Electrical Impedance Spectroscopy: A Complementary Approach Differentiating PID Mechanisms in Photovoltaics

Abstract

Potential-induced degradation (PID) presents a critical reliability issue for solar photovoltaic (PV) modules, with three primary types identified in the literature, namely, PID_s (shunting type), PID_p (polarization type), and PID_c (corrosion type). Electrochemical/electrical impedance spectroscopy (EIS) is a highly effective but underutilized technique for differentiating between these PID mechanisms. When used alongside conventional I–V measurements (e.g., I_sc, V_oc, and FF), EIS offers direct insights into parameters such as R_s, R_p, and C_p, making it a valuable tool for PID type differentiation. In this study, two four-cell glass–glass modules were investigated using p-base PERC monofacial cells with EVA and POE encapsulants. Results indicate that V_oc and FF remained nearly unchanged under +1000 V stress for both EVA and POE modules, suggesting a minimal impact of PID stress on these parameters. However, I_sc was reduced by approximately 8.5% in the EVA module and 10% in the POE module. For the POE module, surface recombination (PID_p) is likely responsible for the I_sc loss, as R_s, R_p, and C_p showed no significant variation. Conversely, in the EVA module, the combined effects of surface recombination and junction recombination (PID_jr) are the probable cause of the I_sc loss, as evidenced by remarkable changes in R_p and C_p. The observed decrease in R_p for the EVA module is attributed to reduced dynamic diode resistance rather than ohmic shunt resistance. This reduction is linked to recombination currents induced by junction trap centers, formed by the positive voltage PID stress in the encapsulant, which contains trace amounts of oxidizable species such as CH₃COOH and/or H₂O. The objective of this study is to evaluate the impact of PID stress on the electrical characteristics of glass–glass PV modules with different encapsulants, utilizing a combined EIS and I–V approach to distinguish between PID mechanisms. The findings highlight the critical role of the encapsulant type in determining PID susceptibility, with the EVA module exhibiting significant degradation linked to junction recombination losses. These insights underscore the necessity of optimizing encapsulant materials to enhance PV module durability and reliability in real-world applications.

1. Introduction

The reliability and durability of photovoltaic (PV) modules have become increasingly critical topics as the solar energy industry seeks to maximize the performance and longevity of renewable energy systems. These lifetime dictating qualities are influenced by various factors, including the materials used in construction, production processes, and exposure to environmental stressors. One significant and persistent challenge affecting PV modules in the field is potential-induced degradation (PID). As highlighted in recent review articles [1,2], PID is closely tied to construction materials—such as solar cell types, anti-reflection coating properties, encapsulant characteristics, and the properties of superstrates and substrates—along with system voltage polarity (e.g., positively or negatively grounded or floating) and environmental conditions, such as humidity and rainfall. Addressing these factors is crucial for advancing the reliability and performance of PV modules in diverse field settings. This paper investigates and presents the impact of two predominant encapsulant types, polyolefin elastomer (POE) and ethylene vinyl acetate (EVA), on PID in glass–glass PV modules utilizing p-base PERC cells.

Recent studies have shown that PID significantly impacts module performance through charge accumulation effects, leading to reduced power output and efficiency losses over time [3,4]. The encapsulant plays a crucial role in mitigating or exacerbating PID effects by influencing the electrical insulation properties and moisture ingress resistance of the module [5]. Depending on the encapsulant type and other construction details including the cell structure and materials, glass–glass module architectures may offer enhanced protection against PID compared to conventional glass–backsheet designs [6]. However, the choice of encapsulant remains one of the decisive factors, as it determines the level of ion migration and leakage currents within the module structure [7]. This study aims to bridge the gap by providing an in-depth comparative analysis of EVA and POE encapsulants in mitigating PID, using the advanced electrochemical characterization technique of EIS. The novelty of this study lies in integrating EIS measurements with I–V measurements to obtain complementary and causal data (R_s, R_p, C_p) from EIS. This complementary approach helps explain the degradation effects on key I–V parameters such as I_sc, V_oc, and FF, while also enabling the deconvolution of various PID mechanisms, including PID_s, PID_p, PID_c, and PID_jr.

Three types of PID have been reported in the literature, namely, PID_s (shunting type), PID_p (polarization type), and PID_c (corrosion type). While the first two types have been extensively studied by multiple research groups, the third type, though possible, has been rarely reported [8,9].

Na-penetration PID (shunting type PID, PID_s) is a distinct degradation mechanism where sodium ions diffuse into the p–n junction, causing shunting through stacking faults. Unlike PID_s, PID_p (polarization) affects charge distribution at the cell surface and PID_c (corrosion) leads to material corrosion/delamination. This paper focuses primarily on the PID_p type and a new PID type, called PID_jr, for junction recombination loss. PID_jr involves increased surface recombination and junction recombination. All these PID mechanisms are discussed in detail in Section 2 of this paper.

Electrochemical/electrical impedance spectroscopy (EIS) is a highly versatile and insightful analytical technique with immense potential yet to be fully realized in the field of silicon photovoltaics [10]. Differentiating PID in solar cells—specifically PID_s, PID_p, PID_c, and PID_jr—requires combining I–V curve data and EIS data. While I–V curves provide performance parameters like I_sc, V_oc, and FF, they lack accurate insights into R_s, R_p (1/R_p = 1/R_sh + 1/R_d, where R_d is the dynamic diode resistance, as explained later), and junction capacitance (C_p) values which can be used to confirm or deconvolute the respective PID mechanisms that are truly responsible for the power loss. PID_jr is distinct from the SRH recombination described in ref. [2]. While ref. [2] attributes increased recombination to Na-induced bulk defects, PID_jr primarily affects surface and junction regions, creating junction recombination centers that lead to I_sc losses. Unlike bulk-driven SRH recombination, PID_jr is strongly influenced by encapsulation chemistry, impacting junction-related parameters in the two-diode model. This distinction is important for developing targeted mitigation strategies, as PID_jr requires junction- and surface-level interventions, whereas Na-induced SRH recombination necessitates limiting Na diffusion into the bulk. While conventional current–voltage (I–V) measurements provide key performance parameters, their insights into internal device behavior are limited. EIS, however, offers complementary and deeper insights into charge transport, re-combination, and capacitance- and resistance-related issues. EIS complements I–V quickly and directly, measuring R_p/R_sh, R_s, and C_p under dark conditions at different bias voltages. Two four-cell glass–glass modules with p-base PERC monofacial cells and EVA and POE encapsulants were investigated for this PID evaluation at +1000 V.

2. Causes for PID Mechanisms

The mechanisms underlying PID_s and PID_p have been extensively studied and documented in the literature [5,6] and are therefore only briefly summarized here. In contrast, the root cause of the PID_c mechanism has been rarely addressed and is examined in greater detail in this section. Additionally, a newly proposed mechanism, PID_jr (junction recombination), is introduced and briefly discussed at the end of this section.

The shunting potential-induced degradation type (PID_s): This degradation mechanism reduces shunt resistance, leading to a decline in both fill factor and open-circuit voltage. It is an electrochemical process driven by the migration of Na⁺ ions from the glass to the silicon nitride surface, followed by their drift to the silicon surface as neutral Na atoms. These neutral atoms subsequently diffuse into the junction, creating shunting pathways through stacking faults [11].

The polarization potential-induced degradation type (PID_p): This degradation mechanism enhances surface recombination, resulting in a large reduction in the short-circuit current and a small reduction in open-circuit voltage. It is primarily caused by the degradation of the passivation effect provided by the AlO_x/SiN_x layer, which leads to increased surface recombination [12].

The corrosion potential-induced degradation type (PID_c with negative polarity): The rarely reported PID_c mechanism could potentially arise from the dissolution of PbO present in the silver metallization when cell is subjected to PID at negative polarity. During PID, neutral sodium (Na) atoms are deposited on the Ag metallization surface of solar cells due to higher electrocatalytic activity compared to other cell surfaces which are in contact with the encapsulant. These Na atoms can interact with PbO, a component in silver metallization that enhances adhesion and reduces contact resistance between Ag and Si. The dissolution of PbO depends significantly on the presence of moisture and/or acetic acid within the encapsulant. In the absence of moisture, Na reacts with acetic acid (CH₃COOH) from the encapsulant, forming sodium acetate (CH₃COONa) and releasing hydrogen gas (H₂), causing local delamination. Sodium acetate, being a weak base, does not effectively dissolve PbO. However, in the presence of trace moisture, Na, being more electronegative than Ag, creates a galvanic cell that protects Ag while forming sodium hydroxide (NaOH). Since NaOH is a strong base, it can strongly react with PbO to form Pb(OH)₂ and Na₂O. The Pb (OH)₂ dissolves and diffuses into the encapsulant, increasing the Ag grid’s contact resistance and decreasing the fill factor (FF) after the PID stress. This degradation could be exacerbated outdoors if moisture infiltrates through laminate edges or coversheets.

The corrosion potential-induced degradation type (PID_c with positive polarity): If the solar cell is held at positive polarity during PID, conventional electrolytic corrosion of the metallization or interconnect can occur in the presence of trace oxidizable impurities (e.g., water, acetic acid) in the encapsulant. Since no evidence of PID_c was observed in this work (i.e., no increase in R_s), the study focuses exclusively on PID_p and PID_jr mechanisms and their impacts on I–V and EIS data.

The junction recombination potential-induced degradation type (PID_jr): This newly identified type of PID is introduced in this study. It involves increased surface recombination and junction recombination, leading to I_sc losses that can be explained using the two-diode model shown in Figure 1. Bulk recombination is not expected to be influenced by any PID type, as PID primarily affects the surface (PID_p), surface and junction (PID_jr), junction (PID_s), or metallization through corrosion (PID_c). PID_jr is hypothesized to originate when oxidized or reduced species, generated during PID stress, reach the junction and form mid-bandgap recombination centers.

In the two-diode model shown in Figure 1, the placement of Sr (surface recombination), Br (bulk recombination), and Jr (junction recombination) is strategically aligned with their respective contributions to carrier losses. Sr and Br are positioned at the high-bias region, indicating their primary influence on carrier recombination at the front surface and within the bulk of the semiconductor material. However, as per the study, bulk recombination (Br) is not expected to be affected by any PID type, suggesting that the recombination losses primarily occur at the surface and junction regions. Junction recombination (Jr) is placed at the low-bias region, where mid-bandgap recombination centers formed by oxidized species due to PID stress contribute to increased recombination within the junction. This leads to a significant reduction in the parallel resistance (R_p), primarily due to a decrease in the dynamic diode resistance rather than an increase in shunt leakage. The two diodes (Rd1 and Rd2) represent recombination mechanisms at different bias levels, with Rd1 being influenced by surface recombination (Sr) and Rd2 being affected by junction recombination (Jr). The positioning of C_p (junction capacitance) further highlights the role of charge accumulation in modifying recombination dynamics, where PID-induced junction degradation alters the capacitance response under varying bias conditions.

3. Methodology

3.1. Construction, DH, and PID Stressing of PV Modules

Three custom-fabricated mini glass–glass PV modules without frames, each containing four p-base PERC monofacial cells, were prepared for this study. To simulate the frame of the PV laminates, conductive aluminum tape was employed to attach the negative electrical lead of the power supply during PID stress testing. These modules were constructed with 3.2 mm-thick glass covers (34.5 cm in length and 37 cm in width). As illustrated in Figure 2, these modules were labeled 144GG, 074GG, and 148GG. The 144GG module served as the control (fresh) module, while the 074GG module (with POE encapsulant on both sides of the cells) and 148GG module (with EVA encapsulant on both sides of the cells) were subjected to damp heat stressing (85 °C/85% RH for 2000 h) followed by +1000 V PID stressing. Damp heat (DH) stressing was selected as a preconditioning step before potential-induced degradation (PID) testing to replicate and accelerate real-world field stress conditions that contribute to module aging and degrada-tion. Exposure to 85 °C temperature and 85% relative humidity accelerates the degradation of encapsulant–glass–cell interfaces, leading to weakened adhesion, increased moisture ingress, and enhanced ion mobility—critical factors that heighten PID susceptibility. This stress condition was chosen based on the IEC 61215 standard, the primary design qualification standard for ensuring basic reliability of PV modules [13]. The 2000-h duration was selected according to the IEC 63209-1 standard [14], which addresses extended stress testing beyond IEC 61215. The prolonged exposure prescribed in IEC 63209-1 is designed to simulate degradation mechanisms observed over decades of field operation. By applying this DH preconditioning approach, the PID testing reflects the conditions experienced by aged modules rather than fresh ones, ensuring a more realistic assessment of PID susceptibility in practical installations.

3.2. Current–Voltage (I–V) Measurements

The I–V curves of all modules were measured outdoors under natural sunlight conditions using an I–V tracer with matched reference cells. These measurements provided critical performance parameters, including the maximum power (P_max), short-circuit current (I_sc), open-circuit voltage (V_oc), and fill factor (FF). The recorded I–V data were subsequently translated to standard test conditions (STC), which standardize the results to 1000 W/m² irradiance (high sun) and 25 °C cell temperature. This translation to STC allowed for a direct comparison of module performance before and after PID stress testing, enabling a precise determination of the relative degradation in each parameter. Low-sun (200 W/m²) I–V data were also obtained to accurately determine the influence of shunt resistance on FF, if any.

3.3. Electroluminescence (EL) Imaging

EL imaging was conducted before and after PID stress testing, using a Sensovation HR-830 camera (manufactured by Sensovation AG, Radolfzell, Germany) with a 60-s exposure time. To effectively distinguish the effects of series resistance from those of shunt resistance, EL images were captured at two current injection levels, 100% I_sc and 10% I_sc. A 60-s exposure time for electroluminescence (EL) imaging was selected to ensure sufficient signal capture at both 100% and 10% I_sc, allowing for a clear distinction between series resistance (R_s) and shunt resistance (R_p) effects. At 100% I_sc, the high current injection enhances luminescence intensity, making it easier to detect series resistance-related variations, such as localized heating or conductive pathway disruptions. Conversely, at 10% I_sc, the reduced injection current minimizes the impact of R_s, allowing shunting-related defects, such as recombination-active regions and leakage pathways, to be more prominently visible. The 60-s exposure time balances signal clarity and noise reduction across both conditions, ensuring that low-intensity regions in 10% I_sc images remain detectable without oversaturation in 100% I_sc images. This exposure duration has been successfully used in previous PID studies to capture degradation patterns effectively, validating its appropriateness for this study.

3.4. Electrochemical/Electrical Impedance Spectroscopy (EIS) Measurements

EIS is a powerful tool for accurately determining the key electrical parameters of a solar cell—series resistance (R_s), parallel/shunt resistance (R_p), and capacitance (C_p)—in RC or Randles circuits. EIS is a technique that analyzes the frequency-dependent response of an electrochemical or semiconductor system to an applied AC voltage, providing insights into charge transport, capacitance, resistance, and reaction kinetics. Physically, it characterizes how a system resists and stores electrical charge, helping to distinguish between different loss mechanisms such as recombination, resistive losses, and degradation effects. By deconvoluting these processes, EIS offers a deeper understanding of internal device behavior, complementing conventional I–V measurements. EIS application in photovoltaics is relatively recent but has demonstrated significant potential in diagnosing degradation mechanisms. This study employs EIS to quantitatively assess degradation in commercial and custom-made bifacial glass–glass (GG) PV modules. A small signal perturbation with a variable frequency (10 mHz to 4 MHz) and an amplitude of 400 mV was applied, allowing for impedance analysis. Due to non-ideal capacitance behavior, EIS spectra often deviate from a perfect semicircle, making the use of the constant phase element (CPE) essential for accurate modeling. The β parameter of the CPE ranges from 0 to 1, indicating behavior from pure capacitance (β ≈ 1) to pure resistance (β ≈ 0) [15]. Impedance analysis provides critical insights into PV module behavior, linking electrical properties to circuit elements. R_s represents ohmic losses, while R_p and C describe the p–n junction characteristics, including transition and diffusion capacitances. A Nyquist plot in EIS is a graphical representation of impedance, where the imaginary component (−Im(Z)) is plotted against the real component (Re(Z)) over a range of AC frequencies as seen in Figure 3. It provides a visual way to analyze resistive and capacitive elements in a system, often displaying semicircles and linear regions that correspond to different electrochemical processes such as charge transfer, diffusion, and recombination. The shape and features of the Nyquist plot help in extracting key parameters like series resistance (R_s), charge transfer resistance (R_p), and capacitance, making it a crucial tool for diagnosing device performance and degradation. Nyquist plots visualize impedance variations, showing higher impedance at low frequencies due to junction resistance and ohmic losses, while at high frequencies, impedance decreases sharply due to parasitic inductive effects from external components such as cables and connectors. This manifests as a vertical shift in the Nyquist plot. Since R_s is frequency-independent, it horizontally shifts the impedance spectrum without altering its shape. This study prioritizes capacitive behavior below 1 MHz while disregarding inductive influences, which were verified to originate from external sources rather than intrinsic PV cell properties [16,17].

The relative degradations of performance parameters obtained from I–V characterizations include I_sc, V_oc, FF, and P_max. These I–V parameters, in combination with EIS and EL data, provide valuable insights into performance degradation, enabling the differentiation of PID mechanisms. When analyzed alongside the EIS data, they were instrumental in identifying and differentiating the underlying causes of performance loss and determining whether these degradations were due to PID_s, PID_p, PID_c, or PID_jr. The EIS spectra of all modules were measured indoors under dark conditions using a BioLogic SP-200 potentiostat (manufactured by BioLogic Sciences Instruments SAS, Seyssinet-Pariset, France) with EIS capabilities. For these four-cell modules, the spectra were recorded at various bias voltages ranging from −0.5 V to +2 V over a frequency range of 10 mHz to 4 MHz, with an applied AC voltage of 400 mV. The resulting EIS spectra were analyzed using ZView^® software Version 4.0h from Scribner Associates Inc., Southern Pines, NC, USA) to extract key electrical parameters of R_s, R_p, and C_p. The percentage of change in the electrical parameters of I_sc, V_oc, FF, R_s, R_sh, and C_p from I–V and EIS was calculated using Equation (1). The selection of frequency and voltage ranges for EIS measurement is critical to accurately capturing the electrical response of photovoltaic modules and diagnosing degradation mechanisms. The chosen frequency range of 10 mHz to 4 MHz ensures a comprehensive analysis of both low-frequency and high-frequency impedance behaviors. At low frequencies (10 mHz to a few Hz), EIS captures recombination effects, charge trapping, and long-term polarization phenomena, which are essential for identifying shunt resistance (R_p) changes due to PID_s and junction recombination losses (PID_jr). Mid-frequency ranges (kHz range) help assess carrier transport properties and interfacial capacitances, while high frequencies (above 1 MHz) allow for the evaluation of series resistance (R_s) and parasitic inductances. The voltage range from −0.5 V to +2 V for this four-cell module enables the analysis of device behavior under different operating conditions, from near open-circuit to forward-bias conditions. The inclusion of negative bias provides insights into leakage pathways and charge accumulation effects, whereas positive bias helps monitor recombination and conduction mechanisms up to 0.5 V per cell. The applied AC voltage of 400 mV ensures a small perturbation to maintain the system within the linear response regime, preventing non-linear distortions while obtaining accurate impedance spectra.

$$\% Change={\displaystyle \frac{Post data-Pre data}{Pre data}}\times 100$$

(1)

4. Results and Discussion

4.1. I–V Results

Figure 4 presents pre-PID and post-PID I–V data for two four-cell modules measured at high/STC and low irradiance levels. A table within the figure summarizes the changes in I_sc, V_oc, and FF caused by PID stress. The data indicate that I_sc is the most significantly affected parameter in both modules, with little or no impact on V_oc and FF. The table in Figure 4 shows that, despite a significant reduction in short-circuit current (I_sc) in both modules (e.g., −8.5% for EVA and −10.1% for POE at STC), V_oc exhibits only minor variations (+0.9% for EVA and +0.4% for POE). This suggests that the dominant degradation mechanisms, such as surface and junction recombination, primarily affect carrier collection rather than the built-in junction potential. Prior studies indicate that V_oc remains stable in PID as long as shunt pathways do not severely impact the depletion region [3].

4.2. EL Results

The EL images obtained before and after PID stress testing at two distinct current injection levels, 100% I_sc and 10% I_sc, are presented in Figure 5.

The images taken at the higher injection level of 100% I_sc were anticipated to reveal any notable increase in series resistance. Conversely, the images acquired at the lower injection level of 10% I_sc were expected to highlight any significant decrease in shunt resistance. A detailed analysis of the EL images using the ImageJ processing tool reveals that the PID stress testing caused no significant changes in either series resistance or shunt resistance; consequently, there was no significant impact on the FF. Therefore, the observed reduction in I_sc following PID stress testing at positive polarity is likely due to factors unrelated to resistance.

As expected, the raw EL images at 100% I_sc exhibited significantly higher overall signal intensity compared to those at 10% I_sc. Brightness normalization was applied using ImageJ software (version 1.54i, June 2023) for a clear comparative analysis and visual representation to quantify the impact of PID stress on solar cell performance, specifically assessing changes in series resistance (R_s) and shunt resistance (R_sh) that could affect the fill factor (FF). Since I–V and EIS results showed minimal variations in R_s and R_sh, EL imaging was used as a complementary tool to visually and quantitatively confirm these findings. The analysis began with image preprocessing, where EL images were loaded into ImageJ, and background noise was removed using image thresholding to enhance defect visibility. Next, the region of interest (ROI) was selected for each of the four solar cells individually to ensure accurate measurement. Using the “Measure” function, the mean light intensity of each ROI was calculated, representing the electroluminescence signal strength, where higher intensity indicates lower recombination losses and lower intensity suggests increased recombination due to PID effects. A comparison of mean intensity values before and after PID stress testing showed no significant reduction, confirming that R_s and R_sh remained largely unaffected. This result aligns with I–V and EIS measurements, which indicated minimal changes in FF, further supporting that junction recombination (PID_jr) was the primary degradation mechanism rather than resistance-related losses. The combination of quantitative EL image analysis and electrical measurements provided strong evidence that PID stress did not significantly impact the FF of the tested modules. By using ImageJ, the study extracted precise quantitative data from EL images, reinforcing that PID-induced degradation was dominated by junction recombination effects rather than changes in resistance.

4.3. EIS Results

EIS spectra for the two four-cell modules were measured at bias voltages from −0.5 V to +2 V, but only Nyquist plots obtained at 0 V and 1 V are shown in Figure 6 to reduce curve crowding. The high-frequency intercept of the Nyquist semicircle gives R_s, while its diameter provides R_p, and C_p is calculated using 1 = 2πf_cR_pC_p, where f_c is the peak/characteristic frequency. Pre- and post-PID data show minimal changes in R_s, R_p, and C_p for the POE module but significant changes for the EVA module across wide bias voltages as high as 0.5 V per cell, where dynamic resistance of the diode decreases significantly [5].

The data analysis from

Table 1. Electrical parameters of fresh and PID-stressed modules at different bias voltages.

Bias Voltage	144GG (Fresh/Control)			074GG, Pre-PID (PID-Stressed)			074GG, Post-PID (PID-Stressed)			148GG, Pre-PID (PID-Stressed)			148GG, Post-PID (PID-Stressed)
	R1 (Ω)	R2 (KΩ)	C1 (µF)	R1 (Ω)	R2 (KΩ)	C1 (µF)	R1 (Ω)	R2 (KΩ)	C1 (µF)	R1 (Ω)	R2 (KΩ)	C1 (µF)	R1 (Ω)	R2 (KΩ)	C1 (µF)
Bias_−0.5	0.050	12.671	2.526	0.020	51.540	2.628	0.081	42.530	3.127	0.055	7.477	2.884	0.136	0.832	5.128
Bias_−0.4	0.051	12.474	2.609	0.057	46.588	3.083	0.080	41.881	3.163	0.060	7.586	2.958	0.136	0.814	5.331
Bias_−0.3	0.051	12.009	2.668	0.079	37.283	3.066	0.080	36.476	3.139	0.066	7.591	3.028	0.135	0.777	5.389
Bias_−0.2	0.051	11.200	2.701	0.080	27.917	3.050	0.082	28.518	3.067	0.069	7.431	3.085	0.135	0.721	5.411
Bias_−0.1	0.028	9.898	2.684	0.082	19.400	3.000	0.086	20.543	3.012	0.075	6.998	3.132	0.137	0.657	5.353
Bias_0	0.046	8.528	2.649	0.086	13.433	2.957	0.084	14.488	2.917	0.092	6.305	3.096	0.142	0.552	5.286
Bias_0.5	0.046	2.258	2.700	0.113	2.070	2.670	0.112	2.053	2.618	0.101	2.014	2.886	0.130	0.235	3.971
Bias_1	0.083	0.359	3.171	0.118	0.381	2.698	0.123	0.377	2.683	0.110	0.361	3.314	0.148	0.090	3.258
Bias_1.5	0.134	0.037	3.967	0.201	0.036	3.467	0.144	0.035	3.454	0.156	0.038	4.086	0.172	0.025	4.830
Bias_1.9	0.165	0.002	39.602	0.352	0.001	50.084	0.252	0.002	137.510	0.268	0.001	16.790	0.255	0.001	15.952
Bias_2	0.161	0.001	24.032	0.397	0.0004	52.016	0.281	0.0004	50.041	0.282	0.0004	31.906	0.274	0.0004	29.673

offers crucial insights into the behavior of PV modules under various voltage biases ranging from −0.5 V to 2 V at a superimposed AC voltage of 400 mV, revealing the effectiveness of EIS in identifying performance issues. Notably, Module 074GG exhibited a higher shunt resistance at negative bias compared to the fresh and 148GG modules. Additionally, capacitance increased at higher forward bias voltages compared to reverse and low positive biases in all three modules. For Module 074GG, the series resistance (R_s) increased by 87.56% compared to the fresh module, and after PID testing, R_s further increased by 41.24%. The shunt resistance (R_p) rose by 106.05% relative to the fresh module, with a 6.82% increase in R_p after PID. Capacitance also saw a 22.27% increase compared to the fresh module, with an additional 18.87% rise post-PID. Similarly, for Module 148GG, R_s increased by 60.68% from the fresh state, and R_s experienced a significant 62.71% rise post-PID. R_p for this module showed a 25.54% increase compared to the fresh module, with post-PID testing revealing a 67.13% rise in R_p. These findings support EIS as an essential diagnostic tool for understanding module health and providing mitigation strategies.

The larger diameter of the high-frequency arc confirms that recombination resistance increases as the reverse bias strengthens. Module 074GG under reverse bias voltages (0 V to 0.5 V) shows a clear increase in parallel resistance (R_p) as reverse bias voltage rises. This is visually represented by the widening of the semicircles on the plot, signaling a rise in recombination resistance and indicating reduced recombination current. For Module 148GG, the EIS analysis under positive bias voltages (0 V to 2 V) reveals a drastic drop in shunt resistance (R_p), from approximately 6 kΩ at 0 V to just 1 Ω at 2 V. Similarly, under reverse bias conditions, R_p drops from 8 kΩ at 0 V to 500 Ω at −0.5 V. This significant reduction in shunt resistance is a hallmark of PID shunting (PID_s), where extensive shunting occurs within the solar cells.

4.4. EIS Complementing I–V for PID Differentiation

The maximum power (P_max) of a PV module is determined by I_sc, V_oc, and FF, as expressed in the equation below:

P_max = I_sc × V_oc × FF

(2)

The impacts on I_sc, V_oc, and FF in I–V data and on R_s and R_p in EIS data due to surface recombination (PID_p) and junction recombination are summarized in Table 1. The I–V results for the 074GG-POE module provided in Table 1 reveal that I_sc decreased by 10.1%, while V_oc and FF remained largely unaffected. This is consistent with the literature on the PID_p mechanism [2,3,4,6]. The EIS results also support the PID_p mechanism, with no change in R_s or R_p, as shown in Figure 4. For the 148GG-EVA module, I_sc was found to be decreased by 8.5%, while V_oc and FF remained largely unaffected. The EIS results indicate a large (91%) reduction in parallel resistance (R_p). Despite this large reduction, the ratio of series resistance (R_s) to R_p remained approximately 1:4000, ensuring that the degradation in R_p had a minimal impact on FF. The large reduction in R_p, affecting FF only minimally, was surprising, and it can probably be attributed to an increase in junction recombination and therefore a decrease in the dynamic resistance of the diode at I_sc and low bias. If oxidizable species are present, then PID under positive polarity is known to generate trace amounts of oxidized species, such as oxygen, at the front silicon surface where the junction is in close proximity. These oxidized species accumulate near or within the junction for the front surface, significantly enhancing junction recombination (PID_jr) in p-type PERC due to formation of trap centers (e.g., B-O complex) without affecting the bulk recombination. This localized recombination at the junction reduces the parallel resistance (R_p), which is basically a combination of the shunt resistance and dynamic resistance of the diode. The reduction in R_p is therefore primarily due to changes in the dynamic diode resistance rather than the ohmic shunt resistance, as evidenced by the combined I–V and EIS data shown in

Table 2. EIS complementing I–V to differentiate PID_p from PID_jr.

Dominant PID Loss Mechanism	Solar Cell Loss Mechanism	Affects Isc in IV *	Affects Voc in IV *	Affects FF in IV *	Affects Rs in EIS	Affects Rp in EIS
PIDp (Polarization)	Surface recombination under +ve 1000V PID of 074GG with POE encapsulant (surface recombination increases)	−10.1%/−2.8%	+0.4%/+1.7%	+2.3%/−0.8%	0%	0%
PIDjr (Junction Recombination)	Junction recombination under +ve 1000V PID of 148GG with EVA encapsulant (diode dynamic resistance decrease)	−8.5%/+0.3%	+0.9%/+0.4%	+2.0%/−2.5%	56% (0.09 → 0.14 Ω)	91% (6.31 → 0.55 kΩ)

* High sun/Low sun illumination.

. The reduction in I_sc can be understood through the relationship shown in Equation (2). Since the oxidized species generated during PID are localized near the junction, junction recombination dominates, leaving bulk recombination negligible. This is reflected in the reduced I_sc, while FF remains largely stable because R_p, despite its significant reduction, still remains in the kilo-ohm range and much higher than R_s. Overall, the data strongly support the assertion that oxidized species generated near the junction during +PID in the EVA module primarily impact I_sc and minimally impact FF by enhancing junction recombination leading to PID_jr.

5. Conclusions

This study’s findings have significant applications and benefits in the field of PV reliability and diagnostics, particularly for early detection and characterization of PID mechanisms. One key application is in quality control and module certification, where manufacturers can use these combined diagnostic methods of I–V, EIS, and EL to detect and assess module susceptibility to various reliability and durability mechanisms, including PID mechanisms. Another major benefit is the ability to optimize cell design by modifying passivation layers or implementing mitigation strategies such as reverse bias recovery or advanced coatings to reduce surface and junction recombination.

EIS proves to be a powerful yet underutilized technique in silicon photovoltaics, offering critical insights into potential-induced degradation (PID) mechanisms, including PID_s, PID_p, PID_c, and the newly identified PID_jr. By complementing conventional current–voltage (I–V) measurements such as I_sc, V_oc, and FF, EIS enables a more comprehensive characterization of key parameters—such as series resistance (R_s), shunt resistance (R_p), and capacitance (C_p)—facilitating a more precise differentiation of PID types. This study reveals that the fill factor (FF) remains relatively stable in both EVA- and POE-encapsulated modules, indicating that the FF is not a major factor in power loss under PID stress. Instead, reductions in short-circuit current (I_sc) are primarily driven by recombination effects. In POE modules, I_sc degradation is likely caused by surface recombination, as evidenced by the stability of R_s, R_p, and C_p. In contrast, the EVA module exhibits combined surface and junction recombination effects, leading to notable variations in R_s, R_p, and C_p. The observed reduction in R_p is attributed to a decrease in dynamic diode resistance rather than ohmic shunt resistance, resulting from recombination currents induced by junction defect centers formed under positive voltage stress in EVA encapsulants. Furthermore, the distinct behavior of R_s and C_p between EVA and POE modules highlights the crucial role of encapsulant chemistry in influencing PID effects. These findings emphasize the importance of encapsulant selection in mitigating PID and demonstrate the effectiveness of EIS as a diagnostic tool for uncovering degradation mechanisms beyond conventional I–V analysis. The insights gained from this study provide a foundation for improved PID-resistant module designs and enhanced long-term reliability of photovoltaic systems.