2.1 Detector probe design principle

Effective dose, E, is defined as follows:

where w_T and w_R are the tissue weighting factor and radiation weighting factor, respectively, H_T is the equivalent dose in an organ or tissue T, D_T,R is the mean absorbed dose in an organ or tissue T from radiation of type R, m_T is the mass of the organ or tissue T, D_R is the absorbed dose in a mass element dm. It is difficult to measure D_T,R directly, and for this reason the effective dose is hard to monitor.

The conversion coefficients between proton fluence and proton effective dose for standard irradiation geometries provided by the ICRP-116 are presented in Fig.1 ^[18].

Fig.1Full-screen View

Effective dose per proton fluence for standard geometries

For a certain amount of incident protons whose energy is X, the corresponding effective dose E(X) can be obtained according to the ICRP-116, the energy deposited in the detector by these protons is Q(X), if the coefficient K(X) between E(X) and Q(X) does not change with energy X, in other words, Q(X) is proportional to E(X), that is,

Then in practice, multiplying the measured deposited energy Q by the coefficient K, we can get the actual effective dose E for astronauts:

Presently, the orbit height of manned spaceflight is mainly in the range of 300–500 km, which is referred to as Low-Earth orbit, or LEO. The energy of protons of LEO are mostly below 400 MeV ^[21,22]. Based on the thirteen conversion coefficients between proton fluence and proton effective dose provided by the ICRP-116 in the energy range of 10 MeV to 400 MeV, our aim is to make the energy deposited in the detector proportional to the corresponding effective dose of the incident protons.

2.2 Detector probe design process

2.2.1 Basic structure and materials

Since 1955, five solar particle events (SPEs) with intensities and energies large enough to jeopardize crew health in spite of normal or even enhanced spacecraft shielding have been observed. The order of magnitude of the integral of the fluence of protons above 10 MeV in these SPEs is 10¹⁰ cm^{–2 [23]}. This is only an order of magnitude smaller than the proton dose threshold (2.0 × 10¹¹ cm^–2) inducing radiation damage in semiconductor detectors. This amount of radiation damage can cause an increase in leakage current, degradation of energy resolution, and a shift of the peak position because of lost charge collection efficiency ^[24]. Consequently, the performance of semiconductor detectors will gradually degrade and eventually lose function after lengthy exposure in space. The proton-stopping power of gas is rather small owing to its low density ^[25], so the volume and pressure of a gas detector would be rather large for the detection of high-energy protons, and this leads to the poor security and practicability of a gas detector in the space environment. The absorbed dose in space measured by the phantom torso experiment in the 9.8-day STS-91 mission (inclination: 51.6^°, altitude: 400 km) was 27.7 mGy, or 2.8 mGy/d ^[12]. Experimental results showed that, at a proton absorbed dose of 1 KGy, the radiation damage reduced the light yield of a plastic scintillator by only 15% ^[26], which reflected the plastic scintillator’s characteristics of stable luminescence efficiency and good radiation hardness. Plastic scintillator material also has the characteristics of good plasticity and high mechanical strength, which enable it to be processed into various shapes ^[27].

Isotropic exposure is the most realistic situation in space. While shielding effects may result in lower isotropic exposure, the movement of the astronauts within the spacecraft counteracts this ^[15], so the proton incidence condition was set as isotropic in the simulation. Based on this, a spherically symmetric detector was desired, so the scintillator was set as a sphere whose radius was 2.25 cm. A cylindrical light guide was used with the scintillator to aid the light collection for the detector. The radius and length of the light guide were 1.0 cm and 10.3 cm, respectively, and the distance from the center of the scintillator to the lower end surface of the light guide was 2.0 cm. Covering the spherical scintillator was a layer of shielding shell, inside which were some truncated conical holes. The energy deposited in the scintillator was modulated by changing the depth and solid angle of these holes. The greater the number of holes, and the more spherical the distribution of the holes in the shielding shell, the better was the spherical symmetry of the detector. Considering the similarity between the geometry of a regular polyhedron and a sphere, the holes in the shielding shell were effectively located at the vertices of a regular polyhedron. In the existing five types of regular polyhedrons, the regular dodecahedron has the greatest number of vertices (20). The two adjacent holes near the light guide were integrated into one to minimize their conflict with the light guide in space, leaving 19 holes in the iron shell. Except for the integrated hole around the light guide (hole 1), the other 18 holes were located at 18 vertices of a regular dodecahedron. The structure of the detector probe and sequence numbers of some of the holes are shown in Fig.2(a). The relative distribution of the 19 holes and their corresponding sequence numbers in the regular dodecahedron are shown in Fig.2(b), where three different colors indicate three different-depth holes.

Fig.2Full-screen View

(Color online) (a) Cross-sectional view of the detector probe. (b) Relative distribution of the 19 holes in the dodecahedron.

Figure 3 shows the range-energy relation and the mass thickness-energy relation of protons in different materials; the data is from NIST’s PSTAR databank (CSDA) ^[28]. From Fig.3 we can see that for protons with a certain energy, the range in Fe is almost the same as that in Cu and Pb, and far less than that in Be and Al. However, the mass thickness in Fe is less than that in Cu and Pb. To reduce the mass and volume of the detector, iron was adopted as the material of the shielding shell.

Fig.3Full-screen View

Penetration depth of proton in different materials. (a) range; (b) mass thickness

2.2.2 Determination of the iron shell thickness

Before digging holes in the iron shell, we needed to determine the iron shell thickness. Figure 4 presents the MCNPX simulation results of the dependence of Q_400MeV on the shell thickness, as the thickness varies from 0 cm to 3 cm, where Q_400MeV is the energy deposited in the scintillator by per fluence of isotropic incident protons of 400 MeV.

Fig.4Full-screen View

Dependence of the deposited energy in the scintillator on the iron shell thickness

As shown in Fig.4, the value of Q_400MeV at d=0 cm is 86% of that at d=3 cm. Q_400MeV remains nearly constant because protons of 400 MeV have strong penetrating power and the energy loss rate in the plastic scintillator is almost the same after it passes through the iron shell of different thicknesses. Consequently, the coefficient K between effective dose and deposited energy is determined:

$K=K_{\text{400MeV}}=E_{\text{400MeV}}/Q_{\text{400MeV}}.$ (4)

By modulating the depth and solid angle of the 19 shell holes, we attempted to make the deposited energy Q at the other 12 twelve energy points, apart from 400 MeV, satisfy the proportional relation with the corresponding effective dose E. Relative deviation $\epsilon$ between Q and E is defined in formula (5) to assess the proportional relation. The smaller the value of $\left|\epsilon \right|$, the better is the proportional relation.

$\epsilon =K\cdot \frac{Q}{E}-1$ (5)

For protons of 150 MeV, 200 MeV, and 300 MeV, Fig.5 presents the MCNPX simulation results of the dependence of $\epsilon$ on the uniform iron shell thickness. It is shown that when d<2.5 cm, ${\epsilon }_{150\text{MeV}}>0$; thus, d must be larger than 2.5 cm. If we want to reduce ${\epsilon }_{150\text{MeV}}$ to zero, we can select a value of d in the range of 2.5 cm to 2.9 cm. Here we chose 2.9 cm to make $\left|{\epsilon }_{150\text{MeV}}\right|$ as small as possible. The method we adopted was to adjust the depth and solid angle of the holes to be dug in the future design, using the positive ${\epsilon }_{150\text{MeV}}$ of the perforated zone (where the covering thickness is less than 2.5 cm) to compensate for the negative ${\epsilon }_{150\text{MeV}}$ (–88%) of the non-perforated zone (where the covering thickness is 2.9 cm). The tendency of the change of ${\epsilon }_{200\text{MeV}}$ with iron shell thickness opposes that of ${\epsilon }_{150\text{MeV}}$, ${\epsilon }_{200\text{MeV}}<0$ when d<0.9 cm, and ${\epsilon }_{200\text{MeV}}>0$ when d>0.9 cm. ${\epsilon }_{300\text{MeV}}$ is very small when d varies from 0 cm to 2.9 cm.

Fig.5Full-screen View

Dependence of the relative deviation between deposited energy and proton effective dose on the iron shell thickness

2.2.3 Determination of depth and solid angles of the holes

First we dug the holes whose sequence numbers are from 1 to 8, with depths and covering thicknesses of 2.9 cm and 0 cm, respectively. For protons from 10 MeV to 80 MeV, Fig.6 shows the MCNPX simulation results of the dependence of $\epsilon$ on $\sum {\Omega }_1$, where $\sum {\Omega }_1$ denotes the total solid angle of these eight holes. From Fig.6 we can see that when $\sum {\Omega }_1$ <3.355 Sr, $\text{max}(\left|{\epsilon }_{15\text{MeV}}\right|,\left|{\epsilon }_{\text{30MeV}}\right|)=-{\epsilon }_{\text{30MeV}}>13\%,$ and when $\sum {\Omega }_1$ >3.355 Sr, $\text{max}(\left|{\epsilon }_{15\text{MeV}}\right|,\left|{\epsilon }_{\text{30MeV}}\right|)={\epsilon }_{15\text{MeV}}>13\%$. $\text{max}(\left|{\epsilon }_{15\text{MeV}}\right|,\left|{\epsilon }_{\text{30MeV}}\right|)$ is minimized to 13% when $\sum {\Omega }_1$ is 3.355 Sr, and $\left|\epsilon \right|$ for protons of 10 MeV, 20 MeV, 40 MeV, 50 MeV, 60 MeV and 80 MeV is within 13% when $\sum {\Omega }_1$ is 3.355 Sr.

Fig.6Full-screen View

Dependence of the relative deviation between deposited energy and proton effective dose on total solid angle of holes 1–8

Next we dug the holes whose sequence numbers are from 9 to 13, with depths and covering thicknesses of 1.9 cm and 1.0 cm, respectively. Modulation of the solid angle of these five holes would not affect the deposited energy of the protons below 80 MeV because these protons cannot penetrate holes 9–13 because their range in iron is less than 1.0 cm. Figure 7 shows the MCNPX simulation results of the dependence of ${\epsilon }_{100\text{MeV}}$ on $\sum {\Omega }_2$, where $\sum {\Omega }_2$ denotes the total solid angle of these five holes. We can see that $\left|{\epsilon }_{\text{100MeV}}\right|$ is minimized to 0.54% when $\sum {\Omega }_2$ is 1.508 Sr. The MCNPX computational accuracy of the deposited energy of protons of 100 MeV is less than 1%, which is comparable to the minimum value of $\left|{\epsilon }_{\text{100MeV}}\right|$.

Fig.7Full-screen View

Dependence of the relative deviation between deposited energy and proton effective dose on total solid angle of holes 9–13

Next, we dug the holes whose sequence numbers are from 14 to 19, with depths and covering thicknesses of 1.4 cm and 1.5 cm, respectively. Modulation of the solid angle of these six holes would not affect the deposited energy of the protons below 100 MeV because these protons cannot penetrate holes 14–19 because their range in iron is less than 1.5 cm. For protons of 150 MeV, 200 MeV and 300 MeV, Fig.8 shows the MCNPX simulation results of the dependence of $\epsilon$ on $\sum {\Omega }_3$, where $\sum {\Omega }_3$ denotes the total solid angle of these six holes. Figure 8 shows that max($\left|{\epsilon }_{150\text{MeV}}\right|$, $\left|{\epsilon }_{\text{200MeV}}\right|$) is minimized to 13% when $\sum {\Omega }_3$ is 2.136 Sr, and $\left|{\epsilon }_{300\text{MeV}}\right|$ is close to zero.

Fig.8Full-screen View

Dependence of the relative deviation between deposited energy and proton effective dose on total solid angle of holes 14–19

2.3 Calculation results and analysis

Geometrical parameters of the 19 holes are shown in Table 1, where ${\theta }_1$/${\theta }_2$/${\theta }_3$ are the angles between the hole axis and the x/y/z axes, respectively. A prototype of the detector probe was constructed by 3D printing technology according to the design scheme described in section 2.2. A stereogram of the detector probe and cross-sectional view of the iron shell are shown in Fig.9(a) and Fig.9(b) respectively. Sequence numbers of some of the holes are marked in the figure.

Fig.9Full-screen View

(a) picture of the detector probe; (b) cross-sectional view of the iron shell

K_400MeV was used as the coefficient K in the simulation process, to reduce the maximum absolute value of $\epsilon$, regarded as a function of the coefficient K at the selected 13 energy points,

where the subscript i from 1 to 13 represents the selected energy point. The minimum of $f(K)$ and the corresponding value K₀ of K when $f(K)$ reaches the minimum can be calculated, and the calculation results are shown in Table 2, from which we can see that $\max {\{\left|\epsilon \right|}_{i=1:13}\}$ is reduced from 13.93% to 13.28% after K_400MeV is replaced by K₀. Fig.10(a) shows the proton effective dose per fluence from the literature and calculation, and the deviation between them is presented in Fig.10(b).

Fig.10Full-screen View

(a) Proton effective dose from literature and calculation; (b) Deviation between literature and calculation

3.1 Detection system

The detection system is composed of a detector probe, phototube, and electrometer. The scintillator is an HND-S2 plastic scintillator whose luminescence spectrum is shown in Fig.11. Its luminous attenuation length is larger than 2 m, and its density is 1.05 g/cm³. Its luminescence efficiency falls by only 15% when the fluence of fast neutrons reaches 1×10¹³ neutrons/cm² and remains nearly constant after the exposure of ⁶⁰Co gamma photons reaches 25.8 C/kg. The scintillator has the characteristics of high transparency, small density, and good radiation hardness. The light guide is made of polymethyl methacrylate (PMMA), which does not scintillate from impinging protons. The phototube is a type GD40H, whose window and cathode materials are borosilicate glass and bialkali (KCsSb) respectively. Its spectral response range is 300–650 nm, with a peak wavelength of 400 nm. The spectral response curve of the phototube has a large overlapping range with the luminescence spectrum curve of the scintillator; the peak wavelengths of the two curves are close to each other, and the phototube can make the best use of the light emitted from the scintillator. The electrometer is a Keithley 6517B electrometer manufactured by Keithley Corporation of America. Its charge measurement sensitivity is 10 fC, and it has the characteristics of high sensitivity and low noise. Deposited energy in the plastic scintillator is converted to electric charge, which is measured by the electrometer.

Fig.11Full-screen View

Luminescence spectrum of the HND-S2 plastic scintillator

3.2 Calibration of the detector photoelectric conversion efficiency

The relation between deposited energy and response of the detector is as follows:

where $q_{\text{H}}$ is the quantity of electric charge measured by the electrometer (C), $Q_{\text{H}}$ is the deposited energy (MeV), $Y_{\text{H}}$ is the scintillator light yield for protons, or the number of photons generated per MeV (MeV^–1). The scintillation light output for the HND-S2 plastic scintillator we adopted is proportional to the proton energy, so Y_H was taken as constant in this work (8.8×10³ MeV^–1, 20 °C). $F_{\text{ph}}$ is the light transfer efficiency in the scintillator and the light guide (dimensionless). $\mu$ is the average quantum efficiency of the incident light on the cathode of the phototube (dimensionless). e is elementary charge (C). The product of $F_{\text{ph}}$ and $\mu$ is defined as the detector’s photoelectric conversion efficiency $\eta$, which can be calibrated on a cobalt source:

where $q_{\text{γ}}$ is the response of the detector when it is exposed to a cobalt source, $Q_{\text{γ}}$ is the deposited energy of gamma rays in the scintillator, and is calculated using the MCNPX program based on the calibration experiment condition. $Y_{\text{γ}}$ is the scintillator light yield for gamma rays, provided by the corporation that manufactured the scintillator. The calibration experiment was performed in a ⁶⁰Co γ-radiation field of Northwest Institute of Nuclear Technology (NINT). The calibration results of the detector photoelectric conversion efficiency are shown in Table 3, where d is the distance from the center of the detector probe to the cobalt source.

3.3 Verification experiment

Isotropic irradiation is an ideal condition. Taking into account the reality, the detector was irradiated in unidirectional proton field. The verification experiment was performed on the platform of the HI-13 tandem accelerator-upgrading project of the China Institute of Atomic Energy. The platform can provide unidirectional proton beams in the energy range of 30 MeV to 100 MeV, with irradiation area of 75 mm×75 mm ^[29]. The detector was irradiated on the platform from four directions: antero-posterior (AP, x), postero-anterior (PA, –x), left lateral (LLAT, –y) and right lateral (RLAT, y). The irradiation directions with respect to the detector probe are defined in Fig.12. The four irradiation directions (PA/AP/LLAT/RLAT) were set normal to the common axis of the light guide and phototube to prevent direct irradiation of the phototube by the proton beam.

Fig.12Full-screen View

Different irradiation directions with respect to the detector probe

Deposited energy $Q_{\text{H}}$ in the scintillator in the experimental proton field can be calculated by the MCNPX program ^[30]. The shielding effect of the light guide and both energy straggling and angle straggling effects of the proton beams had already been considered in the calculation model. Combined with the calibrated detector photoelectric conversion efficiency, the theoretical response $q_{\text{H}}$ of the detector can be obtained by Eq. (7). The theoretical response $q_{\text{H}}$ and the experimental response $q_{\text{H}}{}^{\prime }$ were compared to verify the detector design, as shown in Fig.13.

Fig.13Full-screen View

Total frame diagram of the verification experiment

Fig.14 shows the experiment setup. The deviation $\delta$ between the theoretical response and the experimental response of the detector is presented in Table 4, where $\delta =\left(q_{\text{H}}-q_{\text{H}}\text{'}\right)/q_H\text{'}.$

Fig.14Full-screen View

(Color online) Detector design verification experiment site

The difference between calculation and measurement ranges from –18.70% to 17.53% as shown in Table 4. Considering the stability of the proton beam intensity, temperature effects of the scintillator light yield, and proton radiation damage to the scintillator during the experiment, formula (7) can be transformed into

where t₀ (s) is the duration of each irradiation in Table 4, $I_{\text{H}}(t)$ (protons/s) is the proton beam intensity at time t, $\overline{Q_{\text{H}}}$ (MeV) is the deposited energy in the scintillator of each source proton calculated using the MCNPX program, $T(t)$ (K) is the temperature of the scintillator at time t, and $D(t)$

(Gy) is the accumulated absorbed dose of the scintillator at time t.

If we neglect scintillator heat dissipation in each irradiation process, and assume that all of the deposited energy was converted into heat, which raised the temperature of the scintillator, then the maximum temperature rise of the scintillator would be $\Delta T=I_{\text{H}}\cdot t_0\cdot \overline{Q_{\text{H}}}/(m\cdot c)$=1.56×10^–3 K (PA, 100 MeV), where $c=1330\text{ J}\cdot {\text{kg}}^{\text{-1}}\cdot {\text{K}}^{\text{-1}}$ is the specific heat capacity of the scintillator, $m=0.0497\text{ kg }$ is the mass of the scintillator, $I_{\text{H}}$ = 2.5×10⁹ protons/s, $t_0=60\text{ s}$, and $\overline{Q_{\text{H}}}$=4.30736 MeV. Within –50 to 100 °C, the temperature coefficient of the scintillator light yield is 0.068%/K, therefore, in the experiment, the change in the scintillator light yield caused by its own temperature rise is negligible. In the experiment, if we take the proton beam intensity as constant (2.5×10⁹ protons/s) and neglect its fluctuation (within ±15%), then the accumulated absorbed dose of the scintillator over 32 irradiations is ${\displaystyle \sum D}={\displaystyle \sum I_{\text{H}}\cdot t_0\cdot }\overline{Q_{\text{H}}}/m$=3.6 Gy, which is so small that its influence on the light yield and optical transmittance of the scintillator is negligible ^[26]. The theoretical response of the detector was proportional to the proton beam intensity, whose fluctuation (within ±15%) during the experiment was the dominant cause of the difference between the calculated and measured results.