3D-Printed Photoelectrochemical Cell and its Application in Evaluation of Bismuth Vanadate Photoanodes
3D-Printed Photoelectrochemical Cell and its Application in Evaluation of Bismuth Vanadate Photoanodes
Synthesis, cell design and testing
Bismuth vanadate (BiVO4) is proven to be a promising photocatalyst for water splitting. However, the effect of materials syntheses, electrode preparation and size of photoelectrode on the photocurrent output of BiVO4 photoanodes needs further investigations. In this study, three different BiVO4 nanoparticle synthesis were employed, namely hydrothermal (HT), HT in the presence of ethylene glycol (EG) and HT with the addition of hydrazine hydrate (HH). In addition, two molecular inks (Triton-X and ethyl‐methyl‐imidazole, EMI), were compared for the preparation of BiVO4 photoanodes using a simple doctor-blade technique followed by calcination at 450°C. The photoanodes (9 cm2 active surface) were then compared for their photocurrent density at AM1.5G illumination and 1.2 V (vs . standard hydrogen electrode (SHE)) bias in a specifically designed, three-dimensional (3D)-printed electrochemical cell. The highest photocurrent 0.13 ± 0.1 mA cm–2 was obtained with the EMI ink, whereas tenfold lower photocurrent was obtained with Triton-X due to the higher charge transfer resistance, measured by electric impedance spectroscopy (EIS). The photoresponse was reproducible and relatively stable, with only 8% decrease in five consecutive illumination periods of 1 min.
Sunlight-driven green hydrogen production is emerging as a promising contribution to carbon emission reduction, for which semiconductors as water splitting photocatalysts have arisen as potential materials to reach the worldwide climate goals at a low cost. As photoanode materials for oxygen evolution reaction (OER), bismuth-containing semiconducting metal oxides, such as BiVO4, Bi2WO6 and Bi2MoO6, have shown convincingly visible-light-driven photocatalytic activities due to their well-matching band gaps and redox potentials of valence/conduction band positions (1). In particular, BiVO4 demonstrated formidable photocatalytic performance for water splitting (2). However, BiVO4 often suffers from fast recombination of the photogenerated electron-hole pairs, which limits the electron flow in photoelectrochemical cells (3).
BiVO4 can be synthesised in various morphologies, sizes and crystal structures, which have direct impacts on their photocatalytic properties. It has traditionally been produced through solid-state processes, yielding fast-growing crystals with irregular morphologies and micron-scale sizes. Solution-based technologies, including aqueous, HT and solvothermal processes, have been developed in recent years to synthesise various BiVO4 nanostructures with improved charge transfer capabilities, such as nanoflakes, nanoellipsoids, nanowires, nanofibers, nanosheets, nanoplates and hyperbranched crystals (4–6). The size, shape and crystal structure of the BiVO4 photocatalysts are closely linked to the synthesis conditions, including reaction media, pH value, temperature and reactant concentrations (7). Furthermore, the presence of surfactants or organic additives, such as HH (8) and EG (9), assists BiVO4 formations with the desired polymorph. BiVO4 photoanode films in working conditions suffer from leaching due to their low structural affinity to the electrode. The presence of inks, such as EMI and Triton-X, gives mechanical stability and better attachment to the film (10).
A plethora of BiVO4 synthetic approaches and photoanode preparation methods are reported in the literature (8, 9, 11). The efficiency of such materials for hydrogen evolving photoelectrochemical cells is typically compared in terms of photocurrent densities generated, and by determining the overall solar-to-hydrogen (STH) and incident photon-to-current efficiencies (IPCE) (12), although in many cases the STH values are not correctly provided with no quantitative hydrogen measurement. In addition, a direct comparison of the photoelectrochemical performance is hampered by the different experimental conditions and cell design used in these reports. Standardising the experimental procedures, and in particular the cell design where the electrode materials are tested, is therefore necessary to enable a meaningful comparison of their photoelectrochemical performance.
In this study, a 3D-printed photoelectrochemical cell was purposely designed to allow a standardised comparison of the photocatalytic activity of BiVO4-based photoanodes, further applicable to any other type of photoelectrode. The cell was employed to compare the photocurrent outputs of photoanodes produced using two different methodologies, namely doctor-blading and electrodeposition. For the doctor-blading approach, two inks (Triton-X and EMI) and BiVO4 powders, synthesised in three different methods reported in the literature, namely HT with no additives, HT in the presence of EG and HT with the addition of HH, were mixed in various combinations to prepare the BVO4 films on the photoanodes. A photoanode synthesised through a conventional electrochemical deposition method was used to demonstrate the greater extent of the designed cell. A thorough characterisation of the best performing photoanodes, including powder X-ray diffraction (XRD), scanning electron microscopy (SEM) and electric impedance spectroscopy (EIS), was performed to elucidate key differences affecting their photoelectrochemical behaviours.
Photoelectrochemical Cell Design
The test cell used in this study (Figure 1 and Figure S1 in the Supplementary Information) was specifically designed to ensure a reproducible and standardised analysis of the photoanodes. The frame, which was 3D-printed in ultraviolet (UV)-cured resin materials, consisted of a base and a top part in a square shape (7×7 cm2). The base had a chamber of approximately 20 ml volume hosting two fixed electrodes, a 2×2 cm2 platinum-titanium mesh as the counter electrode and an Ag/AgCl reference electrode. Two holes were also included to introduce the electrolyte and allow operation in flow-cell mode. The top part was designed to host 5×5 cm2 flat photoelectrodes and had a square window of 3×3 cm2 to be exposed to illumination. The two parts were pushed onto each other through a rubber gasket by means of steel rods and plastic screws to ensure leak-proof adhesion of the electrode.
Photoanode Preparation and Testing
The BiVO4 photocatalysts synthesised in various HT conditions are referred to as BVO-HT, BVO-HH and BVO-EG, respectively, whereas the photocatalyst synthesised through the electrochemical deposition method is referred to as BVO-ED from now on. Each of the HT photocatalysts was first mixed with either Triton-X or EMI to form a colloidal paste, which was spread over a fluorine tin oxide (FTO) coated glass slide by doctor-blading technique to form a layer of 60 μm thickness and followed by calcination at 450°C in air, which ensures the formation of BiVO4 in a uniform crystal structure. Higher calcination temperatures (500°C and 600°C) resulted in deteriorated photocatalytic activity (Figure S2 in the Supplementary Information) and were not further investigated. At lower calcination temperatures (<400°C), BiVO4 exhibits a mixture of monoclinic and tetragonal crystal phases, which is not ideal for optimum photocatalytic activities (13), and the photoelectrode displays a dark residue due to the remaining ink that has not been calcined. The BVO-ED photoelectrode was prepared by electrodeposition of BiOI on the FTO followed by dropwise addition of vanadyl acetylacetonate and further calcination (14). The detailed synthetic methods, along with the photoanode preparation and testing methods are available in the Supplementary Information.
Three replicates of each photoanode were tested in the 3D-printed photoelectrochemical cell by linear sweep voltammetry (LSV) and chronoamperometry (CA) in 0.1 M phosphate buffer (pH 7) under dark and illuminated (AM1.5G) conditions. Prior to LSV analyses, each electrode underwent 3–6 cyclic voltammetry (CV) cycles for the same potential range until a stable profile was obtained. Among the photocatalysts synthesised with the HT method, LSV analyses (20 mV s–1 scan rate) showed a better photocurrent generation by the BVO-HH photoanodes prepared with both the Triton-X and EMI inks (Figure 2(a) and (b)). Therefore, BVO-HH based photoanodes became the focus of the study. The highest reproducible photocurrent of 0.44 ± 0.02 mA cm–2 was obtained with the BVO‐HH‐EMI at 1.8 V vs. SHE, whereas the current obtained under dark conditions (0.13 ± 0.02 mA cm−2) was comparable to that of a bare FTO electrode (Figure S3 in the Supplementary Information), strongly suggesting the generation of higher current was due to the BiVO4 photoanode film.
The photocurrent output of the BVO-HH photoanode was compared at 1.2 V vs. SHE in five alternate light/dark cycles of 60 s (Figure 2(c)). Significantly higher photocurrents were obtained using EMI ink (0.13 ± 0.01 mA cm–2) rather than Triton-X (0.014 ± 0.005 mA cm–2). The current output of the BVO-HH-EMI photoanode was twice higher than that obtained by the BVO-ED electrode (Figure S4 in the Supplementary Information), highlighting the potential of this synthesis method. It was also relatively stable over time, with 8.3% decay in the photocurrent after five consecutive illumination cycles (a total illumination period for 300 s) against 15.9% decay of the electrode prepared with Triton-X ink.
A stability test was then performed using the same cell operated in flow-cell mode, where a phosphate buffer solution was constantly recirculated to the cell by a peristaltic pump. CA analyses under prolonged illumination resulted in an irreversible 78% drop of photocurrent in 24 h (Figure S5 in the Supplementary Information), which recovered neither after interrupting and resuming illumination nor after switching off the applied potential overnight and changing the electrolyte. Longer photocurrent stability (over 100 h) has been reported using potentiostatically photopolarised BiVO4 electrodes (15), or sophisticated, multi-material photoanodes, such as plasma-etched NiOOH/BiVO4 (16). The photocurrent output and stability are highly dependent on the BiVO4 morphology and uniformity in the photoanode (17), which were thoroughly investigated by XRD, SEM and EIS measurements (Figure 3).
The XRD pattern of the as-synthesised BiVO4-HH confirms the material is a mixture of monoclinic (Joint Committee on Powder Diffraction Standards (JCPDS) No: 14−0688) and tetragonal (JCPDS No: 14−0133) crystal structures. The corresponding SEM image shows the BiVO4 synthesised through the HH method grew into a micro-spherical morphology with the diameter ranging from 2 μm to 7 μm (Figure 3(a)). A closer observation suggests these microspheres are assemblies of smaller particles. The calcination at 450°C during the preparation promoted a phase transition of tetragonal to monoclinic BiVO4 for both photoanodes as indicated by the absence of tetragonal diffraction peaks (Figure 3(e), BVO-HH-EMI and BVO-HH-Triton-X). In addition, two small diffraction peaks appeared at 27.3 and 27.9 degrees, implying the existence of a small quantity of monoclinic bismite Bi2O3 in both photoanodes. Compared to the as-synthesised BiVO4-HH, calcinations in the presence of EMI or Triton-X both led to destruction of the microspheres into a coral-like porous structures as shown in their SEM images (Figure 3(b) and 3(c)). This type of film structure is known to promote generation of photocurrent (18). The diameter of the ‘coral-bone’ is estimated to be ~120 nm for both photoanodes. The structure of the BVO‐ED powder (Figure S6 in the Supplementary Information) was similar to the one obtained with the HT method after the calcination step, and XRD pattern (Figure S7 in the Supplementary Information) also confirms the presence of a monoclinic phase with small peaks of impurities. Thus, characterisations of the photoanodes so far gave very similar results and cannot explain the differences in the photoresponses in terms of morphological structure.
To get insights into the electron transfer mechanisms, the BVO-HH photoanodes prepared with either the Triton-X or EMI inks, as well as BVO-ED photoanode, were further analysed by EIS under illumination at an applied voltage of 1.2 V vs. SHE. A sinusoidal wave with 10 mV amplitude was applied in the frequency range from 0.1 Hz to 105 Hz (10 steps per decade). The results were visualised as Nyquist plot (Figure 3(f), Figure S4 and Figure S8 in the Supplementary Information) and fitted to a Randles circuit to estimate ohmic drop, charge transfer resistance and pseudo-capacitance, as well as exponent of the constant phase element (CPE) (Table I and Table S2 in the Supplementary Information). The charge transfer resistance of the BVO-HH-EMI photoanode (330 ± 10 Ω) was one order of magnitude lower than that of the BVO-HH-TritonX (1700 ± 800 Ω), which correlates with the different photocurrent output of the two electrodes (Figure 2). It was also substantially lower than the charge transfer resistance of the BVO-ED photoanode (774 Ω). This suggests that the higher photocurrent output obtained with the EMI ink can be attributed to a more efficient electron transfer from the photocatalyst to the FTO electrode, and lower electron-hole recombination than with the Triton-X ink.
|Photoanode||RS, Ω||RCT, Ω||C, μF||α|
|BVO-HH-Triton-X||22 ± 3||1700 ± 800||140 ± 10||0.83 ± 0.04|
|BVO-HH-EMI||20 ± 1||330 ± 10||93 ± 3||0.81 ± 0.01|
The origin of the conductivity disparity between the two HT BVO-HH-EMI and BVO-HH-Triton-X photoanodes might be highly related to the nature of the ink materials. EMI is a type of ionic liquid, which is much more conductive than Triton-X, a non-ionic surfactant. It was indeed previously postulated that the non-volatile imidazolium ring of EMI can help in increasing ionic conductivity and electrochemical stability (19). Therefore, it is plausible that very small amounts of the EMI or Triton-X were maintained in the thin films after the calcination process (10). This implies that it is possible to improve the conductivity of a photoanode by optimising the amount of EMI in future studies.
Furthermore, the electrodes prepared with EMI ink showed good reproducibility of the results (Figure 3(f)), whereas a high deviation was obtained for the electrodes prepared with Triton-X (Figure S8 in the Supplementary Information), particularly for the charge transfer resistance. The pseudo-capacitance of the photoanode prepared with EMI (93 ± 3 μF) was also lower than that obtained with Triton-X (150 ± 10 μF), resulting in a calculated electron lifetime of 0.03 ± 0.00 s and 0.2 ± 0.1 s, respectively. This further suggests that in the photoanodes prepared with EMI, the electrons spend a shorter time on the depletion layer of the semiconductor (20) and thus the probability of electron-hole recombination is lower than in the photoanodes prepared with Triton-X.
After the 24 h stability test, the XRD pattern of the BiVO4 powder scratched off the BVO-HH-EMI contains additional diffraction peaks of unknown materials at 38.3 and 44.5 degrees (Figure 3(e)), suggesting the BiVO4 within the photoanode has been partially decomposed due to the corrosion by the electrolyte solution and photoirradiation (21, 22). This, along with the gradual deconstruction of coral-like porous structure as shown in the corresponding SEM image (Figure 3(d)), account for the significant drop of the photocurrent during the 24 h prolonged stability run (Figure S5 in the Supplementary Information). In the case of for BVO-ED, SEM images were taken after the short-term photocurrent measurements (maximum of 20 min). The porous-like morphology of the as-prepared photoanode evolves to the formation of flower-like structures. The negligible atomic concentration of vanadium in the corresponding EDX results suggests these flower-like structures are bismuth oxide species (Figure S7 in the Supplementary Information). This rapid degradation of the BVO photoanode is not observed even after the 24 h test done with the BVO-HH-EMI sample.
In conclusion, a photoelectrochemical cell was designed and 3D-printed to allow direct comparison of BiVO4-based photoanodes. The BiVO4 synthesised through the HH route (BVO-HH) gave the highest photocurrent output, and EMI outperformed Triton-X as ink for photoanode preparation due to its lower charge transfer resistance. The 3D-printed photoelectrochemical cell showed great robustness in measuring photocatalytic activities with its relatively easy experimental setup no matter its synthesis methodology as was demonstrated. The cell design can potentially be taken as a basic model for customisation, and most importantly, to standardise photocurrent measurements between different photoelectrode materials. The cell needs to be complemented by a protocol that can be replicated in different research laboratories. Further studies are required to improve photocurrent density (for example, by using co-catalysts such as cobalt phosphate) and stability in a long term (for example, adding protective layers), which is essential for industrial application.
X. Zhang, Z. Ai, F. Jia, L. Zhang, X. Fan and Z. Zou, Mater. Chem. Phys., 2007, 103, (1), 77 LINK https://doi.org/10.1016/j.matchemphys.2007.02.008
T. Saison, N. Chemin, C. Chanéac, O. Durupthy, L. Mariey, F. Maugé, V. Brezová and J.-P. Jolivet, J. Phys. Chem. C, 2015, 119, (23), 12967 LINK https://doi.org/10.1021/acs.jpcc.5b01468
C. Cheng, Q. Fang, S. Fernandez-Alberti and R. Long, J. Phys. Chem. Lett., 2021, 12, (14), 3514 LINK https://doi.org/10.1021/acs.jpclett.1c00713
H. He, S. P. Berglund, A. J. E. Rettie, W. D. Chemelewski, P. Xiao, Y. Zhang and C. B. Mullins, J. Mater. Chem. A, 2014, 2, (24), 9371 LINK https://doi.org/10.1039/C4TA00895B
B.-C. Xiao, L.-Y. Lin, J.-Y. Hong, H.-S. Lin and Y.-T. Song, RSC Adv., 2017, 7, (13), 7547 LINK https://doi.org/10.1039/C6RA28262H
F. Q. Zhou, J. C. Fan, Q. J. Xu and Y. L. Min, Appl. Catal. B: Environ., 2017, 201, 77 LINK https://doi.org/10.1016/j.apcatb.2016.08.027
R. A. Rather, A. Mehta, Y. Lu, M. Valant, M. Fang and W. Liu, Int. J. Hydrogen Energy, 2021, 46, (42), 21866 LINK https://doi.org/10.1016/j.ijhydene.2021.04.060
M. Mousavi-Kamazani, J. Mater. Sci.: Mater. Electron., 2019, 30, (19), 17735 LINK https://doi.org/10.1007/s10854-019-02123-0
S. Nikam and S. Joshi, RSC Adv., 2016, 6, (109), 107463 LINK https://doi.org/10.1039/C6RA14700C
E. A. Mohamed, Z. N. Zahran and Y. Naruta, J. Mater. Chem. A, 2017, 5, (15), 6825 LINK https://doi.org/10.1039/C7TA00156H
B.-X. Lei, L.-L. Zeng, P. Zhang, Z.-F. Sun, W. Sun and X.-X. Zhang, Adv. Powder Technol., 2014, 25, (3), 946 LINK https://doi.org/10.1016/j.apt.2014.01.014
Z. Chen, H. N. Dinh and E. Miller, “Photoelectrochemical Water Splitting: Standards, Experimental Methods and Protocol”, SpringerBriefs in Energy Series, Vol. 344, Springer Science and Business Media, New York, USA, 2013, 126 pp
M. F. R. Samsudin, S. Sufian, R. Bashiri, N. M. Mohamed and R. M. Ramli, J. Taiwan Inst. Chem. Eng., 2017, 81, 305 LINK https://doi.org/10.1016/j.jtice.2017.09.045
K. R. Tolod, S. Hernández, M. Castellino, F. A. Deorsola, E. Davarpanah and N. Russo, Int. J. Hydrogen Energy., 2020, 45, (1), 605 LINK https://doi.org/10.1016/j.ijhydene.2019.10.236
R.-T. Gao and L. Wang, Angew. Chem. Int. Ed., 2020, 59, (51), 23094 LINK https://doi.org/10.1002/anie.202010908
R.-T. Gao, D. He, L. Wu, K. Hu, X. Liu, Y. Su and L. Wang, Angew. Chem. Int. Ed., 2020, 59, (15), 6213 LINK https://doi.org/10.1002/anie.201915671
I. Khan, S. Ali, M. Mansha and A. Qurashi, Ultrason. Sonochem., 2017, 36, 386 LINK https://doi.org/10.1016/j.ultsonch.2016.12.014
N. Kiama and C. Ponchio, Surf. Coat. Technol., 2020, 383, 125257 LINK https://doi.org/10.1016/j.surfcoat.2019.125257
A. Syairah, M. H. Khanmirzaei, N. M. Saidi, N. K. Ferhana, S. Ramesh, K. Ramesh and S. Ramesh, Ionics, 2019, 25, (5), 2427 LINK https://doi.org/10.1007/s11581-018-2603-6
J. Ângelo, P. Magalhães, L. Andrade and A. Mendes, Appl. Surf. Sci., 2016, 387, 183 LINK https://doi.org/10.1016/j.apsusc.2016.06.101
D. K. Lee and K.-S. Choi, Nat. Energy, 2017, 3, (1), 53 LINK https://doi.org/10.1038/s41560-017-0057-0
S. Zhang, I. Ahmet, S.-H. Kim, O. Kasian, A. M. Mingers, P. Schnell, M. Kölbach, J. Lim, A. Fischer, K. J. J. Mayrhofer, S. Cherevko, B. Gault, R. van de Krol and C. Scheu, ACS Appl. Energy Mater., 2020, 3, 10, 9523 LINK https://doi.org/10.1021/acsaem.0c01904
This work was performed on the framework of the Science Foundation Ireland (SFI) Pathfinder Award on “Hybrid Bio-Solar Reactors for wastewater treatment and CO2 recycling” (award nr. 19/FIP/ZE/7572PF). Wenming Tong and Pau Farràs acknowledge the financial support from INTERREG Atlantic Area programme (Grant reference EAPA_190_2016). Roberto González acknowledges financial support from INTERREG NPA programme (Grant reference 354). The SEM measurements in this work were performed in the Centre for Microscopy and Imaging at the National University of Ireland Galway, a facility that is funded by National University of Ireland Galway and the Irish Government’s Programme for Research in Third Level Institutions, Cycles 4 and 5, National Development Plan 2007–2013.
Download the Supplementary Information for this article (PDF, 911 KB)
G. Kumaravel Dinesh is working as a postdoctoral researcher at Ryan Institute, School of Chemistry, National University of Ireland Galway. He obtained his PhD in the Department of Chemical Engineering, National Institute of Technology Trichy, India, in 2016. His research work is mainly focused on nanomaterials for energy and environmental remediation, photocatalysis and photoelectrochemical studies of materials.
Paolo Dessì is the principal investigator of the SFI HyBioSol project and works as a postdoctoral researcher at Ryan Institute, School of Chemistry, National University of Ireland Galway. He holds a Masters degree in environmental engineering and a Marie Curie European joint PhD degree in environmental technology. His research focuses on the conversion of waste streams to valuable resources such as green chemicals and biofuels.
Wenming Tong is a postdoctoral researcher in ChemLight Group, School of Chemistry, National University of Ireland Galway. His research interests lie in the growths and optical properties of plasmonic nanoparticles, synthesis and characterisation of metal oxide nanoparticles, electrocatalysis and photocatalysis. He is closely associated with several projects within the research group, including European Union (EU) projects such as FlowPhotoChem, SEAFUEL and SOLAR2CHEM, as well as SFI HyBioSol project.
Roberto González Gómez is a postdoctoral researcher in the ChemLight group at National University of Ireland Galway with 9+ years’ hands-on research experience and academic training. His research is focused on the synthesis and characterisation of nanoparticles for water oxidation. González is currently the workpackage leader of two EU-funded projects HUGE and NEFERTITI.
Pau Farràs is director of the ChemLight group, Lecturer in Inorganic Chemistry and Deputy Lead of the Energy Research Cluster at the Ryan Institute, at National University of Ireland Galway. Farràs is currently the coordinator of three EU-funded projects SEAFUEL, SOLAR2CHEM and FlowPhotoChem. He is co-author of 45 papers with over 1200 citations.