Bubble-induced versus thermodynamic voltage losses during pressurized alkaline water electrolysis

Abstract

Understanding how bubbles influence the efficiency of water electrolysis is crucial to achieve economically competitive hydrogen, generated by renewable energy sources, such as wind and solar power. Water electrolysis is typically performed at high pressures to reduce the cost of energy-intensive mechanical compression of the produced H2. Thus, a better understanding of how the absolute pressure affects electrochemical performance and bubble size is necessary. In general, bubble sizes decrease as the pressure increases. Using different-sized pillar-patterned Ni electrodes generated by Direct Laser Writing, the detached bubble sizes can be modified even at elevated pressures. As the pillar size increases, the bubbles become larger at all pressures investigated from 1 to 6 bar. At a current density of -25 mA/cm2, the cathodic potential increases with pressure according to the thermodynamic voltage losses given by the Nernst equation (~ 23 mV at p = 6 bar). Surprisingly, increasing the current density to 100 mA/cm2 leads to a reduction of the overpotential by up to ~ 60 mV. Reduced bubble sizes at increased pressures minimize the losses caused by the bubbles, thereby compensating for the thermodynamic voltage penalty. Applying the Buckingham Π-theorem enables the derivation of dimensionless numbers to characterize the ratio of bubble-induced and thermodynamic voltage losses

Figures (21)

• Figure 1: Overview of all laser-structured electrodes: (a) Microscopic and (b) confocal images of electrode structures with definition of the spatial period $\Lambda$. (c) Wetting behavior of electrode surfaces with applied water droplet and highlighted contact angles $\theta$
• Figure 2: (a) Bubble size distribution at 1a̅nd -50mAcm depending on the spatial period $\Lambda$, where $\Lambda = 0µm$ corresponds to the non-structured electrode (NSE). (b) Volumetric mean diameter $d_{30}$ as a function of the applied current density $j$ and spatial period $\Lambda$ as well as the variation of the contact angle $\theta$ depending on spatial period $\Lambda$. The lines serve only to guide the reader's eye.
• Figure 3: (a) Snapshots of the bubbles growing at the electrode with $\Lambda = 40µm$ at -100mAcm depending on the absolute cell pressure $p$. (b) Example of the bubble segmentation using StarDist and (c) resulting bubble size distributions at -100mAcm. (d) Nearly monodisperse bubble carpet at the electrode with $\Lambda = 100µm$ at -100mAcm and 6w̅ith highlighted bubble segmentations inside the region of interest. (e) Shift of bimodal to a nearly monodisperse bubble size distribution with increasing pressure for the electrode with $\Lambda = 100µm$ at -100mAcm.
• Figure 4: (a) Volumetric mean diameter $d_{30}$ of each size distribution measured as a function of pressure $p$ and electrode surface at -50mAcm. (b) Comparison of measured H2 volume $V_\mathrm{tot,\ce{H2}}$ of all segmented bubbles with the theoretical H2 volume $V_\mathrm{\ce{H2}, theor}$ calculated using Eq. \ref{['eq:theor_gas_volume']} at -50mAcm. The lines serve only to guide the reader's eye.
• Figure 5: Electrode potential $E$ over time $t$ at all studied current densities $j$ for an example electrode with a spatial period $\Lambda$ of 40µm featuring a switching point at -50mAcm.