Hydrogen Evolution in Battery Electric Vehicle Coolants During Accidental Leakage: The Impact of Corrosion Inhibitors and Electrical Conductivity

Featured Application

The findings of this study can inform the development of safer cooling fluids for battery electric vehicles by highlighting the importance of low electrical conductivity in reducing hydrogen gas generation during accidental coolant leakage.

Abstract

Efficient thermal management is critical to the performance and acceptance of battery electric vehicles (BEVs). In the event of coolant leakage, contact between conventional water–glycol coolants and polarized battery components may induce hydrogen evolution via electrolysis, posing a serious safety hazard. This study investigates the impact of copper corrosion inhibitors and coolant electrical conductivity on hydrogen gas formation through linear sweep voltammetry (LSV) using copper electrodes. Results indicate that commonly used corrosion inhibitors—Tolyltriazole (TTZ), Benzotriazole (BTZ), and Sodium Mercaptobenzothiazole (MBT-Na)—do not significantly reduce hydrogen evolution, even in synergistic combinations. On the other hand, lowering the coolant electrical conductivity markedly decreased hydrogen evolution, with a linear reduction in cathodic current observed in low-conductivity coolants due to the reduced ionic mobility of the electrolyte. Low-conductivity BEV coolant (86 µS/cm) presented a cathodic current density 96% lower than a high-conductivity ICE coolant (2577 µS/cm) at the same overpotential. These findings suggest that optimizing coolant conductivity is a more effective mitigation strategy than relying on corrosion inhibitor formulations.

1. Introduction

Battery electric vehicles (BEV) require a battery thermal management system (BTMS) to control the operating temperature and ensure a uniform heat distribution among the battery cells/packs [1]. Among the various BTMS configurations, indirect liquid cooling—using plates, jackets, and tubes arranged according to pack geometry—is the predominant method in current BEV designs [2,3]. Water–glycol coolants, initially developed for internal combustion engine (ICE) vehicles, have become the predominant thermal management fluids for BEVs due to their desirable heat transfer properties, simplicity, and straightforward integration into vehicle architectures.

In indirect cooling systems, the coolant circulates within plates and chambers without directly contacting the cells [4]. However, under abnormal conditions—such as accidental leakage, inadequate electrical insulation, or mechanical damage—the fluid may come into contact with the battery tabs, as illustrated in Figure 1. Since water–glycol mixtures are electrically conductive, its interaction with the electrical environment of battery packs may introduce safety concerns that are still poorly understood. In a pivotal study, Nilsson and Runefors argued that glycol- and water-based coolants can dissociate under electric fields found in battery systems, generating flammable gases—such as hydrogen—through electrolysis. This mechanism, often overlooked, may contribute to fire ignition prior to thermal runaway, especially when coolant leaks occur in confined spaces where local electrical breakdown or arcing is possible [5].

If the coolant contacts polarized surfaces, such as the battery tabs, two main reactions occur at the cathode. At lower overpotentials, the oxygen reduction reaction (ORR) takes place with the direct or indirect conversion of oxygen into water. At higher overpotentials, the hydrogen evolution reaction (HER) reduces H⁺, resulting in water splitting and H₂ gas generation, as described in Equation (1) [6,7].

$$2{\mathrm{H}}_2\mathrm{O}+{2\mathrm{e}}^{-}\to {\mathrm{H}}_2+{2\mathrm{OH}}^{-}$$

(1)

The amount of H₂ gas generated by HER depends on the material and surface finishing of the tab, as well as the nature of the aqueous solution [8,9]. In alkaline solutions, such as water–glycol coolants, the HER activity of copper is relatively low compared to other systems, since copper weakly binds hydrogen anions, and alkaline solutions slow HER kinetics due to the absence of more suitable proton donors (such as hydronium ions in acid solutions) [7,9]. Despite the slower HER kinetics, the continuous generation of hydrogen at copper tabs in contact with coolants presents a significant safety risk. Even at reduced reaction rates, the continuous gas formation can lead to substantial hydrogen accumulation, elevating the internal pressure of the cooling system, increasing the likelihood of a significant leakage event [10]. Moreover, the accumulation of hydrogen within the battery pack could potentially create an explosive atmosphere, since the energy needed to ignite hydrogen is 10× lower than that of other flammable gases, such as methane or propane, making it prone to ignition even from short-lived sparks [11,12]. These abnormal situations compromise the cooling system’s functionality, resulting in overheating and ultimately triggering an irreversible thermal runaway [13].

Despite growing research interest in glycol-based coolants for BEVs, investigations have predominantly focused on thermal performance and system integration, offering valuable insights into optimizing battery thermal management systems, although without addressing the electrochemical behavior of coolants under fault conditions [14,15]. To the best of our knowledge, no previous work has evaluated hydrogen suppression in water–glycol-based coolants.

This study provides an electrochemical assessment of hydrogen evolution in glycol-based coolants used in BEVs, focusing on the influence of copper corrosion inhibitors and coolant electrical conductivity. Unlike studies that address hydrogen evolution in other environments, this work evaluates HER suppression under high overpotential conditions in complex, commercially relevant coolant formulations. Two HER mitigation strategies are considered: (1) to modify the coolant composition to include HER inhibitors, and (2) to reduce the coolant’s electrical conductivity to inhibit the magnitude of the reaction. One possible approach to the first strategy is to refine the coolant’s corrosion inhibition package, since it commonly includes copper organic corrosion inhibitors such as azole-based compounds including Tolyltriazole (TTZ), Benzotriazole (BTZ), and Sodium Mercaptobenzothiazole (MBT-Na). These corrosion inhibitors are classified as mixed-type corrosion inhibitors, i.e., they are expected to reduce the cathodic corrosion reactions, including HER [16]. Therefore, optimizing the coolant corrosion inhibition package can potentially lower the hydrogen formation, improving the overall safety of the system.

In this context, this paper aims to unravel the contribution of Cu corrosion inhibitors and the coolant’s electrical conductivity on hydrogen formation in Water–glycol coolants. HER was measured through Linear sweep voltammetry (LSV), an electrochemical technique commonly used to assess HER in electrocatalysis and battery applications [17,18,19]. Since HER is a cathodic reaction, a linear potential ramp was applied in the cathodic direction, and the system’s current response was measured, revealing details of HER kinetics and allowing the comparative evaluation across different systems. The investigation was divided into three parts to decouple the influence of the Cu corrosion inhibitor concentration, Cu corrosion inhibitor type, and the coolant electrical conductivity on HER.

2. Materials and Methods

An outline of the three sections of this investigation is presented in Figure 2. Working with commercial glycol-based BEV and ICE coolants allows us to investigate the influence of corrosion inhibitors in two different coolant technologies and analyze a broader range of electrical conductivities, since ICE coolants have an electrical conductivity approximately 25 times larger than BEV coolants.

In the first study, we investigated the influence of different concentrations of Cu corrosion inhibitors on HER in coolants of fixed electrical conductivity. For that, BEV and ICE coolants were manufactured with different amounts of TTZ. In the second study, we analyzed the influence of different Cu organic corrosion inhibitors on HER. For that, a BEV coolant was manufactured with the same final concentration of three distinct Cu corrosion inhibitors: TTZ, BTZ, and MBT-Na, either individually or in combination, to explore any potential synergistic effects among them. Lastly, on the third study, we analyzed the hydrogen evolution in coolants with different electrical conductivities, by using BEV and ICE coolants pure and diluted in monoethylene glycol (MEG). To decouple the effects of electrical conductivity and the quantity of Cu corrosion inhibitors, we adjusted the concentration of Cu corrosion inhibitors in the diluted solutions, to match that of the original coolant.

2.1. Materials Preparation

Copper rods of 0.5 cm² were mounted in epoxy resin, polished to P1200 using SiC paper, cleaned with ultrapure water, and dried with a compressed air gun. Two different commercially available coolants concentrates were used, one designed for ICE and another for BEV thermal management systems; their properties are listed in

Table 1. Summary of ICE and BEV coolants properties.

Property	ICE Coolant	BEV Coolant
Base fluid	Water/Ethylene glycol mixture	Water/Ethylene glycol mixture
Water content (%)	50	50
Freezing point (°C)	−37	−38
Boiling point (°C)	110	112
Electrical conductivity (µS/cm)	2000–8000	<100
Principal corrosion inhibitor	Organic acids	Organic acids
Corrosion inhibitor content (w%)	1–6	<1
Neutralizing base	NaOH	Amine
pH	7.0–9.0	7.0–9.0
Density at 20 °C (g/mL)	1.068	1.067
Kinematic viscosity at 20 °C (mm2/s)	4.3	3.9
Thermal conductivity (W/m·K)	0.42	0.42
Heat capacity (kJ/kg·K)	3.3	3.4

. Both are glycol-based coolants consisting of a MEG base with organic additive technology (OAT) package, a proprietary coolant concentrate based on a mixture of aliphatic diacid and monoacid organic corrosion inhibitors. The coolants were received as concentrated products and were then diluted 50 v/v% in ultra-pure water (0.05 μS/cm, 1.5 ppb of total organic carbon). All reagents used in this research (MEG, TTZ, BTZ, and MBT-Na) were obtained from the antifreeze production suppliers and follow the industry quality standards.

2.2. Solution Characterization

The electrical conductivity of the coolant solutions was measured using a pH meter (SevenExcellence™, Mettler Toledo, Columbus, OH, USA) with an InLab 741-ISM conductivity electrode, following the ASTM D1125 [20]. The pH was determined with a pH meter (Seven2Go pro, Metter Toledo) with an InLab Expert Go-ISM pH electrode, following the ASTM D1287 [21].

2.3. Electrochemical Characterization

The electrochemical characterization was performed in triplicate, using a VersaStat3 potentiostat (AMETEK Princeton Applied Research, Oak ridge, TN, USA). Linear sweep voltammetry (LSV) was performed in a 100 mL three-electrode cell, using a platinum sheet as the counter electrode, an Ag/AgCl (3 M KCl) as the reference electrode (3.247 V vs Li/Li⁺), and a copper rod as the working electrode. All the measurements were performed at 25°C. After the open circuit potential stabilization, the cathodic LSV was carried from OCP to -5 V (vs. Ag/AgCl), with a sweep rate of 10 mV/s. The overpotential range used in this study was selected to approximate localized electrochemical conditions that may occur during coolant leakage or insulation failure in battery electric vehicles (BEVs). The selected range ensures that hydrogen evolution can be distinguished from oxygen reduction, while avoiding the high gas evolution rates that can interfere with accurate current measurements.

2.4. Reproducibility

One representative LSV curve of each coolant solution is presented, while the values of the absolute current recorded at −5 V (vs. Ag/AgCl) are presented along with their standard deviation. All raw linear sweep voltammetry (LSV) data files and corresponding conductivity measurements are available from the authors upon reasonable request to support reproducibility and further analysis.

3. Results and Discussion

3.1. Effect of Copper Corrosion Inhibitor Concentration on HER Suppression

In this study, coolants standard BEV and ICE coolants were manufactured with different amounts of Cu corrosion inhibitor (TTZ), to evaluate the influence of the inhibitor concentration on the suppression of hydrogen evolution. A series of solutions were prepared with TTZ concentrations at 20%, 50%, 100%, 200%, and 430% relative to a standard reference concentration (a baseline against which all relative percentages are compared). The final pH and electrical conductivity of each solution are presented in

Table 2. Summary of BEV and ICE coolants solutions with different TTZ concentrations and their corresponding pH and electrical conductivity at 25 °C.

Coolant	Tolyltriazole (Relative %)	pH	eConductivity at 25 °C (µS/cm)
BEV	20	8.16	89.63
	50	8.12	90.82
	100	8.11	86.43
	200	8.00	92.57
	430	8.01	109.14
ICE	0	7.54	2520.0
	50	7.62	2476.8
	100	7.54	2576.8
	200	7.50	2654.0
	430	8.40	2623.3

. Minimal variations in pH and electrical conductivity are a consequence of TTZ addition.

TTZ is a well-described copper corrosion inhibitor. It forms a chemisorbed protective film on the copper surface, reducing both copper oxidation and the cathodic reactions [22,23]. Thus, different cathodic current evolution in coolants with different TTZ concentrations could be expected. LSV curves of BEV and ICE coolants with different TTZ concentrations are presented in Figure 3a,b, respectively. The current evolution of all tested solutions follows the same pattern: at lower overpotentials, up to—2 V vs. Ag/AgCl, the absolute value of current initially increase due to the oxygen reduction reaction (arrow 1) reaching a plateau due to its limiting diffusion in the electrolyte (arrow 2) [24]. With an increase in overpotential, the absolute value of current increases linearly due to the hydrogen evolution reaction (arrow 3).

In BEV solutions, the variation in TTZ concentration suppressed the oxygen evolution reaction at lower overpotentials. However, at high overpotentials, it was not effective in suppressing or decreasing the hydrogen evolution reaction. In ICE coolants, the solution without TTZ presented a lower onset overpotential for the hydrogen evolution, as shown in the inset of Figure 3b (yellow curve). Solutions containing TTZ present lower current values at −1.4 V (vs. Ag/AgCl), suggesting some cathodic inhibition at lower overpotentials. These results indicate that TTZ contributes to some cathodic inhibition of HER, although only at lower overpotentials. At higher overpotentials, this difference is offset, and the final current value at −5 V is comparable. All the curves presented fluctuations in the current evolution at higher overpotentials due to intense hydrogen evolution on the electrode surface. When comparing the current density values of the coolants with 100% TTZ, the standard BEV coolant exhibited a current 96% lower than that of the standard ICE coolant. This difference can be attributed to the different coolants’ formulation and electrical conductivity, as discussed further in the Section 3.3.

In conclusion, all coolants tested show similar or higher absolute current densities at −5 V (vs. Ag/AgCl) compared to the 100% TTZ solution. If TTZ concentration were effective in suppressing HER at high overpotentials, we would expect a decrease in the absolute current density values, which was not observed. These findings suggest that the protective chemisorbed layer formed by TTZ on Cu alloys is likely at high overpotentials and is not able to prevent the HER.

3.2. Effect of Different Copper Corrosion Inhibitors on HER Suppression

In this study, BEV coolants were prepared with TTZ, BTZ, or MBT-Na, since those compounds are widely used in coolant formulations to prevent Cu corrosion, and their protection mechanism is well stablished [25]. The corrosion inhibition mechanism of the azoles is based on forming a strong adsorbed–chemisorbed protective film that forms complexes with Cu⁺ ions, preventing the metal from reacting with the environment and forming species that lead to copper dissolution (mainly CuCl²⁻) [26]. TTZ is a derivative from BTZ, having a methyl group attached to one of the nitrogen atoms, while MBT-Na has a thiazole ring and a mercapto group.

Table 3. Summary of BEV coolants with different copper inhibitors and their corresponding pH and electrical conductivity at 25 °C.

Coolant	Cu Inhibitors	Inhibitor Content (Relative %)	pH	eConductivity at 25 °C (µS/cm)
BEV	TTZ	100	8.11	86.43
	MBT-Na	100	8.26	108.13
	TTZ	50	8.25	101.98
	MBT-Na	50
	BTZ	100	8.14	97.60
	TTZ	50	8.10	95.31
	BTZ	50

Note: TTZ: Tolyltriazole; BTZ: Benzotriazole; MBT-Na: Sodium Mercaptobenzothiazole.

summarizes the coolants used in this study along with their final pH and electrical conductivity.

In order to obtain coolants with different corrosion inhibitors, but in the same electrical conductivity range, the total amount of corrosion inhibitors was kept constant at 100% relative to the TTZ amount present in the commercial BEV coolant. The synergistic effect of TTZ with BTZ or MBT-Na was also investigated.

The LSV curves of coolants with different copper organic corrosion inhibitors are presented in Figure 4. The current evolution of coolants containing TTZ, BTZ and TTZ + BTZ are similar, indicating comparable suppression of cathodic processes. In coolants containing MBT-Na, a slight decrease in the absolute current associated with limiting oxygen diffusion is noticeable (around −1.5 V). This difference is offset at higher potentials, and in the HER region, coolants containing MBT-Na present the higher absolute current density values among all tested coolants. Additionally, no synergy is observed from the combination of TTZ with BTZ or MBT-Na.

In conclusion, none of the mixed-types copper corrosion inhibitors used in this study—TTZ, BTZ, and MBT-Na—effectively suppressed hydrogen formation, in contrast to what is described in the literature. Previous studies have reported differences in the adsorption ability of TTZ, BTZ and MBT-Na with synergistic effects when combined [27,28,29]. Loo et al. showed that in aqueous solution, TTZ is more strongly absorbed on Cu surfaces than BTZ, possibly due to differences in solubility that regulates the adsorption [27]. Altaf et al. reported a higher corrosion inhibition efficiency of MBT in respect to BTZ due to the structural difference between them, since copper has a higher affinity with towards the sulfur present on the MBT structure [30]. Nonetheless, as our study involves a complex glycol-based medium instead of an aqueous solution, the unique properties of the coolant may alter the adsorption behavior of the inhibitors, necessitating further investigation to understand their performance in this specific environment. Additionally, corrosion inhibitors studies that observe the decrease in cathodic reactions, such as HER, are often conducted at low overpotentials, whereas at high overpotentials, corrosion inhibitors are likely desorbed or corrosion-inhibiting layers are broken, compromising the protective ability of the azole/copper layer, and failing to prevent the hydrogen evolution reaction [31,32].

3.3. Effect of Coolant Electrical Conductivity on HER

In this study, coolants were progressively diluted in MEG, and the amount of Cu corrosion inhibitor in each solution was adjusted to match that of the undiluted coolants. The final pH and electrical conductivity of each solution are presented in

Table 4. Summary of coolant dilutions and their corresponding pH and electrical conductivity at 25 °C.

Coolant	Coolant Concentrate (w%)	Monoethylene Glycol (w%)	pH	eConductivity at 25 °C (µS/cm)
BEV	100	0	8.11	86.43
	75	25	8.05	71.05
	50	50	8.04	50.44
	25	75	7.95	28.70
	10	90	7.80	13.21
ICE	100	0	7.54	2576.80
	75	25	7.51	2040.70
	50	50	7.47	1430.60
	25	75	7.43	712.20
	10	90	7.31	299.10
	2.75	97.25	7.38	86.40

. As the electrical conductivity of a solution depends on its ionic concentration and mobility, diluted solutions showed reduced electrical conductivity [33].

The LSV cathodic scan of the polarized copper electrodes in progressive dilutions of the BEV and ICE coolants is presented in Figure 5a,b, respectively. The onset potential for hydrogen evolution, i.e., the overpotential at which the HER starts slightly increased with the progressive dilution of the coolants. In Figure 5b, it can be observed that ICE solutions containing less than 25% of coolant (electrical conductivity bellow 712 µS/cm) did not present fluctuations in the current evolution at higher overpotentials due to the reduced hydrogen evolution at the electrode surface.

The absolute current density recorded at −5 V (vs Ag/AgCl) was plotted versus the solution’s electrical conductivity, and it is presented in Figure 6. Progressive dilutions from both BEV and ICE coolants resulted in reduced current densities. The relation between the current density at −5 V (vs Ag/AgCl) and the electrical conductivity of the solutions follows a linear correlation for lower conductivities (BEV coolants) and transitions to an exponential relation at higher conductivities (ICE coolants). These results can be explained by the fact that when the copper electrode is polarized in the coolants, the kinetics of HER can be limited by three factors: the activation energy, the concentration polarization, and the ohmic loss due to the low electrical conductivity of the electrolyte [24]. Activation energy, which is the energy necessary for the reactions to initiate, depends on the electrode material and electrolyte nature. For HER in alkaline electrolytes, it involves two electrochemical steps: the adsorption of H_ads and OH_ads (Volmer step) and the recombination of adsorbed H_ads to produce H₂ (Tafel or Heyrovsky step) [34]. The slower step determines the rate of reaction when under activation polarization. The kinetics of HER can also be limited by concentration polarization, where the rate of HER is limited by the availability of reagents in the vicinity of the electrode/electrolyte interface, resulting in an HER rate controlled by mass transport. In our experiment, this is the case for the oxygen reduction reaction but not for HER, as seen in Figure 5. Finally, in electrolytes with low electrical conductivity (i.e., high electrical resistivity), ions that transport the current encounter high resistance, thereby reducing the rate of the reactions. This effect, known as ohmic polarization, is the dominant factor and the responsible for decreasing HER in the diluted coolants, as shown in Figure 5 [35]. On the other hand, on ICE coolants, the current is no longer solely governed by the ohmic polarization. Instead, the activation energy and the concentration polarization become more significant and contribute to determining the final current density, resulting in a non-linear relation, as presented in Figure 6.

These results show that solutions with reduced electrical conductivity considerably decrease the HER and highlight the importance of using low electrical conductivity coolants as a strategy to mitigate the risk associated with hydrogen formation in BEVs.

3.4. Future Research Directions

Future studies should validate our findings under application-relevant conditions. Static immersion test, in which battery cells or modules are submerged in coolant to simulate insulation failure scenarios, recommended by Sandia National Laboratories, allows the evaluation of the coolant electrolysis across different battery designs, states of charge, and overpotentials [36]. Additionally, circulating leak simulations can reproduce flow and pressure conditions typical of real-world systems. These dynamic tests have shown that coolant circulation is a major accelerator for hydrogen generation, highlighting the need for further exploration under operational conditions [37]. The growing recognition of coolant electrical conductivity as a safety parameter, as formalized in the forthcoming Chinese standard GB 29743.2, shows the demand to further study dedicated BEV coolants from laboratory to real-world scales [38].

4. Conclusions

This paper investigated the hydrogen evolution in glycol-based coolants used in battery electric vehicles, focusing on the influence of copper corrosion inhibitors and coolant electrical conductivity. The hydrogen evolution was evaluated by analyzing the cathodic current evolution in a linear sweep voltammetry test.

Results revealed that the use of common copper organic corrosion inhibitors (TTZ, BTZ, and MBT-Na) had no effect in suppressing the hydrogen evolution, regardless of concentration and synergetic combinations. This observation suggests that the effectiveness of these mixed-type corrosion inhibitors may be limited to lower overpotentials, and their behavior may be more complex in glycol-based solutions. In contrast, the coolants’ electrical conductivity has a significant impact on hydrogen formation. When comparing coolants of equal TTZ concentration, the cathodic current density at −5 V (vs. Ag/AgCl) was 96% lower in a low-conductivity BEV coolant (86 µS/cm) than in a high-conductivity ICE coolant (2577 µS/cm). Further, progressive dilutions of both BEV and ICE coolants showed that as electrical conductivity decreased, the cathodic current decreased linearly in low-conductivity ranges and exponentially at higher conductivities. These findings confirm that ohmic polarization, rather than activation energy or mass transport, is the dominant mechanism limiting the hydrogen evolution reaction (HER) under these conditions.

Therefore, while the use of azole-based corrosion inhibitors offers little benefit in suppressing HER at high overpotentials, reducing the electrical conductivity of coolants emerges as a clear and effective strategy to mitigate hydrogen accumulation and the associated safety risks in BEV cooling systems. As in this study all measurements were conducted under static and controlled laboratory conditions, they do not fully replicate the complex and dynamic environments present in operational battery systems. Future works should focus on evaluating hydrogen evolution under realistic conditions and further characterize HER in glycol media.