Numerical simulation of Marangoni flow around a growing hydrogen bubble on a microelectrode

A.M. Meulenbroek (Corresponding author), B.W.J. Bernts, N.G. Deen, A.W. Vreman

Research output: Contribution to journalArticleAcademicpeer-review

2 Citations (Scopus)
34 Downloads (Pure)


In this work, a numerical model for Marangoni flow around a growing hydrogen bubble on a microelectrode is developed and validated. The growth of the bubble is implemented using a body-fitted moving mesh method that moves the bubble interface based on the production of hydrogen at the cathode. The dissolved hydrogen that is produced at the cathode diffuses into the bubble. The geometry of the bubble is a spherical cap with a pinned contact line and a decreasing contact angle as the bubble grows larger. There are two cases: (1) a case in which the bubble grows on a microbubble carpet that grows during the evolution of the bubble, and (2) a case in which the carpet thickness is negligible. The purpose of this model is to simulate and investigate thermocapillary Marangoni convection during almost the entire bubble lifetime (nucleation and detachment are not included). We validate the model with data from two experiments reported in literature (Massing et al. 2019; Bashkatov et al. 2022; Bashkatov, 2022). The evolution of the bubble radius, the temperature profile around the bubble, and the profile of the Marangoni velocity, which is tangential to the bubble surface at 5 μm from the bubble, near the lower half of the bubble are accurately predicted. For the case with a bubble carpet, we investigated models with a growing carpet and fixed carpet thickness. We found that for a given bubble size the Marangoni velocity reduces with increasing carpet thickness. Stagnant bubble models are less complicated and computationally cheaper than the growing bubble model. For the case without a carpet, we compare the growing bubble model with two stagnant bubble models: one using constant current and another one using a transient current. A stagnant bubble model is recommended if the growth rate drb/dt<0.05uM where drb/dt is the instantaneous bubble growth rate and uM the average Marangoni velocity at the bubble interface. This is the case at a relatively late stage of the bubble evolution. At the early stage of the bubble evolution, the absence of bubble growth dynamics in the stagnant bubble models appears to have an effect on the velocity profile and the pressure part of the hydrodynamic force.

Original languageEnglish
Article number143457
Number of pages12
JournalElectrochimica Acta
Publication statusPublished - 20 Dec 2023


We acknowledge the members of the Alkaliboost project, in particular Thijs de Groot and Rodrigo Lira Garcia Barros for fruitful discussions. We acknowledge Gabriela Sanchez Bahoque for proofreading the manuscript. We acknowledge Leon Thijs for sparring over COMSOL. Regarding the implementation of the moving mesh we acknowledge Paul Salden from COMSOL who provided pivotal insight into remeshing procedures. Furthermore, we extend our gratitude to Gerd Mutschke of the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) for the continuous sharing of experimental validation data sets used in Figs. 6 , 8, 9, 11 and 13 . We acknowledge Aleksandr Bashkatov for critical insights in microelectrode experiments, for sharing measurement data used in Figs. 4–7 and for numerous fruitful discussions. We acknowledge the reviewers for their critical review which contributed to a better understanding of the influence of the microbubble carpet on Marangoni convection. This work has been carried out as part of Alkaliboost with topsector energy funding by The Netherlands Enterprise Agency (RVO) and funding by Nobian and HyCC .

FundersFunder number
Rijksdienst voor Ondernemend Nederland


    • Bubble growth
    • Electrolysis
    • Marangoni convection
    • Microelectrode
    • Moving mesh
    • Thermocapillary effect


    Dive into the research topics of 'Numerical simulation of Marangoni flow around a growing hydrogen bubble on a microelectrode'. Together they form a unique fingerprint.

    Cite this