Subscriber access provided by UNIV OF PITTSBURGH
Article
Nano-textured Pillars of Electrosprayed Bismuth Vanadate for Efficient Photoelectrochemical Water Splitting Hyun Yoon, Mukund G Mali, Jae-Young Choi, Min-woo Kim, Sung K Choi, Hyunwoong Park, Salem S. Al-Deyab, Mark T. Swihart, Alexander L. Yarin, and Sam S Yoon Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.5b00486 • Publication Date (Web): 09 Mar 2015 Downloaded from http://pubs.acs.org on March 11, 2015
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
Nano-textured Pillars of Electrosprayed Bismuth Vanadate for Efficient Photoelectrochemical Water Splitting Hyun Yoon1,†, Mukund G. Mali1,†, Jae Young Choi1, Min-woo Kim1,2, Sung Kyu Choi3, Hyunwoong Park4, Salem S. Al-Deyab5, Mark T. Swihart6, Alexander L. Yarin7,8,*, Sam S. Yoon1,* 1
School of Mechanical Engineering, Korea University, Seoul, 136-713, Republic of Korea 2
3 4 5
Green School, Korea University, Seoul, 136-713, Republic of Korea
Department of Physics, Kyungpook National University, Daegu, 702-701, Republic of Korea
School of Energy Engineering, Kyungpook National University, Daegu, 702-701, Republic of Korea
Petrochemicals Research Chair, Dept. of Chemistry, College of Science, King Saud University, Riyadh 11451, Saudi Arabia 6
Department of Chemistry & Biological Engineering, University at Buffalo (SUNY), Buffalo, NY, USA 7
8
College of Engineering, Korea University, Seoul, 137-713, Republic of Korea
Department of Mechanical and Industrial Engineering, University of Illinois at Chicago, Chicago, IL, USA
Abstract We demonstrate, for the first time, electrostatically sprayed bismuth vanadate (BiVO4) thin films for photoelectrochemical water splitting. Characterization of these films by X-ray diffraction, Raman scattering, and high-resolution scanning electron microscopy analyses revealed the formation of nanotextured pillar-like structures of highly photoactive monoclinic scheelite BiVO4. Electrosprayed BiVO4 nanostructured films yielded a photocurrent density of 1.30 and 1.95 mA/cm2 for water and sulfite oxidation, respectively, under 100 mW/cm2 illumination. The optimal film thickness was 3 µm, with an optimal post-annealing temperature of 550 °C. The enhanced photocurrent is facilitated by formation of pillar-like structures in the deposit. We show through modeling that these structures result from electrophoretic motion of submicron particles in the direction parallel to the substrate, as they approach the substrate, along with Brownian diffusion. At the same time, opposing thermophoretic forces slow their approach to the surface. The model of these processes proposed here is in good agreement with the experimental observations.
Keywords: Bismuth vanadate, solar water splitting, electrostatic spray deposition, nanoporous pillar-like structure, photocurrent density, thermophoretically-driven deposition instability *Corresponding author:
[email protected],
[email protected] †These authors contributed equally. 1 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1. Introduction Production of chemical fuels by solar photoelectrochemical (PEC) water splitting is a renewable alternative to traditional fossil fuel-based energy sources. In this process sunlight is used to photoelectrochemically split water and produce hydrogen gas on semiconducting photoelectrodes.1, 2, 3, 4 This, in effect, stores solar energy as stable chemical energy that can later be released in a fuel cell or by combustion. Further chemical processing can produce liquid fuels from this renewable hydrogen in combination with CO2. Relatively inexpensive metal oxides with high stability, such as, TiO2, SrTiO3, BaTiO3, and KTaO3 (UV-light active materials), or Fe2O3, BiVO4, and WO3 (visible-light active materials) have proven effective for PEC water splitting.1, 2, 3, 4, 5, 6, 7 Among these, TiO2 with a band gap of 3.0 eV for rutile and 3.2 eV for anatase phases has been the most popular choice. However, PEC performance of TiO2 is limited, because it is only photoactive at UV wavelengths; it does not absorb at visible wavelengths. To circumvent this shortcoming, TiO2 is often doped with other elements to extend its light absorbance into the visible wavelength range.8 Other semiconductors are photoactive in the visible-light region.5,
6,
7
However, these have other disadvantages, such as susceptibility to
photocorrosion, that have prevented them from displacing TiO2 as the material of choice. Recently, n-type semiconducting bismuth vanadate (BiVO4) has become one of the most promising materials for PEC water splitting with visible light. Furthermore, BiVO4 is nontoxic and chemically stable. BiVO4 exists in three crystalline phases with tetragonal scheelite, monoclinic scheelite and tetragonal zircon structures. The band gaps of tetragonal (scheelite) and monoclinic (scheelite) BiVO4 are 2.9 and 2.4 eV, respectively.9 The lower band gap of monoclinic BiVO4 makes it photoactive in the visible-light region of the solar spectrum.10 Efforts have been made to prepare BiVO4 films by various vacuum and non-vacuum deposition techniques, such as spin coating,11 reactive sputtering,12 spray pyrolysis,13 and metal organic decomposition.14 For pure BiVO4 films, low photocurrent values have been reported for water oxidation,
2 ACS Paragon Plus Environment
Page 2 of 30
Page 3 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
suggesting that they suffer from poor charge carrier separation.15 Hence, novel techniques are required to prepare BiVO4 films with higher quantum efficiencies for charge separation. We report, for the first time, the use of an electrostatic spray deposition (ESD) technique for the fabrication of BiVO4 fluffy pillar-like films. ESD is a non-vacuum coating technique that provides nearly 100% deposition efficiency, with well-dispersed, uniform, wettable, and fine droplets. As a result, ESD offers low material consumption and highly accurate targeting of charged droplets toward a substrate. In addition, the crystallinity, surface texture, film thickness, and deposition rate can be readily tuned by adjusting operating conditions, such as voltage, flow rate, precursor concentration, and substrate temperature.16, 17, 18 The experimental details are described in the Supporting Information and in Figure S1. Here, we have tuned the thickness and morphology of the deposited BiVO4 landscapes by controlling the spraying time and annealing temperature to optimize PEC water splitting performance, as characterized by high photocurrent; see Figure S2 for the water splitting apparatus.
2. Results and Discussion 2.1 Effect of the Effective Film Thickness on Crystallinity BiVO4 films with a Bi:V molar ratio of 1:1 were fabricated by ESD. This 1:1 ratio has been shown to be optimal for PEC water splitting.12 The substrate temperature was kept at 80 °C during ESD. The effective film thickness was varied from 0.1 to 5 µm by increasing the spraying time from 10 to 80 min. The optimal post-deposition annealing temperature was sought in the temperature range of 450 to 600 °C. Four representative samples, prepared using four different spraying times, are discussed here. These are denoted BiVO4-10, BiVO4-20, BiVO4-40 and BiVO4-80, corresponding to spraying times of 10, 20, 40, and 80 minutes, respectively. The effect of film thickness on water splitting was investigated.
3 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The PEC performance of BiVO4 is highly dependent upon its crystalline structure. As mentioned above, BiVO4 can have both zircon and scheelite crystal structures and the scheelite BiVO4 structure can have either tetragonal or monoclinic crystal symmetry.19 The monoclinic scheelite structure shows better visible light photo-activity than the tetragonal structure.20
Figure 1. XRD spectra showing the effect of post-annealing temperature on the crystallinity of BiVO4 thin films deposited by ESD on an ITO substrate at 80°C. Figure 1 shows the XRD pattern of the BiVO4 films obtained at four different annealing temperatures (450, 500, 550, and 600 °C). Diffraction peaks observed at 2θ values of 18.6, 28.6, 34.5, 39.5, 42.1, 45.4, 46.8, 52.9, and 58.2° correspond to the 110, 121, 200, 211, 150, 132, 240, 161, and 123 planes of monoclinic scheelite BiVO4, respectively, in good agreement with JCPDS card No. 14-0688. At 450 and 500 °C annealing temperatures, the heights of the peaks are relatively small, compared to the peaks observed for the 550 and 600 °C annealed samples. The sharper and stronger intensities observed at higher annealing temperatures indicate that these samples have higher crystallinity. Figure S3 shows XRD diffraction patterns with Rietveld refinement fits to BiVO4 films annealed at 450, 500, 550, and 600 °C. Crystallite size and lattice parameter data is summarized in Table S2. The values of unit cell
4 ACS Paragon Plus Environment
Page 4 of 30
Page 5 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
parameters a, b, and c were consistent with literature values for the body-centered monoclinic crystal structure (JCPDS-140688).
Figure 2. Raman spectra showing the effect of post-annealing temperature on the crystalline quality of nanostructured BiVO4 thin films deposited by ESD on an ITO substrate at 80°C.
The monoclinic scheelite crystal structure was further confirmed by the Raman scattering data shown in Figure 2. The Raman peaks found at 124, 210, 324, 363 and 824 cm-1 are typical of the monoclinic phase of BiVO4. The peaks at 124 cm-1 and 210 cm-1 are ascribed to the external modes of BiVO4.21 The peaks at 324 and 363 cm-1 correspond to the asymmetric and symmetric deformation modes of the VO43- tetrahedron, respectively.22 The peak at 824 cm-1 is attributed to the anti-symmetric V-O stretching. While, in Ref. 23, such stretching is observed at both 710 and 826 cm-1, we observed the asymmetric V-O stretching mode at only 824 cm-1. This type of single stretching mode has been previously reported.20 Higher crystallinity is observed at 550 °C rather than at 600 °C. Thus, we used the annealing temperature of 550 °C for the rest of the parametric studies.
5 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 3. Optical extinction spectra of ESD BiVO4 films deposited by ESD on an ITO substrate at 80°C and post-annealed at 550 °C.
Figure 3 shows optical extinction spectra of films produced using four different spraying times (10, 20, 40, and 80 min). The optical extinction of the uncoated ITO substrate (which has the highest transparency) is compared with that of the BiVO4-coated ITO substrates. Optical extinction, which includes contributions from both absorbance and scattering, increases with deposition time, mirroring the increase in film thickness. 2.2 Effect of the Film Thickness on Morphology The surface morphologies of the BiVO4 films (both side and top views) observed by HR-SEM are shown in Figure 4. Each row corresponds to one of the spraying times (10, 20, 40, and 80 min). The effective thicknesses of the films are estimated as 0.1, 0.3, 3, and 5 µm for the aforementioned spraying times, respectively. The side views clearly show that the films consist of nano-textured pillar-like
6 ACS Paragon Plus Environment
Page 6 of 30
Page 7 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
structures that do not seem to be well connected in the lateral direction. This type of structure can improve the photo-activity of the material because the high surface area of such fluffy landscapes provides more sites for H2 generation and shorter transport distances for photogenerated holes to the reactive sites. Thimsen et al. also reported the formation of a similar pillar-like structure via deposition of TiO2 droplets produced by a flame aerosol reactor.24 Their nano-textured pillar-like structure improved the PEC water splitting activity by an order of magnitude when compared with films with polycrystalline and granular type structures. Their work demonstrates the importance of forming nano-structures that drastically increase the surface area of a material. The porous nature of the pillar-like structure shown in Figure 4 provides the samples with an even greater surface area than they would have as dense pillars, further enhancing their photo-activity. A similar porous BiVO4 film was recently reported by Kim and Choi.25 They observed a nano-porous morphology in their BiVO4 films that produced higher surface areas and was effective in the suppression of electron/hole recombination and led to the generation of a higher photocurrent density.
7 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 4. Top view and cross-sectional view SEM images of electrostatically sprayed BiVO4 films deposited at Tsub = 80 °C and post-annealed at Ta=550 °C. The following notation is used: BiVO4-10 to BiVO4-80 corresponds to 10 min to 80 min of deposition, respectively.
8 ACS Paragon Plus Environment
Page 8 of 30
Page 9 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
X-ray photoelectron spectroscopy (XPS, Theta Probe base system, Thermo Fisher Scientific Co.) was carried out before and after PEC experiments, to check for the formation of adventitious photo electrodeposited materials after the PEC testing. Figure S8 compares the XPS of BiVO4 films tested before and after the PEC tests. The XPS spectra of pre-electrolysis films show the Bi 4f7/2 peak at 159.0 eV, V 2p3/2 at 516.8 eV, and O1s at 529.8 eV. These peaks suggest that Bi, V, and O are in their expected +3, +5, and -2 oxidation states, respectively.26 The XPS spectra after electrolysis are same as the preelectrolysis spectra demonstrating the chemical stability of the BiVO4 films and absence of adventitious deposition during electrolysis.
2.3 Effect of Film Thickness on Photocurrent Density To test PEC performance we measured the photocurrent density, a standard measure of water splitting performance, under AM 1.5 illumination for each film. Illumination of the BiVO4 film through the ITO substrate is referred to as “back” illumination while illumination of the film directly is called “front” illumination. In general, back illumination yields higher photocurrents than front illumination,12 and this has been the case in previous studies of BiVO4 films.27, 28, 29 Figures S9 and 5 confirm that back illumination of the BiVO4 produced higher photocurrent than front illumination. In back illumination, charge carriers (e-/h+ pairs) are generated near the ITO contact (deep in the BiVO4 bulk), while in the case of front illumination they are generated near the electrode surface. Holes are used for oxygen formation at the working electrode, while electrons must be transported to the counter-electrode via the ITO contact. The distance over which an electron must travel is, therefore, lower for back illumination. In a porous structure like that generated here, the transport distance for holes is always short, because holes only must reach the solid-electrolyte interface. Thus, reducing the average transport distance for electrons, by using back illumination, is advantageous.
9 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 30
Figure 5. Current-voltage (I-V) curves for the “back”-illuminated thin BiVO4 films measured under A.M. 1.5G simulated solar light (1 sun) in 0.5 M Na2SO4 solution.
The current-voltage (I-V) curves shown in Figures S9 and 5 were recorded at a scan rate of 10 mV/s using a 0.5 M Na2SO4 electrolyte solution. The incident light intensity was 100 mW/cm2. Potentials are shown both vs. RHE (Reversible Hydrogen Electrode) and Ag/AgCl, which are related as ERHE = EAg/AgCl + 0.059 pH + 0.1976, where, ERHE is the potential vs. the reversible hydrogen electrode, EAg/AgCl is the potential vs. the Ag/AgCl electrode, and pH is the pH value of the Na2SO4 electrolyte solution.27 For the BiVO4-10 and BiVO4-20 samples, the photocurrent densities (PCD) at 1.2 V (vs. Ag/AgCl) are 0.091 mA/cm2 and 0.14 mA/cm2, respectively. However, the PCD for the BiVO4-40 sample is dramatically higher, reaching a value of 1.3 mA/cm2 at 1.2 V (vs. Ag/AgCl). For the BiVO4-80 sample, the PCD decreased considerably to 0.30 mA/cm2 at 1.2 V (vs. Ag/AgCl). The optimal BiVO4 thickness results from the trade-off between increasing absorbance of thicker films and increased transport distance for the photogenerated charge carriers. Even though photogenerated holes do not have to be transported 10 ACS Paragon Plus Environment
Page 11 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
across the entire film thickness, as seen in Figure S9, the particles of which the nanoporous pillars are composed are larger for the BiVO4-80 sample than for the BiVO4-40 sample. Rettie et al. reported that a particle size of about 100 nm yields the optimal diffusion length for charge carriers in BiVO430. Here, the diameter of the BiVO4 particles in the BiVO4-80 (thickest) film is more than the optimal transport length of the charge carriers, which could be the reason for this samples’ lower photocurrent density.
Figure 6. Current-voltage (I-V) curves for the back-illuminated thin BiVO4 films measured under A.M. 1.5G simulated solar light (1 sun) in 1 M Na2SO3 sacrificial electrolyte solution.
Furthermore, PEC measurements for all the films were also carried out in the presence of a sacrificial reagent, Na2SO3, which acts as a hole scavenger. Sulfite oxidation is more favorable than water oxidation both thermodynamically and kinetically.25 Figure 6 shows typical current-potential curves that compare the PEC performance using Na2SO3 (1 M) as a sacrificial reagent, instead of Na2SO4. Measurements were conducted under the same experimental conditions as those of the water oxidation 11 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 30
(with Na2SO4). The photocurrent density increases then decreases with increasing coating time from 10 to 80 min. The spraying time of 40 min (BiVO4- 40) again yielded the highest PCD value of 1.9 mA/cm2. The trend of photocurrent vs. deposition time (or film thickness) for sulfite oxidation is similar to that for water oxidation. For the sulfite oxidation, the photocurrent onset potential shifts to a more negative value, to about -0.35 V, compared to -0.1 V for the water oxidation case. For sulfite oxidation, the photocurrent increased rapidly between -0.2 V and 0.2 V vs the reference electrode. At 0.7 V, the PCD value peaked at 1.95 mA/cm2. Table 1. Comparison of the present photocurrent results for BiVO4-40 with literature data. Some of the literature data was acquired under the identical experimental conditions, the intensity of incident light of 100 mW/cm2 and the potential = 1.2 V vs. Ag/AgCl. Incident light
Reactant solution
Photocurrent density
100 mW cm-2
0.5 M Na2SO4
1.30 mA cm-2
100 mW cm-2
1 M Na2SO3
1.95 mA cm-2
Spin coating
100 mW cm-2
0.5 Na2SO4
0.30 mA cm-2
[11]
Reactive sputtering
100 mW cm-2
0.5 M Na2SO4 + 0.1M K2HPO4
~0.60 mA cm-2
[12]
Electro-deposition (1.23V vs RHE)
100 mW cm-2
0.2 M K2HPO4
~1.8 mA cm-2
[25]
100 mW cm-2
0.5 M Na2SO4
0.63 mA cm-2
[31]
100 mW cm-2
0.1M K3PO4 +1 M H3PO4
0.44 mA cm-2
[32]
Electrophoretic deposition
---
0.5 M Na2SO4
~ 0.10 mA cm-2
[32]
Solution-cast Method (~1.7 vs RHE, front illumination)
100 mW cm-2
KOH (pH=13)
~1.1 mA cm-2
[34]
Preparation method
Electrostatic spray deposition
Polymer-assisted direct deposition method Metal–organic decomposition (- 1.0V vs Ag/AgCl)
12 ACS Paragon Plus Environment
Reference Present work: BiVO440
Page 13 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
Table 1 compares the photocurrent density values to those reported in previous studies. The present BiVO4 films exhibited photocurrent density values of up to 1.30 and 1.95 mA/cm2 for water and sulfite oxidation, respectively. These values are substantially higher than the values of the previous reports from Table 1, except the data from Choi’s group, who constructed nano-porous structures to achieve their maximal photocurrent density in the range of 1.8 – 4.5 mA/cm2.25 The impressive PEC performance of the ESD BiVO4 thin films is attributed to their nano-textured pillar-like morphology. The formation of such nano-structured films is not observed in spin coating,11 reactive sputtering,12 polymerassisted direct deposition,31 metal organic decomposition,32 electrophoretic deposition,33 or solution casting.34 As discussed earlier, the development of high surface area BiVO4 nano-structures is favorable for the generation and transport of charge carriers. Hence, ESD is useful for the production of BiVO4 thin films with high PEC efficiency.
3. Modeling One of the major findings in the experiments described in section 3 is the formation of pillar-like structures in electrostatically sprayed BiVO4 films (cf. Figures 4 and S4). The surfaces covered by these structures were found to be highly photoactive, which is most probably related to the enhanced surface area of pillar-like landscapes compared to a flat surface. Because formation of such pillars is highly desirable from the point of view of maximizing photocurrent, elucidating the physical mechanisms responsible for pillar formation is also highly desirable. Such an understanding can be used for optimizing the electrostatic spraying technology to achieve rough pillar-like surfaces with the highest possible surface area, and thus the highest photocurrent, and to maintain this desirable morphology as process changes are inevitably made for scale-up and improved manufacturability. The present section outlines the first steps in this direction, namely, it elucidates the physical mechanisms involved in the electrostatic spraying leading to formation of nanostructured BiVO4 films.
13 ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 30
3.1 Physical Processes and Governing Equations The electrostatic spraying of the 1:1 solutions of bismuth (III) nitrate pentahydrate and vanadium (III) acetylacetonate in acetic acid results in formation of tiny electrically charged droplets moving towards a substrate that is held at an elevated temperature of 80 °C. The droplets evaporate in flight, while the solutes precipitate and form submicron particles seen in Figures 4 and S4. It is not obvious that deposition of such particles should result in formation of a characteristic pillar-like landscape instead of an approximately flat surface. Because the particles are submicron (see Figure 4), their trajectories can be significantly affected by Brownian motion, i.e. thermal diffusion. Moreover, the heated substrate induces a temperature gradient in the gas, which in turn introduces directionality to their Brownian motion, which is associated with thermophoresis of submicron particles35. The highest temperature is at the substrate surface, and thus thermophoretic motion is directed away from the surface. Because the submicron particles are charged, they are also subject to a Coulombic force attracting them to the substrate, in opposition to their thermophoretic motion. In these experiments, no co-flow of gas toward the surface is involved. However, in consideration of future process optimization, it is worth also considering the effect of the aerodynamic entrainment of submicron particles by an air jet impinging onto the substrate, as a potential additional governing parameter. Thus, in computing the drag force on the submicron particles, we do not necessarily assume a zero gas velocity. Of course, the gas velocity must go to zero at the substrate surface. The non-trivial circumstance in the present situation is that the local temperature and electric fields that determine the motion and deposition of submicron particles on the substrate are strongly affected by the growing landscape on its surface. Thus, growth of that landscape is determined by the submicron particle deposition, while the deposited layer, in turn, affects the particle motion and 14 ACS Paragon Plus Environment
Page 15 of 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
deposition. It is precisely this feedback that can produce non-uniformity of the deposited layer and formation of pillar-like structures. Modeling and understanding this effect is the main aim of the present section. Droplet motion toward the substrate is determined by the surrounding gas flow, diffusion, thermophoresis, and electrophoresis. We consider movement toward an initially planar surface, and for simplicity consider motion in two directions (perpendicular and parallel to the surface). Droplet motion toward the substrate located at distance H from the origin is considered in the domain
−∞ ≤ x ≤ ∞, 0 ≤ y ≤ 1 , where the Cartesian coordinates are rendered dimensionless by H. Droplets are deposited onto the substrate, and thus in time the domain evolves into
−∞ ≤ x ≤ ∞, 0 ≤ y ≤ h ( x, t ) ; h ( x,0) = 0; h ( x, t ) 0 to the left from its tip and ∂h / ∂x < 0 to the right from its tip. Therefore, according to Eq. (19) the x-component of the electric drift vC,x has the following signs: vC,x>0 to the left from the pillar tip, and vC,x
