Reverse Electrodialysis-Assisted Solar Water Splitting

Photoelectrochemical (PEC) water splitting provides an attractive route for large-scale solar energy storage, but issues surrounding the efficiency and the stability of photoelectrode materials impose serious restrictions on its advancement. In order to relax one of the photoelectrode criteria, the band gap, a promising strategy involves complementing the conventional PEC setup with additional power sources. Here we introduce a new concept: solar water splitting combined with reverse electrodialysis (RED). RED is a membrane-based power generation technology that produces an electrochemical potential difference from a salinity gradient. In this study, the RED stack serves not only as a separator, but also as an additional tunable power source to compensate for the limited voltage produced by the photoelectrode. A hybrid system, composed of a single-junction p-Si and a RED stack, successfully enables solar water splitting without the need for an external bias. This system provides flexibility in photoelectrode material selection.


Supplementary Figure S1. Schematic representation of reverse electrodialysis (RED). (A) Junction
potential across an ion-exchange membrane. CEM and AEM mean the cation-exchange membrane and the anion-exchange membrane, respectively. HC stands for a high-concentration solution and LC a lowconcentration solution. (B) Structure of the RED stack. An ionic flux (i.e. Na + and Cl -), which is generated inside the RED stack, is converted into an electricity or fuels at electrodes through an electrochemical reaction.
In the PEC-RED hybrid system, the water-splitting reaction takes place at the electrodes and the hydrogen fuel is produced. (C) Flow path of the HC and LC solutions inside the RED stack. The gaskets, between the CEMs and the AEMs, separate the membranes and create a flow path of the HC and the LC solutions. When the HC solution flows in an HC compartment, the LC solution passes the HC compartment through a hole in a gasket, and vice versa. (D) The dimension of the HC and LC compartments inside RED stack.
When solutions with different ion concentrations flow on the opposite sides of an ion-exchange membrane, an electrochemical potential difference develops across the membrane by the chemical potential difference between the solutions.
The electrochemical potential for an ionic species i in phase α is defined as: i 0α is the standard chemical potential, R is the gas constant, T is the absolute temperature, i α and i are the activity and charge of the ionic species i, F is the Faraday constant, and is the electric potential 1 .
The magnitude of the electric potential difference (or the junction potential) created across the ionexchange membrane under the open-circuit condition is expressed as: The transference number ( i ) is defined as the fraction of the current carried by the ion species i to the total electric current across the membrane. The "counter" and "co" mean counter-and co-ion, in the terms of the ion charge, relative to the fixed charge in the membrane 1 .
Activity of the ionic species ( i ) is obtained by multiplying the molar concentration by the activity coefficient. The activity coefficient ( i ) is calculated from the Debye-Hückel extended equation: and are the Debye-Hückel parameters, 0 is the ion size parameter (Å), is the ionic strength in molality ( ), S is the molecular mass of the solvent, and C is the ion-interaction parameter. For aqueous NaCl solution, when T=298.15 K, is 0.5085 mol -1/2 kg 1/2 and is 0.3282 Å -1 mol -1/2 K 1/2 . 0 is 3.78 Å for both Na + and Cl -, S is 18.02 g mol -1 , C is 0.105 kg 2 mol −2 for Na + ion and -0.009 kg 2 mol −2 for Cl -. This Debye-Hückel extended equation is valid up to about 1 mol kg -1 2 .
Assuming ideal ion-exchange membranes ( counter is unity), the junction potential is about 0.19 V for a membrane-pair when the concentrations of HC and LC solutions are 35 g L -1 and 0.7 g L -1 , respectively. Considering the chemical potential difference between the HC and LC solutions is 0.19 V, the chemical potential difference driven by salinity gradient is totally converted into the electric potential. In the actual ion-exchange membranes, the junction potential is predicted to 0.17 is the flow rate of HC and LC solutions (L −1 ) and ∆ is the pressure drop (Pa) over the RED stack 3 .
The H2 chemical power produced ( water splitting ) is expressed as: SC is the short-circuit photocurrent density and F is the faradaic efficiency for hydrogen-and oxygen-evolving reactions 4 .

Supplementary Table S2. Energy conversion efficiency ( ).
Flow rate (mL min -1 ) (%) 9 0.55 6 0.90 3 1.84 Notes: Energy conversion efficiency ( ) is calculated from the ratio of the net produced power to the total input power as: The total system output power ( input ) is the sum of the solar power (100 mW cm -2 ) and the salinity driven power. 3,5 The salinity-driven power is calculated from the change in the Gibbs free energy when the HC and LC are completely mixing as 3

Series connection between NiMo/Si photoelectrode and RED stack
In

Energy diagram of the PEC-RED system
The NiMo/Si photocathode and the Ni foam anode have a leveled Fermi level because they are connected by a conductive wire. Under illumination, the concentration of minority carriers (electrons) increases and thus the quasi-Fermi level of the electrons is separated from that of the holes at the photocathode. When the quasi-Fermi level of the electrons is more negative than the thermodynamic water reduction potential (on the electrochemical scale), the hydrogen evolution reaction occurs. The RED stack builds an electric field in the electrolyte, which changes the relative position of the water reduction/oxidation potential. As a results, the thermodynamic water oxidation potential is more negative than the Fermi level of the electrons in the anode and thus oxygen evolution reaction occurs.

Economic issues of the PEC-RED system
In the PEC-RED system, it is relatively easy to obtain high voltage by increasing the number of membranes. But the power density is low because of the high resistance of the RED stack.
Therefore, the most of the voltage drop occurs across the RED stack when the current increases.
Consequently, this system is not suitable for situations where a high current flows. Moreover, it is estimated that membranes occupy up to 80% of total capitals in salinity gradient-based energy generation systems including RED. 15 Therefore, the feasibility of PEC-RED system depends mainly on the membrane price. Since the currently available commercial membranes are developed for use in electrodialysis, they have high level of robustness and are rather thick. However, the commercial membranes for electrodialysis are overqualified for application in RED because the