A major limitation of the practical implementation of alkaline water electrolysis (AWE) is the substantial increase in overvoltage caused by gas bubbles adhering to, and accumulating on the electrode surface under high current-density operation. Such bubble accumulation inhibits ion transport and blocks active electrochemical sites, resulting in lowered energy efficiency. Because the extent of bubble coverage depends on both the applied current density and the geometric properties of the electrode, the bubble-induced contribution to total overvoltage can vary significantly from system to system. Therefore, when evaluating catalyst or electrode performance, it is essential to consider not only intrinsic catalytic activity, but also to carefully identify and quantitatively separate the bubble-related mass-transfer overvoltage from the overall cell polarization.
Our group previously developed the mass-transfer models in bubble layer based on electrochemical impedance spectroscopy (EIS) using simple Ni wire electrodes, combined with equivalent-circuit analysis for the anode/OER and cathode/HER, respectively [1, 2]. These studies enabled the isolation of the bubble contribution. For OER, the measured i-V curve deviates from the expected Tafel line at current densities above approximately 0.2 A cm
‒2
, indicating the onset of a mass-transfer limitation associated with bubble accumulation. A comparable deviation is observed for HER at current densities exceeding roughly 0.3 A cm
‒2
. In both cases, the mass‐transfer resistance was found to consist of two key components: diffusion resistance and ohmic migration resistance within the bubble layer.
The migration resistance 𝑅
sb
was successfully interpreted in terms of the thickness and void fraction of the bubble layer, the conductivity of the electrolyte, and a tortuosity model characterizing the ion-migration pathways through the inter-bubble gaps. Diffusion transport differs between OER and HER: for OER, the diffusion of hydroxide ions was modeled by a finite-length Warburg impedance element, reflecting the restricted diffusive boundary created by the bubble layer. For HER, two distinct inductive elements were identified at high and low frequency, suggesting that gas evolution dynamically enhances the reaction rate in a frequency-dependent manner. The charge-transfer resistance (𝑅
ct
) at the electrode surface was proven to scale inversely with current density, indicating that 𝑅
ct
becomes relatively insensitive to bubble coverage and that nearly all accessible electrode surface remains electrochemically active even beyond 𝑖
geo
= 0.5 A cm
−2
.
These two mass-transfer models were subsequently extended to a cell-type AWE configuration [3], operated in 2 M KOH at 30 °C. Both CV and EIS measurements were performed under control of anodic potential vs. RHE, allowing simultaneous acquisition of three i-V curves and three Cole-Cole plots corresponding to anodic potential, cathodic potential, and whole-cell voltage. From the three i-V curves, the cell voltage could be decomposed into seven distinct contributions: the theoretical decomposition voltage, the reaction overpotential and mass-transfer overpotential for both anode and cathode, the diaphragm resistance, and the mass-transfer overpotential arising near the diaphragm. Subsequent analysis of the Cole-Cole plots allowed the mass-transfer overpotential to be further separated into diffusion- and ohmic-type components. Through these procedures, we were able to quantitatively demonstrate that bubble-derived mass-transfer overpotential accounts for approximately 30–50% of the total cell overvoltage under high-current density conditions, depending strongly on electrode geometry and gas-release behavior. This finding highlights that bubble management—whether by electrode structuring, surface modification, hydrodynamic control, or cell design—can have an impact comparable to improving intrinsic catalyst performance. Accordingly, the methodology presented here provides a rational framework for diagnosing limiting mechanisms in practical AWE devices and may serve as a useful benchmark when screening electrode materials for high-current industrial hydrogen production.
Acknowledgments
A part of this study was supported by the New Energy and Industrial Technology Development Organization (NEDO, JPNP25002).
References
[1] H. Ikeda, R.Misumi, Y. Nishiki, Y. Kuroda, S. Mitsushima,
Electrochimica Acta
, 430, 141053 2022
[2] D. Kitajima, R. Misumi, Y. Kuroda, S. Mithushima,
Electrochimica Acta
, 502, 144772, 2024
[3] Y. Nishio, R. Misumi, S. Mithushima,
PRiME 2024
, I01B-2773, 2024
R. Misumi, S. Mitsushima· ECS Meeting Abstracts· 0 citations
Introduction
Alkaline water electrolysis (AWE) has attracted considerable attention for hydrogen production; however, the generated bubbles can cover the electrode surface at high current densities, significantly increasing the overpotential
1)
. Further research is required to understand the effects of the geometric surface structure on the bubble generation behavior and electrolysis performance. This study investigated the relationship between the Ni electrode surface structure and the hydrogen bubble generation behavior using electrochemical measurements and a microscopic video camera.
Experimental
Polycrystalline Ni rods (⌀3.0 mm) and hemispherical single-crystal Ni
2)
(⌀3.0 mm) were used as the working electrodes. The polycrystalline Ni rods had different surface roughness, achieved by polishing the bottom surfaces with papers of different grades (
G
p
= 240, 1000, 4000, and 8000). Single-crystal Ni electrodes were prepared with surface orientations of Ni(111), Ni(100), and Ni(110) using a controlled-atmosphere flame fusion method
2)
. The Ni(110) sample with the smoothest surface was used as the electrode in the experiments. The skewness of the probability density of the surface height (
S
sk
/ -) was measured using confocal laser scanning microscopy (VK-X160, Keyence Co., Ltd.).
Electrochemical measurements were performed using three-electrode electrochemical cells with 0.10, 0.50, and 1.0 M KOH solutions as electrolytes. A reversible hydrogen electrode (RHE) and Ni ring (⌀60 mm) were used as reference and counter electrodes, respectively. The current density (
i
/A cm
−2
) was normalized to the projected area of the working electrode surface. After pretreatment, linear sweep voltammetry (LSV) was conducted within a potential window (
E
) of 0.00 to −0.40 V at a scan rate of 0.10 mV s
−1
. The bubble formation behavior on the electrode surface was monitored using an inverted microscope video camera. To detect and track the time evolution of bubble generation and determine the number of bubbles generated, the bubble-generation images were analyzed using an original program
3)
. The
i
value at the onset point of bubble formation,
i
b
, was determined as the point at which the number of generated bubbles began to increase.
Results and Discussion
Fig. 1 shows the polarization curve and bubble behavior at different
i
values ((i)–(iv)) in 1.0 M KOH (
G
p
= 1000). Two linear regions were observed, with the transition point marked by an orange circle. The slope in the higher negative potential range was comparable to the Tafel slope of the Volmer–Heyrovsky reaction (120 mV dec
−1
)
2)
. The red circle in Fig. 1 indicates the onset point of bubble formation, which was situated near the transition point. Hydrogen bubbles were not observed in the lower negative potential range (i) and their generation began at
i
b
(ii). Thereafter, the number and size of bubbles increased (iii). In the higher negative potential range, the bubbles covered most of the electrode surface (iv). The potential at
i
b
is defined as
E
b
. Figure 2 shows the relationship between
E
b
and
G
p
.
E
b
increased with increasing
G
p
, indicating that a rougher electrode surface enhanced bubble nucleation.
S
sk
, which represents the deviation in height distribution from the mean plane of the surface, was introduced to clarify the factors that influenced this result. Fig. 3 illustrates the relationship between
E
b
and
S
sk
, and Fig. 4 shows images of the electrode surface for
S
sk
≈ 0 (rough) and
S
sk
< 0 (smooth). Upon polishing, large convex asperities are removed from the electrode surface, resulting in a smoother surface (
S
sk
< 0). Fig. 3 shows that
E
b
is smaller for electrodes with
S
sk
≈ 0 than for electrodes with
S
sk
< 0. Comparing the surface characteristics of electrodes with
S
sk
≈ 0 and
S
sk
< 0, those with
S
sk
≈ 0 had deeper concave features. Deeper concaves result in slower diffusion of the dissolved hydrogen produced by the electrolytic reaction, leading to a higher degree of supersaturation in the concave regions. Because a higher supersaturation enhances bubble nucleation,
E
b
is lower for electrodes with
S
sk
≈ 0.
Conclusion
The effects of the electrode surface characteristics on bubble generation during the hydrogen evolution reaction were investigated using a Ni electrode. The polarization curve exhibited two linear regions, and the potential at the onset of bubble formation (
E
b
) coincided with the point at which the linear slope changed.
E
b
increased as the convexity of the electrode surface decreased, whereas concave regions enhanced the degree of supersaturation and bubble nucleation.
References
1) D. Kitajima,
et al.
,
Electrochim. Acta
,
502
, 144772, (2024).
2) K. L. Varvaris,
et al
.,
J. Phys. Chem. C
,
127
, 14711-14722 (2023).
3) K. Toyama,
et al
.,
Electrochemistry
,
93(2)
, 027011 (2025).
Figure 1
C. Toba, R. Misumi, Gaurav Verma et al.· ECS Meeting Abstracts· 0 citations
