Impact of Neutron Irradiation on LGADs with a Carbon-Enriched Shallow Multiplication Layer: Degradation of Timing Performance and Gain

Impact of Neutron Irradiation on LGADs with a Carbon-Enriched Shallow Multiplication Layer: Degradation of Timing Performance and Gain E. Navarrete Ramos 1a, J. Duarte-Campderrosa, M. Fern´andeza, G. G´omeza, J. Gonz´aleza, S. Hidalgob, R. Jaramilloa, P. Mart´ınez Ruiz del ´ Arbola, M. Molld, C. Quintanaa, A. K. Sikdarc, I. Vilaa, J. Villegasb aInstituto de F´ısica de Cantabria, IFCA (CSIC-UC), Av. los Castros, Santander, 39005, Spain bInstituto de Microelectr´onica de Barcelona, IMB-CNM (CSIC), C/ dels Til·lers Cerdanyola del Vall`es, Barcelona, 08193, Spain cIndian Institute of Technology Madras, Tamil Nadu, Chennai, 600036, India dOrganisation Europ´enne pour la Recherche Nucl´eaire, CERN, Geneva 23, CH-1211, Switzerland Abstract In this radiation tolerance study, Low Gain Avalanche Detectors (LGADs) with a carbon-enriched broad and shallow multiplication layer were examined in comparison to identical non-carbonated LGADs. Manufactured at IMB-CNM, the sensors underwent neutron irradiation at the TRIGA reactor in Ljubljana, reaching a fluence of 1.5×1015 neq cm−2. The results revealed a smaller deactivation of boron and improved resistance to radiation in carbonated LGADs. The study demonstrated the potential benefits of carbon enrichment in mitigating radiation damage effects, particularly the acceptor removal mechanism, reducing the acceptor removal constant by more than a factor of two. Additionally, time resolution and collected charge degradation due to irradiation were observed, with carbonated samples exhibiting better radiation tolerance. A noise analysis focused on baseline noise and spurious pulses showed the presence of thermal-generated dark counts attributed to a too narrow distance between the gain layer end and the p-stop implant at the periphery of the pad for the characterized LGAD design; however, without significant impact of operation performance. 1Corresponding author: efren.nav[email protected] Preprint submitted to NIM-A April 2, 2025 arXiv:2406.01267v2 [physics.ins-det] 1 Apr 2025 Keywords: Timing detectors, Radiation-hard detectors, Si detectors, carbon enriched gain-layer. 1. Introduction The high-luminosity upgrade of the Large Hadron Collider (HL-LHC) is scheduled to begin in early 2029 and will deliver an integrated luminosity of up to 4000 fb−1over a 10-year period [1]. The HL-LHC will operate at a stable luminosity of 5.0×1034 cm−2s−1, with a possible maximum of 7.5× 1034 cm−2s−1. The main challenge of the HL-LHC will be the superposition of multiple proton-proton collisions per bunch crossing, known as pileup, in a small region. The multiple-collision region will extend to about 50 mm RMS along the beam axis, with an average of 1.6 collisions/mm and up to 200 pp interactions per bunch crossing. In these conditions, disentangling the multiple collisions and correctly associating the reconstructed tracks to their primary production vertex will be a major challenge. To address this, MIP timing sub-detectors have been proposed [2,3], which are targeting a track resolution of 30 ps per track. These detectors are expected to significantly improve the performance of the ATLAS and CMS detectors by disentangling the high number of pileup events. The CMS Endcap Timing Layer (ETL) is a sub-detector proposed to be built using Low Gain Avalanche Detector (LGAD) devices with a pixel size of 1.3×1.3 mm2. The ETL will cover the pseudorapidity range of 1.6<|η|< 3.0, with a total surface area of 14 m2. This sub-detector will be exposed to radiation levels up to 1.5×1015 neq cm−2at |η|= 3.0. However, for 80 % of the ETL area, the fluence is less than 1×1015 neq cm−2. Therefore, these two fluence points are the ones of interest for this radiation tolerance study. LGADs are semiconductor detectors designed for timing applications. They are constructed as n++ −p+−pavalanche diodes, with a highlydoped p+layer introduced to establish a region of very high electric field. This electric field is responsible for initiating the avalanche multiplication of primary electrons, generating additional electron-hole pairs. A schematic cross-section of a standard pad-like LGAD is illustrated in Figure 1. The LGAD structure is carefully engineered to achieve a moderate gain and operate effectively across a wide range of reverse bias voltages before reaching breakdown. We present a radiation tolerance study performed on LGAD with a carbonenriched multiplication layer. The LGAD sensors were manufactured at IMB2 Figure 1: Transversal structure of a Low Gain Avalanche Detector. The multiplication Layer, Collector Ring, P-Stop, Channel Stopper and the Junction Termination Extension (JTE) are visible. Thickness of the active volume and the low resistivity wafer are not scaled. CNM (Institute of Microelectronics of Barcelona, Spain) [4], with the same processing as the one used in the run #12916 [5]2. The LGADs are designed with a shallow gain layer doping profile which is characterized by a maximum of the p+doping concentration in the region close to the n++/p+junction and a relative broad p+implant [6]. Its performance was compared against LGADs with identical layout and manufacturing processing, but without carbon enrichment. The LGADs were irradiated with neutrons at the TRIGA reactor in Ljubljana up to a fluence of 1.5×1015 neq cm−2. The degradation of its timing performance and charge collection with fluence is reported. 2. Samples Description The IMB-CNM carried out a dedicated LGAD production for studying the effects of carbon on the broad multiplication layer that has been used up to now in its LGAD productions. The LGADs were manufactured on 6inch diameter epitaxial wafers with 55 µm active layer thickness and 525 µm support wafer thickness. The handle wafer has a resistivity of 0.001–1 Ω cm and the substrate resistivity is about 2 kΩcm. This run was carried out on epitaxial wafers (Run#15246, 6LG3 process), and is the first run in which the carbon enrichment of the gain layer was implemented by co-implantation as described in [7]. The run included matrices of different number of pads, 1×1 (single diodes), 2×2, 5×5, 16×16 and 16×32 where each pad is 1.3×1.3 mm2. 2The sensor production mentioned in this reference has been developed without carbon enrichment. 3 Results shown in this work refer to wafers W8 and W10. The manufacturing parameters of these two wafers are described in Table 1 including the gain layer depletion voltage measured before dicing the wafer, the boron dose and the Dry Oxidation Time (DOT). It is important to mention that the main difference between these two wafers is the implementation of carbon to the gain layer of wafer W8 in contrast with W10 that has a standard configuration consisting solely of boron implantation. Employing carbon coimplantation on the gain layer for the LGAD sensors manufacturing is beneficial in reducing the effects of the acceptor removal mechanism [8] which is an indicator of the degradation of the gain caused by radiation damage. In subsection 3.3 we show results on the acceptor removal impact on these carbonated sensors. Table 1: Differences in gain layer depletion voltage and carbon and boron doses for the carbonated and standard devices Wafer Carbonated Standard Gain layer depletion Voltage 30 30 Boron dose (1 ×1013 cm−2) 1.9 1.9 Carbon dose (1 ×1013 cm−2) 10 - Dry oxidation time DOT (min) 180 180 Measurements were conducted at Instituto de F´ısica de Cantabria (IFCA) in order to characterize sensors from both wafers. Table 2 shows a summary of the sensors measured in the radioactive source setup. Two carbonated and three non-carbonated (standard) were kept as reference. For radioactive source measurements, samples are arranged in stacks of 3 sensors, where one non-irradiated sensor from W10 was used as a reference. Sensors were irradiated with neutrons to 3 different fluences: 0.6×1015 neq cm−2, 1.0× 1015 neq cm−2and 1.5×1015 neq cm−2, in the 250 kW TRIGA Mark II reactor3 of the Joˇzef Stefan Institute (JSI) [9] at Ljubljana (Slovenia). The standard sensors at the highest fluence were not available for this study. The total number of sensors measured in electrical characterization is higher than the ones measured in the radioactive source setup as we will see in section 3. 3Capable of yield a maximum flux of around 2 ×1013 n cm−2s−1[10]. 4 Table 2: Summary of the LGADs measured in radioactive source setup Fluence (neq cm−2) Carbonated Standard 0 D207, D253 D302, D297 0.6×1015 D238, D217 D228, D259 1.0×1015 D259, D255 D277, D275 1.5×1015 D315, D269 - 3. Electrical Characterization The Current-Voltage (IV) and Capacitance-Voltage (CV) characteristics were measured in a probe station equipped with a thermal chuck. The measurements were performed before and after irradiation. Measurements of non-irradiated devices were conducted at room temperature, while the irradiated devices were measured at a temperature of −25 ◦C. The ohmic contact side (backside) of the sensors was connected to ground, while the cathode and the guard-ring were connected to High-Voltage (HV). For the IV measurement, the guard-ring and main diode currents were determined independently using two different Keithley 2410 sourcemeters [11] that allows supply the High-Voltage for the diode and measure the currrent at the same time. For the CV measurement, the guard-ring and the main diode were connected to HV using Keithley 2410 sourcemeters and read by a Quadtech 1920 LCR-meter [12] connected through a decoupling box. The circuit model used to determine the capacitance was a parallel RC circuit and the measurements were carried out at 1 kHz (100 Hz) frequency before (after) irradiation. In total, 44 LGADs were measured in IV and approximately half of them in CV before irradiation. 3.1. Current-Voltage characteristic Figure 2 displays the main diode current versus the reverse bias at room temperature for both types of sensors before irradiation. Across most of the bias voltage range, the current for both types of sensors is below the nanoampere. At full depletion, for example at 250 V, the leakage current for the carbonated sensors is around 0.91 nA, while for the standard sensors it is about 0.89 nA, representing a small difference of approximately 2%. The breakdown voltage VBD was determined for both carbonated and standard sensors estimating the change in the slope by using the method described in subsection 3.3. A modest increase in the breakdown voltage is 5 0 50 100 150 200 250 300 350 Bias Voltage [V] 10− 10 9− 10 8− 10 7− 10 6− 10 5− 10 4− 10 Pad Leakage Current [A] D213 D217 D228 D238 D243 D255 D259 D265 D269 D277 D294 D315 (a) Carbonated 0 50 100 150 200 250 300 350 Bias Voltage [V] 10− 10 9− 10 8− 10 7− 10 6− 10 5− 10 4− 10 Pad Leakage Current [A] D213 D228 D241 D259 D265 D275 D277 D292 D307 D315 D323 D325 (b) Standard 0 5 10 15 20 25 30 35 40 Bias Voltage [V] 0 0.05 0.1 0.15 0.2 9− 10× Pad Leakage Current [A] D213 D217 D228 D238 D243 D255 D259 D265 D269 D277 D294 D315 (c) Carbonated GL region 0 5 10 15 20 25 30 35 40 Bias Voltage [V] 0 0.05 0.1 0.15 0.2 9− 10× Pad Leakage Current [A] D213 D228 D241 D259 D265 D275 D277 D292 D307 D315 D323 D325 (d) Standard GL region Figure 2: The leakage currents of the main diode are presented as a function of reverse bias before irradiation. (a) shows the results for carbonated sensors, (b) shows the results for standard sensors. (c) and (d) exhibit a zoomed-in view of the bias region where the gain layer is depleted. observed for the carbonated samples compared to the standard ones. The VBD was found to be in the range of 290–350 V for the carbonated samples and 280–340 V for the standard samples. The average VBD is 332 V for the carbonated sensors and 321 V for the standard sensors, with a value dispersion (RMS) of approximately 17 V and 24 V, respectively. It is possible that, due to the relatively small size of the carbonated region compared to the remaining non-carbonated silicon active bulk, not much impact of the carbon on the VBD may be expected. A second electrical characterization at −25 ◦C was carried out after irradiation of the devices, from which two sensors of every type and fluence (0.6×1015 neq cm−2and 1.5×1015 neq cm−2) were characterised. In Figure 3 (a) the pad leakage current of the carbonated sensors can be observed as a 6 0 100 200 300 400 500 600 700 Bias voltage [V] 8− 10 7− 10 6− 10 5− 10 4− 10 Pad Leakage Current [A] =6.00e+14ΦD217, =6.00e+14ΦD238, =1.00e+15ΦD259, =1.00e+15ΦD255, =1.50e+15ΦD269, =1.50e+15ΦD315, (a) Carbonated irradiated 0 100 200 300 400 500 600 700 800 900 Bias voltage [V] 8− 10 7− 10 6− 10 5− 10 4− 10 Pad Leakage Current [A] =6.00e+14ΦD228, =6.00e+14ΦD259, =1.00e+15ΦD275, =1.00e+15ΦD292, =1.50e+15ΦD323, =1.50e+15ΦD325, (b) Standard irradiated Figure 3: Pad leakage currents after irradiation as a function of the reverse bias. The carbonated sensors are presented in (a) and the standard in (b). These IV curves are shown in log scale for Y axis. function of the reverse bias. We did not observe a sharp breakdown in the IV characteristics for the irradiated samples. Instead, the IV curves show a gradual rise in current, indicating a soft breakdown as the bias increases. We avoided applying a bias high enough to reach full breakdown because we did not want to operate the sensor beyond the single event burnout [13] limit. For both carbonated and standard samples, we did not observe a monotonic increase in the leakage current with fluence, suggesting the presence of radiation-induced gain suppression. This gain reduction appears to compensate for the radiation-induced increase in leakage current. 3.2. Capacitance-Voltage characteristic The capacitance of the bare sensors was measured before irradiation at room temperature with the guard-ring connected and at a frequency of 1 kHz on the LCR-meter. The curves of the capacitance versus the reverse bias applied for the carbonated and standard samples, are shown in Figure 4 (a) and Figure 4 (b) respectively. High homogeneity and reproducibility are evident in these curves. The CV curves start with a smooth decrease in capacitance in the gain layer region that ends at approximately 30 V for both carbonated and standard samples, representing the depletion voltage VGL. This is followed by another kink in the curve, indicating the depletion of the bulk, which ends with a final capacitance (Cend) of about 4.0 pF for both types of wafers, at voltages above 32 V (carbonated) and 32.5 V (standard). This Cend is consistent with the fact that all sensors have the same active 7 0 10 20 30 40 50 Bias voltage [V] 0 100 200 300 400 500 600 [pF] p C D213 D226 D228 D241 D243 D259 D265 D275 D277 D292 D294 D307 (a) Carbonated 0 10 20 30 40 50 Bias voltage [V] 0 100 200 300 400 500 600 [pF] p C D213 D226 D228 D241 D243 D259 D265 D275 D277 D292 D294 D307 (b) Standard 28.5 29 29.5 30 30.5 31 31.5 32 32.5 33 Bias voltage [V] 10 20 30 40 50 60 70 [pF] p C D213 D226 D228 D241 D243 D259 D265 D275 D277 D292 D294 D307 (c) Carbonated GL region 28.5 29 29.5 30 30.5 31 31.5 32 32.5 33 Bias voltage [V] 10 20 30 40 50 60 70 [pF] p C D213 D226 D228 D241 D243 D259 D265 D275 D277 D292 D294 D307 (d) Standard GL region Figure 4: Pad capacitance before irradiation as a function of the reverse bias. The carbonated sensors from the carbonated wafer are presented in (a) and the standard sensors in (b). (c) and (d) are zoomed views of the gain layer regions. The characteristic kinks in the curves due to the gain layer and bulk depletion can be observed. area and width. Figure 4 (c) and Figure 4 (d) show an enlarged view of the capacitance curves in the gain layer region, where it can be observed that, in general, the curves of all samples, carbonated and standard, follow a similar shape, but the VGL of the samples is less dispersed in the presence of carbon. To better observe the VGL region [14], these measurements were conducted at a temperature of 10 ◦C with a low frequency of 100 Hz configured in the LCR-meter. The remaining LCR-meter settings remained consistent with those used before irradiation. The pad capacitance after irradiation of the samples is depicted as a function of the applied bias in Figure 5 (a) for the carbonated samples and Figure 5 (b) for the standard devices. After irradiation we can see a peak in the capacitance until a local maximum that has been concluded to be 8 0 10 20 30 40 50 60 Bias voltage [V] 0 50 100 150 200 250 300 [pF] p C =0.0e+00ΦD207, =0.0e+00ΦD253, =6.0e+14ΦD217, =6.0e+14ΦD238, =1.0e+15ΦD255, =1.0e+15ΦD259, =1.5e+15ΦD269, =1.5e+15ΦD315, (a) Carbonated 0 10 20 30 40 50 60 Bias voltage [V] 0 50 100 150 200 250 300 [pF] p C =0.0e+00ΦD287, =0.0e+00ΦD297, =6.0e+14ΦD228, =6.0e+14ΦD259, =1.0e+15ΦD275, =1.0e+15ΦD277, =1.5e+15ΦD323, =1.5e+15ΦD325, (b) Standard Figure 5: Pad capacitance after irradiation as a function of the reverse bias. Non-irradiated samples added for comparison. The carbonated sensors are presented in (a) and the standard sensors in (b). Displacement of the VGL (start of the peak in the curve) as result of the irradiation at the three different fluences is observed. related with the presence of the multiplication layer [15]. The carbonated samples exhibit a noticeable degradation in the gain layer due to irradiation, resulting in a corresponding shift of VGL proportional to the fluence (lower VGL at higher fluences). The standard samples, also present a reduction in VGL that again is evident as a consequence of irradiation. However, the VGL values of the standard samples are lower compared to the carbonated samples. For instance, at the highest fluence, the VGL for the carbonated sensor is approximately 17 V while for the standard sensor, it is around 11 V. From these CV characteristics, we have considered the VGL as the last point before the increase of the capacitance, since this coincides with the VGL extracted from IV characteristics. 3.3. Determination of Acceptor Removal Coefficient It has been shown that LGAD sensors experience a reduction in gain after irradiation with charged hadrons or neutrons [16]. This reduction can be attributed to the initial acceptor removal mechanism, involving the gradual deactivation of acceptors forming the GL [17], specifically boron (B) in this case. As irradiation deactivates the boron implanted in the GL of LGAD sensors, the reverse bias required to fully deplete this gain layer, denoted as VGL, decreases compared to the pre-irradiation state. This reduction in VGL provides an indication of the remaining active boron in the GL. Assuming 9 200 300 400 500 600 700 Bias Voltage [V] 0 10 20 30 40 50 60 70 80 90 [ps] t σ D207, 0E14 D253, 0E14 D238, 6E14 D217, 6E14 D259, 10E14 D255, 10E14 D315, 15E14 D269, 15E14 (a) Carbonated LGADs 200 300 400 500 600 700 Bias Voltage [V] 0 10 20 30 40 50 60 70 80 90 [ps] t σ D302, 0E14 D297, 0E14 D228, 6E14 D259, 6E14 D277, 10E14 D275, 10E14 (b) Standard LGADs Figure 10: Time resolution of both type of sensors: Carbonated (a) and Standard (b), calculated using Equation 3 and with errors calculated with Equation 4. The carbonated sensors show a better behavior after irradiation compared to the standard sensors. All these measurements were performed at a temperature of −25 ◦C. δ1=((σ2,1δ2,1)2+ (σ1,3δ1,3)2+ (σ3,2δ3,2)2)1 2 2σ1 , δ2=((σ2,1δ2,1)2+ (σ1,3δ1,3)2+ (σ3,2δ3,2)2)1 2 2σ2 , δ3=((σ2,1δ2,1)2+ (σ1,3δ1,3)2+ (σ3,2δ3,2)2)1 2 2σ3 , (4) where δi,j is the error in the value σi,j. This method was applied to all samples of this study, maintaining the non-irradiated sensor mentioned in section 4 as reference. The resulting time resolution σtof carbonated and standard sensors is presented in Figure 10 (a) and (b) respectively. As the fluence increases, the voltage needed to achieve the same time resolution also increases. Time resolution improves as bias voltage grows. Finally, the bias voltage needed to reach values below 50 ps of time resolution is about 100 V smaller in carbonated detectors after irradiation. 5. Noise Study To ensure the functionality of the sensors, a noise study was carried out on the carbonated samples. Although this noise affects both carbonated 16 (a) Base-Line distribution example 300 400 500 600 700 Bias Voltage [V] 0.5 0.55 0.6 0.65 0.7 0.75 0.8 Noise [fC] D207, Unirrad D253, Unirrad D238, 6E14 D217, 6E14 D259, 10E14 D255, 10E14 D315, 15E14 D269, 15E14 (b) Base-Line Noise versus bias voltage Figure 11: Example of the distribution of the signals response from a carbonated LGAD irradiated to 1.5×1015 neq cm−2without Radioactive Source in order to extract the BaseLine noise and its gausian fit can be seen in (a). (b) is the summary of the extracted Base-Line Noise of the carbonated samples at the different fluences versus the bias voltage. and non-carbonated samples, this study was not carried out on the standard sensors. Key parameters considered include the baseline noise level and the presence and frequency of micro-discharges (thermally generated spurious pulses). The noise of the carbonated samples was investigated using a random trigger. Two hundred events (waveforms) per bias voltage were collected in a histogram (depicted in Figure 11 (a)), and the Gaussian width of the resulting distribution was taken as the noise value. The increase in noise with voltage was examined for various fluence values and summarized in Figure 11 (b). The increase in noise is more pronounced for the 0.6×1015 neq cm−2 samples, but does not prevent the sensors from reaching operating conditions, that is, the reverse current remains low enough. The fact that the 0.6× 1015 neq cm−2samples present a higher noise pulse rate may result to the fact that at this fluence point the gain suppression is relatively smaller than in the sensors exposed to higher fluences; therefore having a larger excess noise. The measurements were performed at a temperature of −25 ◦C but the study of noise at different temperatures will be carried out in future studies. The amplitude of dark counts or spurious pulses that appear when a high electric field induces micro-avalanches triggered by thermal generated primary electrons, was also measured at voltages close to breakdown for each fluence with a trigger of threshold level of −15 mV, this threshold corresponds to approximately 6 MIPs for a PIN diode (without gain) of 50 µm of active 17 300 400 500 600 700 Bias Voltage [V] 0.3 0.4 0.5 0.6 0.7 0.8 Charge [fC] D207, Unirrad D253, Unirrad D238, 6E14 D217, 6E14 D255, 10E14 D315, 15E14 D269, 15E14 Figure 12: Collected charge of the spurious pulses from the carbonated samples irradiated and non-irradiated. Calculated from the amplitude-charge correlation. Measurements taken at a temperature of −25 ◦C. thickness. Again, no radioactive source was employed for this study. The spurious pulses appeared in all the carbonated samples near the breakdown voltage. Figure 12 shows the equivalent collected charge for these spurious pulses, calculated from the amplitude-charge correlation obtained from the measurements taken in section 4. For the non-irradiated and lowest fluence samples, a scaling trend with the bias voltage can be seen, while for the two higher fluence samples no increase was observed near breakdown. To avoid the pulse rate being limited by the bandwidth of the digital scope, we decided to use NIM [26] electronic modules, specifically a Discriminator, a Timer and a Counter to obtain the pulse rate of the Dark Counts. The minimum threshold of the discriminator was −25 mV that is higher than the −15 mV used with the oscilloscope, resulting in an underestimation of the spurious pulse rate. The resulting rates for the different carbonated samples are shown in Figure 13. The plot starts at the operating voltage for each detector, that is the bias voltage needed to obtain 8 fC by a MIP (the so-called V(8 fC) requirement for the ETL [27]): 240 V,460 V,580 V and 690 V respectively for the nonirradiated, 0.6×1015 neq cm−2, 1.0×1015 neq cm−2and 1.5×1015 neq cm−2 irradiated samples. The early presence of spurious pulses was attributed to the short JTE and the distance between the end of the p+gain layer and the p-stop at the periphery of the pad (see Figure 1). For this production this distance was 23.5µm, which was decided in order to minimize, as much as possible, the inter-pad distance [28], which is one of the main challenges for this technology in multi-pad matrix type LGADs. 18 260 265 270 275 280 285 290 295 Bias Voltage [V] 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Frequency [Hz] D207, Unirrad D253, Unirrad (a) Non-irradiated 510 515 520 525 530 535 540 545 Bias Voltage [V] 0 200 400 600 800 1000 1200 1400 1600 1800 Frequency [Hz] D238, 6E14 D217, 6E14 (b) 0.6×1015 neq cm−2 680 685 690 695 700 705 Bias Voltage [V] 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Frequency [Hz] D259, 10E14 D255, 10E14 (c) 1.0×1015 neq cm−2 770 775 780 785 790 795 800 805 Bias Voltage [V] 0 1000 2000 3000 4000 5000 Frequency [Hz] D315, 15E14 D269, 15E14 (d) 1.5×1015 neq cm−2 Figure 13: Spurious pulse rate versus the bias voltage of the carbonated samples before irradiation (a), and at 0.6×1015 neq cm−2(b), 1.0×1015 neq cm−2(c) and 1.5×1015 neq cm−2 (d) irradiation fluences. Measurements taken in the Radioactive Source setup with NIM electronics with a threshold of −25 mV. 19 6. Conclusions In this study, the first manufacturing run at IMB-CNM of Low Gain Avalanche Detectors with a carbon-enriched multiplication layer was investigated for its radiation tolerance compared to conventional LGADs. The sensors were subjected to neutron irradiation at the TRIGA reactor in Ljubljana, reaching a fluence of 1.5×1015 neq cm−2. The results, reported in terms of degradation in timing performance and charge collection with increasing fluence, demonstrated the potential benefits of carbon enrichment in mitigating radiation damage effects, particularly the acceptor removal mechanism. The acceptor removal constant of the carbonated samples with respect to the standard samples was reduced to less (more) of a half from the results of the IV (CV) curves. Time resolution and the collected charge was studied on the Radioactive Source (RS) setup for samples non-irradiated and irradiated up to fluences of 1.5×1015 neq cm−2. As expected, degradation of the time resolution and the collected charge due to the irradiation was evidenced. The time resolution of the carbonated samples, at a fluence of 10 ×1014 neq cm−2at the maximum bias voltage of 600 V achieved before the breakdown regime, is of 40 ps while at same fluence and bias voltage for the Standard LGADs is about 75 ps, and at a maximum bias voltage of 680 V is around 57 ps. In terms of radiation tolerance, the carbonated samples meet the CMS ETL requirements for the low fluence region (below 10 ×1014 neq cm−2), with a time resolution of less than 50 ps with a collected charge of 8 fC. Additionally, a noise analysis was conducted on the samples. The investigation focused on key parameters, including baseline noise level and the occurrence and frequency of micro-discharges, which may manifest as spurious pulses in silicon detectors due to thermal generation. The noise of carbonated samples was analyzed using a random trigger, measuring signal width without a radioactive source. The resulting noise values were examined across various fluences, as depicted in Figure 12. Despite a more pronounced increase in noise for samples irradiated to 0.6×1015 neq cm−2, the elevated noise levels did not impede the device’s operation. Additionally, spurious, thermally generated pulses were measured beyond the operational voltages, showing a scaling trend with bias voltage for non-irradiated and lower fluence samples, while higher fluence samples exhibited no increase near breakdown. 20 Acknowledgments This work was developed in the framework of the CERN RD50/DRD3 collaboration and has been funded by the Spanish Ministry of Science and Innovation (MCIN/AEI/10.13039/501100011033/) and by the European Union’s ERDF program “A way of making Europe”. Grant references: PID2020113705RB-C31, PID2020-113705RB-C32 and PID2021-124660OB-C22. Also, it was supported by the following European funding programs: the European Union’s Horizon 2020 Research and Innovation (under Grant Agreement No. 101004761, AIDAInnova) and NextGenerationEU (PRTR-C17.I1). This work was also been developed in the framework of the ”Ayudas Maria Zambrano para la atraccion de talento internacional”, co-funded by the Ministry Of University of Spain and the European Union NextGenerationEU, reference code: C21.I4.P1; under the framework of ”Ayudas para contratos predoctorales para la formaci´on de doctores 2019”, co-funded by the European Social Fund program ”El FSE invierte en tu futuro” with grant reference: PRE2019087514; and the Plan Complementario en el ´ Area de Astrof´ısica y F´ısica de Altas Energ´ıas, financed by Next Generation EU funds, including the Recovery and Resilience Mechanism (MRR), the Recovery, Transformation and Resilience Plan (PRTR) and the Autonomous Community of Cantabria. References [1] CERN, High-Luminosity Large Hadron Collider (HL-LHC): Technical design report, CERN Yellow Reports: Monographs, 2020, https://cds. cern.ch/record/2749422. [2] CMS Collaboration, Technical proposal for a MIP timing detector in the CMS experiment phase 2 upgrade, CERN Technical Report, CERNLHCC-2017-027, LHCC-P-009, 2017, https://cds.cern.ch/record/ 2296612. [3] ATLAS Collaboration, Technical Proposal: A High-Granularity Timing Detector for the ATLAS Phase-II Upgrade (CERN-LHCC-2018-023. LHCC-P-012). Geneva: CERN, 2018, https://cds.cern.ch/record/ 2623663. [4] Institute of Microelectronics of Barcelona (IMB-CNM), 2023, https: //www.imb-cnm.csic.es/en. 21 [5] E. Curr´as, A. Doblas, M. Fern´andez, D. Flores, J. Gonz´alez, S. Hidalgo, R. Jaramillo, M. Moll, E. Navarrete, G. Pellegrini, I. Vila, Timing performance and gain degradation after irradiation with protons and neutrons of Low Gain Avalanche Diodes based on a shallow and broad multiplication layer in a float-zone 35 µm and 50 µm thick silicon substrate, Nuclear Instruments and Methods in Physics Research, A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2023, 51–55, https://doi.org/10.1016/j.nima.2023.168522. [6] I. Cort´es, P. Fern´andez-Mart´ınez, D. Flores, S. Hidalgo, J. Rebollo, Gain estimation of RT-APD devices by means of TCAD numerical simulations, Proceedings of the 8th Spanish Conference on Electron Devices, 2011, https://doi.org/10.1109/SCED.2011.5744152. [7] J. Villegas, S. Hidalgo, A. Merlos, N. Moffat, G. Pellegrini, M. Fern´andez, R. Jaramillo, E. Navarrete, A. K. Sikdar, I. Vila, Measurements on last IMB-CNM LGADs production, The 40th RD50 Workshop, 2022, https://indico.cern.ch/event/1157463/ contributions/4922755/. [8] M. Ferrero, R. Arcidiacono, M. Barozzi, M. Boscardin, N. Cartiglia, G. Betta, F. Dalla, Z. Galloway, M. Mandurrino, S. Mazza, G. Paternoster, F. Ficorella, L. Pancheri, W. Sadrozinski, F. Siviero, V. Sola, A. Staiano, A. Seiden, M. Tornago, Y. Zhao, Radiation resistant LGAD design, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2019, https://www.sciencedirect.com/science/ article/pii/S0168900218317741. [9] D. ˇ Zontar, V. Cindro, G. Kramberger, M. Mikuˇz, Time development and flux dependence of neutron-irradiation induced defects in silicon pad detectors, Nuclear Instruments and Methods in Physics Research, A 426, 1999, 51–55, https://www.sciencedirect.com/ science/article/pii/S0168900298014685. [10] European Nuclear Experimental Educational Platform (ENEEP), IJS Ljubljana, 2022, https://www.eneep.org/about/ijs/. [11] Tektronix, Keithley 2400 Standard Series SMU, 2023, https: 22 //www.tek.com/en/products/keithley/source-measure-units/ 2400-standard-series-sourcemeter. [12] IET Labs Inc, IET/QuadTech 1910/1920 1 MHz LCR Meter, 2023, https://www.ietlabs.com/1900-lcr-meter.html. [13] Lipton, J. Ronald, LGAD Single Event Burnout Studies, United States, 2021, https://doi.org/10.2172/1841397. [14] D. Campbell, A. Chilingarov, T. Sloan, Frequency and temperature dependence of the depletion voltage from CV measurements for irradiated si detectors, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 492, 2002, 402–410, https://www.sciencedirect.com/science/ article/pii/S0168900202013530. [15] M. Wiehe, M. Fern´andez, S. Hidalgo, M. Moll, S. Otero, U. Parzefall, G. Pellegrini, A. Barroso, I. Vila, Study of the radiation-induced damage mechanism in proton irradiated low gain avalanche detectors and its thermal annealing dependence, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 986, 2021, https://www.sciencedirect. com/science/article/pii/S0168900220312110. [16] Z. Galloway, V. Fadeyev, P. Freeman, E.-L. Gkougkousis, C. Gee, B. Gruey, C. Labitan, Z. Luce, F. McKinney-Martinez, H. F. Sadrozinski, A. Seiden, E. Spencer, M. Wilder, N. Woods, A. Zatserklyaniy, Y. Zhao, N. Cartiglia, M. Ferrero, M.Mandurrino, A. Staiano, V. Sola, R. Arcidiacono, V. Cindro, G. Kramberger, I. Mandi´c, M. Mikuˇz, M. Zavrtanik, Properties of HPK UFSD after neutron irradiation up to 6e15 n/cm2, Nuclear Instruments and Methods in Physics Research, A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2019, https://www.sciencedirect.com/science/ article/pii/S0168900219306278. [17] G. Kramberger, M. Carulla, E.Cavallaro, V. Cindro, D. Flores, Z. Galloway, S. Grinstein, S. Hidalgo, V. Fadeyev, J. Lange, I. Mandi´c, G. Medin, A. Merlos, F. McKinney-Martinez, M. Mikuˇz, D. Quirion, G. 23 Pellegrini, M. Petek, H. Sadrozinski, A. Seiden, M. Zavrtanik, Radiation hardness of thin Low Gain Avalanche Detectors, Nuclear Instruments and Methods in Physics Research, A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2018, https://www. sciencedirect.com/science/article/pii/S0168900218301682. [18] G. Kramberger, V. Cindro, I. Mandi´c, M. Mikuˇz, M. Zavrtanik, Studies initial acceptor removal in p-type silicon, 26th RD50 Workshop (Santander), 2015, https://indico.cern.ch/event/381195/ contributions/905665/. [19] C. Besleaga, A. Kuncser, A. Nitescu, G. Kramberger, M. Moll, I. Pintilie Bistability of the BiOi complex and its implications on evaluating the “acceptor removal” process in p-type silicon, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2021, https://www. sciencedirect.com/science/article/pii/S0168900221007944. [20] K. Wu, X. Jia, T. Yang, M. Li, W. Wang, M. Zhao, Z. Liang, J. Guimaraes, Y. Fan, H. Cui, A. Howard, G. Kramberger, X. Shi, Y. Heng, Y. Tan, B. Liu, Y. Feng, S. Li, M. Li, C. Yu, X. Yang, M. Zhai, G. Xu, G. Yan, Q. Zhai, M. Ding, J. Luo, H. Yin, J. Li, Design and testing of LGAD sensor with shallow carbon implantation, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2023, https://www. sciencedirect.com/science/article/pii/S0168900222009895. [21] N. Bacchetta, D. Bisello, A. Candelori, M. Da Rold, M. Descovich, A. Kaminski, A. Messineo, F. Rizzo, G. Verzellesi, Improvement in breakdown characteristics with multiguard structures in microstrip silicon detectors for CMS, 2001, https://www.sciencedirect.com/science/ article/pii/S0168900200012079. [22] M. Fernandez, Status report on the radiation tolerance assessment of CNM AIDA2020v2 and HPK-P2 LGADs., 16th Trento Workshop on Advanced Silicon Radiation Detectors, 2021, https://indico.cern. ch/event/983068/contributions/4223223/. [23] E.-L. Gkougkousis, Radiation hardness of 6” SoI CNM LGADs, 24 35th RD50 workshop, 2019, https://indico.cern.ch/event/855994/ contributions/3636943/. [24] CIVIDEC Instrumentation, TCT Amplifier for Detector Physics, integrated Bias-Tee, 2023, https://cividec.at/electronics-C2-TCT. html. [25] P. McKarris, M. C. Vignali, M. O. Wiehe, Commissioning of a beta setup for time resolution measurements, DT 2019 Summer students, 2019, https://indico.cern.ch/event/840877/. [26] Costrell, et al., Standard NIM Instrumentation System 1990, https: //www.osti.gov/biblio/7120327. [27] CERN, Market Survey - Technical Description Supply of LGAD sensors for the CMS MIP Timing Detector Endcap 2021, https://cds.cern. ch/record/2777890/files/MS-4689_Technical_Description.pdf. [28] P. Fern´andez-Mart´ınez, D. Flores, S. Hidalgo, V. Greco, A. Merlos, G. Pellegrini, D. Quirion, Design and fabrication of an optimum peripheral region for low gain avalanche detectors Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 821, 2016, https://www. sciencedirect.com/science/article/pii/S0168900216300663. 25