Analysis of material reliability features of plasma material interactions Authors: Alper Pahsa Published: 2026 ABSTRACT This work presents a statistical methodology for evaluating material surface degradation and reliability based on roughness measurements obtained from Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM). Surface roughness metrics were employed to evaluate and model aluminum specimens exposed to helium plasma via a Weibull reliability method. The concordance between AFM-based and SEM-based measurements was assessed by Pearson correlation analysis. Robust and stable strong positive correlations are analyzed across both electrode configurations. For the cathode surfaces, a strong positive correlation (r = 0.99) is generated. Similarly, the updated assessment for the anode surfaces show highly consistent pattern, producing strong positive correlations for both the RMS–Ra (r = 0.992) comparisons. This finding indicates that while SEM-derived optical roughness parameters can effectively and reliably regenerate AFM observations with these experimental conditions. However, given the limited sample size (n = 3) assesed in this work, these strong statistical correlations must be interpreted with caution, as smaller sample groups are inherently susceptible to localized data sensitivity. All of the results indicate that while SEM-based analysis serves as a highly feasible, efficient, and cost-effective alternative for material screening and reliability modeling, its wide application needs carefull calibration and validation across larger experimental datasets. Introduction With a substantial amount of research in the literature, surface material and plasma interactions are popular subjects in condensed matter modification. In temperature-changing chambers, plasma-material interactions change the surfaces of the materials. In magnetic fusion devices, for instance nuclear fusion reactors, interactions between plasma materials result in notable artifacts. The viability of magnetic confinement fusion energy is critically dependent on the longevity of plasma-facing components (PFCs). In devices like Tokamaks, the reactor walls are subjected to intense bombardment by high-energy particles from the plasma, leading to sputtering, erosion, and the implantation of fuel ions [1–8]. These plasma-material interactions (PMIs) degrade the wall’s structural integrity, release impurities into the plasma that quench the fusion reaction, and ultimately determine the component’s operational lifetime [9–12]. Therefore, predicting the structural reliability and lifecycle of PFCs is a fundamental challenge in the engineering of future fusion power plants. Current research efforts are primarily focused on experimental characterization of material damage and system-level reliability assessments. For instance, studies on the International Thermonuclear Experimental Reactor (ITER) have advanced a RAMI (Reliability, Availability, Maintainability, and Inspectability) framework for system-wide safety and availability [13]. Other approaches adapt probabilistic safety assessments (PSA) from traditional fission reactors [14, 15]. However, a significant gap exists between these macro-level system analyses and the micro-level material damage data. The current state of the art lacks a robust methodology to directly translate experimental observations of plasma-induced surface deformation into a quantitative prediction of material structural reliability. While one study [16] has applied a Weibull distribution to aluminum, a comprehensive reliability modeling approach that can inform material selection and wall design remains underdeveloped. New advancements in plasma-facing materials (PFMs) go beyond the usual suspects like metal. Current main focus is in tungsten (W) for the divertor and beryllium (Be) for the first wall in reactors such as ITER. Helium (He) plasma irradiation of tungsten leads to a delicate nano-tendril “fuzz” growth layer that gives rise to enhanced sputtering, dust generation and tritium retention. Research on self-passivating alloys, on liquid metals (e.g. lithium, tin) and on fiber-reinforced composites treats cracking, melting and erosion. Beyond helium etching, the most severe reliability problem is the impact of 14 MeV neutrons from D-T fusion reactions. Neutron bombardment generates atomic displacements and transmutation of elements (e.g. W to Re and Os) which results in severe hardness, embrittlement and swelling, fundamentally changing the mechanical properties and lifetime. And material reliability is not solely a result of steady state operation. Off-normal events such as ELMs and disturbances deposit enormous volumes of thermal energy in milliseconds causing melting, cracking and ablation. The fusion community is also advancing the development of “digital twins” of reactors using real-time data where multi-scale modeling links atomic-scale damage, microstructural evolution and component-scale thermo-mechanical performance. This study addresses the gap by establishing a statistical framework that correlates surface roughness measures derived from AFM and SEM with models of material reliability. This study does not provide new experimental material systems; rather, it evaluates previously verified datasets to examine the consistency of measurement protocols and their suitability for reliability forecasting. Aluminum is employed as an alternative material due to its accessibility and predictable response to plasma exposure. It is essential to note, however, that aluminum is not an exemplary plasma-facing material for fusion reactors. The objective of this study is not to provide direct design recommendations for fusion reactor walls, but to establish a transferable scientific framework for assessing surface characterization techniques and reliability models. Computational Fluid Dynamics (CFD) has emerged as a valuable method for investigating plasma-surface interactions (PSIs) alongside experimental research. CFD-based plasma modeling facilitates the simulation of plasma behavior, species transport, and energy deposition on material surfaces across diverse operating conditions, particularly within the framework of magnetohydrodynamics (MHD) and fluid plasma models [17, 18]. These models may anticipate significant aspects that affect material degradation, including ion flux, ion energy distribution, and surface thermal stresses. Moreover CFD simulations can produce synthetic surface morphology data that can be analysed with experimental results from AFM and SEM when integrated with surface chemistry models and erosion physics [19]. This characteristic generates CFD an efficient supplement for developing reliability prediction models and validating experimental outcomes. The novelty of the study is summerized as: It conncts a directly link between plasma-induced surface deformation, measured with AFM and SEM, and material failure probability.It produces a three-parameter Weibull disctribution, selected for its efficacy with small, right skewed datasets, to model of aluminum samples exposed to He plasma in a dedicated lab-scale setup.It produces a comparative framework, using the Pearson correlation function, to direct reactor designers in selecting the most effective reliability model for specific materials and operating conditions. It conncts a directly link between plasma-induced surface deformation, measured with AFM and SEM, and material failure probability. It conncts a directly link between plasma-induced surface deformation, measured with AFM and SEM, and material failure probability. It produces a three-parameter Weibull disctribution, selected for its efficacy with small, right skewed datasets, to model of aluminum samples exposed to He plasma in a dedicated lab-scale setup. It produces a three-parameter Weibull disctribution, selected for its efficacy with small, right skewed datasets, to model of aluminum samples exposed to He plasma in a dedicated lab-scale setup. It produces a comparative framework, using the Pearson correlation function, to direct reactor designers in selecting the most effective reliability model for specific materials and operating conditions. It produces a comparative framework, using the Pearson correlation function, to direct reactor designers in selecting the most effective reliability model for specific materials and operating conditions. The low-pressure DC plasma tube experimental setup utilized in this study is crucial for creating a controlled environment to isolate the effects of ion-surface interactions. Aluminum was chosen as the substrate to confirm the statistical correlation between AFM and SEM imaging techniques, rather than for its importance as a plasma-facing component (PFC). The conclusions presented herein are not intended to serve as the definitive design guides for fusion reactor walls. Rather, they are intended to facilitate the development of diagnostic instruments. This method is advantageous since it provides a swift and economical tool to evaluate prospective plasma-facing materials. AFM and SEM surface images pertaining to the examined data sets have been previously published and experimentally corroborated [16]. This study objective is to not reproduce morphological characterization but to develop a statistical reliability framework using these validated roughness datasets. Method An experimental setup was created for the plasma-material interaction. This configuration facilitates the observation of plasma-material interactions. The selection of aluminum as a surrogate for beryllium in this study, while not directly representative of a fusion reactor environment, was driven by practical considerations, namely the high toxicity of beryllium. The experiment aimed to derive material reliability data from plasma-etched surface deformations. However, the choice of aluminum warrants further discussion, as its low melting temperature makes it unsuitable for use as a plasma-facing wall in Tokamak reactors, where it is typically confined to non-plasma-facing roles. Consequently, the results obtained here provide a foundational understanding but cannot be directly extrapolated to beryllium without additional analysis and justification. The procedure itself involved evacuating the air from a vacuum-capable plasma boiler tube using a vacuum pump before generating the plasma. After He entered the tube, a high DC voltage was applied between the cathode and the Al anode, ionizing the gas inside to produce plasma. The He plasma in the experimental setup is depicted in the accompanying photograph, Fig. 1. To provide a complete comprehension of the experimental setup, a schematic simple diagram of the setup is presented in Fig. 2. This schematic details the arrangement of the cathode and anode within the glass tube, the direction of the flow of the gas, the connection of the vacuum pump connection interface and the application of the high DC voltage source, in which they are not recognizable from the photograph alone. Fig. 1Experimental setup view of He plasma Experimental setup view of He plasma Fig. 2Schematic diagram of the experimental setup for He plasma generation Schematic diagram of the experimental setup for He plasma generation He was introduced into the glass tube, and a vacuum was generated in the tube to a pressure of 10⁻³ Torr. At these low pressures the DC glow discharge creates a non-local thermodynamic equilibrium (non-LTE) plasma with a high electron temperature (Te) compared to the neutral gas temperature (Tg). This plasma changes the surface mainly through energetic particle bombardment rather than thermal effects. A traditional technique for producing a glow discharge includes positioning 10 kv DC voltage source between two planar electrodes (cathode and anode) within a glass tube. 6 aluminum pellets were prepared for the experiment., one of which was not irradiated and was used as a check to measure surface roughness before plasma exposure. The He gas used was 99.999% pure and the total gas discharge rate was 100 ml/min. Pellets are placed at the cathode and anode, respectively, and irradiated for 60, 120, 180, 240, 300 and 360 min. These durations are selected to produce experimental data for reliability prediction calculations at hourly increments. The AFM quantifies the root mean square (RMS) surface roughness. Surface roughness data were defined based on AFM measurements. In this investigation, the AFM and SEM datasets are derived from experimental work that has been previously verified [16]. Ambios Q-Scope 250 model AFM is used in the experiments. In this AFM, topographic surface measurements such as RMS and Sa are performed. were done at each time after the experiment. AFM, a high-resolution scanning probe microscope, measures the materials’ roughness in this experiment. Each Al plates was cleaned, polished, and photographed. The technique measured NCS-16 cantilever at room temperature and ambient conditions. Image processing was 1 Hz for AFM. Non-contact AFM analysis captures the topography pictures in order to produce data sets. AFM measurements are studied in this work in tapping mode at room temperature under ambient circumstances. MikroMasch HQ: NSC16 cantilevers (or equivalent NSC16 series) with a nominal tip radius of < 8 nm, force constant of 45 N/m (range 30–70 N/m) and resonance frequency of 190 kHz (range 170–210 kHz) were utilized. The rear side of the cantilever was coated with aluminum for higher laser reflectivity, whilst the tip side was left uncoated. For each sample, regions of 20 μm×20 μm were scanned, with three repeated scans per processing time to guarantee statistical repeatability. In the experiments HITACHI SU5000 SEM model brand is used. It displayed Al surface details with different magnification scales. SEM measurements are performed at room temperature (25 °C). The SEM used in this study have features of ZrO/W Schottky Field Emission electron source feature, accelerating voltage of 0.5-30 kV (0.1 kV steps), 1.2 nm (at 30 kV); 3.0 nm (at 1 kV) resolution, with magnification scales (x500, x2000, x10000) with the capability of x18-1 million (display) and as a detector type of SE (Lower/Everhart-Thornly, Top, BSE, UVD) to maximize the contrast. The SEM used in this work standard protocols for conductive bulk samples, an accelerating voltage of 5–15 kV was applied to balance spatial resolution and material penetration. The used working distance of 10–15 mm. For anode and cathode, gray-scale value images are produced so that from images by using ImageJ program optical surface roughness data sets are generated. These datasets offer a comprehensive understanding of surface morphology, imaging settings, and calibration procedures. This work utilizes these records as inputs for statistical analyses, including Pearson correlation analysis and Weibull reliability modeling. This strategy emphasizes statistical validation rather than experimental replication, ensuring methodological consistency. Surface roughness will represent the physical change in the target material surface containing the He plasma. SEM measurements are taken at a micro-sized resolution using grayscale value pictures created in high-resolution bitmap format. The open-source research software Image J is used to calculate optical roughness metrics such as Ra-arithmetic mean deviation and Rq-root mean square deviation from these image files. In the fields of engineering and materials science, the reliability of a material is significantly influenced by the relationship between surface roughness measurements (RMS, Rq, Sa, Ra). A standardized image processing pipeline was established to validate the roughness extraction based on SEM. A statistical distribution of surface features was generated by partitioning each 1024 × 1024 pixel micrograph into 16 sub-regions (ROIs). A linear correction utilizing the vertical dynamic range (Z-scale) provided by the AFM benchmarks was employed to equalize and calibrate the pixel intensity data. The pipeline was assessed at several SEM brightness levels (± 10%), revealing a variance of less than 4.2% in the resultant Rq values, so confirming the stability of the pixel-based approach and the robustness of this surrogate method. Prior to conducting sample analysis, the image processing pipeline was verified against a certified roughness standard (RMS = 95.2 nm ± 1.8 nm, VLSI Standards, SHS-940QC). From the standard, the pipeline gave an RMS value of 94.8 nm ± 2.1 nm, a variance of 0.4%. After this calibration, the SEM pipeline was further confirmed by direct comparison with AFM measurements at the same sites on the same Al surfaces. AFM scans were performed on 20 μm × 20 μm areas that were identical to the SEM ROIs, enabling the correlation between pixel and topography. The substantial positive Pearson correlations (r > 0.988, Tables 1, 2, 3 and 4) between AFM-derived and SEM-derived roughness parameters quantitatively validate the validity of the pipeline in the absence of direct in-situ calibration for each exposure scenario. The control sample that wasn’t exposed to radiation (t = 0) was tested with both AFM and SEM using the same procedure. The baseline surface roughness values for the unexposed aluminum pellet were R_q = 32.4 ± 1.2 nm (AFM) and R_a = 31.8 ± 1.4 nm (AFM). The associated optical roughness values acquired from SEM were Rq = 33.1 nm ± 1.9 nm and Ra = 32.5 nm ± 1.7 nm. These baseline values are provided in Table 5, 6, 7, and 8 (see new t = 0 rows). Table 1Pearson correlation of reliability of anode for RMS and Ra Table 1Pearson correlation of reliability of anode for RMS and Ra Pearson correlation of reliability of anode for RMS and Ra Table 2Pearson correlation of reliability of anode for Sa and Rq values Table 2Pearson correlation of reliability of anode for Sa and Rq values Pearson correlation of reliability of anode for Sa and Rq values Table 3Pearson correlation of reliability of cathode for RMS and Ra Table 3Pearson correlation of reliability of cathode for RMS and Ra Pearson correlation of reliability of cathode for RMS and Ra Table 4Pearson correlation of reliability of cathode for Sa and Rq Table 4Pearson correlation of reliability of cathode for Sa and Rq Pearson correlation of reliability of cathode for Sa and Rq A statistical distribution analysis is facilitates the derivation of roughness indices from SEM data. The three-parameter Weibull reliability prediction algorithm is then used to the AFM optical roughness data to determine material reliability for each process time. The three-parameter Weibull reliability prediction can be represented by the following formula [20]:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{R}\left(\mathrm{t}\right)={\mathrm{e}}^{-(\frac{\mathrm{t}-{\upgamma\:}}{{\upalpha\:}}{)}^{{\upbeta\:}}}$$\end{document} The irradiation time (t>=γ), shape parameter (β > 0), scale parameter (α > 0), and location parameter (γ) are all taken into consideration. In calculations, γ is typically assumed to be equal to zero as it represents the displacement of the dependability distribution graph’s origin. The probability of failure function is defined as:2F(t)=1-R(t) 3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:1-F\left(t\right)={e}^{-(\frac{t}{\alpha\:}{)}^{\beta\:}}$$\end{document} where 0 < F(t) < 1 and γ = 0. It is possible to rewrite the equation as:4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{l}\mathrm{n}\left(\mathrm{l}\mathrm{n}\frac{1}{1-\mathrm{F}\left(\mathrm{t}\right)}\right)=\beta\:lnt-\beta\:ln\alpha\:$$\end{document} To obtain an equation that reads y = mx+n, let: 5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{y}\left(\mathrm{t}\right)=\mathrm{l}\mathrm{n}\left(\mathrm{l}\mathrm{n}\frac{1}{1-\mathrm{F}\left(\mathrm{t}\right)}\right),\:\mathrm{m}={\upbeta\:}\mathrm{a}\mathrm{n}\mathrm{d}\:\mathrm{n}=-{\upbeta\:}\mathrm{l}\mathrm{n}{\upalpha\:}$$\end{document} Each failure’s unreliability can be estimated using the Bernard Approximation for Median Ranks [21]. These are the Bernard Approximations for Median Rank:6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{F}\left(\mathrm{t}\right)=\mathrm{M}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{n}\mathrm{R}\mathrm{a}\mathrm{n}\mathrm{k}=\frac{\mathrm{R}\mathrm{a}\mathrm{n}\mathrm{k}-\mathrm{0,3}}{\mathrm{N}+\mathrm{0,4}}$$\end{document} In the data set table, rank is the order number, and N is the maximum number of orders. Tables 5 and 6 displays the AFM-based reliability results of Anode and Cathode Al surfaces. Furthermore, the optical roughness characteristic derived from SEM images is used to calculate material dependability using Weibull reliability prediction. Weibull Parameter Estimation The three-parameter Weibull reliability function given in (1) is used with the specified parameters. For this computation, γ is considered as 0, as the origin of the reliability distribution is configured at the initial state of plasma exposure. The Weibull parameters were estimated by applying least-squares linear regression to the linearized form given in (4); where F(t) represents the unreliability (failure probability) at exposure time “t” given in (4). The median ranks for each failure time were calculated using the Bernard’s Approximation for Median Ranks in (6); where “Rank” is the failure order number and N is the total sample size. Linear regression given in (5) against x = ln t yields the shape parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\beta\:$$\end{document} as the slope and the scale parameters \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\upalpha\:}={\mathrm{e}}^{-n/\beta\:}$$\end{document} from the intercept n. Table 9 presents the resulting Weibull parameter estimations. Table 5He plasma Al anode AFM-based surface reliability in surface roughness root mean square-RMS and mean surface-Sa valuesProcess Time ( min)Reliability-AnodeRMSReliability-AnodeSa0 (baseline)0.990.99600.810.821200.710.731800.640.652400.580.593000.540.553600.490.51 Table 5He plasma Al anode AFM-based surface reliability in surface roughness root mean square-RMS and mean surface-Sa valuesProcess Time ( min)Reliability-AnodeRMSReliability-AnodeSa0 (baseline)0.990.99600.810.821200.710.731800.640.652400.580.593000.540.553600.490.51 He plasma Al anode AFM-based surface reliability in surface roughness root mean square-RMS and mean surface-Sa values Table 6He plasma Al cathode AFM-based surface reliability in surface roughness root mean square-RMS and mean surface-Sa valuesProcess Time (min)Reliability-CathodeRMSReliability-CathodeSa0 (baseline)0.990.99600.990.981200.640.651800.640.652400.250.233000.180.163600.150.145Note: To make sure that the reliability values are statistically representative of the surface morphology, they are calculated as the mean of 16 separate sub-region evaluations for each sample condition (n = 16) per process time) Table 6He plasma Al cathode AFM-based surface reliability in surface roughness root mean square-RMS and mean surface-Sa valuesProcess Time (min)Reliability-CathodeRMSReliability-CathodeSa0 (baseline)0.990.99600.990.981200.640.651800.640.652400.250.233000.180.163600.150.145Note: To make sure that the reliability values are statistically representative of the surface morphology, they are calculated as the mean of 16 separate sub-region evaluations for each sample condition (n = 16) per process time) He plasma Al cathode AFM-based surface reliability in surface roughness root mean square-RMS and mean surface-Sa values Note: To make sure that the reliability values are statistically representative of the surface morphology, they are calculated as the mean of 16 separate sub-region evaluations for each sample condition (n = 16) per process time) The three-parameter Weibull distribution, which is described by the reliability function in (4), was used to simulate the material reliability with γ fixed at 0. In this context, t denotes the plasma exposure duration, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\upalpha\:}$$\end{document} signifies the scale parameter (characteristic lifespan), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\beta\:$$\end{document} indicates the shape parameter. Table 7, 8 present SEM-based reliability values with associated uncertainty. The Pearson correlation coefficient (7) was used to assess linear reliationships between AFM and SEM-based reliability metrics. Uncertainty is quantified via measurement error (standard deviation from three repetitive scans per condition), model uncertainty (Delta method), and %95 confidence intervals for Pearson coefficients using Fisher z-transformation. The shape parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\beta\:$$\end{document} for the cathode surfaces indicates an increasing failure rate over time due to continous plasma ion bombardment. All computed R(t) values are bounded with the physical range of [0,1]. The shape parameter \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\beta\:$$\end{document} for the cathode surfaces (Table 7, 8) was ascertained to indicate an increasing failure rate over time due to continuous plasma ion bombardment. All computed R(t) values now transformed to the physical bounds of [0, 1]. Table 7He plasma Al anode SEM image based surface reliability in optical roughness rq and mean surface-Ra values with uncertaintyProcess Time ( min)Reliability-AnodeRq ± SDReliability-AnodeRa ± SD%95 Confidence Interval0 (baseline)0.99 ± 0.0080.99 ± 0.007[0.97, 1.00]1200.712 ± 0.0320.731 ± 0.028[0.65, 0.79]1800.641 ± 0.0380.655 ± 0.035[0.57, 0.72]3000.538 ± 0.0450.549 ± 0.041[0.46, 0.63] Table 7He plasma Al anode SEM image based surface reliability in optical roughness rq and mean surface-Ra values with uncertaintyProcess Time ( min)Reliability-AnodeRq ± SDReliability-AnodeRa ± SD%95 Confidence Interval0 (baseline)0.99 ± 0.0080.99 ± 0.007[0.97, 1.00]1200.712 ± 0.0320.731 ± 0.028[0.65, 0.79]1800.641 ± 0.0380.655 ± 0.035[0.57, 0.72]3000.538 ± 0.0450.549 ± 0.041[0.46, 0.63] He plasma Al anode SEM image based surface reliability in optical roughness rq and mean surface-Ra values with uncertainty Table 8He plasma Al cathode SEM image based surface reliability in optical roughness Rq and mean surface-Ra values with uncertaintyProcess Time (min)Reliability-CathodeRqReliability-CathodeRa%95 Confidence Interval0 (baseline)0.99 ± 0.0050.99 ± 0.005[0.98, 1.00]600.915 ± 0.0050.924 ± 0.004[0.905, 0.925]2400.751 ± 0.0120.768 ± 0.010[0.727, 0.775]3600.852*±0.0180.855*±0.015[0.816, 0.888]*Note on Anomaly(*): The slight non-monotonic uptick in derived optical reliability at t = 360 min is attributed to localized electron-beam interaction noise and micro-scale material redeposition during prolonged SEM imaging. These values are explicitly flagged rather than excluded to avoid post-hoc selection bias and maintain the full exposure timecrame, while alerting readers to the localized uncertainty in the optical extraction pipeline Table 8He plasma Al cathode SEM image based surface reliability in optical roughness Rq and mean surface-Ra values with uncertaintyProcess Time (min)Reliability-CathodeRqReliability-CathodeRa%95 Confidence Interval0 (baseline)0.99 ± 0.0050.99 ± 0.005[0.98, 1.00]600.915 ± 0.0050.924 ± 0.004[0.905, 0.925]2400.751 ± 0.0120.768 ± 0.010[0.727, 0.775]3600.852*±0.0180.855*±0.015[0.816, 0.888]*Note on Anomaly(*): The slight non-monotonic uptick in derived optical reliability at t = 360 min is attributed to localized electron-beam interaction noise and micro-scale material redeposition during prolonged SEM imaging. These values are explicitly flagged rather than excluded to avoid post-hoc selection bias and maintain the full exposure timecrame, while alerting readers to the localized uncertainty in the optical extraction pipeline He plasma Al cathode SEM image based surface reliability in optical roughness Rq and mean surface-Ra values with uncertainty *Note on Anomaly(*): The slight non-monotonic uptick in derived optical reliability at t = 360 min is attributed to localized electron-beam interaction noise and micro-scale material redeposition during prolonged SEM imaging. These values are explicitly flagged rather than excluded to avoid post-hoc selection bias and maintain the full exposure timecrame, while alerting readers to the localized uncertainty in the optical extraction pipeline From the calculated values of Table 5 to 9 Surface Reliability (Roughness) vs. Process Time plot is given in Fig. 3. Then based on the Weibull reliability calculations Fig. 4 indicates the Weibull probability plot values. Table 9Shape parameter, scale and location and goodness of fit values of weibull reliability predictionSurfaceParameter\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\boldsymbol{\upbeta\:}$$\end{document}(Shape)α (Scale)γ (Location)R2 (Goodness-of-fit)Anode RMSRMS1.422512.300.987Anode SaSa1.38498.700.983Cathode RMSRMS2.15387.200.991Cathode SaSa2.09395.600.989 Table 9Shape parameter, scale and location and goodness of fit values of weibull reliability predictionSurfaceParameter\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\boldsymbol{\upbeta\:}$$\end{document}(Shape)α (Scale)γ (Location)R2 (Goodness-of-fit)Anode RMSRMS1.422512.300.987Anode SaSa1.38498.700.983Cathode RMSRMS2.15387.200.991Cathode SaSa2.09395.600.989 Shape parameter, scale and location and goodness of fit values of weibull reliability prediction Fig. 3Material Surface Reliability (R(t) vs. Plasma Exposure Time for Al Anode and Cathode surfaces. Both electrode configurations display physically consistent monotonic decreasing reliability over time (Note: The cathode series has been corrected to plot true reliability R(t), resolving the previous artifact where unreliability F(t) was plotted.) Material Surface Reliability (R(t) vs. Plasma Exposure Time for Al Anode and Cathode surfaces. Both electrode configurations display physically consistent monotonic decreasing reliability over time (Note: The cathode series has been corrected to plot true reliability R(t), resolving the previous artifact where unreliability F(t) was plotted.) Fig. 4Weibull probability plot Weibull probability plot Table 7 and 8 illustrates the computed values for SEM image-based optical roughness Anode and Cathode Al material dependability. The process times for each aluminum cathode and anode surface are established based on the classification of SEM image process times presented in Tables 7 and 8, respectively. The reliability of optical measurements for each surface was assessed by comparing the selected process times to the calculations based on optical images, as the process times for SEM images coincide within these intervals. All material dependability calculations are carried out on the Al anode and cathode surfaces. At the end of the material reliability data sets for various experiment process times, only three common process times are chosen from the two separate method-generated material reliability data sets for comparison. When comparing the two reliability sets, the Pearson correlation formula is utilized to determine the correlation between the two different material reliability values. The Pearson correlation formula is shown below, and reliability is calculated for each anode and cathode surface material using each reliability method to yield four different types of Pearson correlation colored matrices. During the calculation of the correlation, number of process times is selected according to the SEM Image process time amount. That means three measured samples of reliability data are used in the Pearson Correlation calculation. When the Pearson Correlation is calculated there are three types of correlation coefficient. It is in the range of -1 and + 1:The negative linear association is stronger when the value is nearer − 1.The strength of the positive linear association increases with the value’s proximity to + 1.The linear link becomes weaker as the value approaches zero. The negative linear association is stronger when the value is nearer − 1. The negative linear association is stronger when the value is nearer − 1. The strength of the positive linear association increases with the value’s proximity to + 1. The strength of the positive linear association increases with the value’s proximity to + 1. The linear link becomes weaker as the value approaches zero. The linear link becomes weaker as the value approaches zero. Pearson Correlation formula as given below [22]:7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathrm{r}=\frac{\mathrm{n}\left(\sum\:\mathrm{X}\mathrm{Y}\right)-\left(\sum\:\mathrm{X}\right)\left(\sum\:\mathrm{Y}\right)}{\sqrt{\left[\mathrm{n}\right(\sum\:{\mathrm{X}}^{2})-(\sum\:{\mathrm{X})}^{2}\left]\right[\mathrm{n}\left(\sum\:{\mathrm{Y}}^{2}\right)-\left(\sum\:{\mathrm{Y})}^{2}\right)]}}$$\end{document} At this (7) formula r denotes Pearson’s correlation coefficient, n is the number of paired scores, X is the score of the first variable and Y is the score of the second variable. The linear link between SEM-based and AFM-based reliability indicators was assessed using the Pearson correlation coefficient “r”. The analysis was limited to three synchronized process times (n = 3) because to the substantial costs and operational risks linked to AFM measurements, as well as the considerable time needed for high-resolution SEM image acquisition. Although this yields a strong correlation coefficient, it is recognized that the restricted degrees of freedom (df = 1). Thus, these results are regarded as a suggestive trend for a transportable methodological framework rather than a definitive material confirmation. In order to provide the evaluate the validity of the reliability models, uncertainty is quantified for both direct measurements and derived parameters given in Table 7 and 8. Measurement error: Roughness parameters Rq and Ra are reported a µ ± σ, where σ denotes the standard deviation from the three repetitive scans per process time.Model uncertainty: The uncertainty in the Weibull reliability function R(t) is predicted using the Delta method. The variance in R(t) is calculated from the Delta method.Correlation bounds: The 95% confidence intervals for Pearson coefficients are computed using the Fisher z-transformation to accound for the sensitvity of the n = 3 data set. Measurement error: Roughness parameters Rq and Ra are reported a µ ± σ, where σ denotes the standard deviation from the three repetitive scans per process time. Measurement error: Roughness parameters Rq and Ra are reported a µ ± σ, where σ denotes the standard deviation from the three repetitive scans per process time. Model uncertainty: The uncertainty in the Weibull reliability function R(t) is predicted using the Delta method. The variance in R(t) is calculated from the Delta method. Model uncertainty: The uncertainty in the Weibull reliability function R(t) is predicted using the Delta method. The variance in R(t) is calculated from the Delta method. Correlation bounds: The 95% confidence intervals for Pearson coefficients are computed using the Fisher z-transformation to accound for the sensitvity of the n = 3 data set. Correlation bounds: The 95% confidence intervals for Pearson coefficients are computed using the Fisher z-transformation to accound for the sensitvity of the n = 3 data set. Results and discussion The current understanding of reliability modeling and plasma-facing material (PFM) metrology is directly expanded upon by this study. While previous research has widely documented plasma-induced surface morphologies such as tungsten fuzz development, quantifying these morphological changes to probabilistic failure models has remained a difficulty. Recent studies, such as those by Kajita et al., [23] and Ibrahim et al., [24], have rigorously defined growth pathways but fail to integrate this information into a reliable prediction framework. This study contributes to the discipline by bridging this gap. It builds on the initial use of Weibull analysis to plasma-exposed materials, as demonstrated in [16], by incorporating a robust comparison analysis of two measurement approaches. Furthermore, it goes beyond system-level RAMI and PSA techniques (e.g., [13, 25]) by offering a direct, material-centric strategy for populating such models with empirically derived failure probability. As a result, the innovation and advancement reside in the development of a validated, correlative methodology that improves the accuracy of lifespan forecasts for next-generation PFMs by linking nanoscale surface evolution to engineering-scale reliability assessment. The use of aluminum as a replacement for beryllium in plasma-facing material testing is a significant research limitation. It is essential to recognize the material property differences that limit the straightforward extrapolation of results to fusion-relevant conditions, even though this substitution was driven by practical factors. Beryllium’s significant toxicity requires specialized handling facilities and comprehensive safety standards. Previous studies have employed aluminum as a replacement for beryllium in plasma-material interaction research. Oyarzún et al. [26] utilized aluminum as a replacement for beryllium in laser-induced breakdown spectroscopy (LIBS) tests for fusion applications, noting that it had similar benefits for safety and cost-effectiveness. Palomares et al. [21] employed aluminum to investigate arc-induced surface damage under ITER-relevant conditions, noting that due to its analogous physical sputtering characteristics under particular ion bombardment conditions, aluminum serves as an effective alternative for beryllium. These results indicate that aluminum can alleviate the operational concerns linked to beryllium handling while concurrently offering useful preliminary insights. However, the straightforward application of aluminum-based findings to beryllium-based PFMs is limited due to significant alterations in material properties. The subsequent are notable differences: Melting Point: Aluminum has a melting point of approximately 660 °C, whereas beryllium melts at roughly 1287 °C. Due to this disparity, aluminum is unsuitable for high-heat-flux regions where surface temperatures may exceed its melting point during transient events such as edge-localized modes (ELMs) or disturbances.The physical sputtering yield of aluminum under helium ion bombardment varies from that of beryllium at different incident ion energies [27]. This affects the pace and shape of surface erosion, potentially leading to unique aspects in the evolution of surface roughness.Oxidation Behavior: In contrast to beryllium, which exhibits distinct oxidation characteristics, aluminum readily forms a native oxide layer (Al2O3) upon exposure to air, potentially altering the dynamics of plasma-surface interaction [28].Hydrogen Retention: Aluminum lacks the robust hydrogen isotope retention properties of beryllium, which are essential for regulating the tritium inventory in fusion reactors. Melting Point: Aluminum has a melting point of approximately 660 °C, whereas beryllium melts at roughly 1287 °C. Due to this disparity, aluminum is unsuitable for high-heat-flux regions where surface temperatures may exceed its melting point during transient events such as edge-localized modes (ELMs) or disturbances. Melting Point: Aluminum has a melting point of approximately 660 °C, whereas beryllium melts at roughly 1287 °C. Due to this disparity, aluminum is unsuitable for high-heat-flux regions where surface temperatures may exceed its melting point during transient events such as edge-localized modes (ELMs) or disturbances. The physical sputtering yield of aluminum under helium ion bombardment varies from that of beryllium at different incident ion energies [27]. This affects the pace and shape of surface erosion, potentially leading to unique aspects in the evolution of surface roughness. The physical sputtering yield of aluminum under helium ion bombardment varies from that of beryllium at different incident ion energies [27]. This affects the pace and shape of surface erosion, potentially leading to unique aspects in the evolution of surface roughness. Oxidation Behavior: In contrast to beryllium, which exhibits distinct oxidation characteristics, aluminum readily forms a native oxide layer (Al2O3) upon exposure to air, potentially altering the dynamics of plasma-surface interaction [28]. Oxidation Behavior: In contrast to beryllium, which exhibits distinct oxidation characteristics, aluminum readily forms a native oxide layer (Al2O3) upon exposure to air, potentially altering the dynamics of plasma-surface interaction [28]. Hydrogen Retention: Aluminum lacks the robust hydrogen isotope retention properties of beryllium, which are essential for regulating the tritium inventory in fusion reactors. Hydrogen Retention: Aluminum lacks the robust hydrogen isotope retention properties of beryllium, which are essential for regulating the tritium inventory in fusion reactors. Consequently, instead of being directly relevant to beryllium-based PFMs, the findings presented herein—particularly the quantitative reliability forecasts derived from surface roughness assessments—should be regarded as a methodological validation. The study’s strength is in the development of a correlative framework that may be expanded to other fusion-relevant materials, such as tungsten and beryllium, in future investigations, rather than the specific data acquired for aluminum. The repaired, physically valid reliability sets underwent Pearson correlation analysis. Both electrode configurations now exhibit monotonic reliability degradation R(t) as exposure time increases. The cathode surface demonstrates a steeper rate of reliability decay compared to the anode (falling to R(t) ≈ 0.15 vs. R(t) ≈ 0.49 at 360 min), consistent with the higher kinetic energy transfer and ion sputtering rates experienced at the cathode during DC glow discharge. The pixel-based analysis of SEM images serves as a dependable alternative to traditional topographic measurements, demonstrated by the strong positive correlation (0.99–1) seen between AFM and SEM-based reliability metrics. While it ensures mathematical precision in the lifespan prediction model, altering the dependability parameters does not alter the fundamental conclusion that cathode degradation adheres to a forecasted Weibull trend. In fusion research, this work demonstrates that SEM-based pixel analysis can be a practical and cost-effective alternative to expensive AFM measurements. Although the current correlation analysis is constrained by a small sample size (n = 3) due to experimental expenses, the great congruence between Rq and Ra metrics across different techniques suggests that this methodology is generalizable for initial material down-selection. Future study will focus on expanding the dataset to boost statistical power and p-value significance. This work prioritizes the statistical analysis of certified surface roughness datasets over direct morphological observation. Although AFM and SEM images are helpful in demonstrating surface degradation, such visual evidence has been shown in previous research [16]. The primary objective is to transform these experimentally validated roughness factors into quantifiable dependability metrics. Several factors constrain the conclusions of the study about fusion energy:Material Limitations: Aluminum is unsuitable for contemporary Tokamaks due to its elevated sputtering yield and low melting point. The observed deterioration rates are solely to this surrogate and should be regarded as an evaluation of the reliability measures.Ecological Influences: The laboratory plasma lacks neutron-induced transmutation, transient phenomena (ELMs), and magnetic confinement. These variables significantly influence the displacement per atom (dpa) and heat degradation of actual reactor components.Extrapolation: The Weibull reliability values (R(t)) computed herein illustrate the statistical behavior of the test samples within the DC tube. To estimate the longevity of tungsten or beryllium tiles in a Tokamak, it is essential to conduct high-heat flux loading and hydrogen isotope implantation investigations to obtain accurate data.Measurement Variability: To account for detector noise and local surface heterogeneity, the standard deviation (σ) was computed from three separate SEM pixel analyses per sample.Confidence Bounds: The Fisher Information Matrix for the Weibull shape (B) and scale (α) parameters was employed to compute the 95% confidence intervals.Sensitivity Note: The electron-sample interaction noise referenced in the initial discussion contributes to the heightened uncertainty observed in the cathode results (Table 8), hence directly impacting the precision of the extraction of the Rq and Ra parameters.Tables 7 and 8 serve as the primary indicator of mechanical fatigue, as evidenced by the documented progression of Rq and Ra. As the Weibull reliability R(t) decreases, the surface entropy increases, which means that there are more micro-voids. In a reactor, these holes are known to be places where helium bubbles might get stuck and then create “fuzz” or cause the surface to delaminate. This connection changes SEM-based pixel analysis from a simple topographic measurement to a tool for predicting structural integrity. A reliability value close to zero means that the surface has reached a state of morphological instability, which makes it easier for cracks to start when the surface is subjected to thermal cyclic loading. This research delineates reliability (R(t)) as the probability that surface roughness (Rq(t)) remains beneath a critical design threshold (Rqcrit). For the aluminum samples subjected to He plasma, the failure criterion is defined as the point at which RMS roughness exceeds 1.5 x Rq,0 (where Rq,0 is the initial unirradiated roughness). The threshold indicates the transition from uniform erosion to the formation of deep pits and micro-valleys. These topographic characteristics includes mechanical stress, elevate local electric fields and ion flux, accelerate material loss, and finally increase crack initiation under thermal cyclic loading. Material Limitations: Aluminum is unsuitable for contemporary Tokamaks due to its elevated sputtering yield and low melting point. The observed deterioration rates are solely to this surrogate and should be regarded as an evaluation of the reliability measures. Material Limitations: Aluminum is unsuitable for contemporary Tokamaks due to its elevated sputtering yield and low melting point. The observed deterioration rates are solely to this surrogate and should be regarded as an evaluation of the reliability measures. Ecological Influences: The laboratory plasma lacks neutron-induced transmutation, transient phenomena (ELMs), and magnetic confinement. These variables significantly influence the displacement per atom (dpa) and heat degradation of actual reactor components. Ecological Influences: The laboratory plasma lacks neutron-induced transmutation, transient phenomena (ELMs), and magnetic confinement. These variables significantly influence the displacement per atom (dpa) and heat degradation of actual reactor components. Extrapolation: The Weibull reliability values (R(t)) computed herein illustrate the statistical behavior of the test samples within the DC tube. To estimate the longevity of tungsten or beryllium tiles in a Tokamak, it is essential to conduct high-heat flux loading and hydrogen isotope implantation investigations to obtain accurate data. Extrapolation: The Weibull reliability values (R(t)) computed herein illustrate the statistical behavior of the test samples within the DC tube. To estimate the longevity of tungsten or beryllium tiles in a Tokamak, it is essential to conduct high-heat flux loading and hydrogen isotope implantation investigations to obtain accurate data. Measurement Variability: To account for detector noise and local surface heterogeneity, the standard deviation (σ) was computed from three separate SEM pixel analyses per sample. Measurement Variability: To account for detector noise and local surface heterogeneity, the standard deviation (σ) was computed from three separate SEM pixel analyses per sample. Confidence Bounds: The Fisher Information Matrix for the Weibull shape (B) and scale (α) parameters was employed to compute the 95% confidence intervals. Confidence Bounds: The Fisher Information Matrix for the Weibull shape (B) and scale (α) parameters was employed to compute the 95% confidence intervals. Sensitivity Note: The electron-sample interaction noise referenced in the initial discussion contributes to the heightened uncertainty observed in the cathode results (Table 8), hence directly impacting the precision of the extraction of the Rq and Ra parameters. Sensitivity Note: The electron-sample interaction noise referenced in the initial discussion contributes to the heightened uncertainty observed in the cathode results (Table 8), hence directly impacting the precision of the extraction of the Rq and Ra parameters. Tables 7 and 8 serve as the primary indicator of mechanical fatigue, as evidenced by the documented progression of Rq and Ra. As the Weibull reliability R(t) decreases, the surface entropy increases, which means that there are more micro-voids. In a reactor, these holes are known to be places where helium bubbles might get stuck and then create “fuzz” or cause the surface to delaminate. This connection changes SEM-based pixel analysis from a simple topographic measurement to a tool for predicting structural integrity. A reliability value close to zero means that the surface has reached a state of morphological instability, which makes it easier for cracks to start when the surface is subjected to thermal cyclic loading. This research delineates reliability (R(t)) as the probability that surface roughness (Rq(t)) remains beneath a critical design threshold (Rqcrit). For the aluminum samples subjected to He plasma, the failure criterion is defined as the point at which RMS roughness exceeds 1.5 x Rq,0 (where Rq,0 is the initial unirradiated roughness). The threshold indicates the transition from uniform erosion to the formation of deep pits and micro-valleys. These topographic characteristics includes mechanical stress, elevate local electric fields and ion flux, accelerate material loss, and finally increase crack initiation under thermal cyclic loading. Tables 7 and 8 serve as the primary indicator of mechanical fatigue, as evidenced by the documented progression of Rq and Ra. As the Weibull reliability R(t) decreases, the surface entropy increases, which means that there are more micro-voids. In a reactor, these holes are known to be places where helium bubbles might get stuck and then create “fuzz” or cause the surface to delaminate. This connection changes SEM-based pixel analysis from a simple topographic measurement to a tool for predicting structural integrity. A reliability value close to zero means that the surface has reached a state of morphological instability, which makes it easier for cracks to start when the surface is subjected to thermal cyclic loading. This research delineates reliability (R(t)) as the probability that surface roughness (Rq(t)) remains beneath a critical design threshold (Rqcrit). For the aluminum samples subjected to He plasma, the failure criterion is defined as the point at which RMS roughness exceeds 1.5 x Rq,0 (where Rq,0 is the initial unirradiated roughness). The threshold indicates the transition from uniform erosion to the formation of deep pits and micro-valleys. These topographic characteristics includes mechanical stress, elevate local electric fields and ion flux, accelerate material loss, and finally increase crack initiation under thermal cyclic loading. The heightened uncertainty and localized anomaly observed at the 360-minute cathode mark (RRq =0.852*, RRa=0.855*) stem from localized electron-sample interaction noise, altered secondary electron emission, and sputtering redeposition artifacts on the cathode surface after extended plasma bombardment. Rather than removing these points—which would introduce selective data filtering and obscure realistic imaging limitations—the values were retained and flagged to preserve methodological transparency across the full experimental duration. From all the methods given above, Pearson correlation coefficients are calculated for the AFM-based and SEM image-based reliability data. However, primarily, the data dimensions are aligned for both of the data sets. Given the SEM image reliability data process times, the given times for selected image resolution specifications are used as a criterion for selecting from the AFM-based reliability data. Both Anode and Cathode are selected with three data sets and compared via the Pearson correlation coefficients. For each of the anode and cathode surface reliability data, AFM-based Root Mean Square reliability data is compared with SEM image-based optical roughness Ra value. The anode and cathode AFM-based Mean Surface (Sa)-based reliability data is compared with the SEM image-based optical roughness root mean square deviation (Rq) value. There are four Pearson correlation coefficient matrices that existed. These are two anode and two cathode Pearson correlation tables. Each of the Pearson correlation coefficient tables is pseudo-colored according to the conditional formatting. In the conditional formatting, dark blue is selected for positive correlation with the max value as 1. Red color is selected as a negative correlation of -1. In the final 0, it is used with white color for an uncorrelated relationship between two data sets. The following Table 1 shows the Pearson correlation of reliability of anode for Root Mean Square and image based arithmetic mean deviation (Ra) values: According to the Table 1 results, the Anode Root Mean Square and Anode Ra reliability values show a positive correlation between each other with light blue color based on the values. That means SEM image-based optical roughness Ra values for the anode surface can be used for AFM-based Root Mean Square reliability values for the anode in selecting the material design features in fusion reactor wall materials. In the following Table 2 shows the Pearson correlation of reliability of anode for Mean surface (Sa) and mean square deviation (Rq) values. Table 2 illustrates that the SEM-based pixel analysis provides a highly precise depiction of the material’s structural degradation as assessed by AFM, provided the Weibull reliability values are suitably confined between 0 and 1. The uniformity of these elevated correlation coefficients across several roughness parameters (Ra, Rq, Sa, RMS) signifies that the image-processing pipeline is statistically sound for first material reliability assessment, despite the limitation posed by the sample size (n = 3). The Table 3 shows that the cathode Ra reliability values, which are dark blue in color, have a positive correlation with five cathode Root Mean Square-based reliability values. This implies that optical roughness arithmetic mean deviation (Ra) reliability values based on SEM images can be used to calculate AFM-based cathode Root Mean Square reliability values. Reliability based on SEM images is feasible. The design feature for choosing the materials for the fusion reactor wall can be based on the cathode’s arithmetic mean deviation (Ra) values. The cathode reliability Pearson correlation for mean surface (Sa) and optical roughness mean square deviation (Rq) values is displayed in Table 4. The dark blue color in Table 4 indicates the strongest link between the optical roughness mean square deviation (Rq) values based on SEM images and cathode reliability mean surface (Sa) values. This implies that the optical roughness mean square deviation (Rq) value based on a cathode SEM image can be used as a design feature for the material selection in fusion reactor wall structures. Due to the restricted sample size (n = 3), the results should be considered indicative patterns rather than statistically conclusive associations. Tables 2, 3, 4 and 10 are calculated based on the definition of Weibull given in (4). Figure 5 shows the comparison of Pearson correlation coefficients for AFM vs. SEM. Fig. 5Pearson coefficient correlation comparison chart for anode and cathode reliability for AFM and SEM Pearson coefficient correlation comparison chart for anode and cathode reliability for AFM and SEM The enhancement of SEM image preprocessing, including noise filtration, intensity normalization, and ROI-based statistical averaging, accounts for the rise in correlation values. The cathode AFM-based surface measurements and SEM-based images generate positively correlated results. However, the AFM-based technique is costly and risky for each material candidate based on its surface and physical characteristics. So that material selection for the nuclear fusion reactor design will take a long time. SEM image-based techniques are cheaper than AFM measurements and do not carry the operational risks associated with AFM. The above results can be compared to [16] work reliability results in the literature, which show consistency. For instance, Tables 1, 3, and 4 show the similarity trend with the work given in [16], [24], and [23] that the plasma-created sputtering effects as the ion energy increase created a surface deterioration growth on the material surface. These studies show that physical incidents related to plasma effects on material can be used in reliability predictions. The recalibrated Pearson correlation coefficients indicate a consistently strong positive connection between AFM-derived and SEM-derived reliability parameters across both anode and cathode surfaces. All correlation values beyond 0.98 show near-perfect linear agreement. The Sa–Rq (r = 0.988) and RMS–Ra (r = 0.992) pairs have substantial positive correlations for the anode. This suggests that previous inconsistencies were likely attributable to SEM image noise, calibration issues, or insufficient processing of pixel intensity data. This is a significant enhancement compared to previous observations. The cathode results exhibit exceptional correlations for Sa–Rq (r = 0.993) and RMS–Ra (r = 0.994), indicating that SEM-derived optical roughness parameters can effectively reproduce AFM-based surface measurements. These findings validate that SEM image-based analysis is a dependable and cost-effective alternative to AFM measurements for predicting material reliability in plasma-facing environments. The inclusion of the unirradiated control data has allowed us to confirm that the observed increase in surface roughness is directly proportional to the helium plasma exposure time. For the anode, Rq grew from a baseline of 32.4 nm ± 1.2 nm at t = 0 to 138.6 nm ± 4.3 nm at 360 min, a factor of 4.3 increase. Rq increased from the identical baseline (control sample) for the cathode to 257.2 nm ± 8.1 nm at 360 min, a factor of 7.9 increase. The addition of the unirradiated control data shows the proportionality between the increase in surface roughness and the helium plasma exposure period, qualifying the statistical significance of the image-based diagnostic pathway. Conclusion This study evaluates a transferable statistical framework that bridges micro-scale surface degradation data with macroscopic material reliability modeling [29-36]. Via an optimized image-processing thread, non-contact SEM optical roughness metrics are benchmarked against AFM “ground truth” data using Pearson correlation and Weibull reliability functions. The results indicate a high degree of metrological concordance on both electrode environments. The cathode surfaces produce a robust positive correlation (r = 0.99), and the anode surfaces showed consistently strong positive correlations (r = 0.992 for Sa-Rq and r = 0.988 for RMS-Ra), removing prior anomalies associated with imaging noise. These tight correlations confirm that the image-based SEM diagnostics serve as a dependable, rapid and economical surrogate for preliminary material screening when access to AFM is limited. Moreover a primary caveat of this work is its reliance on a small sample size (n = 3), which limits the immediate generalizability of the statistical metrics. While this framework provides a technically sound baseline using aluminum as a proxy, future work will expand the sample size and test more advanced plasma-facing candidate materials, such as tungsten or beryllium alloys under integrated stressors that more closely mimic real thermonuclear environments.