Dynamic Photoelectrochemical Device Using an Electrolyte

Subscriber access provided by UNIV OF DURHAM

Article

Dynamic photoelectrochemical device using electrolyte-permeable NiOx/SiO2/Si photocathode with an open circuit potential of 0.75 V Jin-Young Jung, Jin-Young Yu, and Jung-Ho Lee ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b16918 • Publication Date (Web): 07 Feb 2018 Downloaded from http://pubs.acs.org on February 15, 2018

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 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 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.

ACS Applied Materials & Interfaces 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 28 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

ACS Applied Materials & Interfaces

Dynamic photoelectrochemical device using electrolyte-permeable NiOx/SiO2/Si photocathode with an open circuit potential of 0.75 V Jin-Young Jung, Jin-Young Yu and Jung-Ho Lee* Departments of Materials Science and Chemical Engineering, Hanyang University, 55 Hanyangdaehak-ro, Sangnok-gu, Ansan, Kyeonggi-do 15588, Republic of Korea KEYWORDS Photoelectrochemical cells, Water splitting, Silicon, NiOx, Semiconductor junction

ABSTRACT As a thermodynamic driving force obtained from sunlight, the open circuit potential (OCP) in photoelectrochemical cells is typically limited by the photovoltage (Vph). In this work, we establish that OCP can exceed the value of Vph when an electrolyte-permeable NiOx thin film is employed as an electrocatalyst in a Si photocathode. The built-in potential developed at the NiOx/Si junction is adjusted in-situ according to the progress of the NiOx hydration for the hydrogen evolution reaction (HER). As a result of the decoupling of OCP from Vph, a high OCP value of 0.75 V (vs. reversible hydrogen electrode) is obtained after 1 hour operation of HER in an alkaline electrolyte (pH = 14), thus outperforming the highest value (0.64 V) reported to date with conventional Si photoelectrodes. This finding might offer insight into novel photocathode designs such as those based on tandem water splitting systems.

1 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 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 2 of 28

INTRODUCTION One of the fundamental configurations in photoelectrochemical (PEC) devices is the built-in potential (Vbi) at the semiconductor junctions, which results in a thermodynamic separation of electron and hole charge carriers towards water-reducing and water-oxidizing electrodes, respectively.1‒5 In particular, the Vbi at the semiconductor/electrolyte junctions represents the relative difference between the Fermi-level of the photoelectrode and the redox-level of the electrolyte.6‒10 Higher Vbi values could be realized by employing a variety of solid-state junctions (also known as buried junctions) such as metal/semiconductor Schottky junctions,11‒16 p-n homojunctions,17‒22 and semiconductor heterojunctions,23‒28 which exploit the large difference in Fermi-levels between two different solid states. While buried junctions generally exhibit fixed Vbi values during PEC operation, Lin and Boettcher have recently demonstrated a new type of semiconductor junctions capable of dynamic change in Vbi (the so-called adaptive junctions) based on the integration of an electrolyte-permeable NiOx electrocatalyst on TiOx photoanode.29‒31 The electrolyte-permeability of the porous NiOx most likely allows NiOx to be transformed into NiOOH (or Ni(OH)2) throughout the entire layer during oxidation/reduction reactions, thereby adjusting the average state of oxidation/reduction as well as the Fermi-levels. The Vbi values in the NiOx-coated TiOx photoanode are therefore modified in situ as the oxygen evolution reaction progresses.29 Increased Vbi values in the adaptive junctions could improve the open circuit potential (OCP), which is the maximum thermodynamic potential generated in a PEC device. But, a fundamental understanding of the PEC operation principles pertaining to this system that facilitate improvements in OCP remains currently unclear. The PEC operation employing buried junctions normally follows the conventional principles of photovoltaic operation. The ideal OCP should be


Page 3 of 28 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


equal to a value of Vbi that is determined by the Fermi-level difference of the two materials.32,33 Nevertheless, the actual OCP values display a photovoltage (Vph) that is determined by quasiFermi-level splitting, and is affected by photovoltaic losses arising because of charge generation, separation, and recombination.17,34,35 In this regard, a dynamic change in Vbi observed in adaptive junctions can modify the quasi-equilibrium states of the resulting PEC device, leading to coupled modulations of OCP. If this hypothesis is true, it should be possible for OCP to exceed Vph. Here, in order to observe the adaptive junction behavior at the NiOx/p-Si interface, we have integrated the electrolyte-permeable structure of NiOx thin film on SiO2 grown p-Si photocathodes. Two key factors in this work were the permeability of NiOx and interfacial SiO2 passivation. Interestingly, systematic investigation of the relationship between OCP and Vph reveal that the behavior of OCP is decoupled from that of Vph. As a result, we can obtain the highest value of OCP (0.75 V) reported to date, which corresponds closely to the theoretical maximum value (0.75– 0.80 V) estimated for Si photovoltaics.36 In addition, exploring the charge transfer kinetics unravels the overpotential loss suffered as a result of serial resistances at the thin SiO2 tunneling layer and NiOx electrocatalysts. RESULTS AND DISCUSSION Utilizing a commonly employed architecture for efficient PEC water oxidation and reduction, a 10-nm-thin NiOx film was electron-beam deposited onto a 2-nm-thin, rapidly thermally oxidized (RTO) Si substrate to produce the electrocatalyst/insulator/semiconductor configuration.37,38 The ultrathin RTO film was anticipated to play essential roles in defect passivation (by a high-quality SiO2/Si interface)11 and as a diffusion barrier (for Ni into Si). When the NiOx was electron-beam deposited onto a hydrogen-terminated Si substrate, we observed interfacial formation of nickel



Page 4 of 28

disilicides (Ni:Si=1:2) and SiO2 using scanning transmission electron microscopy (STEM), energy-dispersive X-ray spectroscopy (EDXS), and electron energy loss spectroscopy (EELS) (Figures 1a, S1, and S2). This formation stemmed from the diffusion of activated Ni into Si during the slight oxidization of the Si surface by the electron-beam deposited NiOx. Ultimately, however, the presence of a thin RTO layer was found to prevent the formation of Ni-silicides (Figure 1b). Linear sweep voltammetry (LSV) responses of hydrogen-terminated and SiO2-grown Si photocathodes in 1 M KOH under dark/illumination conditions were compared (Figure 1c). While the hydrogen-terminated photocathode showed a high dark (leakage) current (10.87 mA/cm2 at −1 V vs. RHE) because of the Ni-silicide acting as a recombination site, the SiO2-grown photocathode exhibited a significantly lower dark current (0.01 mA/cm2). As a result, the saturated photocurrent density, which was measured at −1 V vs. RHE, improved by ~5 mA/cm2 when the values obtained under dark and illuminated conditions were compared. The high photocurrent level (~31 mA/cm2) of the SiO2-grown photocathode resulted from high external quantum efficiency (see Figure S3). In addition, we also observed a large anodic shift (~500 mV) in the LSV curve at the point of formation of interfacial SiO2 instead of the presence of Ni-silicide. This observation was attributed to the change in the Vbi value because of the development of NiOx/p-Si junctions instead of Nisilicide/p-Si junctions. For the hydrogen-terminated photocathode, the predominance of Nisilicide/p-Si junctions where Fermi-level of Ni-silicide is more negative than that of p-Si most likely caused negative Vbi, which induced bending of the upward band at the Si band edge (Figure S4a), acting as a barrier for electron transfer. This required considerably larger overpotential and located the LSV curve at a negative potential region. For the SiO2-grown photocathode, on the other hand, the NiOx/p-Si junctions represent the dominant factor inducing the downward bending and a positive Vbi, (Figure S4b), thereby shifting the LSV curve to a positive potential region. To


Page 5 of 28 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


confirm these predictions, we employed the Mott–Schottky (M–S) measurements in 1 M KOH under dark conditions, where the x-intercepts of the M–S plots provided the flat band potential (Vfb, the level of which is approximately equal to that of Vbi) (Figure 1d).39 As we expected, the SiO2-grown photocathode exhibited a more positive Vfb (0.5 V vs. RHE) than the hydrogenterminated photocathode (−0.24 V vs. RHE).

Figure 1. Characterization of evaporated NiOx on H-terminated and SiO2-grown Si photocathodes. STEM images and EDXS profiles of a NiOx thin film coating on (a) H-terminated Si (NiOx/Si) and (b) SiO2-grown Si (NiOx/SiO2/Si). (c) LSV response and (d) M–S plots for PEC devices constructed from NiOx/Si (black) and NiOx/SiO2/Si (red), operated in 1 M KOH electrolyte (pH = 14). Dotted and solid lines in (c) represent dark and illuminated conditions, respectively. The linear



Page 6 of 28

x-intercept in each M–S curve (the position of vertical line) indicates the value of Vfb (i.e., Vbi). The arrows in (a) and (b) indicate the increasing cycle number. NiOx is typically known as a p-type semiconductor with a more negative Fermi-level than that of p-Si, which induces a negative Vbi value at the NiOx/p-Si junction (see M–S plot of solid-state NiOx/SiO2/Si in Figure S5). For this reason, the positive Vbi value (0.5 V) observed for our NiOx/SiO2/Si photocathode represents an unprecedented result. In addition, when the LSV and M– S characteristics were examined in sequence over several consecutive runs, interestingly, Vbi and the LSV curve shifted gradually to the anodic direction (see red curves in Figures 1c and d). From these results, we can suggest that the HER operation (i.e., the reduction reaction) provokes changes in the Fermi-level of NiOx and thus also the Vbi of the NiOx/p-Si junction. To verify this relationship, we characterized the Vbi values at the NiOx/electrolyte junction with varying reduction reaction rates by implementing the M–S measurements for the electrolyte/NiOx/n+-Si at various frequencies (Figure S6). The results revealed cathodic shifts of Vbi with decreasing frequency. At a very low frequency (0.1 Hz), the Vbi could not be measured. These behaviors imply that the Fermi-level of our NiOx thin film changes together with its reduction rate, and a significant reduction reaction can dissipate the potential drop at the Helmholtz layer and space-charge region for the NiOx/electrolyte junction. Therefore, we can demonstrate that the Fermi-level of NiOx can reach the redox-level of H+/H2 (EHER). According to previous work by Lin and Boettcher, the Fermi-level shift could be realized by the exploitation of the electrolyte-permeable, porous structure of the redox-active NiOx thin films.3032

It is generally known that the electronic density of states of NiOx can be potentially extended

owing to its broad redox waves of Ni2+/Ni3+ and the oxidation/reduction reactions that lead to the transformation of NiOx into NiOOH/ Ni(OH)2—a process that is accompanied by a transition of


Page 7 of 28 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


its oxidation/reduction states.40-42 If this transition occurs throughout the entire layer of electrolytepermeable NiOx, its Fermi-level can be effectively changed. By contrast, this level would remain unchanged for an electrolyte-impermeable, dense NiOx.30 For our NiOx/SiO2/Si photocathode, transfer of photon-generated electron from the Si into the NiOx represents the driving force for the reduction of NiOx. The transformation during PEC reaction was verified from a oxidation wave at anodic current region in illuminated LSV responses of NiOx/SiO2/p-Si, and the reduction reactions of our NiOx thin film are confirmed by the presence of a reduction wave in cyclic voltammetry (CV) responses of NiOx/SiO2/p+-Si in the dark (Figure S7). In addition, the growth and shift for both reduction and oxidation waves with the repetitive cycling of CV are ascribed to the increase in the reduction area,30,40 demonstrating that our NiOx thin film possesses an electrolyte-permeable, porous structure. The 40 nm-thick NiOx thin films clearly reveal porous, poly-crystalline structure, where the reduction reaction can occur throughout its grain boundaries (Figure S8). The grain size (2–5 nm) observed for the thin films was relatively small, causing the electrical properties of the reduced surface to deviate substantially from those of NiOx bulk. During the PEC tests, we also observed two results that evidenced the electrolyte-permeability. Firstly, the dense NiOx with a thickness of 40 nm is known to prevent the infiltration of KOH ions into the Si surface, with good long-term electrochemical stability over 100 hours.43,44 By contrast, our NiOx is only stable for 6 hours (Figure S9). When a 15-nm-thick Ti layer was inserted between NiOx and SiO2, stable operation for >6 hours was achieved because the inserted Ti layer was effective to hinder the electrolyte permeation into the SiO2 which is corroding readily in alkaline electrolyte. Secondly, when we implemented an activation process via chronopotentiometry at −10 mA/cm2 to further investigate the transition of NiOx into Ni(OH)2 as a function of the HER, the activation induced an anodic



Page 8 of 28

shift in the potential levels (Figure 2a). In addition, we found that the activation time for the saturation of the potential level was dependent on the NiOx thickness (Figure S10). These results suggest that the transition requires significant time for the reduction to arise over the entire layer. The transition after the activation was clearly verified by several distinct changes in the X-ray photoelectron spectroscopy (XPS). In the Ni 2p3/2 spectra, the ratio of Ni(OH)2 and NiOx peak areas increased from 9.16 to 12.81 (Figure 2b). In the O 1s spectra, the ratio of the Ni–OH and Ni– O–Ni peak areas increased from 2.3 to 3.4 (Figure 2c). The XPS survey spectra showed that the atomic ratio of O and Ni changed from 1.62 to 1.98 (Figure S11). In addition to the changes in the main peak area ratios, all the spectral peaks were found to shift to higher binding energies (by ~0.3 eV), thus indicating that the transition is accompanied by a significant change in the energy level of the material. This change may cause the Fermi-level to shift towards the conduction band. The unchanged XPS results obtained after three cycles of LSV scanning imply that the activation enables the transition to take place (Figure S12). In C 1s spectra, we could observe the peak broadening of C=O bonding along with a blue peak shift during PEC reaction (see Figure S12c), implying that carbon contamination might affect the chemical binding behavior of nickel and oxygen. However, we could observe the O 1s peak shift only at the activation (not at cycling of LSV) so that carbon is not associated with the peak shift behaviors of Ni 2p3/2 and O 1s spectra.


Page 9 of 28 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 2. Activation of electrolyte-permeable NiOx by operation of PEC-HER. (a) The activation process was performed using chronopotentiometry for NiOx samples with thicknesses of 5 (black), 10 (red), and 20 nm (blue) under a constant current density of −10 mA/cm2. Some noise and drastic change in the chronopotentiometry plots stemmed from the accumulation and release of H2 bubbles. (b) Ni 2p and (c) O 1s XPS spectra of the NiOx thin film in pristine (top) and test samples after activation process for 1 hour (bottom). The yellow and green peaks in the Ni 2p spectra are associated with Ni(OH)2 and NiOx, respectively. The gray peaks represent satellite features. The purple, blue, and gray peaks in the O 1s spectra are associated with Ni–O–Ni, Ni–O–H, and adsorbed H2O, respectively. The vertical dotted lines indicate the peak positions for the samples before (black) and after (red) activation.



Page 10 of 28

We characterized the LSV responses of NiOx/SiO2/Si under pristine (as-deposited thin film), activated, deactivated, and reactivated NiOx conditions, as shown in Figures 3a, b. The activation implemented by PEC operation at 10 mA/cm2 for 1 hour led to considerable anodic shifts in the LSV curve and OCP. During subsequent measurements of the LSV response after the activation, cathodic shifts in the LSV curve and OCP were observed as a result of the deactivation of NiOx caused by the predominant drift of the majority carrier holes in Si into NiOx upon application of anodic current. The deactivation was also observed under unbiased dark conditions after 1 hour owing to the diffusion of the holes. By contrast, the reactivation process resulted in anodic shifts, again. Similar patterns in behavior were also observed in the Vbi results (Figure 3c), which imply that PEC performances were affected mainly by the development of adaptive junction. A high Vbi value of 0.85 V might facilitate a negatively charged inversion layer, which is beneficial both in terms of thermodynamics for the charge split and kinetics for charge transfer (see Figure S13a). The high Vbi reduced the majority carrier (hole) density at the band edge of Si, as identified from the increased slope of the M–S plot (see Figure 3c).12 This reduction can efficiently prevent recombination of charge carriers at the Si/SiO2 interface and the NiOx layer by blocking the transport of holes from Si to NiOx. In turn, this behavior facilitated a decrease in the dark (leakage) current and thus an increase of the Vph (see Figures S13b, c). A high Vbi can also be the driving force for accelerating the charge transfer during the HER. Moreover, the reduced interfacial recombination kinetics prevent the formation of charge carrier trapping pathways,34 and the increased electron density at the Si band edge is favorable for facile tunnelling conduction through the SiO2 layer.37 These advantages can reduce the charge transfer resistance for the HER (see electrochemical impedance spectroscopy results shown in Figure S13d).


Page 11 of 28 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


In order to compare the adaptive junction with a conventional buried junction, we also characterized the PEC performance of a NiOx/Ti/SiO2/Si photocathode that can generate a high fixed Vbi as a result of a buried Schottky Ti/Si junction,11 while maintaining the electrocatalytic activity of the NiOx layer (see blue curve in Figures 3a to c). Unlike the adaptive junction, both the LSV curve and the Vbi of the conventional buried junction remained essentially unchanged. As a result of the activation, the adaptive NiOx/Si junction showed better Jph and OCP with a higher Vbi when compared to the results obtained for the buried Ti/Si junction. We also explored the Vph values by means of momentary potential difference measurements in darkness and under illumination during open circuit conditions in 1M KOH (Figure 3d). When measured several times consecutively, the Vph value for the NiOx/SiO2/Si changed while that for the NiOx/Ti/SiO2/Si remained unchanged. These results clearly manifest the differences between the adaptive NiOx/Si junction and the buried Ti/Si junction. In addition, the change in Vph proves to be an adaptive NiOx/p-Si junction rather than a mixed equilibrium potential caused by two partial reactions of NiOx and KOH electrolyte3 (More detailed information and description for the Vph changes are provided Figures S14 and S15).



Page 12 of 28

Figure 3. Comparison of PEC performances for the adaptive NiOx/Si junction and buried Ti/Si junction. (a) and (b) LSV response, (c) Mott–Schottky plots, and (d) Vph measurements for NiOx/SiO2/Si (red) and NiOx/Ti/SiO2/Si (blue) photocathodes. The plot in (b) represents the low photocurrent density level range (

Dynamic Photoelectrochemical Device Using an Electrolyte

Recommend Documents