（1）Data cuts to remove background events

 In the experiments, the neutron capture cross-section is derived from measuring the γ-rays emitted in such reaction. Nonetheless, γ-rays emitted from other reactions and sources can also be recorded inducing intro background noise. To improve the signal-to-noise ratio, data cuts were applied to the detected γ-rays based on its energy. When a neutron capture reaction occurs, γ-rays are emitted with energies relative to the excited state of the measured isotope. In the case of ²⁴¹Am, γ-rays are emitted after a capture event have energies up to about 5.6 MeV. Hence, a data cut at 5.6 MeV was applied to the measured events, since γ-rays over such energy necessarily come from a background source. This is particularly important in the present experimental setup, since neutron captured in the I present in the NaI(Tl) detector are a large source of γ-ray noise with energies between 6.5 and 7 MeV. Moreover, events with energies of 478 and 511 keV are another large source of background events. The former originates from neutrons capture in boron, extensively used as neutron shielding, whereas 511 keV is the energy equivalent of an electron, which are generated by the interaction of γ-rays with surrounding materials. Hence, to reduce the influence of background events and improve the experimental accuracy, data cuts were performed at 0.6 and 6 MeV, removing the influence of high and lower energy γ-rays. These peaks, as well as the data cuts applied, can be seen clearly in the raw γ-ray spectra of the different measured samples shown in Fig. 2.

Fig. 2 Measured γ-ray from the different samples together with the data cuts applied

(2) Background estimation using resonances

 When measurements are performed, a large amount of background γ-rays are detected together with capture γ-rays coming from the measured sample. This background level becomes quite significant when the sample employed has a very small mass, and also is sealed in case, which was the case for the ²⁴¹Am, since it is highly radioactive. This is background is estimated and then removed through the measurement of other samples and a non-sample measurement. The most important background for ²⁴¹Am, was that caused by the sample case along with a small amount of Yttrium that was used as binder. The typical approach would be to rely in the Yttrium mass amounts provided by the supplier. However, these values carry a substantial amount of uncertainty since they are hard to quantify. Hence, in the present analysis, for a more accurate assessment of the background events caused by the sample case, the background normalization relied in physical phenomena, in this case the resonances caused by neutrons interacting with Yttrium, that display an extremely high degree of correlation to the amount of Yttrium irradiated, and also accounts for any flux attenuation within the sample. The measured spectra for both the ²⁴¹Am and dummy case samples is displayed in Fig. 3, with the Yttrium resonance used to normalize the background shown in the in-plot.

Fig. 3 Dummy spectra normalized using the resonance of Yttrium

(3) Normalization independent from nuclear data

 The cross-section is determined as the number of capture events induced in the sample per number incident neutron, and accounting for the area density of the sample. However, the final results have to be normalized since the capture events and the incident neutrons are determined from different measurements with distinct detection efficiencies. The most common approach involves the use of a well-known cross-section value present in evaluated nuclear data libraries, such as JENDL-5. Nonetheless, this procedure introduces the uncertainty present in nuclear data libraries, which is non-negligible. To avoid that, the present results were normalized using the saturated resonance technique, which is based on measuring an element that has a large capture resonance with a thick enough sample for that resonance to be completely saturated. This means that all neutrons with the energy of that resonance interact with the sample and, hence, the event rate attained at that resonance is equal to the incident neutron flux, accounting for 2 % of the neutrons being backscattered. The saturation achieved with a 0.1 mm thick ¹⁹⁷Au sample is shown in Fig. 4, as well as the normalized neutron flux to that resonance. Thus, by relying on a physical constrain, the final cross-section results were attained independent from other nuclear data.

Fig. 4 Normalized neutron flux to the saturated resonance of ¹⁹⁷Au

(4) Correction for missed γ-rays

 To reduce the influence of the background events, events with energies below 0.6 MeV were not considered in the analysis, as shown above. This means that capture γ-rays with energies below 0.6 MeV were also missed, but this fact does not introduce any energy-dependent bias in the cross-section results. However, since the results were normalized using the capture data of ¹⁹⁷Au with the saturated resonance technique, the ratio of the missed γ-ray below 0.6 MeV relative to the total γ-rays emitted ratio is different between ²⁴¹Am and ¹⁹⁷Au and needs to be taken into account. ¹⁹⁷Au emits capture γ-rays up to 6.5 MeV, much higher than the 5.6 MeV limit for ²⁴¹Am. This difference was corrected using the γ-rays spectra calculated using the nuclear reaction model code CCONE for both ²⁴¹Am and ¹⁹⁷Au. The calculated γ-ray spectra from CCONE for both isotopes are shown in comparison with that derived from the experimental work in Fig. 5.

Fig. 5 Experimental and calculated capture γ-ray spectra from ²⁴¹Am and ¹⁹⁷Au