A pilot-scale sustainable hydrogen production system using reverse electrodialysis (RED) technology was launched. The system is based on direct conversion of salinity gradient energy (SGE) between seawater (SW) and sewage treated water (STW) to hydrogen production by water electrolysis. The hydrogen production rate was almost the same as the theoretical value. This indicates that the RED hydrogen production system can convert SGE between SW and STW to hydrogen energy at high current efficiency.

Utilization of renewable energy and its diversification is quite important to ensure energy safety for human's sustainability. Since renewable energy such as solar- and wind-based energy has intermittent and fluctuating characteristics, power-to-gas technology has a great interest to control electricity grid by utilizing electricity to split water into hydrogen (H2) and oxygen (O2) (Götz et al. 2016). This H2 can be used as efficient gaseous fuel for fuel cell vehicle (Mori & Hirose 2009), fuel cell-based power supplying system (Reiter & Lindorfer 2015), injection into a natural gas pipeline network (Walker et al. 2016) and hydrogen-fueled combustion turbine for electric power generation (Taamallah et al. 2015) without CO2 emission. Therefore, to overcome the intermittent and fluctuating nature of renewable energy, a combination of renewable energy-based electricity and hydrogen are promising to be useful in future.

Salinity gradient energy (SGE) is a sustainable and environmentally friendly energy related to the electrochemical potential between two solutions with different salinities (Giacalone et al. 2018). It has been theoretically estimated that about 0.5 kWh of energy can be potentially captured by mixing 1 m3 of seawater (SW) with 1 m3 of river water (Veerman et al. 2009), and the estimated global potential is 2,000 TWh/year (Logan & Elimelech 2012). Because Japan is an island nation, there are many opportunities to access fresh water and SW simultaneously along the coastline, such as various manufacturing factories, thermal power plants, municipal wastewater plants and so on, where fresh wastewater is discharged into the sea directly after adequate treatment. Therefore, SGE between these fresh water and SW streams potentially can be utilized.

Reverse electrodialysis (RED) is one of the emerging technologies for capturing SGE by applying ion exchange membranes (IEMs) (Daniilidis et al. 2014; Hong et al. 2015). In RED, low salinity (LS) and high salinity (HS) solutions flow alternatively between anion-exchange membranes (AEMs) and cation-exchange membranes (CEMs). Anions and cations in the HS solution move to the LS solution through AEMs and CEMs in opposite directions under their respective concentration gradients, which can be converted to electricity by redox reaction on the electrodes (Holladay et al. 2009; Geise et al. 2013; Tedesco et al. 2015). If the optimum electrode and electrode solution is chosen, H2 and O2 are generated through water electrolysis reaction. Therefore, the RED system enables to directly convert SGE into useful gaseous fuels.

In this study, we demonstrate a direct hydrogen production system from SGE using RED technology, here called a RED-H2 system, using SW and sewage treated water (STW). In Japan, about 17% of the sewage treatment plants directly dispose their treated wastewater into the sea, and therefore, there is a nice opportunity to generate H2 (also O2) using the system. Moreover, wastewater treatment plants exist in all communities and are often located in cities, which have a high H2 demand, or their environs. This means that excess H2 transportation is not needed and the RED-H2 system can be easily used through a pipeline network in the respective local communities. Therefore, a RED-H2 system in a sewage treatment plant near the sea will contribute to constructing decentralized energy systems with high social energy security and diversity in the future.

Figure 1 shows a schematic illustration of the RED-H2 system. The system has number of pairs of energy converting cells stacked between two electrode compartments. Each cell has a CEM, an AEM and two flow channels of SW and STW, and will generate ca. 0.15 V when SW and STW are fed to the channels. Hence, for an example, the stack of 20 pairs of the cells will give ca. 3 V of voltage between the electrodes. Hydrogen and oxygen gases generate at the electrodes by water electrolysis driven by the voltage.

Figure 1

A schematic diagram of a RED hydrogen production system.

Figure 1

A schematic diagram of a RED hydrogen production system.

Close modal

Figure 2 shows the flow chart of the RED-H2 system built in Fukuoka city. In the system, Neosepta® AMX and CMX (Astom Co. Japan) were used as AEM and CEM, respectively. The fundamental properties of the membranes are listed in Table 1. These membranes are standard IEMs for electrodialysis (ED) applications. In this study, a commercial ED stack was used as a RED-H2 stack. The characteristics of the stack are listed in Table 2. The inter-membrane distance between CEM and AEM in the cell is 600 μm. The effective membrane area is 1,000 cm2 (20 cm × 50 cm), and number of pairs of the cell was 200. Hence, the total membrane area of the stack was 40 m2. The electrode material of the two electrodes was Pt coated titanium plates, and 5 wt% Na2SO4 was fed to the electrode compartments as electrode solutions. The electrodes were connected to an electrochemical measuring device (PLZ 164 W/Kikusui Electronics Co.) to change the load resistance for controlling the voltage-current relationship of the system. Generated H2 and O2 gases at the respective electrode were trapped using gas trapping instruments to measure the volume at a predetermined time interval.

Table 1

Fundamental properties of the membranes used in the RED-H2 stack

MembraneThickness [μm]Electric resistance [Ωcm2]Transport number [–]Water content [–]
Neosepta® CMX 170–190 2.3 0.98< 0.25–0.30 
Neosepta® AMX 160–180 3.0 0.98< 0.25–0.30 
MembraneThickness [μm]Electric resistance [Ωcm2]Transport number [–]Water content [–]
Neosepta® CMX 170–190 2.3 0.98< 0.25–0.30 
Neosepta® AMX 160–180 3.0 0.98< 0.25–0.30 
Table 2

Characteristics of RED-H2 stack

Membranes CMX/AMX 
Active membrane area 0.1 m2 
Inter-membrane distance 600 μm 
No. of pair 200 
Total membrane area 40 m2 
Electrodes Ti-Pt coating 
Electrode solution 5 wt.% Na2SO4 
Membranes CMX/AMX 
Active membrane area 0.1 m2 
Inter-membrane distance 600 μm 
No. of pair 200 
Total membrane area 40 m2 
Electrodes Ti-Pt coating 
Electrode solution 5 wt.% Na2SO4 
Figure 2

A flow chart of the RED-H2 system.

Figure 2

A flow chart of the RED-H2 system.

Close modal

STW was obtained from Wagiro water treatment center (Fukuoka, Japan). It was difficult for us to get SW directly from Hakata bay. Hence, brine from a SW desalination center (Fukuoka, Japan) with ca. 90 mS/cm of ionic conductivity and STW were mixed to get salt water with 50 mS/cm of ionic conductivity, which is the same concentration as SW at Hakata bay. Here, the salt solution is called SW. Both the brine and STW were filtered by cartridge filter and fiber filter (pore size: 10 μm), respectively, and adequately mixed to control their ionic conductivity. After this, they were fed to RED-H2 stack. Because the brine was already treated by sand filtration in the desalination plant, we chose cartridge filter for the brine. The STW was mainly treated by anaerobic-oxix activated sludge (A/O) process in Wagiro water treatment center. Ionic conductivity, hydraulic pressure, flow rate and temperature of SW and STW were monitored with measurement instruments at the indicated points in Figure 2. The experiment conditions in the test were listed in Table 3.

Table 3

Experiment conditions of the RED-H2 system

Brine conductivity 78–96 mS/cm 
SW conductivity 50 mS/cm 
STW conductivity 1.0–2.5 mS/cm 
SW flow rate 2–8 L/min 
STW flow rate 2–8 L/min 
Electric current 0.5–3.0 A 
Water temperature 18–28.5 °C 
Brine conductivity 78–96 mS/cm 
SW conductivity 50 mS/cm 
STW conductivity 1.0–2.5 mS/cm 
SW flow rate 2–8 L/min 
STW flow rate 2–8 L/min 
Electric current 0.5–3.0 A 
Water temperature 18–28.5 °C 
In an RED stack, theoretical maximum voltage (open circuit voltage: VOC) is given in terms of the following equation (Mei & Tang 2018):
formula
(1)
where F, R, z and T are Faraday constant, gas constant, valance of ions and temperature, respectively. t+ and t are transport number of CEM and AEM, respectively. γ, x and C are activity coefficient, molar fraction of salt and salt concentration, respectively.
Theoretical VOC under the experimental conditions is about 35.4 volt under an approximation that ion conductivity was converted into NaCl concentration. Figure 3 shows an example of the voltage-current curve during the RED test. The open circuit voltage (current = 0) of the system showed 28.6 V, meaning that the stack voltage was about 20% less than the theoretical voltage. There will be two factors which cause VOC reduction: (1) the effect of divalent ions in SW such as Mg2+, Ca2+ and SO42− (Post et al. 2009), and (2) the decrease in the concentration ratio inside the stack even at zero current condition. The voltage decreased with increasing current, and the slope of the voltage-current curve then gives the internal electric resistance of the RED-H2 stack. Theoretical internal electric resistance of the stack is given in terms of the following equations (Długołȩcki et al. 2009).
formula
(2)
where R, N, Sm, d and EC are resistance, number of cell pair, effective area, inter-membrane distance between CEM and AEM, and electric conductivity of the solutions, respectively. The subscripts of el, AEM, CEM, HS and LS are electrode, AEM, CEM, HS compartment and LS compartment, respectively. Subscript of m2 means resistance per 1 m2. Theoretical calculated resistance without electrode from salt concentrations of SW and STW at the inlet was 13.3 Ω, which was higher than the evaluated resistance (10 Ω) from the experimental data. In Rint, the RLS is dominant (about 90% of Rint without Rel), and the RLS decreased within the stack because electric conductivity of LS solution increased due to the salt diffusion from HS to LS along the current. Therefore, the Rint was slightly decreased with increasing the current. The output power became maximum of 15.3 W at 1.4 A of the current.
Figure 3

The voltage and output power of the stack as a function of electric current. Solid curve: stack voltage; broken curve, output power. Flow rate of SW and STW was 4.3 L/min., water temperature, 28.5 °C.

Figure 3

The voltage and output power of the stack as a function of electric current. Solid curve: stack voltage; broken curve, output power. Flow rate of SW and STW was 4.3 L/min., water temperature, 28.5 °C.

Close modal

Figure 4 shows hydrogen production of the system at a constant current operation of 1.5 A as a function of time. The hydrogen production was directly proportional to the electric quantity (electric current × time). The hydrogen production rate calculated from the slope was 0.90 L/h, which is almost the same as the theoretical value. This indicates that the electric current conversion efficiency of hydrogen production in the system is ca. 100%.

Figure 4

Hydrogen production of the system as a function of time. Solid circles experiments; broken line, theoretical value obtained from the electric quantity. Flow rate of SW and STW was 4.3 L/min., water temperature, 28.5 °C.

Figure 4

Hydrogen production of the system as a function of time. Solid circles experiments; broken line, theoretical value obtained from the electric quantity. Flow rate of SW and STW was 4.3 L/min., water temperature, 28.5 °C.

Close modal

Figure 5 shows the seasonal effect on water temperature and subsequent power output performance of the RED-H2 system. During 7 months (Sep. 2016–Feb 2017), water temperature decreased from 28 °C to 18 °C, and about 30% of the power output decreased in this period. Because water temperature influences both VOC and Rint, a higher temperature is suitable to obtain a higher energy output. From this evaluation, temperature dependence was estimated as about 3%/°C.

Figure 5

Water temperature (SW and SWT) and power output performance during 7 months. Hydrogen production of the system as a function of time. Solid circles experiments; broken line, theoretical value obtained from the electric quantity. Flow rate of SW and STW was 4.3 L/min., water temperature, 28.5 °C. Flow rate of SW and STW was 4.3 L/min.

Figure 5

Water temperature (SW and SWT) and power output performance during 7 months. Hydrogen production of the system as a function of time. Solid circles experiments; broken line, theoretical value obtained from the electric quantity. Flow rate of SW and STW was 4.3 L/min., water temperature, 28.5 °C. Flow rate of SW and STW was 4.3 L/min.

Close modal

To estimate the possibility of long-time operation without pre-filtration, the long-time performance of the system without fiber filtration was evaluated. The converted power output (at 23 °C from the 3%/°C relation) is shown in Figure 6. We evaluated the converted power output without fiber filter by using STW directly. In this case, reduction of power output was not observed in 300 h even in the absence of fiber filter. However, after then, power output gradually decreased, and 80% reduction was observed at 800 h. The visual observation inside the stack after disassembling the stack revealed that this performance reduction would be caused by clogging of the foulants especially at the STW compartment. To investigate the performance recovery, we measured the performance after simple cleaning (without chemical treatment) inside of the disassembled stack. The converted power output of the system after cleaning was also shown in Figure 6. The result revealed that the simple cleaning allowed the power output performance (voltage and current) completely recovered. Moreover, we also confirmed the resistance and thickness of the IEMs were same to those of initial values. Therefore, this results clearly supported the clogging of the foulants which was easy removable by simple cleaning was the main reason of performance reduction, and the membrane fouling was not severe (at least reversible). In addition, optimum method for cleaning in place (i.e. cleaning inside of the RED-H2 stack without disassembly) also should be developed in the future, to achieve stable H2 (and O2) production by utilizing this power output directly in water electrolysis.

Figure 6

Power output (converted to those at 23 °C) in long time operation without fiber filtration for STW. Flow rate of SW and STW was 4.3 L/min.

Figure 6

Power output (converted to those at 23 °C) in long time operation without fiber filtration for STW. Flow rate of SW and STW was 4.3 L/min.

Close modal

In this research, we have developed the direct H2 production system from SGE utilizing RED technology, called the RED-H2 system, at pilot-scale. The RED-H2 system can convert SGE between SW and STW to hydrogen at high current efficiency. Long-time evaluation of the RED-H2 system revealed that temperature dependence of the performance was about 3%/°C, and stable operation in 1,100 h was achieved by using STW treated by fiber filtration (pore size: 10 μm). Because Japan has many opportunities, with 17% of sewage treatment plants directly disposing STW into the sea, the RED-H2 system will contribute to developing a H2 decentralized energy system network for high energy security and diversity in the future.

This work was supported by feasibility study of the Breakthrough by Dynamic Approach in Sewage High Technology (B-DASH) Project conducted by the Ministry of Land, Infrastructure, Transport and Tourism.

Daniilidis
A.
,
Vermaas
D. A.
,
Herber
R.
&
Nijmeijer
K.
2014
Experimentally obtainable energy from mixing river water, seawater or brines with reverse electrodialysis
.
Renewable Energy
64
,
123
131
.
Długołȩcki
P.
,
Gambier
A.
,
Nijmeijer
K.
&
Wessling
M.
2009
Practical potential of reverse electrodialysis as process for sustainable energy generation
.
Environmental Science & Technology
43
,
6888
6894
.
Geise
G. M.
,
Curtis
A. J.
,
Hatzell
M. C.
,
Hickner
M. A.
&
Logan
B. E.
2013
Salt concentration differences alter membrane resistance in reverse electrodialysis stacks
.
Environmental Science & Technology Letters
1
(
1
),
36
39
.
Giacalone
F.
,
Catrini
P.
,
Tamburini
A.
,
Cipollina
A.
,
Piacentino
A.
&
Micale
G.
2018
Exergy analysis of reverse electrodialysis
.
Energy Conversion and Management
164
,
588
602
.
Götz
M.
,
Lefebvre
J.
,
Mörs
F.
,
Koch
A. M.
,
Graf
F.
,
Bajohr
S.
,
Reimert
R.
&
Kolb
T.
2016
Renewable Power-to-Gas: a technological and economic review
.
Renewable Energy
85
,
1371
1390
.
Holladay
J. D.
,
Hu
J.
,
King
D. L.
&
Wang
Y.
2009
An overview of hydrogen production technologies
.
Catalysis Today
139
,
244
260
.
Hong
J. G.
,
Zhang
B.
,
Glabman
S.
,
Uzal
N.
,
Dou
X.
,
Zhang
H.
,
Wei
X.
&
Chen
Y.
2015
Potential ion exchange membranes and system performance in reverse electrodialysis for power generation: a review
.
Journal of Membrane Science
486
,
71
88
.
Mori
D.
&
Hirose
K.
2009
Recent challenges of hydrogen storage technologies for fuel cell vehicles
.
International Journal of Hydrogen Energy
34
,
4569
4574
.
Reiter
G.
&
Lindorfer
J.
2015
Global warming potential of hydrogen and methane production from renewable electricity via power-to-gas technology
.
The International Journal of Life Cycle Assessment
20
(
4
),
477
489
.
Taamallah
S.
,
Vogiatzaki
K.
,
Alzahrani
F. M.
,
Mokheimer
E. M. A.
,
Habib
M. A.
&
Ghoniem
A. F.
2015
Fuel flexibility, stability and emissions in premixed hydrogen-rich gas turbine combustion: technology, fundamentals, and numerical simulations
.
Applied Energy
154
,
1020
1047
.
Tedesco
M.
,
Brauns
E.
,
Cipollina
A.
,
Micale
G.
,
Modica
P.
,
Russo
G.
&
Helsen
J.
2015
Reverse electrodialysis with saline waters and concentrated brines: a laboratory investigation towards technology scale-up
.
Journal of Membrane Science
492
,
9
20
.
Veerman
J.
,
Saakes
M.
,
Metz
S. J.
&
Harmsen
G. J.
2009
Reverse electrodialysis: performance of a stack with 50 cells on the mixing of sea and river water
.
Journal of Membrane Science
327
(
1–2
),
136
144
.
Walker
S. B.
,
Mukherjee
U.
,
Fowler
M.
&
Elkamel
A.
2016
Benchmarking and selection of Power-to-Gas utilizing electrolytic hydrogen as an energy storage alternative
.
International Journal of Hydrogen
41
,
7717
7731
.