Feasibility of Using Rainwater for Hydrogen Production via Electrolysis: Experimental Evaluation and Ionic Analysis

Abstract

This study evaluates the feasibility of employing rainwater as an alternative feedstock for hydrogen production via electrolysis. While conventional systems typically rely on high-purity water—such as deionized or distilled variants—these can be cost-prohibitive and environmentally intensive. Rainwater, being naturally available and minimally treated, presents a potential sustainable alternative. In this work, a series of comparative experiments was conducted using a proton exchange membrane electrolyzer system operating with both deionized water and rainwater collected from different Austrian locations. The chemical composition of rainwater samples was assessed through inductively coupled plasma, ion chromatography and visual rapid tests to identify impurities and ionic profiles. The electrolyzer’s performance was evaluated under equivalent operating conditions. Results indicate that rainwater, in some cases, yielded comparable or marginally superior efficiency compared to deionized water, attributed to its inherent ionic content. The study also examines the operational risks linked to trace contaminants and explores possible strategies for their mitigation.

1. Introduction

As global efforts intensify to reduce carbon emissions, hydrogen has emerged as a key component in energy strategies, valued for its clean energy profile, versatility across applications, and ability to be stored in various forms for later use—such as in underground salt caverns, as a cryogenic liquid, or within solid-state materials like metal hydrides [1]. Besides that, hydrogen is increasingly recognized as a critical enabler of the energy transition due to its high gravimetric energy density, zero-carbon emissions at the point of use, and versatility across sectors [2,3,4]. When produced from renewable electricity, hydrogen offers a clean alternative to fossil fuels, particularly in hard-to-electrify sectors such as heavy industry, long-haul transport, and seasonal energy storage. Jeje et al. [5] emphasize hydrogen’s role in decarbonizing these sectors while improving energy security and enabling sector coupling through power-to-gas applications. The authors of [6] highlight that hydrogen compares favorably to conventional hydrocarbons in energy density and greenhouse gas emissions, especially when used in fuel cells instead of combustion engines. Compared to fossil fuels, hydrogen combustion or use in fuel cells produces minimal pollutants like CO₂, SOₓ, or particulate matter, making it significantly cleaner for climate and air quality [4,6].

Hydrogen can be produced through various techniques, including photocatalytic and photoelectrochemical (PEC) water splitting, as well as formic acid decomposition and ammonia decomposition. Formic acid, in particular, has gained attention due to its liquid-state storage and potential for on-demand hydrogen release over metal-based catalysts [7,8]. Photocatalysis offers the promise of directly converting sunlight to hydrogen, but currently suffers from low efficiency, limited catalyst stability, and scale-up challenges [9,10]. Ammonia decomposition has also attracted interest due to ammonia’s high hydrogen density and easier storage, yet typically requires high temperatures and complex reactor designs [11,12]. In contrast, water electrolysis—particularly PEM and alkaline systems powered by renewables—offers higher maturity, commercial availability, and well-established performance, achieving practical efficiencies in the range of 55–80% (LHV) under favorable operating conditions [13,14,15,16]. Additionally, ammonia decomposition and thermal routes may introduce contaminants such as nitrogen or CO, posing challenges for high-purity hydrogen applications [17]. While the compatibility of alternative methods with rainwater remains largely unexplored, electrolysis remains the most robust and scalable option for decentralized green hydrogen production, even more when driven by electricity from renewable sources and without releasing carbon emissions [18,19]. The electrolysis process (especially the PEM) splits water into hydrogen and oxygen using electricity. This process involves two half-reactions: at the anode, water is oxidized to produce oxygen, protons, and electrons; at the cathode, the protons combine with electrons to form hydrogen gas. PEM allows only protons to pass through while keeping gases and electrons separated, enabling efficient and high-purity hydrogen generation. The overall reaction can be observed in Equation (1). When this process is driven entirely by renewable energy sources such as wind or solar, the resulting hydrogen is classified as green hydrogen, distinguishing it from other forms like blue hydrogen (produced from natural gas with carbon capture and storage) or turquoise hydrogen (derived from methane pyrolysis with solid carbon byproducts) [20]. However, the widespread adoption of electrolysis faces key obstacles, including the substantial requirement for purified water, typically distilled or deionized, which is both resource- and energy-intensive to produce [21].

H₂O _(liquid) + Energy → H_{2 (gas)} + ½ O_{2 (gas)}

(1)

Water quality plays a critical role in hydrogen production by electrolysis, especially for PEM systems where conductivity, organics, and ions might affect performance and maintenance. Research showed that marginal water sources like rainwater may reduce treatment needs compared to seawater or wastewater [22]. Also, certain ions can be tolerated at low concentrations, though purification remains essential [23]. This conventional demand for high-quality water not only adds to the operational cost but also raises concerns about long-term sustainability, particularly in regions where freshwater is scarce. As such, the search for alternative water sources that require less processing and result in lower costs is of growing interest. In this context, rainwater emerges as a practical alternative—it is naturally renewed through the hydrological cycle, often cleaner than surface water or seawater, and widely accessible in many parts of the world. Yet, its viability for direct use in electrolyzers has not been systematically evaluated.

Although the effects of water impurities on electrolyzer performance and durability have been explored in prior studies [22,23,24], relatively little attention has been given to characterizing rainwater as a potential input. This may stem from the longstanding emphasis on using ultrapure water in a conservative attempt to prevent membrane fouling or catalyst degradation in Proton Exchange Membrane (PEM) systems. As a result, naturally occurring, minimally treated water sources like rainwater have often been overlooked in research and practice. However, rainwater presents a unique opportunity: it is renewable, locally available, and often requires significantly less treatment than surface or seawater. Simões et al. [25] highlight that rainwater harvesting can offer a technically and economically viable feedstock for hydrogen production in decentralized settings—especially where water treatment infrastructure is limited. Their study also stresses that water sourcing is a critical but underrepresented dimension in hydrogen system planning, and that quality metrics must be considered alongside availability. While other untreated sources such as tap or seawater require pre-treatment to avoid electrode degradation of electrolyzer components [26], rainwater may strike a balance between availability and performance—particularly when combined with basic filtration or quality monitoring strategies.

Nevertheless, while rainwater as an electrolysis feedstock has been explored in some prior works, there remains a lack of experimental data directly comparing efficiency, hydrogen output, and practical impurity thresholds between rainwater and deionized water in a PEM electrolyzer [27]. To address this research gap, the present work conducts an experimental evaluation of rainwater as a feedstock in hydrogen-producing electrolysis systems. By comparing its performance against laboratory-grade deionized water, we evaluate not only the hydrogen yield and efficiency under different operating conditions, but also the effects of ionic composition on system behavior. Additional analyzes such as ion chromatography (IC) and inductively coupled plasma (ICP) testing provide insight into the rainwater’s chemical profile and its relevance for practical use.

The results suggest that rainwater can offer comparable—and in some cases even slightly superior—electrolysis performance when contrasted with deionized water. These findings could have significant implications for the decentralization and cost-effectiveness of hydrogen production, particularly in rural or off-grid areas where water treatment infrastructure is limited.

Overall, this study supports the advancement of sustainable hydrogen production methods by proposing a low-cost, locally available water source that could alleviate the dependence on purified inputs. The findings encourage further research into system durability, contaminant management, and scalability of rainwater-based electrolysis systems.

2. Materials and Methods

2.1. Investigated Water Sources

Electrolysis-driven hydrogen generation relies heavily on the availability and quality of water. High-purity water sources are typically used to avoid efficiency losses and component degradation in electrolyzers. This chapter explores and compares two such sources—deionized water and rainwater—investigated experimentally to evaluate their performance in a PEM electrolysis system.

2.1.1. Research Methodology

The methodology employed was a combination of bibliographic review and empirical analysis. The main objective was to evaluate how viable and effective alternative water sources—particularly rainwater—are in comparison to deionized water. Rainwater was collected and subjected to chemical analysis and electrolysis testing using a custom PEM system designed to track energy input and hydrogen output. Microsoft Excel 365 (Microsoft Corporation, Redmond, WA, USA) and Python 3.13.5 (Python Software Foundation, Wilmington, DE, USA) were used for processing and visualizing the data, while quality control measures were based on established laboratory procedures.

2.1.2. Deionized Water

Deionized (DI) water was prepared under controlled laboratory conditions at the University of Applied Sciences Upper Austria. The deionization process involves ion exchange resins [28] that replace unwanted cations and anions with hydrogen (H⁺) and hydroxide (OH⁻) ions, effectively purifying the water to remove conductivity-causing ions. The resins require regular regeneration to maintain performance. The investigated DI water offers very low levels of conductivity (around 1.16 µS/cm), aligning with typical DI water standards (ISO 3696:1987 – Grade 2/3) [29], which specify conductivity around 1 µS/cm and most important, below 5 µS/cm (the standard for water equilibrated with air under lab conditions). It also presents minimal organic content (0.318 mg/L), right within U.S. EPA guidance, which specifies the TOC upper limit as 0.35 mg/L regarding general laboratory activities [30], making it highly suitable for electrolysis in terms of efficiency and minimal wear on system components.

2.1.3. Rainwater

Rainwater samples were collected in Salzburg (Sample A, collected on the 10 April 2024) and Wels (Sample B, collected on the 12 April 2024 and Sample C, collected on the 13 April 2024), Austria. Samples were captured using laboratory-cleaned glass collectors and stored in poly-ethylene containers, following protocols from Texas State University on glassware sanitation [31]. A preliminary filtration removed coarse impurities using MN 617—110 mm filter paper. More advanced approaches such as activated carbon or sub-micron filtration could further reduce organic matter or residual ions, at the cost of added complexity and expense to the method. [32]

To evaluate the ionic composition of the rainwater samples and to compare their composition with deionized water, two complementary analytical techniques were used: Ion Chromatography (IC) and Inductively Coupled Plasma Optical Emission Spectroscopy (ICP). The IC analysis was performed using a Metrohm system equipped with a Metrosep A Supp 17—150/4.0 chromatography column (Metrohm AG, Herisau, Switzerland), specifically designed for the separation of common anions. The IC test was employed to quantify common anions, such as chloride (Cl⁻), sulfate (SO₄²⁻), nitrate (NO₃⁻), and fluoride (F⁻). Calibration was performed using standard solutions with known concentration:

• Nitrate Standard for IC, 1000 mg/L, 100 mL, catalog no. 74246—from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany), used for nitrate identification.
• Sulfate Standard for IC, 1000 mg/L, 100 mL, catalog no. 90071—from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany), used for sulfate identification.
• Fluoride Standard for IC, 1000 mg/L, 100 mL, catalog no. 77365—from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany), used for fluoride identification.
• Chloride Standard for IC, 1000 mg/L, 100 mL, catalog no. 39883—from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany), used for chloride identification.

The results were presented as chromatograms where each peak corresponds to a specific anion.

The retention time identifies the ion, while the peak area is proportional to its concentration. In contrast, ICP test used a Thermo Scientific iCAP 7200 Duo (Thermo Fisher Scientific, Waltham, MA, USA), and it was used to analyze cations, including sodium (Na⁺), potassium (K⁺), calcium (Ca²⁺), magnesium (Mg²⁺), and aluminum (Al³⁺). For accurate detection, samples were acidified with 2% nitric acid to preserve the stability of metal ions and to prevent precipitation. Calibration standards for ICP ranged from 1 part per billion (ppb) to 1 part per million (ppm), enabling precise quantification of trace metal concentrations.

Additional analyzes were also performed to complement the ionic characterization. The Total Organic Carbon (TOC) test was carried out using the HACH LCK 380 cuvette test method (Hach Company, Loveland, CO, USA), which quantifies total carbon (TC) and total inorganic carbon (TIC) and TOC is calculated as the difference among them. The resulting gas passes through a membrane and is detected photometrically, providing a measure of the total organic matter present in the sample. pH measurements were conducted using both visual test strips (Instruments Direct, Manchester, UK) and potentiometric analysis with the MultiLine 3620 IDS (Xylem Analytics, Weilheim, Germany), ensuring accurate assessment of the rainwater’s acidity levels.

For comparison and reference purposes, the deionized water sample was subjected to the corresponding applicable tests.

2.2. The Electrolyzer System

The core of the experimental setup used for evaluating water sources in electrolysis was a compact yet robust PEM electrolyzer system assembled in the laboratory. The system operated with a fixed voltage of 4.2 V, while the current ranged between 15 and 26 A during testing. The water flow rate ranged from 215 to 380 mLn/min (3.6–6.6 mLn/s), scaling with the applied current (and thus power input). The system integrates various components that ensure accurate performance measurement, operational safety, and high-purity of the produced hydrogen. This section outlines the system architecture, its main parts, and their specific roles in supporting the electrolysis process.

The electrolysis system—as schematically represented in Figure 1 and photographically shown in Figure 2 below—is composed of several key subsystems, including:

• Water Tank (1): The water tank serves as the primary reservoir for the electrolysis process, storing either deionized or rainwater depending on the test. It ensures a consistent water supply to the PEM electrolyzer, acting as the initial stage of the system’s circulation loop. Made of polyethylene, the tank offers durability and chemical resistance necessary for maintaining water purity throughout the process. Its role is not limited to storage; it also supports the return of unused or condensed water from the recirculation line, promoting efficiency and reducing water consumption during repeated testing cycles.
• Pump (2): The pump is responsible for maintaining a steady and controlled flow of water from the tank through the filter and into the PEM stack. It plays a crucial role in maintaining system performance, especially during variations in load or water quality. In this setup, with the aim of guaranteed hydration, heat dissipation, and recirculation needs of the system, a peristaltic pump with a rated flow of 300 mL/min was used.
• Filter (3): Positioned between the water tank and the electrolyzer, the filter safeguards the PEM stack by removing particulates and potential ionic contaminants. The filter used in this setup is made of Acetal Homopolymer with NBR (nitrile butadiene rubber) seals, possessing an internal volume of 30 cm³, a maximum flow capacity of 2 L/min and has a particulate retention of less than 10 microns. Operating effectively between 5 °C and 50 °C, it ensures that no damaging impurities reach the sensitive membrane and catalyst layers of the electrolyzer. This is particularly important when testing with rainwater, which may contain residual organic or inorganic matter even after pre-filtration.
• PEM Stack (Electrolyzer): The PEM stack (4) is the central component where the electrochemical reaction occurs, splitting water into hydrogen and oxygen gases. The system comprises two electrolysis cells, each requiring 1.8–2.2 V for operation, and is equipped with Nafion 115 membrane (DuPont, Wilmington, DE, USA). This configuration allows for hydrogen production rates between 300 and 400 mL/min. Regarding electrocatalysts, the configuration utilize platinum-based cathodes and iridium-based anodes—optimized for HER and OER, respectively. The PEM stack must remain free from contamination to preserve high proton conductivity and membrane integrity, making the upstream water purification stages critically important.
• Recirculation Line: The recirculation line is designed to return unused water from the electrolyzer back to the tank, minimizing waste and improving overall system efficiency. It also recovers water vapor that is transported along with the generated hydrogen gas. This closed-loop setup allows the same volume of water to be used repeatedly, significantly lowering consumption and ensuring consistent input quality. In rainwater scenarios, where water conservation is part of the sustainable objective, the recirculation line is especially valuable.
• Dryer (5): After electrolysis, the produced hydrogen gas may carry residual moisture. The dryer removes this moisture to ensure the hydrogen output is dry and of higher purity. It functions by passing the humid hydrogen through a bed of silica gel (desiccant material), which absorbs water molecules onto its porous surface. This absorption process reduces the water vapor content in the gas stream, achieving lower dew point. This step is vital for protecting downstream measurement equipment and ensuring accurate quantification of hydrogen production. Additionally, dry hydrogen is preferable in real-world applications like fuel cells, where water content can interfere with performance and safety.
• DC Power Supply: The DC power unit delivers the necessary electrical input to initiate and sustain the electrolysis reaction. Operating in a controlled mode, it delivers up to 4.2 V while the current is gradually increased to a maximum of 26 A during testing. Voltage and current regulation are essential to manage the electrolyzer’s performance curve and efficiency. The stability of this power source ensures reproducible testing conditions across all water types and operating scenarios.
• Control Unit: The control unit acts as the central management system for the electrolyzer setup. FlowSuite software, version 2.91 (Bronkhorst High-Tech B.V., Ruurlo, The Netherlands) was used to monitor operational parameters such as voltage, current, flow rates, and safety conditions. By monitoring these parameters, the control unit maintains efficient operation and safeguards the system from potential harm. It also contributes significantly to data collection, transmitting performance indicators to the FlowSuite software responsible for analyzing results and generating efficiency curves. The complementary activities of the control system were carried out manually, setting the pump and the input power manually.
• Mass Flow Controller (MFC): The selected MFC (6), EL-FLOW Prestige FG-111BP (Bronkhorst High-Tech B.V., Ruurlo, The Netherlands), Bronkhorst monitors and controls the hydrogen gas output rate during operation. It allows real-time monitoring of gas output and ensures that hydrogen production data is consistent and reliable. By integrating directly with the control unit and FlowSuite data logging software, version 2.91, the MFC enables precise calculation of electrolyzer efficiency, correlating gas output to electrical input during performance testing.
• Gas Leak Detector (7): Given the flammable nature of hydrogen, the system includes a gas leak detector as a safety mechanism, 500GD Multi-Gas Detector (MRU Instruments, Neckarsulm, Germany), which was installed in the experimental room near the electrolyzer. The detector monitors any unintended hydrogen release and if a leak is detected, it allows operators to shut down the system immediately, preventing possible hazards. Its inclusion was particularly important since the experiments were conducted in confined indoor spaces.

Although the system integrated a drying unit to reduce moisture, hydrogen purity was not analyzed according to ISO 14687:2019 [33]. Future work should include trace contaminant analysis (e.g., CO, NH₃, sulfur compounds), since these are critical for applications such as fuel cells.

All of these elements are essential for enabling the electrolysis process to run reliably, efficiently, and safely during experimental testing. The water, whether deionized or rain-collected, was initially stored in a polyethylene tank and continuously recirculated to minimize waste. The implemented set-up is depicted in Figure 2.

2.3. Overview of Conducted Tests

To assess the feasibility of using rainwater as a substitute for deionized water in hydrogen production via electrolysis, two stages of experimental testing were conducted. The first focused on chemical characterization and parameter evaluation of the water samples, while the second measured electrolyzer performance under operational conditions.

The testing procedure aimed at answering two core questions:

• How do the physical and chemical characteristics of rainwater compare to those of deionized water?
• Does rainwater influence the electrolysis system efficiency and hydrogen output differently than deionized water?

To address these, rainwater samples were collected, filtered, and analyzed through multiple laboratory tests. These results formed the foundation for subsequent electrolysis efficiency comparisons under identical electrical and operational conditions.

Electrolyzer Performance Evaluation

Hydrogen output and electrical input were logged at different current levels, keeping the voltage fixed at 4.2 V. Each water type (DI and rainwater) was tested under identical conditions. The results were compared, and the efficiency calculated to different points, as represented by Equations (2)–(5).

Power Input (source):

$$Power Input \left(W\right)=Voltage of Source \left(V\right)\times Current of Source \left(A\right)$$

(2)

Energy Content of Hydrogen:

$${EnH}_2=Hydrogen HHV\times Hydrogen density$$

(3)

$${EnH}_2=33.33{\displaystyle \frac{kWh}{kg}}\times 0.0899{\displaystyle \frac{kg}{m^3}}=3{\displaystyle \frac{kWh}{m^3}}=3{\displaystyle \frac{Wh}{L}}$$

Power Generated by Electrolyzer:

$$Power Electrolyser \left(W\right)=Hydrogen produced \left({\displaystyle \frac{L}{h}}\right)\times 3{\displaystyle \frac{Wh}{L}}$$

(4)

Efficiency:

$$Eff \left(\%\right)={\displaystyle \frac{Power Electrolyser \left(W\right)}{Power Input \left(W\right)}}\times 100$$

(5)

Another important performance metric in electrolysis is the Faradaic efficiency (FE), which quantifies the proportion of electrical charge that is effectively converted into hydrogen gas, based on Faraday’s laws of electrolysis [34]. This parameter is essential for understanding whether the current supplied is solely contributing to the desired hydrogen evolution reaction (HER) or also driving parasitic side reactions, such as water oxidation inefficiencies or crossover losses through the membrane [35]. In this study, however, FE was not directly measured due to the absence of gas-tight compartments and real-time hydrogen quantification tools such as gas chromatography. Instead, system-level efficiency (based on power input vs. system output) was assessed under identical experimental conditions for both deionized and rainwater trials, as previously mentioned, and it was deemed sufficient to evaluate performance impacts from varying water purity.

3. Results

This section presents the key results from both test phases. The first phase was the one focused on characterizing the water samples’ chemical composition, while the second assessed the performance of a PEM electrolyzer using rainwater and deionized water under identical operating conditions. The results provide insights into the effects of water composition on electrolyzer efficiency, safety, and hydrogen purity.

3.1. Water Quality and Composition

The detailed characterization of rainwater and deionized water samples yielded critical insights into ionic, organic, and physicochemical differences. These factors directly impact electrolyzer durability and hydrogen output quality.

Ionic Composition, Organic Content and pH

Although rainwater is formed via natural distillation, trace amounts of ions may still be present due to atmospheric deposition processes, such as the absorption of sea salt aerosols, industrial emissions, or urban dust during precipitation. Similar low-level ionic concentrations have been reported in previous rainwater studies [36,37,38].

The IC chromatogram depicts the separation of anions based on their respective retention times within the IC system (see an example in Figure 3 for sample B). Each peak on the graph represents a specific anion, with its retention time marked along the horizontal axis in minutes. The height or area of each peak correlates with the anion’s concentration in the analyzed sample.

Table 1. Summary of the first tests results and literature comparison.

Ion	Sample A	Sample B	Sample C	Rainwater STD	Deionized Water	Tap Water (Lit.)	Seawater (Lit.)
Cl− (mg/L)	9.816	6.975	6.558	1.773	-	73.9 [39] 79 [23]	19,162 [40] 200 [23]
F− (mg/L)	0.042	0.012	0.036	0.016	-	0.000 [23]	1.3 [40] 1 [23]
SO42− (mg/L)	3.304	2.750	3.070	0.278	-	120 [39] 47 [23]	2680 [40] 2800 [23]
NO3− (mg/L)	1.310	2.242	0.940	0.671	-	20.6 [39] 0.82 [23]	0.000 [23]
Na+ (mg/L)	1.197	1.607	3.535	1.248	0.008	49.1 [39] 62 [23]	10,679 [40] 11,000 [23]
Ca2+ (mg/L)	0.851	0.561	0.455	0.205	0.000	102 [39] 51 [23]	410 [40] 400 [23]
Mg2+ (mg/L)	0.309	0.235	0.164	0.073	0.000	8.81 [39] 7 [23]	1278 [40] 130 [23]
Al3+ (mg/L)	0.007	0.011	0.033	0.014	0.000	0.004 [23]	0.000 [23]
K+ (mg/L)	5.936	1.514	1.336	2.606	0.009	0.000 [39] 0.000 [23]	395 [40] 400 [23]
Organic Matter (mg/L)	7.130	7.680	5.980	0.867	0.318	0.1–10 [41]	<1 [42]
Conductivity (μS/cm)	16.260	25.500	14.480	5.916	1.160	50–500 [43]	50,000 [43] 44,000–58,000 [40]
pH (-)	6.755	6.563	6.068	0.354	5.539	6.5–8.5 [44] 6.5–8.0 [45]	8.05–8.15 [46] 7.6–8.4 [40]

summarizes the ion concentrations in the three rainwater samples and the deionized water control.

As expected, deionized water exhibited negligible ionic and organic content, with a conductivity of 1.16 µS/cm, pH of 5.54, and TOC concentration of 0.318 mg/L. In contrast, rainwater samples collected presented low but measurable levels of impurities, including both ionic species and organic matter. Despite being less pure than DI water, the rainwater samples remained within acceptable bounds for short-term electrolysis applications, particularly after filtration.

The conductivity of rainwater ranged from 14.48 to 25.50 µS/cm, representing less than 2.6% of the 1000 µS/cm upper limit commonly set for PEM electrolyzer feedwater. Commercial PEM systems typically require conductivity levels below 1 mS/cm, often maintained via ion exchange resins in recirculation loops [23]. Rainwater pH values were slightly acidic, between 6.07 and 6.76, which falls within the generally accepted operational window for PEM electrolysis (pH 5–8) [23]. Deviations outside this range could impair proton exchange capacity and catalyst stability. For comparison, tap water from the same laboratory had a conductivity of 364 µS/cm and a pH of 7.4, emphasizing rainwater’s relative suitability.

Organic content in rainwater was higher than in deionized water, with TOC values between 5.98 and 7.68 mg/L, but still below levels typically linked to immediate degradation of PEM components. Organic matter can nonetheless pose risks to system integrity over time, as it may adsorb onto catalyst surfaces–reducing electrochemical activity—or be oxidized into CO or CO₂, potentially impairing catalyst stability and accelerating degradation [23]. These values, however, might raise questions about whether organic compounds present in the water could contribute directly to hydrogen production during electrolysis. Prior studies have demonstrated that measurable hydrogen generation from organic species—such as formic acid [12,47,48,49], formaldehyde [50,51], or ethanol [52]—occurs only at much higher concentrations—typically above 100 mg/L—and under specific catalytic conditions [47,53]. Given that the TOC levels in this study are approximately one to two orders of magnitude lower, and no dedicated catalytic pathways for organic oxidation are present in the system, this study will not explore this possibility further understanding that there is no strong evidence suggesting otherwise, but it may be addressed in future studies focused on long-term degradation or gas purity impacts.

Certain ions observed—such as magnesium (Mg²⁺) and fluoride (F⁻)—are generally considered less harmful in PEM electrolyzer operation. As noted by Becker et al. [23], Mg²⁺ at levels up to 1 ppm shows little to no impact on membrane conductivity due to its non-redox nature and low mobility. Likewise, F⁻ is a weakly interacting anion with minimal influence on electrode kinetics. However, even ions considered chemically inert can influence product gas purity, especially in high-efficiency or recirculating systems where small amounts may accumulate over time. Although ISO 14687 [33] does not impose limits on feedwater quality, it sets stringent thresholds for contaminants in the hydrogen gas product—such as CO < 0.2 ppm and NH₃ < 0.1 ppm—highlighting the importance of proper system-level purification and monitoring.

In contrast, more reactive ions like Cl⁻ and Na⁺ are associated with membrane fouling, electrochemical corrosion, and catalyst degradation. Their long-term presence, even at moderate concentrations, can reduce component lifespan, particularly in systems with high recirculation and limited water refreshment. To mitigate these risks, a deionizer filter was integrated into the experimental system to reduce contaminant load before entering the PEM stack.

Taken together, these results emphasize that evaluating rainwater for electrolysis applications requires attention not only to total impurity levels, but also to the specific chemical behavior and reactivity of the ions and organic compounds present. While short-term performance may be unaffected, long-term operation with untreated or minimally treated rainwater could lead to performance losses or irreversible damage without appropriate mitigation strategies.

3.2. Electrolyzer Performance

The second test series evaluated how each water type performed in actual hydrogen production. The tests used a fixed-voltage, variable-current profile (up to 26 A at 4.2 V) to calculate hydrogen output and system efficiency at multiple load points.

3.2.1. Efficiency Calculations and Comparative Results

The efficiency was calculated following Equations (2)–(5) in Section 2.3. These metrics enable a robust comparison of energy conversion rates for both water types.

The efficiency curve in Figure 4 plots hydrogen production efficiency as a function of applied power for both water types.

It is possible to see that the deionized water had its efficiency ranging from 59.23 to 65.60% with an average of 62.39%. The rainwater, on the other hand, had a slightly higher average value, reaching 63.25% and having its efficiency ranging from 59.90 to 68.01%.

3.2.2. Performance Interpretation

Even though most literature does not support explicitly beneficial doping effects from typical waterborne ions, the higher efficiency recorded for rainwater, despite its greater impurity content, suggests that trace levels of specific ions may act as mild conductivity enhancers. Becker et al. for example emphasize that cations such as Na⁺, Mg²⁺, and Ca²⁺ generally degrade membrane performance by increasing ohmic resistance and accelerating corrosion [23], which might not represent an immediate harm, but a midterm point of attention.

On the other hand, the presence of chloride (Cl⁻) and sodium (Na⁺) ions—though measured within moderate levels—can pose long-term degradation risks if not mitigated through filtration or ion exchange. Becker et al. [23] report that chloride concentrations above 500 ppm significantly accelerate platinum dissolution, causing a measurable loss in electrochemical surface area and long-term catalyst degradation. Although the Cl⁻ levels detected in the present study were around 10 mg/L, this remains far below the reported threshold of close to 500 mg/L at which platinum dissolution and significant catalyst degradation have been observed [23].

Sodium ions (Na⁺), similarly, may compromise membrane performance if present in elevated concentrations. According to Becker et al. [23], Na⁺ at concentrations exceeding 10 ppm can begin to displace protons in the ion exchange membrane, resulting in decreased conductivity and increased cell voltage. While the sodium levels found in the rainwater samples analyzed here were far below that threshold, accumulation over repeated cycles could lead to membrane fouling and efficiency losses.

4. Discussion

The comparative performance of deionized water and untreated rainwater in a PEM electrolysis system revealed promising results, particularly with respect to the feasibility of rainwater as a substitute feedstock. While deionized water remains the industrial standard due to its predictably low conductivity and lack of impurities [18], this study demonstrated that filtered rainwater, despite its variable chemical composition, can deliver comparable operational efficiency when compared to the more expensive deionized water.

Limited to short-term experiments, long-term durability remains a key uncertainty, since impurities such as Na⁺ and Cl⁻ can gradually accumulate and lead to membrane fouling, catalyst corrosion, or reduced proton conductivity [23]. Future work should include extended operation to capture degradation pathways.

In direct response to the first research question, the rainwater samples exhibited higher conductivity and higher total organic carbon than deionized water, as well as measurable concentrations of chloride, nitrate, sodium, and sulfate ions. These characteristics reflect greater ionic and organic presence but remain within acceptable ranges for short-term PEM electrolyzer operation.

The increased performance observed with rainwater samples—most notably the slight decrease in required voltage and increased hydrogen yield—can be attributed to the presence of dissolved ions, which enhanced the conductivity of the electrolyte medium [23]. This outcome aligns with theoretical principles, as ionic species facilitate charge transport across the membrane and thus reduce the overall energy barrier required for electrolysis.

Thus, with respect to the second research question, the findings indicate that rainwater does influence system performance by improving conductivity and enabling marginal efficiency gains in hydrogen production under the tested conditions. Despite having more impurities than deionized water, rainwater’s ionic content positively contributed to electrochemical activity, challenging the assumption that only ultrapure water is viable for PEM electrolysis.

However, while the presence of ions can offer electrical advantages, it also introduces risks to the long-term integrity of electrolyzer components. As highlighted in the rainwater tests, the formation of foams and residues—especially on the cathode—indicates that organic matter or trace metals may accumulate on electrode surfaces or clog internal flow channels over time. These effects, though minimal in short-term trials, could lead to decreased membrane performance, corrosion of metallic components, or scaling if not mitigated [21].

Additionally, the variability in rainwater composition-both geographically and temporally—raises questions about standardization and control. The conductivity of the rainwater samples tested in this study ranged from 43 to 261 µS/cm, and ion chromatography results showed fluctuations in nitrate, sulfate, calcium, and magnesium content. In industrial applications, such variability could introduce inconsistencies in efficiency, system behavior, and maintenance requirements. Therefore, future implementation of rainwater-fed electrolysis would likely require preliminary water profiling and potentially modular pre-treatment systems that can adapt to local environmental conditions. About the geography, only samples from Austria were analyzed. Since rainwater composition can vary significantly depending on industrial pollution, marine proximity, or seasonal effects [25], feasibility studies should be carry on to guarantee the site specificity or a broader sampling could be conducted to improve the generalizability of the findings.

Despite these limitations, the economic and environmental implications of using rainwater are significant. The need for high-purity water sources has been cited as a barrier to large-scale electrolytic hydrogen production, particularly in arid regions or isolated installations where distilled or deionized water is either unavailable or costly to produce [18]. Rainwater, being naturally available and often cleaner than surface or seawater, represents a low-cost, low-carbon alternative that aligns well with the decentralized deployment of renewable hydrogen production systems.

Moreover, the results support the hypothesis that electrolyzer technology may tolerate a broader range of water qualities than previously assumed, provided that basic filtration is ensured and contaminants remain within manageable limits. This challenges the conventional dependence on ultrapure water in electrolysis design and invites future exploration of system robustness under less-than-ideal conditions.

In recirculating electrolyzer systems, impurities may concentrate over repeated cycles. While this effect was not measured here, accumulation of Na⁺, Ca²⁺, and Cl⁻ could eventually exceed safe thresholds and reduce efficiency [23], which should be explored in a long-term analysis.

From an economic perspective, rainwater represents a low-cost alternative to deionized water. Industrial supply costs for DI water can be in the order of $120 per m³ when purchased in smaller volumes, according to U.S. water service providers [54]. By contrast, harvested rainwater requires only minimal infrastructure and basic filtration, leading to substantially lower operating expenditures. A detailed cost–benefit analysis was beyond the scope of this study, but the qualitative advantage of rainwater is clear, particularly in decentralized or rural applications.

The findings of this study suggest that filtered rainwater could be integrated into electrolyzer systems without significant sacrifice in performance—particularly for small-scale, off-grid, or rural hydrogen applications where purification infrastructure is limited. However, the chemical complexity and variability of rainwater highlight the importance of continued monitoring and adaptive filtration to ensure long-term reliability. Further research is necessary to assess the long-term operational impacts of rainwater-fed systems, including studies on membrane degradation, fouling kinetics, and the effect of specific trace ions or organics on system lifetime and safety.

Finally, this work also raises the broader question of resource decentralization in hydrogen production: whether adapting systems to locally available water sources—rain, graywater, or treated municipal supplies—could play a role in the scalability of green hydrogen without overwhelming freshwater demand.