Nanometer-Resolved Operando Photo-Response of Faceted BiVO4 Semiconductor Nanoparticles

Photo(electro)catalysis with semiconducting nanoparticles (NPs) is an attractive approach to convert abundant but intermittent renewable electricity into stable chemical fuels. However, our understanding of the microscopic processes governing the performance of the materials has been hampered by the lack of operando characterization techniques with sufficient lateral resolution. Here, we demonstrate that the local surface potentials of NPs of bismuth vanadate (BiVO4) and their response to illumination differ between adjacent facets and depend strongly on the pH of the ambient electrolyte. The isoelectric points of the dominant {010} basal plane and the adjacent {110} side facets differ by 1.5 pH units. Upon illumination, both facets accumulate positive charges and display a maximum surface photoresponse of +55 mV, much stronger than reported in the literature for the surface photo voltage of BiVO4 NPs in air. High resolution images reveal the presence of numerous surface defects ranging from vacancies of a few atoms, to single unit cell steps, to microfacets of variable orientation and degree of disorder. These defects typically carry a highly localized negative surface charge density and display an opposite photoresponse compared to the adjacent facets. Strategies to model and optimize the performance of photocatalyst NPs, therefore, require an understanding of the distribution of surface defects, including the interaction with ambient electrolyte.


d) height profile along dashed line in c). e) force gradient maps without and with illumination, respectively. g-h) local charge density along dashed lines (profile 1 and profile 2) in e), f) without (black and green charge profiles) and with (red and pink charge profiles) illumination.
For a tip with a flat bottom with radius  2 , the tip-sample interaction area is   =  2 2 .The 2D diffuse layer charge density   reported throughout this work is extracted based on the assumption that the charge density on the sample is homogeneous within the interaction area.
Hence, the total charge detected by the tip is   =     .Assuming that a unit cell step can be described as a sharp one-dimensional charge defect, we estimate that the total charge experienced by the tip on top of the defect is composed of two contributions, namely the one from the 1D-line charge density  along the defect and the contribution from the adjacent terraces, i.e.
Hence, we can estimate the line charge density as: From Fig. 7b and c, we can extract    ≈ −0.1 / 2 right on top of the step and    ≈ −0.05 / 2 in the middle of the facet in the dark.Under illumination (red curve in Fig. 7c), we have    ≈ −0.03 / 2 and    ≈ +0.08 / 2 .The 3D tip radius in the present experiments is   = 27.Taking  2 and the radius of the part of the tip that is within one Debye screening length of the surface (  = 3 ), we find: Using  2 = 27 , this leads to estimated line charge densities of   ≈ −1 / and   = −2.1 /, corresponding to distances between of adjacent elementary charges of approximately 1nm and 0.5nm, respectively.I.e. the one dimensional line charge density becomes more negative upon illumination.The fact that the local minimum above the step shifts upward in Fig. 7c is thus a consequence of the accumulation of positive charge on the facets whereas the defect actually accumulates more electrons.
The electrostatic double layer contribution contains the required information on the surface charge and was obtained by solving the full Poisson-Boltzmann equation with a boundary condition that involves a constant regulation 4,[8][9][10] .For a 1-1 electrolyte it is given by:   () =    ∑ ρ , (exp (−     () where   , , ϵ, ρ , ,   ,   are Boltzmann constant, the temperature vacuum permittivity, bulk number density of i-th ionic species, elementary charge, valency of i-th ionic species and electrostatic potential, respectively  ∞ the bulk ion concentration and  the elementary charge; () is the electric potential for 0 <  < .
Calculation of the electric double layer contribution requires knowledge of the potential () in the electrolyte.Here the potential  and its derivative are obtained by solving the Poisson-Boltzmann equation (PB) 6 : with the boundary condition The subscript s denotes quantities calculated at the surface.Here  ̂ is a unit normal vector and σ is the surface charge density, which we describe here using a constant-regulation model, α and β are fit parameters which minimize the function / (10)   where   (  ) and  , (  ) are theoretically calculated and experimentally measured electrostatic forces, respectively, for a separation of   between two surfaces.
For an isolated solid-electrolyte interface, the equation 7 reduces to the Grahame equation, which relates diffuse layer charge density to diffuse layer potential: σ = √8 ∞  0      sinh (   q e of Twente, P.O.Box 217, 7500 AE Enschede, The Netherlands.‡ These authors contributed equally to this work.

Figure
Figure S1.(a) SEM image of BiVO 4 nanoparticles on a silicon wafer.(b) Typical SEM image of AFM tip after experiment.SEM image of AFM tip (MikroMash NSC36/Cr-Au BS) used for AFM spectroscopy measurements shown in Figure 7.The radius of the AFM tip (R tip =26.95) is obtained by fitting a circle to the hemispherical tip.

Figure S2 .
Figure S2.Amplitude modulation -AFM topography (a), amplitude (b) and phase (c) images of BiVO 4 nanoparticle adsorbed on sapphire in a 10 mM NaCl, pH 5.6 solution.These images correspond to the location where 2D-force map shown in Figure 1 c is collected.d)-f) height profile along solid lines in a).AFM tip parameters: MikroMash NSC36 with silicon tip and cantilever: Q factor = 3.7; resonance frequency = 31.65kHz; spring constant = 1.3 N m −1 ; tip radius = 26.9 ± 2 nm).

Figure
Figure S3.(a) Color-coded 2D total interaction stiffness (k TOT ) map or total force gradient (-dF Tot /dz) map shows tip-BiVO 4 NP nanoparticle interactions in 10 mM NaCl solution at pH = 5.8.The 2D map was extracted from a 3D force versus distance (FD) volume plot when the tip is 2.5 nm away from the

Figure S4 .
Figure S4.Average force gradient versus distance curves obtained on one single point of {010} facet of BiVO 4 and sapphire (Al 2 O 3 ) substrate during the one round of illumination and in dark.The average is for forces recorded during 8 minutes at 1 Hz ramp rate.Illumination reduces the electrostatic repulsion on the {010} facet while the sapphire substrate control surface displays no light response.AFM tip parameters: MikroMash NSC36 with silicon tip and cantilever: Q factor = 4.1; resonance frequency = 30.163kHz; spring constant = 1.22 N m −1 ; tip radius = 22.5 ± 2 nm).

Figure
Figure S5.AM-AFM topography, amplitude, and phase images of BiVO 4 nanoparticle adsorbed on sapphire in 10 mM NaCl, pH 5.6.These images correspond to the location where the force maps shown in Fig. 2 are collected.(a, b, and c) correspond to images before the first 3D force map is recorded, while (d, e, and f) are images of the BiVO 4 particle after completion of the 3D force maps (at pH 4, 6, and 9 with 10 mM NaCl, in the dark and under illumination), as shown in Figure 2. The data indicates that the BiVO 4 nanoparticle did not change (dissolve or degrade) over the course of all experiments.The order of the experiments was as follows: pH 5.6, 10 mM NaCl (dark); pH 8.5, 10 mM NaCl (dark); pH 4.5, 10 mM NaCl (dark); pH 4.5, 10 mM NaCl (illuminated); pH 5.6, 10 mM NaCl (illuminated); and pH 8.5, 10 mM NaCl (illuminated).Tip parameters: Q factor= 3; resonance frequency= 19.110 kHz; spring constant = 0.65 N m −1 ; tip radius = 14.5 nm ± 2 nm.

Figure S6 .
Figure S6.Electrical (potential) properties of {110} and {010} facets BiVO 4 nanoparticles a) Average diffuse layer potential (left Y axis) and surface charge (left Y axis) of {010} and {110} facets of BiVO 4 nanoparticle as a function of pH in 10 mM NaCl with and without illumination.Data reveal facetdependent isoelectric points of  {010} ≈ 4.5 and  {110} ≈ 6. b) Measured zeta potential of BiVO 4 nanoparticle suspension as a function of pH in 10 mM NaCl.Error bars are statistical standard deviations from 3 separate measurements (green dash line is to guide the eyes).Suspensions of our BiVO 4 NPs display an (average) IEP of pH 3 in electrokinetic measurements of the ζ-potential.Average local diffuse layer surface potential as a function of pH in 10 mM NaCl with and without illumination for {010} facet (panel c) and {110} facet (panel d).Error bars are statistical standard deviations from 6-10 independent measurements per condition.The thick, dashed lines are a guide for the eye.Surface potential values are converted from the surface charge values in Figure 4 using the Graham equation.

Figure S7 .
Figure S7.Surface photo-response of {010} and {110} facets of BiVO 4 NPs in electrolyte of variable pH.Left side of the graph are data extracted from Figure 4 and Figure S6 c and d.The photoresponse depends strongly on the specific facet and on the ambient pH.Right part of the graph: surface (photo)voltage (SPV) measured in air for {010} and {011} facets of BiVO 4 (data from Zhu et al., NanoLett.2017 1 ).

Figure S8 .
Figure S8.AFM images a) Height; b) Amplitude; c) Phase of the BiVO 4 nanoparticle shown in Figure 6a.AFM images d) Height; e) Amplitude; f) Phase corresponding to the location where the twodimensional 2D force map measurement shown in Figure 6b is performed.AFM tip parameters used to acquire AFM images and the force map shown in Figure 6b: MikroMash NSC36 with silicon tip: Q factor = 3.7; resonance frequency = 31.65kHz; spring constant =1.3 N m −1 ; tip radius = 22.14 ± 2 nm.

Figure
Figure S9.AFM images a) Height; b) Amplitude; c) Phase of BiVO 4 nanoparticle shown in Figure 7a.AFM images correspond to the location where two dimension (2D) force maps measurement shown in Figure 7d and e are performed.d)height profile along dashed line in c).e) force gradient maps without and with illumination, respectively.g-h) local charge density along dashed lines (profile 1 and profile 2) in e), f) without (black and green charge profiles) and with (red and pink charge profiles) illumination.
intersection between equation 6 and the Grahame equation gives the isolated surface charge and potential.Once the  is known, we can then quantify the charge regulation behavior using the regulation parameter, , that can then be calculated as =     +  (12)where,   and   are the diffuse layer capacitance and inner layer capacitance, respectively. 1and  ∞ are the Debye length (  ) and bulk solution ionic strength The Debye length is given by,