Magnetic Field-Assisted Control of Phase Composition
and Texture in Photocatalytic Hematite Films
Myeongwhun Pyeon, Vanessa Rauch, Daniel Stadler, Mehmet Gürsoy, Meenal Deo,
Yakup Gönüllü, Thomas Fischer, Taejin Hwang, and Sanjay Mathur*
We report the influence of external magnetic fields applied parallel or perpendicular to the substrate during plasma chemical vapor deposition (PECVD) of hematite (α-Fe2O3) nanostructures. Hematite films grown from
iron precursors show pronounced changes in phase composition (pure hematite vs. coexistence of hematite and magnetite) and crystallographic textures depending upon whether PECVD is performed with or without the influence of external magnetic field. Static magnetic fields created by rod-type (RTMs) or disk-rod-type magnets (DTMs) results in hematite films with anisotropic or equiaxed grains, respectively. Using RTMs, a superior photo-electrochemical (PEC) performance is obtained for hematite photoanodes synthesized under perpendicularly applied magnetic field (with respect to substrate), whereas parallel magnetic field results in the most efficient hematite photoanode in the case of DTM. The experimental data on microstructure and functional properties of hematite films show that application of magnetic fields has a significant effect on the crystallite size and texture with preferred growth and/or suppression of grains with specific texture in Fe2O3films. Investigations on the water splitting
properties of the hematite films in a photoelectrochemical reactor reveal that photocurrent values of hematite photoanodes are remarkably different for films deposited with (0.659 mA cm2) or without (0.484 mA cm2) external magnetic field.
1. Introduction
Electromagnetic field assisted synthesis and processing of advanced materials offers a less explored terrain despite the promise of generating materials with unprecedented microstructure and func-tionality. Static external magnetic fields have been studied as alternative fabrication and structuring pathways to optimize the functional materials characteristics.[1–6]For instance, magnetic Co and Ni nanostruc-tures synthesized in external magnetic fields resulted in enhanced magnetic properties due to a coherently aligned domain structure.[7] Similarly, the in flu-ence offield strength on the morphology of Ni particles could be demonstrated for samples synthesized under external mag-neticfields.[8]In case of gas phase deposi-tion, the reported data are related to chemical vapor deposition (CVD) of carbon nanomaterials that exhibited increasing film crystallinity under external magnetic fields[4,9,10]or on influencing the properties
of transition metals used as catalysts in carbon nanotube (CNT) growth.[2,3]For the
latter, a deformation of the metal (e.g., Ni) nanodroplets formed during the process leads to promoted growth kinetics in direction of the applied magnetic field and thus enabling a more defined colinear arrangement of CNTs. Moreover, self-assembly of iron nano-particles formed by hotfilament CVD of Fe(CO)5resulting in the
formation of anisotropic iron nanoclusters have been reported as well.[11] The role of external magneticfields on plasmas was
exploited at differentfield strengths, resulting in confinement and contouring of plasma sheaths at smallfield strengths used for p-type doping in 2D materials[12] or on larger scales for plasma fusion reactors.[13] In addition, particle formation and
motion in confined plasmas is extensively studied mostly from a theoretical viewpoint focusing on dust particle movement and accumulation.[14–16]For instance, investigations on the mobility of paramagnetic clusters in a magnetically confined plasma revealed field strength dependent location either below (magnetic field flux 40 mT) or above (magnetic field flux 120 mT) the main plasma cloud.[17] In this study, a
magnetically confined oxygen plasma was used for the decomposition of Fe(CO)5 to deposit Fe2O3[18,19] on F:SnO2
Dr. M. Pyeon, V. Rauch, D. Stadler, Dr. M. Deo, Dr. Y. Gönüllü, Dr. T. Fischer, Prof. S. Mathur
Institute of Inorganic Chemistry, University of Cologne
50939 Cologne, Germany
E-mail: sanjay.mathur@uni-koeln.de M. Gürsoy
Chemical Engineering Department SelScuk €Universitesi
42075 SelScuklu, Konya, Turkey Dr. T. Hwang
Heat Treatment R&D Group
Korea Institute of Industrial Technology (KITECH) 113-58, Siheung-si
Gyeonggi-do, Republic of Korea
The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adem.201900195.
DOI: 10.1002/adem.201900195
(FTO) substrates. Systematic variations of film growth times and magnetic field orientations performed to examine the field effects on the morphology of as-deposited films revealed that a preferred growth was initialized through the field-assisted deposition already in amorphous thin films. X-ray diffraction data of the annealed film showed that the applied magnetic fields exert a strong effect on the microstructure of hematite deposits that modulated the performance of iron oxide thin films as photoanode in PEC water splitting reactions.[18,19]
2. Experimental Section
Iron pentacarbonyl (Fe(CO)5, Acros Organics) was used as iron
source without further purification. Pure oxygen was fed into the reaction chamber as the reactive plasma gas. Operation parameters of PECVD were adopted and modified from earlier studies.[18,19]Typically, 20 sccm of O2were fed into the reaction
chamber and plasma power of 20 W was applied in all cases. Film growth time periods of 1, 2, and 5 min were used for experiments carried out with and without magnetic fields (parallel and perpendicular to the substrate). Under these experimental conditions, a growth rate of 57(4) nm min1 was achieved.[20] As-deposited amorphous iron oxidesfilmswere annealedat750C for 2 h in air using a tube furnace to obtain crystalline hematite phase. Two types of experimental setups were designed where rod-type magnets (RTMs,Ø10 50 mm2, N45, nickel-plated) and disc-type magnets (DTMs, Ø30 6 mm2, Y30) were employed (purchased from MagnetMax, dogeo GmbH, Germany) in geometrical alignment illustrated in Scheme 1. In case of the RTMs, deposition time of PECVD was varied (1, 2, and 5 min) to elucidate growth behavior offilms under magnetic field assisted PECVD. In a second set of experiments, DTMs were employed to investigate different magnetic field geometries with distance between stacks of magnets maintained at 24 mm. This experi-mental configuration allowed fabricating thin-films under comparable gas phase conditions, with varying magnetic field intensities and orientations. A portable magnetometer (KOSH-AVA 5, WUNTRONIC GmbH, Germany) was used to determine magnetic field strength and orientation at the sample position (Table 1).
The vapor phase decomposition of Fe(CO)5in oxygen plasma
was carried out with and without static magnetic fields. Since plasma profile is significantly affected by reaction chamber geometry, thus influencing the film growth, reference samples
were produced in a zero field control experiment having the same geometries as their magnetic field-assisted counterpart. PEC activity of the prepared hematitefilms was evaluated in a 3-electrode PEC cell with 1.0 M NaOH (aq) as reported earlier.[18]
Current generated from the samples were recorded as a function of applied voltage by a potentiostat (Versastat 4, AMETEK) with and without illumination (Xenon lamp, 150 W, Oriel). The obtained current values were normalized to current density (active area0.64 cm2) and the applied voltages (SCE, saturated calomel electrode reference electrode) were converted to reversible hydrogen electrode (RHE). Morphology and crystal-linity of the prepared hematitefilms were analyzed by scanning electron microscopy (SEM, Zeiss Neon 40 CrossBeam) and X-ray diffractometery (XRD, STOE-STADI MP). For the phase analysis of PECVD deposits, Cu-Kαradiation (λ¼ 0.15418 nm) and Mo-Kα radiation (λ¼ 0.07107 nm) were used for experiments carried out with RTM and DTM, respectively.
3. Results and Discussion
3.1. External Field Influence of Rod-Type Magnets (RTM) Figure 1 presents photographs of a glow discharge generated with magneticfields present inside the PECVD chamber. Plasma generated with a radio frequency (r.f.) power of 20 W exhibited distinct differences in the shape of plasma plume depending on the magneticfield geometry, which consequently affected local plasma density. While RTM-par demonstrated a parallel magneticfield contribution of 40(1) mT and 0(1) mT perpendic-ular to the substrate, RTM-perp revealed a zero field point centered on the substrate with increasing perpendicular magnetic field contribution up to 10(1) mT (parallel and perpendicular) at the corner of the substrate. Hence, it appeared that the plasma on the substrate under parallelfield orientation (RTM-par, Figure 1a) was more localized than under
Scheme 1. Illustrations of the experimental setups in PECVD processes, employing (a) rod-type and (b) disk-type magnets.
Table 1. Sample nomenclature with respect to field geometry and setup presented in this work.
Field orientation Rod-type magnet (RTM) Disk-type magnet (DTM)
Zero field (Reference) RTM-0 DTM-0
ParallelkSubstrate RTM-par DTM-par
perpendicular field orientation (RTM-perp, Figure 1b). These observations together with describedfield strengths are in good agreement with qualitative magneticfield profiles obtained from thefinite element investigations performed using finite element method magnetics (FEMM).[21] The optical micrographs of
obtained films are presented in supplementary information (Figure S1, Supporting Information).
Hematite (Fe2O3)films deposited on transparent conductive
oxides (TCOs) are widely reported as working electrodes for PEC[22,23] and various efforts were made to improve their
Figure 1. Photographs of plasma generated under (a) parallel and (b) perpendicular field geometry, induced by rod-type magnets with corresponding qualitative FEMM plots demonstrated for RTM-par (c) and RTM-perp (d).
Figure 2.J–V curves plotted as a function of (a) magnetic fields at a fixed deposition time and (b) deposition times under specific magnetic field geometries.
performance through electronic doping[24–27]or modulation of surface properties.[28–30]The PEC activities of hematitefilms, deposited under different external magnetic field geometries, were comparatively analyzed in a 3-electrode system as a function of magneticfields and deposition times (Figure 2). In a short deposition process (1 min), the highest photocurrent density at 1.23 V vs. RHE was obtained from zerofield control sample (RTM-0, 0.406 mA cm2) compared to those deposited under the parallel field (RTM-par, 0.206 mA cm2) and the perpendicular magnetic field (RTM-perp, 0.105 mA cm2). However, when the deposition time was increased to 2 min, the perpendicular magnetic field produced the most efficient hematitefilms exhibiting high photocurrent of 0.659 mA cm2 (0.484 mA cm2for zero field experiment and 0.278 mA cm2 for parallelfield orientation). On the contrary, no favorable effect of magnetic fields on PEC performance was observed from 5 min process (0.406, 0.208, and 0.105 mA cm2 for RTM-0, RTM-par, and RTM-perp, respectively).
Field-dependent differences in morphology were evidently observed in the SEM analyses of films deposited for 2 min (Figure 3). While particles with irregular shapes and sizes were formed on zerofield sample, homogeneous grains of average size 70 nm were formed under parallel magnetic field
alignment, leading to a denser microstructure. A perpendicular magneticfield orientation produced facetted hematite particles (white arrows in Figure 3c) with dispersed crystallites and large intergranular voids. Moreover, the average grain size infilms grown in RTM-perp mode was larger than those grown in RTM-par. According to theoretical investigations, an accumu-lation of paramagnetic nanoparticles below the plasma cloud is likely during the application of small magnetic fields with perpendicular orientation to the substrate.[17]In this work,field
orientation and strength are in comparable regime and higher concentration of larger particles seemed to be present near the surface. Also, an alignment of magnetic nanoparticles along the field lines is probable as already reported for hot filament CVD processes.[11] It is widely accepted that high surface area of photoelectrode provides better performance in PEC water splitting that explains the need of nanostructured morphology with high specific surface area.[31] As expected, the PEC
performances decreased with gradually decreasing film porosity from RTM-perp through RTM-0 to RPM-par. More-over, the lowest onset potential (0.7 V vs. RHE) observed from the sample prepared under perpendicularfield is ascribed to enhanced water oxidation kinetics, supported by the lowest dark onset potential.
Figure 3. Surface morphologies of the hematite films deposited for 2 min (a) zero field (RTM-0), b) under parallel field (RTM-par), and c) under perpendicular field (RTM-perp).
Figure 4. Surface morphologies of the hematite films deposited under (a) parallel field and (b) perpendicular field with deposition time of 5 min. Tilted angle images of the films deposited under the perpendicular field for (c) 2 min and (d) 5 min.
While nearly no changes in PEC performances were observed for samples grown with increased deposition times in RTM-0 and RTM-par configurations, samples grown in RTM-perp revealed significant changes in PEC performance with respect to the process time. Irrespective of thefield orientation, larger crystallites were observed upon prolonging the deposition time to 5 min (Figure 4). However, crystallites measured for RTM-perp experiment were found to be larger than those observed for RTM-par experiments, emphasizing the existence of particle-size gradient within the magnetically confined plasma. Furthermore, an orthogonal growth direction colinear to applied magnetic field lines was visible after varying the deposition time from 2 to 5 min for RTM-perp configuration (Figure 4c and d). In contrast to 2 min deposition time, no crystalline facets were observed for longer deposition periods (Figure 3c and Figure 4b). Counterintuitively, prolonged deposition time with perpendicu-larfield orientation did not lead to increased PEC performances and instead showed a detrimental influence on the photocurrent values possibly due to charge accumulation and recombination at the grain boundaries[23]that became more pronounced due to grain coarsening observed during longer process time (Figure 4d). Consequently, a decreased photocurrent was observed for 5 min. Nevertheless, all samples prepared in perpendicular field orientation had almost same dark onset potentials, ruling out morphological features (increased rough-ness, etc.) as the origin for the increased PEC performance of films obtained with 2 min deposition time.
Given the most significant influence of magnetic fields on PEC performance observed in samples deposited for 2 min, XRD analysis was performed to exclude possible phase impurities as explanation, as recently reported.[32]The XRD profiles of all heat-treated hematite films showed typical crystalline features of hematite (PDF 84–0308) without any other secondary phases. One striking observation was the preferred growth along (110) crystallographic plane in RTM-perp samples that was corrobo-rated by the relative intensity of this reflexion (Figure 5). The superior PEC performance of thesefilms is attributed to the better electrical conductivity along (110) plane that is, e.g., four times higher than that along (104) plane.[33–35]In contrast, the
absence of (110) plane under the parallel field adjustment explained lowest photocurrent density of thefilm, compared to the other photoanodes. The RTM-0 control sample demon-strated phase pure hematite reflexions with the literature known intensity ratios, with PEC performance in the median range of analyzed samples.
3.2. External Field Influence of Disk-Type Magnets (DTM) For a better understanding of the influence of external magnetic fields on the hematite film growth, both parallel and perpendicularfield orientations were applied during the same process. For DTM-par and DTM-perp, the strength of perpendicular and parallel magnetic fields was recorded with a magnetometer and found to be 19.0(1) mT and 3.3(1) mT, respectively (Figure 6a and b). Therefore, the dominantfield was present in the arrangement of the DTMs parallel to the substrate.
Figure 5. XRD patterns of the hematite films deposited for 2 min under different magnetic fields. Asterisks () represent SnO2 (PDF 41-1445) from FTO substrates.
For DTM-perp, no parallelfield contribution was observed, while a perpendicular magnetic field of 33.4(1) mT was visible (Figure 6d and c).
A photograph of the PECVD process, carried out with the set of magnet stacks revealed a good agreement with qualitative finite element simulation of the used field geometry (Figure 7). In contrast to RTM-par field geometry, DTM-par field demonstrated a cone-like curvature in between the DTMs. In case of perpendicular field orientation, the applied magnetic field demonstrated larger inhomogeneity (Figure 7a).
The resulting surface morphologies of hematitefilms grown on FTO substrates are gathered along with an image obtained from zerofield control in Figure 8. In general, no remarkable morphological differences were observed for DTM-0 compared to RTM-0. Thus, all differences discussed in the following text are attributable to applied magneticfield effects. As observed for films grown in RTM-perp configuration, DTM-perp revealed increased amount of grain boundaries, however, no significant changes in crystallite sizes were observed between DTM-perp and DTM-par arrangements. Nevertheless, when the
Figure 7. a) A photograph taken during a PECVD process with the set of magnet stacks and (b) a qualitative finite element analysis of field geometry. The position of substrates and magnets are schematically inserted in (a).
Figure 8. Surface morphologies of the hematite films prepared on FTO substrates under different plasma environment, formed by the designed configuration with magnet stacks; a) zero field (DTM-0), b) parallel field (DTM-par), and c) perpendicular field (DTM-perp).
morphologies of the films produced under attractive and repulsive interactions are compared, the topography of the aggregates were found to be different. Films formed under parallelfield exhibited rather larger aggregates resulting in fewer grain boundaries (Figure 8b), more aggregates based on densified crystallites were observed in films grown under perpendicularfield alignment (Figure 8c).
XRD patterns of the hematite films on silicon wafers (incidence angle of X-ray beam¼ 2.5) demonstrated the formation of phase pure hematite in all experiments (Figure 9). However, DTM-par and DTM-perp samples showed a decrease in the intensity of reflexion corresponding to (110) plane, whereas DTM-0 exhibited both reflexions due to the (104) and (110) planes. Thus, a magnetic field-assisted suppression of (110) textured grains and thus a control over microstructure was achieved infilms grown in RTM-par, DTM-perp, and DTM-par configurations, while RTM-perp increased I(110)/I(104) ratio. For
RTM-0 and DTM-0 no changes in crystallographic structure were observed.
The photocurrent data observed from PEC experiments showed that the hematite films formed in the presence of magnetic fields out-performed the films obtained under zero field deposition. Specifically, a photocurrent density of 0.50 mA cm2 (DTM-par), 0.35 mA cm2 (DTM-perp), and 0.05 mA cm2 (DTM-0) were detected at 1.23 V vs RHE (Figure 10). Since the porosity of the films deposited under parallel and perpendicular magnetic fields appeared to be comparable, the smooth surface (Figure 8b) might be advanta-geous for charge transportation with reduced charge recombi-nation in the bulk whereas clear boundaries between primary particles (Figure 8c) could act as efficient recombination sites, possibly leading to decreased PEC activity.[30]
A comparative overview on recent developments in thefield of anisotropic hematite nanostructures is displayed in Table 2. Given an evident shift in the onset potential from 0.8 to 0.7 V (vs. RHE) and enhanced photocurrent density of 0.61 mA cm2
(against 0.484 mA cm2), RTM-perp sample were found to exhibit improved performances for non-doped or reduced iron oxide catalysts.
4. Conclusion
This study demonstrates for the first time, the influence of external magneticfields on the hematite films grown in a low-temperature plasma-enhanced CVD. Preliminary results on the crystallinity and photoelectrochemical properties of hematite films showed that the application of external magnetic fields affects the grain growth and densification processes manifested in the surface morphology and optical properties of film deposited in the presence and absence of external magneticfield. Moreover, the alignment of an external magneticfield parallel or perpendicular to the substrate apparently altered the precursor flux to the substrate that ultimately influenced the film morphology. Despite an evident increase in the length of columnar grains, higher growth rate in the presence of magnetic field cannot be presumed since this might be due to a preferential growth of anisotropic structures and may not correlate to an overall higher growth rates requiring more
Figure 9. The XRD patterns of DTM-0 (black), DTM-par (red), and DTM-perp (blue).
Figure 10. PEC performances of the hematite films grown with and without external magnetic field in PECVD process.
Table 2. Summary of recent publications focusing on PEC water splitting using anisotropic hematite nanostructures in 1 M KOH.
Photoanode Photocurrent density at 1.23 V [mA cm2] Onset potential [VRHE] Fe2O3Nanorods on FTO[25] 0.50 0.80
Sn-Fe2O3Nanotubes on FTO[25] 1.50 0.80
Fe2O3Nanorods on FTO[36] 0.31 0.80
Sn-Fe2O3Nanorods on FTO[36] 0.94 0.95
Hydrogen treated Fe2O3on FTO[37] 2.50 0.90
RTM-perp (this work) 0.61 0.70
precursorflux being converted into the solid film. Application of magnetic fields parallel and perpendicular to the substrate showed a significant effect on the grain texture with preferred growth and/or suppression of grains with specific texture differing in atomic densities in Fe2O3lattice. The work reported
here shows the importance of externalfields in influencing the microstructure and crystallographic properties of the hematite deposits that was demonstrated by significant enhancement in the photocurrent values infilms deposited under the influence of magnetic field when compared to zero field PECVD experiments. A deconvolution of plasma and magnetic field effects was not possible in the presented experimental approach, however; we are carrying out additional studies on thermal CVD to understand thefield influences. In addition, the variation of field strengths and use of non-magnetic precursors/ions is currently under progress that can provide further and new insights into the influence of external magnetic fields.
Supporting Information
Supporting Information is available from the Wiley Online or from the author.
Acknowledgements
The authors gratefully acknowledge the University of Cologne for providing the infrastructure and the German Science Foundation (DFG) for financial support in the frame of the priority program “Manipulation of matter controlled by electric and magnetic field: Towards novel synthesis and processing routes of inorganic materials” (SPP 1959). M.G. acknowledges The Scientific and Technological Research Council of Turkey (T €UBITAK) BIDEB 2212 and 2214/A for financial support. Dr. M.D. also acknowledges Science and Engineering Research Board (SERB), India for fellowship under SERB-Overseas postdoc fellowship scheme. D.S. acknowledges the financial support from the Evangelisches Studienwerk Villigst e.V.
Conflict of Interest
The authors declare no conflict of interest.
Keywords
hematite, magnetic field assisted PECVD, metal oxide semiconductor, nanostructures, PEC water splitting
Received: February 20, 2019 Revised: April 1, 2019 Published online: May 6, 2019
[1] J. Zhang, C. Pan,J. Phys. Chem. C 2008, 112, 13470. [2] C. Luo, Q. Fu, C. Pan,Sci. Rep. 2015, 5, 9062.
[3] C. S. Han, T. H. Lee, G. M. Kim, D. Y. Lee, Y. S. Cho,J. Korean Ceram. Soc. 2016, 53, 167.
[4] Y. Wang, J. Li, H. Long, H. Luo, B. Zhou, Y. Xie, S. Li, Q. Wei, Z. Yu, Surf. Coat. Tech. 2016, 292, 49.
[5] O. Guillon, C. Elsässer, O. Gutfleisch, J. Janek, S. Korte-Kerzel, D. Raabe, C. A. Volkert,Mater. Today 2018, 21, 527.
[6] J. S. Kim, B. C. Mohanty, C. S. Han, S. J. Han, G. H. Ha, L. Lin, Y. S. Cho,ACS Appl. Mater. Interfaces 2013, 6, 116.
[7] H. Zhao, B. Zhang, J. Zhang, L. Zhang, X. Han, P. Xu, Y. Zhou,J. Phys. Chem. C 2010, 114, 21214.
[8] H. Hu, K. Sugawara,Chem. Phys. Lett. 2009, 477, 184.
[9] Y. Wang, J. Li, N. Hu, Y. Jiang, Q. Wei, Z. Yu, H. Long, H. Zhu, Y. Xie, L. Ma, C.-T. Lin, W. Su,Mater. Res. Express 2018, 5, 035009. [10] X. You, Z. Yu, L. Shi, L. Wang,J. Cryst. Growth 2009, 311, 4675. [11] D.-H. Kim, H.-S. Jang, H.-R. Lee, C.-D. Kim, H.-D. Kang,Appl. Phys.
Lett. 2004, 85, 109.
[12] S. Wu, Y. Zeng, X. Zeng, S. Wang, Y. Hu, W. Wang, S. Yin, G. Zhou, W. Jin, T. Ren,2D Mater. 2019, 6, 025007.
[13] J. Ongena, R. Koch, R. Wolf, H. Zohm,Nat. Phys. 2016, 12, 398. [14] M. Puttscher, A. Melzer,Phys. Plasmas 2014, 21, 123704. [15] M. Puttscher, A. Melzer,Phys. Plasmas 2015, 22, 073701. [16] M. Puttscher, A. Melzer, U. Konopka, S. LeBlanc, B. Lynch,
E. Thomas,Phys. Plasmas 2017, 24, 013701.
[17] V. V. Yaroshenko, G. E. Morfill, D. Samsonov,Phys. Rev. E 2004, 69, 016410.
[18] A. Mettenbörger, T. Singh, A. P. Singh, T. T. Järvi, M. Moseler, M. Valldor, S. Mathur,Int. J. Hydrog. Energy 2014, 39, 4828. [19] A. P. Singh, A. Mettenbörger, P. Golus, S. Mathur, Int. J. Hydrog.
Energy 2012, 37, 13983.
[20] A. Mettenbörger, Y. Gönüllü, T. Fischer, T. Heisig, A. Sasinska, C. Maccato, G. Carraro, C. Sada, D. Barreca, L. Mayrhofer, Nano Energy 2016, 19, 415.
[21] D. C. Meeker, Finite Element Method Magnetics, Ver. 4.2 (25 Feb 2018), http://www.femm.info.
[22] D. A. Grave, N. Yatom, D. S. Ellis, M. C. Toroker, A. Rothschild,Adv. Mater. 2018, 30, 1706577.
[23] S. Chu, W. Li, Y. Yan, T. Hamann, I. Shih, D. Wang, Z. Mi,Nano Futures 2017, 1, 022001.
[24] Gurudayal, S. Y. Chiam, M. H. Kumar, P. S. Bassi, H. L. Seng, J. Barber, L. H. Wong,ACS Appl Mater Interfaces 2014, 6, 5852. [25] C. Li, A. Li, Z. Luo, J. Zhang, X. Chang, Z. Huang, T. Wang, J. Gong,
Angew. Chem. Int. Ed. Engl. 2017, 56, 4150; Angew. Chem. 2017, 56, 4214.
[26] Y. Ling, G. Wang, D. A. Wheeler, J. Z. Zhang, Y. Li,Nano Lett. 2011, 11, 2119.
[27] A. Kaouk, T.-P. Ruoko, M. Pyeon, Y. Gönüllü, K. Kaunisto, H. Lemmetyinen, S. Mathur,J. Phys. Chem. C 2016, 120, 28345. [28] D. Jiang, Q. Yue, S. Tang, L. Zhang, L. Zhu, P. Du,J. Catal. 2018, 366,
275.
[29] Z. Luo, T. Wang, J. Zhang, C. Li, H. Li, J. Gong,Angew. Chem. Int. Ed. Engl. 2017, 56, 12878; Angew. Chem., 2017, 56, 13058.
[30] M. Wang, H. Wang, Q. Wu, C. Zhang, S. Xue,Int. J. Hydrog. Energy 2016,41, 6211.
[31] M. Fang, G. Dong, R. Wei, J. C. Ho,Adv. Energy Mater. 2017, 7, 1700559.
[32] J. Leduc, Y. Goenuellue, P. Ghamgosar, S. You, J. Mouzon, H. Choi, A. Vomiero, M. Grosch, S. Mathur,ACS Appl. Nano Mater. 2019, 2, 334.
[33] S. Kment, P. Schmuki, Z. Hubicka, L. Machala, R. Kirchgeorg, N. Liu, L. Wang, K. Lee, J. Olejnicek, M. Cada,ACS Nano 2015, 9, 7113.
[34] O. Zandi, A. R. Schon, H. Hajibabaei, T. W. Hamann,Chem. Mater. 2016,28, 765.
[35] S. Chu, Y. Cui, N. Liu,Nat. Mater. 2017, 16, 16.
[36] J. Wang, C. Du, Q. Peng, J. Yang, Y. Wen, B. Shan, R. Chen,Int. J. Hydrog. Energy 2017, 42, 29140.
