Daniel V. Esposito

H2 evolution at Si-based metal–insulator–semiconductor photoelectrodes enhanced by inversion channel charge collection and H spillover

Photoelectrochemical (PEC) water splitting represents a promising route for renewable production of hydrogen, but trade-offs between photoelectrode stability and efficiency have greatly limited the performance of PEC devices. In this work, we employ a metal-insulator-semiconductor (MIS) photoelectrode architecture that allows for stable and efficient water splitting using narrow bandgap semiconductors. Substantial improvement in the performance of Si-based MIS photocathodes is demonstrated through a combination of a high-quality thermal SiO2 layer and the use of bilayer metal catalysts. Scanning probe techniques were used to simultaneously map the photovoltaic and catalytic properties of the MIS surface and reveal the spillover-assisted evolution of hydrogen off the SiO2 surface and lateral photovoltage driven minority carrier transport over distances that can exceed 2 cm. The latter finding is explained by the photo- and electrolyte-induced formation of an inversion channel immediately beneath the SiO2/Si interface. These findings have important implications for further development of MIS photoelectrodes and offer the possibility of highly efficient PEC water splitting.

Methods of photoelectrode characterization with high spatial and temporal resolution

Materials and photoelectrode architectures that are highly efficient, extremely stable, and made from low cost materials are required for commercially viable photoelectrochemical (PEC) water-splitting technology. A key challenge is the heterogeneous nature of real-world materials, which often possess spatial variation in their crystal structure, morphology, and/or composition at the nano-, micro-, or macro-scale. Different structures and compositions can have vastly different properties and can therefore strongly influence the overall performance of the photoelectrode through complex structure–property relationships. A complete understanding of photoelectrode materials would also involve elucidation of processes such as carrier collection and electrochemical charge transfer that occur at very fast time scales. We present herein an overview of a broad suite of experimental and computational tools that can be used to define the structure–property relationships of photoelectrode materials at small dimensions and on fast time scales. A major focus is on in situ scanning-probe measurement (SPM) techniques that possess the ability to measure differences in optical, electronic, catalytic, and physical properties with nano- or micro-scale spatial resolution. In situ ultrafast spectroscopic techniques, used to probe carrier dynamics involved with processes such as carrier generation, recombination, and interfacial charge transport, are also discussed. Complementing all of these experimental techniques are computational atomistic modeling tools, which can be invaluable for interpreting experimental results, aiding in materials discovery, and interrogating PEC processes at length and time scales not currently accessible by experiment. In addition to reviewing the basic capabilities of these experimental and computational techniques, we highlight key opportunities and limitations of applying these tools for the development of PEC materials.

A scanning probe investigation of the role of surface motifs in the behavior of p-WSe2 photocathodes

The spatial variation in the photoelectrochemical performance for the reduction of an aqueous one-electron redox couple, Ru(NH_3)_6^(3+/2+), and for the evolution of H_2(g) from 0.5 M H_2SO_4(aq) at the surface of bare or Pt-decorated p-type WSe_2 photocathodes has been investigated in situ using scanning photocurrent microscopy (SPCM). The measurements revealed significant differences in the charge-collection performance (quantified by the values of external quantum yields, Φ_(ext)) on various macroscopic terraces. Local spectral response measurements indicated a variation in the local electronic structure among the terraces, which was consistent with a non-uniform spatial distribution of sub-band-gap states within the crystals. The photoconversion efficiencies of Pt-decorated p-WSe_2 photocathodes were greater for the evolution of H_2(g) from 0.5 M H_2SO_4 than for the reduction of Ru(NH_3)_6^(3+/2+), and terraces that exhibited relatively low values of Φ_(ext) for the reduction of Ru(NH_3)_6^(3+/2+) could in some cases yield values of Φ_(ext) for the evolution of H_2(g) comparable to the values of Φ_(ext) yielded by the highest-performing terraces. Although the spatial resolution of the techniques used in this work frequently did not result in observation of the effect of edge sites on photocurrent efficiency, some edge effects were observed in the measurements; however the observed edge effects differed among edges, and did not appear to determine the performance of the electrodes.

Deconvoluting the influences of 3D structure on the performance of photoelectrodes for solar-driven water splitting

Three-dimensionally (3D) structured photoelectrodes offer a number of potential benefits for solar fuel production compared to conventional planar photoelectrodes, including decreased optical losses, higher surface area for catalysis, easier removal of product species, and enhanced carrier collection efficiency. However, 3D structures can also present challenges, such as lower photovoltage and larger surface recombination. Quantifying and understanding the advantages and disadvantages of 3D structuring can maximize benefits, but this goal is not trivial because the factors that affect photoelectrode performance are intertwined. In this article, we provide an overview of the benefits and challenges of using 3D photoelectrode structures and present a systematic approach for deconvoluting the most common effects of 3D structure on photoelectrode performance. As a basis for this study, metal–insulator–semiconductor (MIS) photoelectrodes consisting of p-Si micro-pillar arrays with well-defined diameter, pitch, and height were fabricated by reactive ion etching (RIE). A general framework for modeling the influences of 3D structure on photoelectrode current–potential performance is presented, and a comparison of the loss mechanisms in 3D and planar photoelectrodes is illustrated using loss analysis diagrams. We expect that most of the measurements and analyses that we demonstrate for MIS photoelectrodes can be applied with equal success to liquid-junction and p–n junction 3D structured photoelectrodes.

Strain Engineering and Raman Spectroscopy of Monolayer Transition Metal Dichalcogenides

We describe a facile technique based on polymer encapsulation to apply several percent controllable strains to monolayer and few-layer Transition Metal Dichalcogenides (TMDs). We use this technique to study the lattice response to strain via polarized Raman spectroscopy in monolayer WSe2 and WS2. The application of strain causes mode-dependent redshifts, with larger shift rates observed for in-plane modes. We observe a splitting of the degeneracy of the in-plane E modes in both materials and measure the Grüneisen parameters. At large strain, we observe that the reduction of crystal symmetry can lead to a change in the polarization response of the A mode in WS2. While both WSe2 and WS2 exhibit similar qualitative changes in the phonon structure with strain, we observe much larger changes in mode positions and intensities with strain in WS2. These differences can be explained simply by the degree of iconicity of the metal-chalcogen bond. One of the iconic characteristics of monolayer 2D materials is their incredible stretchability which allows them to be subjected to several percent strain before yielding [1]. The application of moderate (~1%) strains is expected to change the anharmonicity of interatomic potentials [2, 3], phonon frequencies [4, 5] and effective masses [6, 7]. At larger strains, topological electronic[8] [9] and semiconductor-metal structural phase changes have been predicted [10-13]. Important technological applications such as piezoelectricity can be explored by the application of systematic strain [14, 15]. One of the chief problems in achieving reproducible strain is the intrinsic nature of 2D materials as single layer sheets they need to be held to a flexible substrate which is then stretched or compressed. Previous experiments [16-19] have used flexible polymers as substrates and metal or polymer caps in order to constrain the 2D material. Using these techniques, approximate strains up to 4% have been reported so far in the literature, but independent verification of the applied strain has been lacking. Achieving large reproducible strains in engineered geometries will allow us to probe these exciting properties of individual 2D materials and their heterostructures [4, 17, 20-26]. In this work, we develop a new strain platform to apply large range accurate uniaxial tensile strains on monolayer and fewlayer materials. One of our chief innovations is the development of a novel polymer-based encapsulation method to enable the application of large strain to 2D materials. Here, we apply this technique to study the strain-dependent properties of monolayer WSe2 and WS2 grown by Chemical Vapor Deposition (CVD) on SiO2/Si substrates [27-29]. We use cellulose acetate butyrate (CAB) to lift the monolayers from the SiO2/Si substrates and transfer to polycarbonate substrates. The two polymers are then bonded to produce encapsulated monolayers and multilayers. The key to achieving good bonding is perfect control over the temperature, time and pressure during the bonding process. Additionally, polymer layers that are in the amorphous phase cause nonlinear strain-deflection behavior which is not desirable in our experiments. To resolve this issue, we crystallize the polymer stacks by annealing near the glass transition temperature followed by slow cooling. The crystallized polymers are fully flexible, elastic and springy substrates as shown in Fig. 1(a). After all of our processing steps, we find that the polymer stacks enter into the plastic regime at 7% strain. We find that strains up to this value are perfectly transferred to the encapsulated 2D material as described below. Our strain method adopts the extra-neutral axis bending technique – Fig. 1(b) in which areas above the neutral axis undergo tensile strain while those below the axis experience compressive strain. In our method, we use a screw-driven vertical translation stage to apply strain to the polymer stacks. We solve the Euler-Bernoulli equation for our geometry in order to achieve an accurate relation between the vertical displacement δ of the translation stage and the strain ε of the 2D material. For a fully isotropic, linear and elastic material, the strain-displacement relation is derived as: ε = 3tδ a(3b + 2a) ⁄ where t is the substrate thickness, b and a are center support and cantilever lengths respectively. In our experiments, the use of a fine adjustment screw gives us a resolution of 0.05% strain for 0.5 mm substrates, with essentially no limit to the maximum strain that can be applied. More details are provided in the Supplemental Material. Shown in Fig. 1(c) is an optical image of triangular flakes of WSe2 encapsulated by this process. We adjust the CVD process to produce triangular flakes in order to easily identify the crystallographic directions of the grown monolayers. Since the strains achievable in our experiments are large, we can directly verify from optical measurements that the strain being applied to 2D layer is the calculated value. This is illustrated in Fig. 1(d). Each of these images is obtained by overlaying two images, one at zero strain and one at a fixed value of strain (4.2% and 6.5% respectively). Only the edges of the triangles are shown in the images, which are lined up to be at the same vertical height at the top vertex of the triangle. We can directly see by inspection that the length of the triangle along the strain direction is larger when strained as one expects. A pixel-height measurement of the edge-detected images gives us a direct experimental measure of the applied strain, which can be compared to the calculated strain based on the screw displacement. It is found that the two measurements match within 0.1% absolute strain. Thus, our technique allows for the application of uniform, highly repeatable and independently measurable strain on TMD monolayers and heterostructures. In order to probe the effects of strain on our samples, we choose to characterize with Raman spectroscopy a simple yet powerful way to measure lattice properties and their coupling to the electronic degrees of freedom. Strains were applied in both zigzag and armchair directions (Y and X axes in Fig. 1(e) ) in our experiments. Our Raman setup with 532 nm excitation wavelength is shown in Fig. 1(f). The measurements were performed while controlling for the incident light’s polarization (Ei) direction (θ in Fig. 1(e) ). For each experiment, Raman spectra were collected in both the parallel(Es || Ei) and crosspolarized (Es ⊥ Ei) detector geometries, shown with standard notations Z(YY)Z and Z(YX)Z respectively. In our experiments, we found no dependence of the Raman spectra on the angle of incidence relative to the crystallographic axis at zero strain. We therefore fix our incidence angle to the Y direction, and measure the unpolarized, parallel-polarized and cross-polarized Raman spectra at each value of strain which is applied in the X direction. We first discuss the properties of monolayer WSe2. Shown in Fig. 2(a) are a sequence of spectra taken at different values of strain in the unpolarized, parallel and cross polarization geometries. Previous Raman spectroscopy measurements performed on monolayer WSe2 have identified three vibrational modes [30-32] termed A, E and 2LA. A is an out-of-plane phonon mode in which the top and bottom chalcogen atoms vibrate in opposing directions; while E is in-plane mode where the metal atoms vibrate out-of-phase with the chalcogen atoms [33]. The 2LA mode results from a double resonance process involving two phonons from the LA branch. Second order processes can in general give rise to a complex lineshape in the Raman spectrum; yet, in the case of WSe2 we find that a single Lorentzian can be used to model well the 2LA mode lineshape. Although Aand E modes are nearly degenerate, they can be distinguished from each other by polarization dependency of their intensities. The out of plane, symmetric A mode disappears due to its symmetry in the cross polarization geometry, leaving behind only the E mode. Our spectra in the cross-polarization geometry can thus be modeled well as the sum of two Lorentzian peaks corresponding to E and 2LA modes. Information of the E mode position can then be used to fit the spectra seen in the parallel polarization geometry in order to extract the nearlyoverlapping A mode position. Having understood the polarization-dependent Raman spectra of unstrained monolayer WSe2, we apply uniaxial strains and measure the Raman response. The effects of uniaxial strain up to 1% on monolayer WSe2 has previously been experimentally investigated via unpolarized Raman [17] and absorption spectroscopy [34]. Raman spectra under increasing uniaxial strain up to 3% are shown in Fig. 2(a). A close examination of spectral lineshapes in the cross polarization geometry shows that the E mode becomes broader with increasing strain. In general, we expect that the initially doubly degenerate E mode splits on the application of strain into E and E. The displacement eigenvector of the E mode is orthogonal to the direction of strain, while it is parallel for the Emode, as has previously been observed for MoS2 and graphene [3, 16, 21]. While we cannot observe a complete separation of the E and E modes in our data, it is nevertheless straightforward to fit the lineshape to two Lorentzian functions and extract the splitting as a function of strain, as shown in Fig. 2(e). The splitting of the E mode under tensile strain due to the anharmonictiy of molecular potentials can be described by Grüneisen parameter γ = (|∆ωE′+| + |∆ωE′−|) 2ωE′(1 − v) ⁄ and the shear deformation potential β = (||∆ωE′+| − |∆ωE′−||) 2ωE′(1 + v) ⁄ where ωE′ is the frequency of E mode, ∆ωE′+ and ∆ωE′− are the frequency shifts of split modes per unit percent strain and v is Poisson’s ratio which is 0.27 for our substrates. We obtain values of γ = 0.38 , β = 0.10 for WSe2 which are smaller than those reported for graphene [2, 3]. Using t

Structure-Property Relationships Describing the Buried Interface Between Silicon Oxide Overlayers and Electrocatalytic Platinum Thin Films

Encapsulation of an active electrocatalyst with a permeable overlayer is an attractive approach to simultaneously enhance its stability, activity, and selectivity. However, the structure–property relationships that govern the performance of encapsulated electrocatalysts are poorly understood, especially those describing the electrocatalytic behavior of the buried interface between the overlayer and active electrocatalyst. Using planar silicon oxide (SiOx)-encapsulated platinum (Pt)/titanium (Ti) bilayer thin films as model electrodes, the present study investigates the physical and electrochemical properties of the SiOx|Pt buried interface. Through a combination of X-ray photoelectron spectroscopy and electroanalytical measurements, it is revealed that a platinum oxide (PtOx) interlayer can exist between the SiOx overlayer and Pt thin film. The thickness and properties of the PtOx interlayer can be altered by modifying (i) the thickness of the SiOx overlayer or (ii) the thickness of the Pt layer, which may expose the buried interface to oxophilic Ti. Importantly, SiOx|Pt electrodes based on ultrathin Pt/Ti bilayers possess thinner PtOx interlayers while exhibiting reduced permeabilities for Cu2+ and H+ and enhanced stability during cycling in 0.5 M H2SO4. These findings highlight the tunability of buried interfaces while providing new insights that are needed to guide the design of complex electrocatalysts that contain them.

Scanning Line Probe Microscopy: Beyond the Point Probe

Scanning probe microscopy (SPM) techniques have become indispensable tools for studying nano- and microscale materials and processes but suffer from a trade-off between resolution and areal scan rate that limits their utility for a number of applications and sample types. Here, we present a novel approach to SPM imaging based on combining nonlocal scanning line probes with compressed sensing (CS) signal analysis methods. Using scanning electrochemical microscopy (SECM) as an exemplar SPM technique, we demonstrate this approach using continuous microband electrodes, or line probes, which are used to perform chemical imaging of electrocatalytic Pt discs deposited on an inert substrate. These results demonstrate the potential to achieve high areal SPM imaging rates using nonlocal scanning probes and CS image reconstruction.

The role of ultra-thin SiO2 layers in metal-insulator-semiconductor (MIS) photoelectrochemical devices (Presentation Recording)

Solid-state junctions based on a metal-insulator-semiconductor (MIS) architecture are of great interest for a number of optoelectronic applications such as photovoltaics, photoelectrochemical cells, and photodetection. One major advantage of the MIS junction compared to the closely related metal-semiconductor junction, or Schottky junction, is that the thin insulating layer (1-3 nm thick) that separates the metal and semiconductor can significantly reduce the density of undesirable interfacial mid-gap states. The reduction in mid-gap states helps “un-pin” the junction, allowing for significantly higher built-in-voltages to be achieved. A second major advantage of the MIS junction is that the thin insulating layer can also protect the underlying semiconductor from corrosion in an electrochemical environment, making the MIS architecture well-suited for application in (photo)electrochemical applications. In this presentation, discontinuous Si-based MIS junctions immersed in electrolyte are explored for use as i.) photoelectrodes for solar-water splitting in photoelectrochemical cells (PECs) and ii.) position-sensitive photodetectors. The development and optimization of MIS photoelectrodes for both of these applications relies heavily on understanding how processing of the thin SiO2 layer impacts the properties of nano- and micro-scale MIS junctions, as well as the interactions of the insulating layer with the electrolyte. In this work, we systematically explore the effects of insulator thickness, synthesis method, and chemical treatment on the photoelectrochemical and electrochemical properties of these MIS devices. It is shown that electrolyte-induced inversion plays a critical role in determining the charge carrier dynamics within the MIS photoelectrodes for both applications.