3-1-Physicochemical characterization
The gadolinium oxide particle size distribution was obtained using SLS analysis (Fig. 2-b). As it is seen, most of the particles were between 1 and 10 micrometers. Additionally, the SEM image (Fig. 2-a) displays the structure and particle size of Gd2O3 powder, confirming the SLS results. The EDX analysis of the sample was also performed (Fig. 3), and the results indicated that the particles were made of pure gadolinium oxide.
Figure 4 demonstrates the elemental map of gadolinium within the composite, indicated by the green color dots. The absorber particles were dispersed in the polymeric matrix in an entirely random manner. Such homogeneous dispersion of particles was observed at different concentrations of Gd2O3. The neutron shield composite, including epoxy and dispersed Gd2O3, is demonstrated in Fig. 5-a. As precipitation of denser gadolinium oxide might occur during the manufacturing of the composite, it was necessary to check if the Gd2O3 particles were homogeneously dispersed in the cured samples. For this purpose, we took the SEM images of the surface (Fig. 5-b) and cross section of the broken composite (Fig. 5-c). The results demonstrated a uniform distribution of particles throughout the body of the samples.
Figure` 6- XRD spectrum of a) Epoxy, b) Gd 2 O 3 , c) Epoxy / Gd2O3 composite
To study the molecular structure of the neat epoxy resin, and the potential effect of composite formation on the functional groups of the polymer, samples of epoxy and epoxy/gadolinium oxide composite were analyzed by the FTIR ATR method.
In both spectra, the peak around 3200 cm− 1 corresponds to the stretching vibration of O-H bond, the regions between 2800 and 2950 cm− 1 are attributed to symmetric and asymmetric stretching of C-H bonds in CH2 and CH3 groups, the peak at 1720 cm− 1 is related to the stretching vibration of carbonyl group C = O, and the observed peaks at 1630 and 1504 cm− 1 refers to the stretching of C-C and C = C bonds in aromatic compounds. Also, the absorptions in the regions of 1288, 1220, and 1105 cm− 1 are respectively ascribed to the stretching vibrations of C-O-C, C-O, and C-OH bonds. According to Fig. 7, the functional groups of the epoxy have not undergone significant changes due to the presence of Gd2O3 particles, so the addition of gadolinium oxide did not affect the molecular structure of the polymer, an observation that is in agreement with the previous studies 34,35. In the ATR analysis, the radiation used has a limited penetration depth in the sample. When a composite sample is present with a filler on its surface, the incident beam, at some points, might less effectively penetrate the surface, which in turn, might cause lower intensities of the composite peaks.
In the applications of the prepared composites, often, mechanical strength is one of the most significant requirements. In order to evaluate this property, a tensile test was performed on the Gd2O3 -Epoxy composite. According to ISO 527 standard (Fig. 8), three dumbbell-shaped specimens were made from the foregoing composite, a tensile test was carried out on all of the specimens, and the average value of strengths was used in the calculations. The test was repeated on the composites with different Gd2O3 concentrations. The gauge length has been considered to be 75 mm according to the ISO tensile standard. Initially, the pure epoxy sample was tested as the control sample, and the other samples were measured relative to it. The stress-strain curves are exhibited in Fig. 9. Also, the calculated values of the elastic modulus, tensile strength, and percentage of elongation at break are displayed in Fig. 10 for comparison.
In Fig. 9, the stress-strain curves of all composite samples change almost linearly up to the point of fracture. This trend indicates the rigid nature of the fabricated composites, which has also been observed in other studies, which compounded tantalum oxide 36, bismuth oxide 37 and gadolinium oxide 38 with epoxy in the composites. Based on the comparison of the tensile characteristics of different samples (Fig. 10), it can be concluded that by increasing of the percentage of Gd2O3, at first, elastic modulus and tensile strength increase, but above 2% Gd2O3 content, they start diminishing. Such behavior can be explained by the following mechanisms: 1) At lower concentration of Gd2O3, where the epoxy matrix is still homogenous, satisfactory adhesion of resin and Gd2O3, causes effective transferring of the mechanical load to the stiffer gadolinium oxide, thereby enhancing its mechanical strength, and 2) Higher Gd2O3 content, ruins the uniformity of the matrix and the obtained composite cannot act as an integrated and cohesive material 39,40. Note that the non-uniform distribution of metallic oxides at higher Gd2O3 concentrations may lead to the accumulation of particles in some points of the matrix. Such an effect is evidenced in Fig. 11. The points marked with yellow circles are the location of the particle accumulation in the 10% Gd2O3 sample. Apparently, sample fracture begins from these agglomerations. Besides the aforementioned effects, the percentage of elongation at the breakpoint showed a decreasing trend with the increasing Gd2O3 content.
The results of Fig. 10 are comparable with those reported by Prabhu et al. for the Ta2O5-reinforced composite 36. Based on the results of the mechanical tests of the current study, the optimal concentration of Gd2O3 in the epoxy composite was found to be around 2%.
Thermogravimetric Analysis (TGA) was performed on the plain and Gd2O3-containing composites to scrutinize their thermal resistances, and the potential changes, imposed on Gd2O3-containing samples during the fabrication. Our TG tests were started at the ambient temperature, and reached to700 oC, with a 10 \(\frac{{{}^{o}C}}{{\hbox{min} }}\) growing rate, in the nitrogen atmosphere.
According to Fig. 12, the highest percentage of the degradation occurred at 350 oC, but samples were completely decomposed at 700 oC. Adding the gadolinium oxide to the polymeric matrix, had a positive effect on the thermal stability of the composite (Fig. 12), likely because compounding with Gd2O3 increases the thermal conductivity of the samples, which in turn leads to a better dissipation of destructive thermal energy from the composite specimens. Similar results were reported by the previous researchers 41–43.
3 − 2 Neutron shielding performance
Epoxy samples with 0.5, 2, 5, and 10% Gd2O3 content underwent neutron attenuation experiments. For each of the aforesaid weight percentages, the neutronic test was carried out on the specimens with thicknesses ranging from 1 to 4 cm. All experiments and simulations were repeated three times, and their average values are reported, here. Neutron shielding simulation was carried out by the Monte Carlo particle transport model using the MCNPX code. At the large sample thicknesses, neutron moderation is significant, and scattering and absorption cross-sections increase as its consequence. This change in energy of the neutron beam leads to a greater effect of changes in the gadolinium mass fraction in thicker samples compared to the thinner ones. Note that at the fast and epithermal energies, any reduction in the neutron energy increases the absorption rate much more than the scattering rate. According to the experiments, all composites with 0.5%, 2%, 5%, and 10% Gd2O3 content performed better than the neat epoxy in terms of neutron absorption (Figs. 13–17). Most notably, the composite with 10 wt% gadolinium oxide could absorb 70% of the incident neutrons at a 4 cm thickness. The results indicate that Gd2O3-bearing composites are effective materials for neutron shielding. Here, there is a point that deserves attention: Although neutron capture increases with increasing both thickness and absorber content, absorption of the neutron in 157Gd isotope is accompanied by the production of gamma rays, and characteristic X-rays, which on the other hand increase the total dose of the personnel. As a result, the optimum values of Gd2O3 and composite thickness are those values that minimize the total dose of "neutron + gamma", and simultaneously meet the economical considerations. Engaging in such optimization processes, that require detailed information about the neutron reactions, and cost analysis is beyond the scope of the current study.
Table 2
Attenuation factor through experiment and simulation
| Thickness (cm) | Attenuation factor (ΔI/I0 × 100) |
| 0.0% Absorber | 0.5% Absorber | 2.0% Absorber | 5.0% Absorber | 10.0% Absorber |
| Exp. | Sim. | Exp. | Sim. | Exp. | Sim. | Exp. | Sim. | Exp. | Sim. |
| 1 | 7% | 10% | 20% | 23% | 34% | 30% | 42% | 36% | 47% | 42% |
| 2 | 19% | 16% | 34% | 30% | 48% | 39% | 52% | 46% | 58% | 49% |
| 3 | 30% | 24% | 46% | 42% | 56% | 48% | 59% | 54% | 66% | 57% |
| 4 | 37% | 35% | 54% | 55% | 63% | 59% | 66% | 62% | 70% | 68% |
| Average absolute error (AAE, %) | 12.8 |
Table 2 presents the experimental and computational values of the attenuation factor (Af ) for different examined samples. The last row of the table indicates an average absolute error (AAE, %), that is defined as:
$$AAE,\% =\frac{{\sum\limits_{1}^{n} {\frac{{\left| {A{f_{cal.}} - A{f_{\exp .}}} \right|}}{{A{f_{\exp .}}}}} }}{n} \times 100$$
6
where indices "cal." and "exp." denote the calculated and experimental values of "Af ", respectively; and "n" refers to the total number of the experiment. According to Table (2), (AAE, %) is about 12.5% which shows a relatively fair agreement between simulation and experiment.
In addition to the attenuation factor (Af ), the experimental data and calculation resulted in the effective macroscopic absorption cross-section (Σa). The data presented in Table 3 indicate that the macroscopic absorption cross-section increases nonlinearly with the Gd2O3 content. This observation could be attributed to the increased likelihood of microbubble and porosity formation when higher filler concentrations are used. Additionally, the formation of agglomerates becomes more prevalent at higher concentrations, which may have a detrimental effect on the shielding performance.
The average error of MCNP in the simulation of the experiments was about 14% which is fairly acceptable (Table 3).
Table 3
Macroscopic absorption cross-section, based on experiment and simulation
| Macroscopic Absorption Cross Section (Σa, cm− 1) |
| Gd2O3 (w/w, %) | Experiment | MCNP |
| 0.5 | 0.242 | 0.261 |
| 2 | 0.425 | 0.356 |
| 5 | 0.541 | 0.446 |
| 10 | 0.643 | 0.544 |
| Average absolute error (AAE, %) | 14.26 |