetic fields from 1 MeV to 10 MeV or two monoenergetic fields and a broad source (\(^{252}\)Cf or \(^{241}\)Am-Be). In alternative, “alleviated” criteria with realistic fields are allowed. Exposures were performed in terms of H\(_{p}\)(10, 0\(^{\circ }\)) from about 20 up to 500 \(\mu \)Sv and H\(_{p}\)(10) rates from about 30 up to about 1000 \(\mu \)Sv h\(^{-1}\). Angular tests were only performed at the highest energy, corresponding to the quality of the \(^{241}\)Am-Be field. More extensive angular testing is planned for the near future. Additional measurements with reference X-ray fields were performed at ENEA Bologna to evaluate the photon rejection capability of DOIN. The main parameters of the irradiations are resumed in Table 1.

Table 1 Main parameters of the reference fields used for DOIN calibration
Full size table

5 Results and discussion

The data acquired by DOIN were processed as follow:

  • Determination of the calibration curve: for every neutron field, the experimental values for \(q \) and \(L \) were measured by the dosemeter. The calibration coefficient was derived as \(k = \frac{L}{H_p(10)}\). Thus, \(k \) was put in relation to \(q \) in Fig. 3. By fitting the experimental data, the reported calibration curve (solid line) was obtained.

Fig. 3
figure 3

Relation between the energy index \(q \) and the calibration coefficient \(k \) obtained by calibration in reference neutron fields. The fit is the calibration curve. Uncertainties are comparable with graphical symbols

Full size image

  • The calibration curve was applied to the dosemeter output (L and q) by obtaining H, i.e., the best estimation of the personal dose equivalent.

The ratio \(R =\frac{ H}{H_{p}(10)}\), or dosemeter response, was compared with the minimum and maximum acceptable limits recommended by IEC 61526, namely: \(0.65<R < 4.0\) from thermal energies up to 100 keV; \(0.65< R < 2.22\) from 100 keV to 10 MeV. These criteria apply to the combined energy and angle response.

The results are shown in Fig. 4, where \(R \) is plotted as a function of the fluence-average energy of the neutron fields. The IEC limits are also reported. The calibration curve was also applied to the data acquired at different incidence angles with the \(^{241}\)Am-Be. The corresponding \(R \)-values ranged from 1.3 to 1.6, again satisfying the IEC criterion. Additional tests with reference X-ray fields at ENEA Bologna showed an X-ray sensitivity lower than 2 mSv\(^{-1}\) in the range from 48 to 205 keV (fluence-averaged photon energy). The delivered value of photon \(H_{p}\)(10, 0\(^{\circ }\)) was 0.5 mSv for photon measurements.

Fig. 4
figure 4

Experimental response of DOIN compared to the minimum and maximum acceptable limits recommended by IEC 61526 Standard

Full size image

Although the response of DOIN is not literally flat in energy and in angle, its variability for the neutron fields used in this work is limited to the range \(0.7<R<1.5\), complying with the IEC criteria. If compared with the previous works shown in Sect. 1, DOIN exhibits better dose response and larger sensitivity. In addition to the measurements here described, others will be performed in near future to complete the device characterization in view of its industrialization and commercialization, namely: extension of the energy testing up to 14.8 MeV, angular tests at different energies than \(^{241}\)Am-Be, linearity test as a function of the dose rate, response to pulsed fields and photon tests in a wider range of energy and dose rate.

6 Conclusions

A new electronic personal dosemeter for neutrons, called DOIN, was developed and tested in a variety of reference neutron fields. Thanks to an innovative, patented combination of detectors and geometry, the response of this dosemeter complies with relevant requirements from thermal neutrons up to the quality of the \(^{241}\)Am-Be. This, together with the large neutron sensitivity and the good photon rejection, make DOIN superior to the state of art and attractive for operational neutron monitoring.