nLGAD gain response to low-penetrating particles

Contents lists available at ScienceDirect Nuclear Inst. and Methods in Physics Research, A journal homepage: www.elsevier.com/locate/nima Full Length Article nLGAD gain response to low-penetrating particles Jairo Villegas a,∗, Carmen Torres b, Milos Manojlovic a, M. Carmen Jimenez-Ramos b,c, Neil Moffat a, Javier Garcia Lopez b,d, Salvador Hidalgo a aInstituto de Microelectronica de Barcelona, IMB-CNM-CSIC, 08193 Cerdanyola del Valles, Barcelona, Spain bCentro Nacional de Aceleradores, CNA, 41092 Sevilla, Spain cDepartamento de Fisica Aplicada II, Universidad de Sevilla, 41012 Sevilla, Spain dDepartamento de Fisica Atomica, Molecular y Nuclear, Universidad de Sevilla, 41012 Sevilla, Spain ARTICLE INFO Keywords: LGAD X-rays and charged particle detectors IBIC ABSTRACT Detecting low-penetrating particles is fundamental for a wide range of applications in the fields of industry, medicine, and pure research. The Low Gain Avalanche Detector built on n-type substrates (nLGAD) is considered a good candidate for that, as it showed to have the potential to detect such particles with high sensitivity while avoiding the high noise levels associated with a traditional avalanche photodetector. An overview of the electrical characterization of the latest nLGAD batch fabricated at the Institute of Microelectronics of Barcelona is presented in this work. Additionally, gain response measurements for light with varying penetration depths in silicon and 600 keV protons are also presented. The results confirm the aforementioned potential of nLGADs to detect low-penetrating particles in silicon. Additionally, it was observed that the gain response of the devices is not only dependent on the penetration depth of the studied particles, but also on the nature and flux of the source. 1. Introduction Low Gain Avalanche Detectors (LGAD) were first conceptualized and fabricated at the Institute of Microelectronics of Barcelona (IMBCNM) [1]. As they were originally thought for timing applications for high-energy physics experiments, their design and manufacturing was targeted to detect charged particles with a high range in silicon (mips) [2]. In other words, a good gain response of a traditional IMBCNM LGAD is limited to high penetrating particles in silicon, and their overall performance was found to be drastically reduced when detecting low-penetrating ones. This effect originates from the depth at which a highor low-penetrating particle deposits its energy within the detector [3]. In order to overcome this disadvantage for applications requiring the detection of low-penetrating particles, the nLGAD was developed at IMB-CNM [3,4] as an evolution of the LGAD used for high energy physics experiments, and based on the n-type Avalanche Photodetector for low-penetrating photons [5]. The functionality of an nLGAD is alike to that of a traditional LGAD. That is, the charge multiplication phenomenon is achieved by diffusing a moderately doped multiplication layer in between an electrode and a high resistivity substrate, which creates a very localized high electric field around the PN junction when the detector is biased. The main difference of a nLGAD with respect to an LGAD is the conductivity ∗Corresponding author. E-mail address: [email protected] (J. Villegas). type of the substrate and the multiplication layer, which are n-type instead of p-type, as well as the conductivity type of the electrodes. This has certain implications, with the key point being the change in drift direction for electrons and holes. The first IMB-CNM nLGAD prototypes demonstrated a gain of 15–20 for low-penetrating visible photons (wavelength of 404 nm) [3], while the gain response of typical IMB-CNM LGADs for such particles was below 2. An extensive review of the physical mechanisms that make the construction of nLGADs optimal for low-penetrating particles can be found in [3]. The nLGAD is designed for applications requiring enhanced sensitivity to low-penetrating particles in silicon, such as soft X-rays or UV photons. While some strategies focus on reducing the entrance window in p-type LGADs [6] or standard n-type PiN detectors [7], the nLGAD can also enhance the signal amplitude through avalanche multiplication, potentially achieving a gain of 15–20 for these low-penetrating particles. 2. Fabrication and electrical characterization of the studied nLGADs The studied devices were fabricated on 100 mm n-type high resistivity wafers (CNM-4nLG1 technology), which corresponds to the https://doi.org/10.1016/j.nima.2025.170208 Received 3 August 2024; Received in revised form 23 November 2024; Accepted 6 January 2025 Nuclear Instruments and Methods in Physics Research A 1072 (2025) 170208 Available online 15 January 2025 0168-9002/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). J. Villegas et al. Fig. 1. Schematics of the front view and cross section of the studied nLGADs. Fig. 2. Doping concentration around the PN junction of the studied nLGADs, extracted via SRP technique [8]. second batch of IMB-CNM manufactured nLGADs, being the first one comprehensively studied in [3]. The devices have an active area and thickness of 1.3 ×1.3 mm2and 270 μm, respectively (Fig. 1). Within the active area, the n-type multiplication region has an area of 1×1 mm2. During the fabrication, a circular entrance window of 0.7 mm diameter is etched on the metal surface covering the active area, to avoid light reflection during gain response measurements. The surface of the entrance window is composed of a layer of 0.7 μm of silicon nitride over a0.4 μm layer of silicon dioxide. An nPiN wafer was also fabricated in this batch, so its diodes have the same geometry as the one depicted in Fig. 1. The net doping concentration of the PN junction composing the nLGADs was investigated via Spreading Resistance Profiling (SRP) technique (Fig. 2). An overview of this reverse engineering technique can be found in [8]. Such measurements allowed us to measure the thickness of the 𝑃++ layer (∼1 μm), which never gets depleted and hence composes, along with the passivation layers, the dead entrance window of the device. Capacitance vs. bias voltage measurements were performed on more than 20 devices distributed across different places on the wafer. The CV curves were obtained with a Keysight Agilent 4284 A LCR-meter in parallel mode, at 20 ◦C, 10 kHz and 500 mV AC. Fig. 3displays the average CV curve of such measurements, showing a very uniform depletion voltage of the gain layer (𝑉𝑔 𝑙≃ 28 V) across the wafer. Alongside, leakage current vs. bias voltage measurements were carried out for the same samples, at 20 ◦C and with a Keithley 2410. The backside of the detectors was grounded and the current was read from the frontside. The guard ring (Fig. 1) was connected in every measurement to separate the surface current from the leakage current in the detector bulk. The average leakage current curve is shown in Fig. 3, highlighting a leakage current (in the operational voltage region) of the order of the nA, and a mean breakdown voltage of 𝑉𝐵 𝐷≃ 225 V. 3. Gain response to ultraviolet, visible and infrared light Three nLGAD detectors were investigated for their gain response when exposed to ultraviolet (UV) light of 369 nm, visible light of 404 nm and infrared (IR) light of 1064 nm. The measurements were carried out at room temperature with a Transient Current Technique (TCT) setup [9]. The laser beam was focused around the center of the active area (Fig. 1), at normal incidence with respect to it and with a projected area of ≈(𝜋∕4)60 ⋅60 μm2on the surface of the entrance window (of ≈(𝜋∕4)700 ⋅700 μm2). Such center was found by scanning the signal response of the devices in X and Y using steps of 2 μm. The laser pulse frequency (1 kHz) was kept constant for all wavelengths and both nLGAD and reference nPiN during the measurements. The pulse width was also fixed to the minimum value of the laser diodes, though it differed for each wavelength: ≈100 ps for 369 nm and ≈300 ps for 404 and 1064 nm. The laser-generated signal is read from the device frontisde (as shown in Fig. 1), goes through a C2-HV Broadband Cividec amplifier, with an analog bandwidth of 2 GHz, 40 dB gain and an integrated Bias-Tee, before being analyzed using a DRS4 Evaluation Board Oscilloscope. For every device, wavelength and voltage point, 10 000 waveforms were averaged to extract the gain, that was inferred by dividing the integrated amplified output signal (Fig. 4) of the nLGADs by that of a nPiN of the same fabrication batch. As the 369 nm UV light is the least penetrating in silicon of all the studied wavelengths [10], the laser intensity was first controlled to achieve an acceptable Signal to Noise Ratio (SNR) in both the reference nPiN and the nLGAD. It was found that a reasonable SNR for the reference PiN was only obtained with the laser diode power at its maximum value (5 mW), for which SNR ranged between 2 and 4 (Fig. 4) within the studied voltage range. For lower values of the laser diode power, the UV signals in the nPiN were degraded (SNR < 2) or even lost (SNR ≤1), which would have impeded a gain response analysis. This maximum laser diode power value was maintained when evaluating the nLGADs gain for light pulses of 404 and 1064 nm. For these wavelengths, the maximum power values were 50 mW and 100 mW, respectively. Nevertheless, the number of photons being absorbed within the active volume of the detector was always smaller than those composing the actual laser pulse, regardless of the used wavelength. This can be understood by evaluating the reference PiN signal (Fig. 4) and the Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 2 J. Villegas et al. Fig. 3. Average curves (with standard deviation errors) of the capacitance and leakage current measurements at 20 ◦C. In the average IV curve, the bar errors are only displayed up to the mean breakdown voltage of 𝑉𝐵 𝐷≃ 225 V, as the relative error in the current measurements grows exponentially after such bias point. Fig. 4. TCT output signals for illumination of an nLGAD and an nPiN with light of 369, 404 and 1064 nm wavelength. The Y axis is adjusted for every wavelength to better distinguished the differences between the nLGAD and the nPiN signals. penetration depth in silicon for every photon species. For IR light of 1064 nm, with a mean penetration depth of ≈1000 μm [10] (>270 μm of active thickness), most of the photons will cross the detector without interacting with it, and only a negligible number of them will be absorbed close to the PN junction, hence triggering the avalanche mechanism for electrons. For visible light of 404 nm, with a penetration Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 3 J. Villegas et al. Fig. 5. (1): Gain response to TCT light pulses of 369, 404 and 1064 nm wavelength for the three studied nLGADs. (2): Average gain (with standard deviation error bars) at 100, 150 and 200 V of reverse bias as a function of the penetration depth [10] of the photon species. depth of ≈0.1 μm [10] (<270 μm), the largest part of the photons will be absorbed within the dead entrance window (passivation and 𝑃++ layers) without causing a detectable signal, and just a few of them will reach the surface of the depletion width to do so. The situation is even more drastic for UV light of 369 nm, with a penetration depth of ≈0.01 μm [10]. As a result, the actual intensity within the active volume of the detectors varied between wavelengths, despite the laser diodes being operated at their maximum power. By evaluating the nPiN signals in Fig. 4, we may infer that the highest intensity within the active volume occurs for IR light, followed by visible and UV light. It is worth noticing that the duration of the laser-generated signals is higher for IR than for visible or UV light. This is a direct consequence of the depth at which the charge carriers are generated. For the IR laser beam, the photons are absorbed with the same probability all along the active thickness of the nLGAD, so both electrons and holes have to cover a larger distance to the electrodes. Moreover, these primary holes (with a smaller mobility than electrons) will trigger the avalanche mechanism as they cross the PN junction in their motion from the bulk to the P electrode. In turn, this generates an extra bunch of electrons that delays the fall of the signal even further as they traverse the 270 μm thickness to the 𝑁electrode. In contrast, visible and UV laser pulses are absorbed very close to the PN junction. In this situation, holes are swiftly collected at the P electrode, leaving only the faster electrons to contribute to the signal duration as they drift to the 𝑁one. The potential of the nLGAD is highlighted when we examine the gain response (Fig. 5) and the signal amplitude (in contrast to the nPiN ones) for every one of the aforementioned photon wavelengths. The inner multiplication mechanism of the nLGAD allows for a greater SNR for all the studied wavelengths. When compared to an nPiN, such improvement of the SNR is boosted for UV light, with the lowest penetration depth in silicon of all the photon species investigated via TCT. It is significant to mention that the error associated to the reference PiN signals when illuminated with UV light was quite large due to their low SNR (Fig. 4). In turn, this caused an impact on how accurate the gain for such wavelength could be determined. 3.1. Gain response to visible light as a function of the beam flux The penetration depth of the target particle to be detected is not the only parameter that determines the gain response of an nLGAD. Evidence of gain reduction in p-type LGADs for high TCT beam fluxes was observed in [11] for IR photons of 1064 nm. Similar gain response measurements as a function of the beam flux of the 404 nm wavelength TCT laser were conducted on one nLGAD device (labeled as LG13 in Fig. 6. Charge integral of the laser-generated nPiN signals (CC) dependence on 𝑉𝑡ℎ within a linear range, at 100 V and 20 ◦C. The projected area 𝐴≈ (𝜋∕4)155 ⋅136 μm2 was unchanged during the measurements. Fig. 5). By definition, the beam flux is expressed as 𝐼=𝑁𝛾 𝐴(1) where 𝑁𝛾is the number of incident photons per unit time (that is, the intensity) and 𝐴is the projected area of the laser beam on the device surface. With a TCT setup, 𝑁𝛾can be tuned by adjusting a Digital to Analog Converter (DAC) threshold 𝑉𝑡ℎ, that controls the amount of power that is fed to the laser diode. The smaller the 𝑉𝑡ℎ value is, the greater the number of photons composing the output laser beam pulse [12]. The reference nPiN detector was first measured to find values of 𝑉𝑡ℎ where the integral of the laser-generated signal (e.g. the Collected Charge or CC) was linear. That is, to find a𝑉𝑡ℎ range where we can assume that 𝐶 𝐶(𝑛𝑃 𝑖𝑁) ∼𝑁𝛾∝ −𝑉𝑡ℎ. The results are shown in Fig. 6. On the other hand, 𝐴can be adjusted by displacing the board where the nLGAD is mounted in the z position (the perpendicular to the beam direction). By moving the detector towards or opposite to the opening of the laser source, the projection of the beam on the center of the detector surface de-focuses, so 𝐴can grow larger or smaller. As in the previous section, 𝐴was estimated by assuming an elliptical projection of the beam on the detector surface, which axes length are inferred by scanning in X and Y in steps of 2 μm. Larger values of 𝐴translate into having a bigger spatial spread of the incident photons onto the detector, hence reducing the beam flux. For a given 𝑉𝑡ℎ, it was observed that the reference nPiN signal did not change with 𝐴. In contrast, the nLGAD signal did, as Fig. 7depicts. This result showed the first evidence of Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 4 J. Villegas et al. Fig. 7. Charge integral of the laser-generated nLGAD and nPiN signals (CC) dependence on the projected area 𝐴, at 100 V and 20 ◦C. The DAC threshold 𝑉𝑡ℎ = 1056 mV was unchanged during the measurements. The CC is normalized, for every device, with respect to its maximum value across 𝐴, in order to better distinguish the trend for the nLGAD and the reference nPiN. Table 1 DAC, CC(nPiN) at 20 ◦C and 100 V, axes length in X and Y for the beam projection, estimated A and relative flux. DAC (mV) CC(nPiN) (arb) X (μm) Y (μm) A=(𝜋∕4)XY (μm2) Rel. flux (%) 1353 109.7 150 144 16 953 3.7 1254 143.5 162 135 17 143 4.8 1056 235.7 161 148 18 706 7.2 1056 235.7 133 122 12 801 10.5 825 318.9 144 138 15 675 11.6 1056 235.7 108 97 8285 16.2 825 318.9 115 113 10 203 17.8 1056 235.7 83 73 4775 28.2 825 318.9 91 90 6388 28.5 825 318.9 69 68 3686 49.4 1056 235.7 60 52 2415 55.7 825 318.9 48 48 1819 100 gain suppression in nLGADs, as both the signal height and its CC got reduced for smaller values of 𝐴(i.e. for larger values of the beam flux). Both the DAC threshold 𝑉𝑡ℎ and the projected area 𝐴were swept to investigate the nLGAD gain response dependence on the 404 nm wavelength beam flux, which was estimated to be proportional to 𝐶 𝐶(𝑛𝑃 𝑖𝑁 , 𝑉𝑡ℎ)∕𝐴.Table 1shows the DAC, CC(nPiN) and A values that were used during the measurements. The measurements were carried out at 100 V and 20 ◦C. 𝑉𝑡ℎ was swept within the linear range presented in Fig. 6, while 𝐴was always kept smaller than half the area of the entrance window (≈(𝜋∕4)700⋅700 μm2) to ensure that there was no reflection of the incident photons with the metal. The beam was always pointed at the center of the entrance window and at normal incidence. For every flux point, the gain was inferred by dividing the integrated output signal of the nLGAD by the nPiN one. The results are shown in Fig. 8, where a clear gain reduction with increasing beam flux is observed. In this figure, the flux was scaled to its maximum value, which corresponds to the minimum value of both 𝑉𝑡ℎ and 𝐴. This result highlights that the gain response of an nLGAD for low-penetrating photons is subjected to both the penetration depth of the species and the beam flux of the source. 4. Gain response to low-energy protons Gain response to low-energy protons measurements were performed using the nuclear microprobe beamline at the Centro Nacional de Aceleradores (CNA, Seville) using a 600 keV proton beam. Protons with such energy are low-penetrating in silicon, as their range is ≈10 Fig. 8. nLGAD gain response, at 100 V and 20 ◦C, as a function of the 404 nm wavelength laser beam flux (as defined in Eq. (1)). The intensity is scaled to its maximum value, which corresponds to the minimum value of 𝑉𝑡ℎ and 𝐴in Table 1. μm [13]. The microprobe is connected to a 3 MV tandem accelerator that provides different species and ion energies [14]. It showcases a set of micrometer slits designed to avoid radiation damage during Ion Beam Induced Current (IBIC) characterization and to prevent the ion beam halo, which decreases the spatial resolution. The quadrupole lens system allows for focusing the ion beam to micrometer dimensions. This setup is further complemented by a 2D scanning system synchronized with the acquisition one, allowing spatial information by enabling mapping with excellent spatial resolution for areas of up to a few millimeters. A more detailed description of the nuclear microprobe beamline is provided in [15]. IBIC measurements can be conducted with the ion beam incident at different angles with respect to the normal of the detector. This is enabled by a special sample holder capable of rotating under vacuum with an accuracy of 1 degree. Furthermore, the signal acquisition and subsequent processing were carried out using a conventional electronic chain, comprising an HV source (model NHR 22 20X), a CANBERRA pre-amplifier (Model 2003BT), a Tennelec TC 245 amplifier, and connected to the OMDAQ acquisition software. One nLGAD and one reference nPiN detector were tested via IBIC. The nLGAD under test was previously measured via TCT, and corresponds to the one labeled as LG13 in Fig. 4. The reference nPiN was also the same as the one used during the TCT measurements reported in the previous section. Fig. 9shows a typical energy spectrum obtained when conducting an IBIC 2D scan on the nLGAD under test. The energy deposited by a 600 keV proton in the nLGAD sample generates an electrical signal, that is processed by the aforementioned electronics chain. The energy spectrum can further be imaged as a 2D mapping, which allows to evaluate the channel range where the counts correspond to protons crossing the multiplication layer (Fig. 10). Once such channel range is identified, the collected charge (CC) in the device can be defined as its centroid (Fig. 9), obtained via Gaussian fit. The same procedure is then done with the reference nPiN, which allows us to estimate the gain by dividing the CC of the nLGAD by the nPiN one. The CC experimental error is statistically inferred by dividing the Full Width at Half Maximum (FWHM) of the Gaussian fit by the square root of number of counts within its channel range. In Fig. 10, the spectra of the nPiN and nLGAD were obtained at normal incidence, 50 V and room temperature. The 2D sweep in the nPiN case was done around the active area (discarding periphery elements), reason why only a peak is observed in the energy spectrum. It is worth noticing that the energy peak around the nLGAD active area is way broader than that of the nPiN. This is due to statistical fluctuations in the number of avalanche-generated charge carriers. The Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 5 J. Villegas et al. Fig. 9. Energy spectrum, at normal incidence, 50 V and room temperature, for 600 keV protons impinging on the nLGAD under test. The left picture shows the spectrum in logarithmic scale, where counts in the nLGAD periphery elements can be grasped. The right picture shows, in linear scale, only the energy range around the active area, along with its Gaussian centroid. Fig. 10. (1): Spectra of the nPiN and nLGAD, at normal incidence, 50 V and room temperature. (2): Image reconstruction of the nLGAD spectrum in (1), where both the active area and the periphery elements (1) can be distinguished. The channel number has been scaled, in both pictures, to the CC of the reference nPiN, so a gain spectrum and map are obtained. energy resolution at 50 V, estimated as 𝐹 𝑊 𝐻 𝑀∕𝐶 𝐶, was found to be 0.5% for the nPiN and 17.7% for the nLGAD. 4.1. Gain response vs. reverse bias at normal incidence With the aforementioned methodology, gain response measurements were conducted on the devices under test. The first set of measurements were carried out at room temperature, normal incidence and sweeping the bias voltage. The gain response was evaluated both by impinging the 600 keV proton beam through the frontside (𝑃++ in Fig. 1) and the backside (𝑁++ in Fig. 1) of the devices. Taking into account the thickness of the passivation, 𝑃++ and multiplication layers (Figs. 1and 2), a 600 keV proton has its Bragg peak within the highresistivity substrate when the beam is incident through the nLGAD frontside. In particular, the position of such peak occurs at a depth of ≈6.5 μm, setting the zero depth at the PN junction, as displayed in Fig. 2. Given this scenario, holes will be the main contributor to trigger the avalanche mechanism as they are drifted to the 𝑃++, leaving just a few electrons to do so on the opposite direction. On the other hand, impinging the proton beam through the backside leave us with holes as the sole contributor to trigger the avalanche mechanism. The results are shown in Fig. 11, along with the gain response to 1064 nm IR TCT beam pulses already presented in Fig. 5. As in the case of having the 600 keV beam incident to the device frontside, most of the avalanche-generated charge is caused to the drift of holes for TCT IR photons under the same incident conditions. In turn, Fig. 11. nLGAD gain response to 600 keV protons (with the beam incident to the device front and backside) and 1064 nm IR TCT beam pulses. this would result in a similar gain response for both particle species, as confirmed by the results in Fig. 11. It is worth noting that gain suppression effects still need to be investigated both for TCT IR photons and 600 keV protons. Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 6 J. Villegas et al. Fig. 12. (1): Bragg peak depth dependence on rotation angle, inferred via SRIM simulations, for 600 keV protons impinging on the studied nLGAD. (2): nLGAD gain response, at 50 V and room temperature, to 600 keV protons as a function of the Bragg peak depth. 4.2. Gain response dependence on Bragg peak depth The interest in rotating the nLGAD during the measurements is to modify the depth of the Bragg peak of the incident protons within the detector. In other words, the more we rotate the nLGAD holder, the closer to the detector surface the ions deposit the majority of their energy. This occurs due to the longer track within the dead entrance window layers that protons need to traverse as the rotation angle increases. Taking into account the thicknesses of the dead entrance window layers that cover the nLGAD multiplication layer (Figs. 1and 2), the Bragg peak depth dependence on rotation angle for 600 keV protons can be estimated via SRIM simulation [16], as shown in Fig. 12(1). The gain response results obtained by varying the incidence angle (i.e. the Bragg peak depth) at room temperature are also shown in Fig. 12(2). The reverse bias was fixed at 50 V because, at higher voltages and incidence angles (corresponding to high inner gain values), the noise in the pre-amplifier increased significantly. The analysis of the results reveals that there are three gain response regions based on the protons Bragg peak depth within the device. When the Bragg peak is located within the bulk, the gain is nearly constant, and holes are the major contributor to cause impact ionization. When the Bragg peak is located within the multiplication layer, more electrons start to be involved in the avalanche mechanism, so the gain increases steadily from ≈2 to ≈4. Lastly, once the Bragg peak is located within the 𝑃++ layer, the gain response is boosted from ≈4 to ≈15. On one hand, this is due to the larger number of electrons triggering avalanche multiplication. On the other hand, the ionization charge density within the PN junction drastically drops once the Bragg peak is located within the 𝑃++, as the generated charge is quickly recombined in this layer. In turn, this may reduce potential gain suppression effects. 4.3. Potential applications Overall, the results in Fig. 12 set a stepping stone for the potential use of nLGADs in applications requiring efficient detection of lowpenetrating charged particles, such as neutron detection systems [17], plasma heating for nuclear fusion experiments [18] or the study of the natural 𝛽-decay of neutrons [19,20]. Neutrons are neutral particles, making them difficult to detect directly with silicon detectors. However, by coating the entrance window with specific materials, neutrons can be detected indirectly. Such materials include Li-6 or B-10, that are able to capture neutrons, a reaction that produces other charged particles, including alphas of ∼1 MeV [17]. An efficient detection of these low-penetrating alpha particles, with a range in silicon of ∼1 μm [13], is essential in these experiments. On the other hand, plasma heating for nuclear fusion experiments is often performed by injection of protons accelerated at ≈100 keV [18]. A proper particle detection system for these protons, with a range in silicon of ∼1 μm [13] is needed for plasma diagnosis during its heating. Finally, the study of the natural 𝛽-decay of neutrons [19,20] requires to detect and analyze the properties of the remnant proton of such reaction, with an energy of ≈1 keV and a range in silicon of ∼ 0.01 μm. 5. Conclusions The studies presented in this work demonstrate that the nLGAD gain response is enhanced for low-penetrating photons detection, a result already observed in [3,4]. Moreover, they also highlight that the flux of the beam source is a key parameter to consider when evaluating the performance of such devices, as higher fluxes lead to lower gain values. This is a direct consequence of having a higher ionization charge density around the PN junction (i.e. where the electric field is at its maximum value), an effect already observed in [11]. Additionally, IBIC measurements to 600 keV protons suggest that the gain response is enhanced in the case of low-penetrating charged particles detection, as long as they deposit their energy close to the device surface. As a case in point, it has been observed that relatively moderate levels of gain (5–20) are obtained only when electrons are the major contributor to trigger the avalanche mechanism. The gain suppression mechanism may also be present in nLGADs when detecting charged particles, as found in [21] when evaluating the gain response of p-type LGADs. However, further measurements are required to fully confirm this hypothesis. Overall, the gain response results establish a basis for the potential use of the nLGAD in applications that require efficient detection of low-penetrating photons or charged particles. CRediT authorship contribution statement Jairo Villegas: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Carmen Torres: Methodology, Formal analysis, Data curation. Milos Manojlovic: Formal analysis, Data curation. M. Carmen JimenezRamos: Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis. Neil Moffat: Methodology, Investigation. Javier Garcia Lopez: Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis. Salvador Hidalgo: Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 7 J. Villegas et al. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This work has been funded by the Spanish Ministry of Science, Innovation and Universities (MICIU/AEI/10.13039/501100011033/) and by the European Union’s ERDF program ‘‘A way of making Europe’’. Grant references: PID2020-113705RB-C32, PID2021-124660OBC22, PDC2021-121718-C32 and PDC2023-145925-C32. Also, it was supported by the following European funding programs: European Union’s Horizon 2020 Research and Innovation (under Grant Agreement No. 101004761, AIDAInnova) and NextGenerationEU (PRTRC17.I1). M. Carmen Jimenez-Ramos acknowledges the support of this work through a VI PPIT-US contract. This investigation has also been partially financed by the project ref. ASTRO21/1.1/1 with financing from the European UnionNextGenerationEU, the Ministry of Science, Innovation and Universities, Recovery Plan, Transformation and Resilience, Spain, the Department of University, Research and Innovation of the Junta de Andalucia and the University of Sevilla, Spain. Data availability Data will be made available on request. References [1] G. Pellegrini, et al., Technology developments and first measurements of low gain avalanche detectors (LGAD) for high energy physics applications, Nucl. Instrum. Methods Phys. Res. A 765 (2014) 12–16, http://dx.doi.org/10.1016/j.nima.2014. 06.008. [2] C. Grieco, et al., Overview of CNM LGAD results: Boron Si-on-Si and epitaxial wafers, J. Instrum. 17 (C09021) (2022) http://dx.doi.org/10.1088/1748-0221/ 17/09/C09021. [3] J. Villegas, et al., Gain measurements on NLGAD detectors, Nucl. Instrum. Methods Phys. Res. A 1055 (168377) (2023) http://dx.doi.org/10.1016/j.nima. 2023.168377. [4] W. Khaled, et al., First results for the pLGAD sensor for low-penetrating particles, Nucl. Instrum. Methods Phys. Res. A 1040 (167220) (2022) http://dx.doi.org/ 10.1016/j.nima.2022.167220. [5] Hamamatsu Si APD catalog, 2024, [Online]. Available: https://www.hamamatsu. com/content/dam/hamamatsu-photonics/sites/documents/99_SALES_LIBRARY/ ssd/si_apd_kapd0001e.pdf. [6] V. Hinger, et al., Resolving soft X-ray photons with a high-rate hybrid pixel detector, Front. Phys. 12 (2024) http://dx.doi.org/10.3389/fphy.2024.1352134. [7] J. Segal, et al., Thin-entrance window process for soft X-Ray sensors, Front. Phys. 9 (2021) http://dx.doi.org/10.3389/fphy.2021.618390. [8] R.G. Mazur, D.H. Dickey, A spreading resistance technique for resistivity measurements on silicon, J. Electrochem. Soc. 113 (255) (1966) http://dx.doi.org/ 10.1149/1.2423927. [9] G. Kramberger, Advanced transient current technique systems, in: The 23rd International Workshop on Vertex Detectors, 2015, http://dx.doi.org/10.22323/ 1.227.0032. [10] M.A. Green, Self-consistent optical parameters of intrinsic silicon at 300 K including temperature coefficients, Sol. Energy Mater. Sol. Cells 92 (4) (2008) 1305–1310, http://dx.doi.org/10.1016/j.solmat.2008.06.009. [11] E. Curras, et al., Gain suppression mechanism observed in low gain avalanche detectors, Nucl. Instrum. Methods Phys. Res. A 1031 (166530) (2022) http: //dx.doi.org/10.1016/j.nima.2022.166530. [12] Particulars laser user manual, version 2.0, 2015, [Online]. Available: https: //www.particulars.si/index.php. [13] M. Berger, et al., Stopping-power and range tables for electrons, protons, and helium ions, NISTIR (4999) (2017) http://dx.doi.org/10.18434/T4NC7P. [14] J. Gomez-Camacho, et al., Research facilities and highlights at the centro nacional de aceleradores (CNA), Eur. Phys. J. Plus 136 (273) (2021) http: //dx.doi.org/10.1140/epjp/s13360-021-01253-x. [15] M.A. Ontalba, et al., External microbeam set-up at the CNA (sevilla) and its application to the study of tartesic jewellery, Nucl. Instrum. Methods Phys. Res. B 181 (2001) 664–669, http://dx.doi.org/10.1016/S0168-583X(01)00355-X. [16] J.F. Ziegler, et al., SRIM – the stopping and range of ions in matter, Nucl. Instrum. Methods Phys. Res. B 268 (11–12) (2010) 1818.–1823, http://dx.doi. org/10.1016/j.nimb.2010.02.091. [17] B.F. Phlips, et al., Neutron detection using large area silicon detectors, Nucl. Instrum. Methods Phys. Res. A 579 (1) (2007) 173–176, http://dx.doi.org/10. 1016/j.nima.2007.04.033. [18] H. Han, et al., A sustained high-temperature fusion plasma regime facilitated by fast ions, Nature 609 (2022) 269—-275, http://dx.doi.org/10.1038/s41586-02205008-1. [19] L.J. Broussard, et al., Using nab to determine correlations in unpolarized neutron decay, Hyperfine Interact. 240 (1) (2019) http://dx.doi.org/10.1007/s10751018-1538-7. [20] D. Moser, et al., NoMoS: An R×B drift momentum spectrometer for beta decay studies, in: International Workshop on Particle Physics at Neutron Sources, PPNS 2018, no. 04003, 2018, http://dx.doi.org/10.1051/epjconf/201921904003. [21] M.C. Jimenez, et al., Study of ionization charge density-induced gain suppression in LGADs, Sensors 22 (3) (2022) http://dx.doi.org/10.3390/s22031080. Nuclear Inst. and Methods in Physics Research, A 1072 (2025) 170208 8