3.1. Backscattering Proton Spectra

In our experiment we demonstrate that the counting rate of backscattered protons is too high when the beam current exceeds 10 nA. In order to restrict the dead time of the proton detector to less than 3%, we attached an aluminum collimator (thickness 100 μm, diameter 0.7 mm) to the detector.

First, we placed a 6.3-μm blank Mylar film on the sample holder. We recorded the spectra S under a 9-nA beam current for 1.20 μC Faraday cup charge integration. Figure 2 shows the proton spectrum and the SIMNRA simulation. The colored curves in Figure 2 are the general and elements-specific simulations. Peak ① indicates the protons backscattered by the helium atoms. The actual height of the helium (He) peak is lower than the simulation height and diminishes more rapidly in lower-energy regions. This can be attributed to the fact that when the distance from the exit window increases, the detector angle to the He volume decreases. The stacking of Peak②(carbon), Peak ③ (nitrogen), and Peak ④ (oxygen) indicates the protons backscattered by the Kapton film (Kapton film contains carbon). The air impurity in the He flow resulted in shifts of the nitrogen (Peak ⑤) and oxygen (Peak ⑥) peaks in lower-energy regions. Peak⑤ and Peak⑥ increased with decreases in the helium flow. The protons backscattered by the Mylar film were carbon (Peak ⑦) and oxygen (Peak ⑧). Because of the shelter provided by the exit device, the solid angle of the Mylar film was very small, which resulted in the peak height being much lower than those in the simulation.

Figure 2.Full-screen View

The Mylar film backscattering proton spectrum and simulation in the initial solid angle arrangement.

In order to confirm the sample effect on the spectra, we conducted our measurements using a thick lead plate sample. Figure 3(a) clearly shows that a lead (Pb) peak significantly affects the spectrum.

Figure 3.Full-screen View

Backscattering proton spectra: (a) In the initial solid angle arrangement. (b) After the Au-Si surface barrier proton detector was adjusted to have no solid angle to the sample.

Moreover, our detection and simulation results demonstrate that the air impurity in the He flow affected the nitrogen (N) and oxygen (O) peaks in low energy regions. In the initial solid angle arrangement, the samples significantly impacted the Kapton film backscattering proton peak. In other words, spectra were highly sample-dependent.

In order to eliminate the effect of samples on the proton backscattering spectra, we adjusted the position of the Au-Si surface barrier proton detector so that it had no solid angle in relation to the sample. We then re-recorded spectra under sample vacancy, Mylar film, and lead (Pb) plate sample situations. Figure 3(b) clearly shows the effect of this change: no protons backscattered by samples were detected. Additionally, the background level for the Pb plate sample was slightly and continuously heightened. This difference is more noticeable in low energy sections (most likely because of multiple scattered protons).

Beam-induced thinning of the exit window resulted in a shift of the trailing edges. ^[9] Furthermore, the PIGE cross-section was sensitive to beam energy, which was related to the thickness of the Kapton film. We artificially reduced the density of the Kapton film by 5% in the simulation, which resulted in shifts in both the He and Kapton peaks. The energy decrease at half of the maximum of the He peak was approximately 21KeV. Thus, in the future, beam-induced thinning of the exit window can be revealed by a shift in the He backscattering proton peak, as well as the FWHM of the exit window peak.

Based on the above results and in order to minimize the effects of air impurity and exit window thinning, we selected the peak area in the channel region (450~555) for proton total count integration.

3.2 Reproducibility under Different Sample Situations

After adjusting the detector position, we recorded proton backscattering spectra under three types of sample situations: Sample vacancy, Mylar film, and Teflon film. The beam current was approximately 9 nA, and the Faraday cup beam charge integration was 1.20 μC for every sample. We recorded the spectra five times under each sample scenario.

We integrated the total backscattered protons in channel 450~555. Figure 4(a) displays high levels of consistency for the specific results. Teflon film spectra were almost identical to Mylar film spectra; however, Teflon film was more fragile. In the sample vacancy situation, the helium atmosphere was diluted by air, and the N and O peaks in the low energy regions were sharpened. Significantly, different sample situations had no significant effect on factor C_Q when the proton detector had no solid angle in relation to the sample. For comparison to Figure 4(a), the integration of channel 400~555 is shown in Figure 4(b).

Figure 4.Full-screen View

Specific results of C_Q under the three types of sample situations. (a) Integration region is channel 450-555. (b) Integration region is channel 400~555.

3.3 Beam Current Dependence

Based on the above results, we only studied the beam current dependence of C_Q under the Mylar sample situation. The beam current increased from 2 nA to 20 nA, which covers the commonly used beam current levels in our laboratory. We recorded five spectra under each beam current, and the beam charge integration was 1.20 μC for every measurement. As shown in Figure 5, factor C_Q varied from 27.60 kc /μC to 29.91 kc /μC. The mean value was 28.95±0.6 kc /μC, with a RSD of 2.0%. In this study, the primary sources for errors are detection errors and charge integration errors.

Figure 5.Full-screen View

C_Q (the ratio between integral backscattering proton counts and integral beam charges) as a function of beam current.

We observed proton-beam induced thinning of the Kapton film exit window with the naked eye. Although this proton-beam induced thinning has not yet resulted in significant variations in the spectra or factor C_Q, the Kapton film should be replaced periodically to prevent it from bursting and to maintain stable proton energy loss.

Ultimately, our study determined that the value of C_Q can change with different detector solid angles, exit window types, and beam energies; however, the value of C_Q remains constant under a specific geometric arrangement and parameter setting. Additionally, factor C_Q can be revised before analysis if necessary. In the future, we will study the precise measurement of exit window thinning and the beam energy dependence of C_Q.