Hae-Kyun Park, Dong-Hyuk Park, Bum-Jin Chung*

Department of Nuclear Engineering Kyung Hee University, 1732 Deogyeong-daero, Giheung-gu, Yongin-si, Gyeonggi-do 17104, Korea

Keywords:

ABSTRACT

1. Introduction

Hydrogen is one of the promising energy carriers replacing the conventional CO2emitting fossil fuels[1–4].The water electrolysis has been drawing attention as the carbon-free hydrogen production method, since it can use clean energy sources such as solar,wind and nuclear powers [5]. Electrochemical reaction generates hydrogen at the cathode during normal electrolysis. As shown in Fig. 1, beyond the range of normal electrolysis, the formation of partial(unstable)hydrogen film impedes the electrochemical reaction acting as an insulator. Then the electrolysis transforms into the partial contact glow discharge electrolysis (CGDE) where the current density decreases as the cell potential increases. Upon a further increase in the cell potential, the electrolysis experiences complete CGDE,which emits plasma at the cathode[6–9].The current density and the cell potential at the partial CGDE are known as critical current density (CCD) and breakdown voltage (Vb), respectively.Hence,the electrolysis with higher CCD or lower Vbis desirable for the efficient production of hydrogen [10]. It has been reported that the parameters affecting the CCD and the Vbare electrolyte conductivity(κ),electrolyte type,electrode geometry,temperature,etc.[7,8].The present work investigated the influences of κ on the CCD and the Vb, varying the concentration of H2SO4.Photographic analyses were performed to analyze hydrogen bubble behaviors throughout the experiments from normal electrolysis to partial CGDE.

2. Backgrounds

2.1. Nucleation of hydrogen bubble

Hydrogen bubble starts to nucleate at the cathode surface when the liquid solution reaches supersaturation condition during electrolysis[11,12].The supersaturation increases with increased overpotential,which leads to an increase of Gibbs free energy[13].The maximum Gibbs free energy is defined as the activation energy barrier for nucleation [14]. Hence, the increase of overpotential leads to an increase of nucleation sites[15].In a practical situation,hydrogen bubbles are formed at heterogeneous interface such as a small cavity at the cathode as depicted in Fig. 2, where the activation energy barrier for nucleation is low [16–20]. These hydrogen bubble behaviors at the cathode surface can affect the mass transfer rate at an electrolysis [14,21]. More detailed theory for the nucleation at gas-evolving electrode can be found in these papers[14–16].

Fig. 1. Typical cell potential–current curve of water electrolysis [7].

2.2. Influence of electrolyte conductivity on breakdown voltage

Several studies [7,8] have investigated the influence of κ (or conductance) on the onset of partial CGDE. Alteri et al. [7] measured the Vbusing a cylindrical tungsten varying the electrolytes;H2SO4, KOH, K2CO3and KCl. They reported that the Vbdecreased with the κ regardless of the electrolyte type. Gupta and Singh [8]also reported a decreasing trend of the Vband increasing trend of CCD according to the electrolyte conductance using a platinum(Pt) wire cathode and KHSO4solution. However, those existing works did not explain the phenomenological mechanism leading to the partial CGDE in the absence of the explanation in terms of the bubble behavior or observation. Instead, they either provided the other experimental parameters affecting the Vband the CCD such as electrolyte temperature,usage of surfactant,and electrode material[8]or rather focused on the complete CGDE phenomenon accompanying plasma emission beyond the Vb[7].

3. Experimental

3.1. Consist of electrolysis and experimental circuit

A two-electrode system was designed for the water electrolysis as shown in Fig.3.An upward-facing copper disk cathode of 0.01 m diameter was used as the working electrode generating hydrogen gas. As the counter electrode, a rectangular copper anode of 0.1 m× 0.2 m was used which was located 0.25 m away from the cathode. Then, the cell constant used in the present experiments was calculated as 25.55 m-1by averaging surface areas of anode and cathode. The experiments were carried out under atmospheric pressure and at room temperature (294 K). Sulfuric acid of 95%assay(CAS No.7664–93-9,extra pure grade)was used for the electrolyte. Concentration of the sulfuric acid was varied to vary the conductivity of the electrolyte as in Table 1. The electrolyte conductance (Siemens) data were taken from Ref. [22] according to the concentration of H2SO4.

The experiments were conducted under a potentiostatic condition by employing DC Power supply (Keysight N8952A). A data acquisition(DAQ) system was used to record and monitor the cell potential and the current. The cell potential was measured using NI9225 (National Instruments) and the large current was estimated by measuring the potential drop using NI9238 (National Instruments) across the shunt register (0.1 mΩ) instead of direct measurement.The electrolysis was performed with maximum cell potential of 106.23 V, which facilitated the electrolysis to experience the partial CGDE at 85.26 kA?m-2. The electric data were recorded with time interval of 10 ms. During the experiments,the high-speed camera (Phantom Lab 111 6GMono, Vision Research) captured the hydrogen bubble behaviors with 1000 frame per second.Three independent repetitive experiments were conducted for a single measurement. The average CCD and Vbvalues were used for the results, which had small discrepancies;within 4.3% and 2.4% for the CCD and the Vb, respectively.

3.2. Design of cathode electrode

Fig. 4 presents the cross-sectional diagram and top-view of the cathode electrode,which is identical to the authors’previous work[23]. A K-type thermocouple was imbedded close to the cathode surface (1 mm) to measure the temperature during the experiments. The cathode surface was polished using 2000 grit emery paper and the measured surface roughness (Ra) was 0.30 μm via atomic force microscopy (AFM, Veeco, Dektak 150), so that it can be regarded as smooth surface. In order to maintain the surface condition, the cathode surface was cleaned with deionized water and isopropyl alcohol(IPA)in sequence just before all the individual experiments.

3.3. Uncertainty analysis

Uncertainties of the measured CCD values were estimated applying data-reduction techniques [24]. The uncertainty of the CCD can be expressed as:

Fig. 2. Nucleation, growth and detachment of hydrogen bubble, reproduced from Ref. [14].

Fig. 3. Experimental apparatus and electric circuit.

Table 1 Conductivity of electrolyte according to H2SO4 concentration (293.5 K)

The UAwas calculated as 3.93×10-6m2,since the resolution of the milling machine was 2.5 × 10-4m. The UIwas calculated applying Ohm’s law:

where URand UEwere provided by the specifications of the devices,SCRD-R0001-5.0-H and NI9238 respectively. The measurement uncertainty,UEfor the Vbcan be directly calculated by applying gain and offset errors of NI9225.Consequently,the maximum fractional uncertainties of CCD and Vbwere 6.28% and 0.14%, respectively.

4. Results and Discussion

4.1. Detection of breakdown voltage and CCD

Fig. 5 shows E–I curve according to the time steps. During the normal electrolysis, the current density linearly increased as the cell potential increased.Hydrogen bubble size increased as shown in Fig. 5(a)–(d) due to the coalescence of hydrogen bubbles near the cathode surface with the increased hydrogen generation rate.However at a certain high cell potential, at around point (e), the current density reached a saturation. The current density and the cell potential at point (e) are the CCD and Vbrespectively, which are generally known as the maximum operation limits at the electrolysis[23].After the CCD,at region(f),the hydrogen film covered the cathode surface. Repetition of rapid formation and collapse of hydrogen film after the CCD caused the fluctuation of the current density. Fig. 6 synchronizes the recorded current density with the captured the hydrogen behaviors during the fluctuation regime at region(f).The formation of the hydrogen film hindered the additional hydrogen generation acting as an insulator, which significantly decreased the current density near zero. However, the current density soon recovered due to the replenishment of electrolyte as the hydrogen film collapsed. Then, the current density was maintained as the CCD level showing similar bubble behavior to the point (e), until subsequent hydrogen film was formed. This repetitive phenomenon threshed the current density, which commonly appeared in all the cases. The temperatures measured by the thermocouple imbedded at the cathode were 358–366 K at the CCD point and 366–369 K at region (f) according to the different overpotential of the experiments, which are small variations.Hence,the temperature effect of the present work can be regarded as a control variable.

Fig. 4. Schematic design and photograph of test apparatus.

4.2. Hydrogen bubble behaviors at normal electrolysis

Fig. 7 shows bubble behaviors at normal electrolysis according to the κ for a fixed current density of 49.02 kA?m-2.As the current densities were the same,the bubble generation rate was identical.However, as the κ decreased, the applied cell potential increased and the overpotential also increased due to the increased Ohmic loss. Then, the increased Gibbs free energy due to the increased overpotential additionally activated the smaller surface cavity resulting in the increased nucleation site.Consequently,the hydrogen bubble coalescence phenomenon near the cathode surface occurred easily due to the increased nucleation site as the κ decreased. This phenomenon facilitates the bubble rich layer to form the hydrogen film easily leading to the breakdown of normal electrolysis (partial CGDE).

Fig. 8 shows the bubble diameter with respect to the κ, which was measured using ‘Image J’ software for the randomly selected five still cuts from each case.The maximum relative standard deviation was 4.39%,which is good precision(case#2,red symbol).As the κ increased, at the initial stage, a rapid decrease in the bubble diameter was observed, and then the trend reduced to a pseudoasymptote, which is affected by the bubble rich layer. The rapid decreasing initial two points(circle and triangle)of Fig.8 are visualized as Fig.7(a)and(b)respectively,which showed thicker bubble rich layers than the others. Fig. 9 shows bubble population according to the κ. The maximum relative standard deviation was somewhat higher than that of Fig. 8 as 12.84%(case #1, black symbol), but still showed increasing trend clearly. The bubbles above the bubble rich layer in Fig.7 were counted manually to calculate the bubble population. Although the side-view as in Fig. 7 cannot identify the overlapped bubbles, the authors believed that it is enough to assess qualitative trends according to the κ because the bubbles above the bubble rich layer are selected as specimen.Figs. 8 and 9 are symmetric as the measurements were made at a constant current condition.

Fig. 5. Detection of the CCD and Vb together with hydrogen bubble behaviors according to the current density; case #4.

Fig. 6. Hydrogen bubble behaviors at fluctuation regime of current density.

Fig. 7. Still cuts of hydrogen bubbles varying the κ with fixed current density of 49.02 kA?m-2.

4.3. Influences of electrolyte conductivity on CCD and Vb

The CCD was increased with the increasing κ as shown in Fig. 10. Since the hydrogen film, which leads to the partial CGDE was formed at the higher current density as κ increased due to the reduced number of nucleation sites as discussed in the previous section. However, the increasing trend of the CCD seems to be saturated with ～104% augmentation of maximum hydrogen generation rate (CCD) for the given range of Fig. 10. Therefore, it is inferred that the bubble size and population have close relationship to the CCD.Fig.11 shows bubble behaviors at the CCD according to the κ. Bubble rich layers were commonly observed, which appeared just before the formation of the hydrogen film, and the actual hydrogen generation rate was higher at the higher κ case.

Fig. 8. Variation of hydrogen bubble diameter according to κ.

Fig. 9. Variation of hydrogen bubble population according to κ.

Fig. 10. Variation of CCD according to κ.

Fig. 11. Hydrogen behaviors at CCD according to κ.

Fig. 12. Variation of Vb according to κ.

The decreasing trend of Vbwas observed with the increasing κ as shown in Fig.12.The absolute values are different due to the different experimental conditions;electrolyte type,geometry of cathode, and system temperature, etc. However, the decreasing trends according to the κ are the same.Generally,the cell potential should be proportional to the current density.However,when the κ of the system decreases, the proportionality changes and the cell potential maintains high value even with the decreased current density.Thus, the measured Vbis still higher despite at the lower CCD due to the lower κ of the system.The decreasing trend of Vbalso seems to be saturated showing maximum drop of 74%.

5. Conclusions

The influence of κ on the CCD and the Vbin the water electrolysis were investigated varying concentration of H2SO4. In situ observations of hydrogen bubbles were performed to probe into the mechanism of the phenomena, which have been rarely done in the existing works.

The observation of hydrogen bubble behaviors showed that the coalescence of hydrogen bubble at to the cathode electrode directly affected the onset of partial CGDE. The variation of the current density is completely synchronized with formation and collapse of hydrogen film.

The maximum hydrogen generation rate, CCD increased as the κ increased, while decreasing trend of Vbwas observed. Despite under the identical current density condition, the overpotential increased with the lower κ, which increases the hydrogen nucleation site and augmented the bubble coalescence.The bubble coalescence characteristics were assessed indirectly by quantifying the size and the population of detached hydrogen bubbles. The bubble size and population curves according to the κ elucidated the trend of the CCD, which is closely related to the formation of bubble rich layer.In other words,the augmentation of bubble coalescence(high overpotential or low κ)caused the partial CGDE at a lower current density condition(reduced CCD).It is concluded that the κ affects the formation of bubble rich layer,which gave rise to increased CCD with the increased κ.

Both the tendencies of CCD and Vbshowed asymptote within the investigated range of the κ. Thus, these results may provide the optimization ranges of CCD and Vbat the water electrolysis.These results would only elucidate the phenomena at the horizontal upward-facing configuration of cathode.

Data Availability

Data will be made available on request.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was sponsored by the Korean Ministry of Science and ICT(MSIT)and was supported by nuclear Research&Development program grant funded by the National Research Foundation (NRF)(2021M2D1A1084838).

Nomenclature