Photoelectrochemical Demonstrator Device for Solar Hydrogen Generation

Transcription

1 Photoelectrochemical Demonstrator Device for Solar Hydrogen Generation Project Deliverable Report D1.5 and MS8 Deliverable title: Deliverable nature: Dissemination level: Lead beneficiary: Contracted date of delivery: Actual date of delivery: Author(s): Second-generation hybrid PEC-PV device with an efficiency of 10% Report Public HZB Mar-17 Mar-17 Fatwa F. Abdi, Yimeng Ma, Ji-Wook Jang, Ibbi Ahmet, Roel van de Krol, Bernd Stannowski (HZB); Matthew Mayer, Min-Kyu Son (EPFL); Avner Rothschild, Hen Dotan, Yifat Piekner, Avigail Landman (Technion); Adelio Mendes, Tania Lopes, Antonio Vilanova (U Porto); Michael Wullenkord (DLR); Artjom Maljusch (Evonik) PECDEMO is a Collaborative Project co-funded by FCH JU under the call SP1-JTI-FCH GA n : Start date: April 1 st, Duration: 36 months.

2 TABLE OF CONTENTS 1. Executive Summary Detailed Report on the Deliverable Background and Objectives Results and Discussion BiVO 4-HIT Silicon tandem device Cu 2O-HIT Silicon tandem device Conclusions and next steps References Appendices... 9

3 1. EXECUTIVE SUMMARY PECDEMO consortium aims to further strengthen Europe s leading position in the field of solar fuels by focusing at the development of a hybrid photoelectrochemicalphotovoltaic (PEC-PV) tandem device for solar water splitting. The project s aim is to demonstrate a large scale device (50 cm 2 ) with high solar-to-hydrogen (STH) efficiency of 8-10% and long term stability (1000 hours). The first work package (WP1) focuses on small-area devices (< 1 cm 2 ) based on the three metal oxides of choice (Fe2O3, Cu2O and BiVO4) and PV devices (e.g., multi-junction Si, perovskite). The results from this work package are expected to be translated to the final large scale device. Within the three years of the work package, we have accomplished almost all the set deliverables on photocurrent, efficiency, photon management and stability. The objective of this final deliverable of the work package is then to combine all the accomplishments to date, in order to show a 2 nd generation device (i.e., PEC-PV tandem using a photon management strategy) that shows 10% STH efficiency. BiVO4-based devices, although not meeting the deliverable target, came very close. STH efficiencies of 7.5% and 9.2% are obtained for the combination of a dual BiVO4 photoanode with a HIT series silicon solar cell and a dual BiVO4-Fe2O3 photoanode with a HIT series silicon solar cell, respectively. The deliverable target is achieved by combining a Cu2O-Ga2O3 NW photocathode with a HIT-series silicon solar cell. STH efficiencies of 10.2% and 16.3% are obtained for the combination with 2- and 3-HIT series silicon solar cells, respectively. The results therefore represent and summarize the successful efforts of the work package, and PECDEMO project in general. Finally, the report of milestone 8 is also included in this report, which identifies the most promising hybrid PEC-PV device concept. Page 1

4 2. DETAILED REPORT ON THE DELIVERABLE 2.1. Background and Objectives Solar water splitting with appreciable solar-to-hydrogen (STH) efficiencies have only been achieved with tandem devices, i.e., two (or more) light absorbing materials having different bandgaps are used. The large bandgap material is usually used as a photoelectrode, and the small bandgap material is typically a solar cell placed behind the photoelectrode. 1-3 This tandem configuration allows the shortwavelength part of the solar spectrum to be absorbed by the photoelectrode, while providing the transmitted spectrum for the solar cell to provide the bias potential needed by the photoelectrode. The solar spectrum can thus be utilized more efficiently (as compared to single-absorber configuration), and that no external bias potential is needed to perform solar water splitting (i.e., all photocurrents are translated to STH efficiency). Since the attainable STH efficiency is normally limited by the photocurrent of the photoelectrode, PECDEMO initially focused on improving the performance of our selected photoelectrodes: Fe2O3, BiVO4 and Cu2O. We have successfully improved the performance by achieving the D1.1 target: a photocurrent of 5.4 ma/cm 2 was obtained with Cu2O photocathode and ~5 ma/cm 2 with BiVO4 photoanode. The next step was to meet the device target at the midterm of the PECDEMO project: 1 st generation device with 8% STH efficiency (D1.2). Here, 1 st generation device corresponds to a stacked tandem device configuration, where no specific photon management strategy is used. Using a combination of a BiVO4 photoanode and a HIT series solar cell, we achieved 7.5% STH efficiency, which is very close to the deliverable target. Beyond this point, we have attempted to improve the STH efficiency of our tandem devices, by taking advantage of photon management strategies (the resulting tandem devices are then referred to as 2 nd generation devices ). Several photon management strategies, including distributed Bragg reflector (DBR) and dichroic mirror that have been described in D1.3, were then applied to the PEC-PV tandem devices. The deliverable reported here (D1.5) describes these efforts, with a goal of achieving 10% STH efficiency using a 2 nd generation device. Page 2

5 2.2. Results and Discussion BiVO4-HIT Silicon tandem device In the midterm of the project, we have achieved 7.5% water oxidation on two tandem BiVO4 photoanodes combined with silicon solar cells. For such high efficiency, efficient collection of photons below the absorption band (<500 nm) BiVO4 is essential, especially for the photons between 450 nm and 500 nm where the absorbed photon to current efficiency is still modest. 4 Distributed Bragg reflector (DBR) can be used to efficiently collect photons and convert these photons to photocurrent. As a result, we can possibly attain a solar water splitting device with the same (or better) efficiency without the need of having two separate BiVO4 photoanodes, as shown in Figure 1a. Figure 1. (a) Schematic presentation of design strategy of employing DBR to couple two BiVO4 photoanodes into one integrated photoelectrode. (b) AM1.5 photocurrent-voltage curve of H2-treated gradient-doped W:BiVO4 in a single- and dual-electrode configuration, in an electrolyte containing Na2SO3 as a hole scavenger. (c) Photocurrent improvements by using DBR in 100 nm and 200 nm thick BiVO4 photoanodes. Page 3

6 We have shown that the photocurrent improvement that is expected from a single to a dual BiVO4 photoanode to be ~20-30%, as shown in Figure 1b. Indeed, Figure 1c shows that on a 100 nm thick BiVO4 photoanode, the improvement in photocurrent is c.a. 20 %. However, when employing a 200 nm thick BiVO4 photoanode, the improvement in photocurrent with DBR was modest. We attribute this to a modest amount scattering that our films have (as observed in our UV-vis spectrum), due to the roughness of our films. Better optical optimization is therefore needed before we can successfully use this strategy in our BiVO4. In order to improve the STH efficiency closer to the 10% target of this deliverable, we have attempted to address the fundamental limitation of BiVO4, which is the relatively large bandgap of 2.4 ev. While our efforts in decreasing the bandgap of BiVO4 is still ongoing, we have simply extended the light harvesting ability of the device by combining a BiVO4 and a Fe2O3 photoanode into a dualphotoanode configuration (Figure 2). As a result, we obtained STH efficiency of 9.2%, which is very close to the Deliverable 1.5 target. Detailed results on this dual BiVO4- Fe2O3 photoanode is recently published in Nature Communications. 5 Figure 2. Schematic illustration of a dual BiVO4-Fe2O3 photoelectrode used in tandem configuration with a silicon solar cell. An STH efficiency of 9.2% has been achieved with this configuration Cu2O-HIT Silicon tandem device We have previously demonstrated that photocathodes based on Cu2O Ga2O3 core shell nanowires (NW) show both enhanced photocurrent density and early onset potential. Notably, the onset potential around 1.0 V vs RHE represents a benchmark photovoltage from Cu2O photocathodes and enables the devices to be Page 4

7 integrated into PEC-PV tandems for standalone solar-driven water splitting. Based on the previous estimation, the combination of a Cu2O Ga2O3 NW photocathode and HIT PV cell could achieve a solar-to-hydrogen (STH) efficiency above 10 %, which is a final goal of Deliverable 1.5 and WP1 in PECDEMO project. To demonstrate this, we assembled a hybrid PEC-PV tandem cell employing a Cu2O Ga2O3 NW photocathode, a HIT PV cell and an IrO2 anode. For this demonstration, we adopted the tandem configuration with a 600 nm cut-off wavelength dichroic mirror (as shown in Figure 3). This configuration has been identified in our Milestone 3 (Midterm assessment of best photoelectrode and PV candidates for 2 nd generation devices). The Cu2O NW photocathode (active area: cm 2 ) was prepared using optimal conditions with a 20 nm thick Ga2O3 overlayer, a 20 nm thick TiO2 protection layer and RuOx HER catalyst. Two HIT PV cells produced by HZB were used: one is a 2 series connected HIT PV cell (2 HIT) and the other is a 3 series connected HIT PV cell (3 HIT), each having an active area of 1 cm 2 per individual HIT cell. Figure 3. Schematic setup of our PEC-PV tandem configuration, consisting of a Cu2O photocathode, an IrO2 anode, a HIT solar cell and a dichroic mirror. Figure 4 shows the results of this demonstration, which was tested in near neutral solution (ph 5). The Cu2O Ga2O3 NW photocathode shows an onset potential of +0.9 V vs RHE and reaches a photocurrent density of 10 ma cm -2 at 0 V vs RHE (Fig. 4a), while 2 HIT/3 HIT PV cells behind the dichroic mirror show an open circuit voltage of 1.2 V and 1.8 V, respectively (Fig. 4b). Based on the intersection of current voltage curves of each component (Fig. 4c), an expected operating short-circuit photocurrent is 0.65 ma (6.98 ma cm -2 ) with 2 HIT PV cell and 1.02 ma (10.96 ma cm - 2 ) with 3 HIT PV cell, respectively. Indeed, these are observed in the short-circuit Page 5

8 measurements of the real PEC-PV tandem device under AM 1.5 illumination (Fig. 4d). The corresponding STH efficiencies are 10.3 % (2 HIT PV) and 16.2% (3 HIT PV). Although we still have a remaining issue that the active area was not the same between the Cu2O NW photocathode and the HIT PVs, the STH efficiency of PEC-PV tandem device with the Cu2O Ga2O3 NW photocathode, 2 HIT/3 HIT PV cells and the dichroic mirror fully satisfied the goals targeted in Deliverable 1.5 and WP1 of the PECDEMO project. We note that even if we corrected the area of the photoelectrode and the solar cell to their respective area, we obtained STH efficiencies of 8.4% (2 HIT PV) and 11.8% (3 HIT PV), thereby also satisfying the deliverable and work package target. Figure 4. (a) One-sun (AM 1.5) photocurrent density-voltage curve of Cu2O Ga2O3 NW photocathode in ph 5 solution. (b) AM 1.5 photocurrent density-voltage curves of 2 HIT and 3 HIT PVs behind the dichroic mirror (600 nm wavelength cut-off). (c) Individual current voltage curves of Cu2O Ga2O3 NW photocathode, 2 HIT/3 HIT PV cells and IrO2 OER anode. (d) Chopped AM 1.5 short-circuit photocurrent density and calculated STH efficiency of the real PEC-PV tandem device. The green dashed line is the 10% STH efficiency target. Page 6

9 2.3. Conclusions and next steps We successfully achieved the goal of Deliverable 1.5 (Second-generation hybrid PEC-PV device with an efficiency of 10%) using a Cu2O Ga2O3 NW photocathode and 2 HIT/3 HIT PV cells with a 600 nm cut-off wavelength dichroic mirror (10.3 % with 2 HIT PV cell and 16.2% with 3 HIT PV cell). However, scale-up of Cu2O Ga2O3 NW photocathodes with high activity and long-term stability remains a challenge. Therefore, future works on the PEC-PV tandem device with this combination will be focused on understanding and improving the stability of Cu2O Ga2O3 NW photocathodes fabricated on larger scales. In addition, the devices based on BiVO4, which have already shown efficiencies close to the 10% target will be further developed. Then, we can obtain the durable hybrid PEC-PV tandem device with a STH efficiency above 10% in the near future. Page 7

10 3. REFERENCES (1) Brillet, J.; Yum, J. H.; Cornuz, M.; Hisatomi, T.; Solarska, R.; Augustynski, J.; Graetzel, M.; Sivula, K. Nat Photon 2012, 6, 824. (2) Abdi, F. F.; Han, L.; Smets, A. H. M.; Zeman, M.; Dam, B.; van de Krol, R. Nat. Commun. 2013, 4. (3) Pihosh, Y.; Turkevych, I.; Mawatari, K.; Uemura, J.; Kazoe, Y.; Kosar, S.; Makita, K.; Sugaya, T.; Matsui, T.; Fujita, D. Scientific Reports 2015, 5. (4) Abdi, F. F.; Firet, N.; van de Krol, R. ChemCatChem 2013, 5, 490. (5) Kim, J. H.; Jang, J.-W.; Jo, Y. H.; Abdi, F. F.; Lee, Y. H.; van de Krol, R.; Lee, J. S. Nat. Commun. 2016, 7, (6) Rothschild, A.; Dotan, H. ACS Energy Letters 2017, 2, 45. (7) Landman, A.; Dotan, H.; Shter, G. E.; Wullenkord, M.; Houaijia, A.; Maljusch, A.; Grader, G. S.; Rothschild, A. Nat Mater 2017, advance online publication. Page 8

11 4. APPENDICES Milestone 8 Report: Identification of most promising hybrid PEC-PV device concept PECDEMO consortium has investigated PEC-PV tandem solar water splitting device in the past three years. Three metal oxides (Fe2O3, BiVO4 and Cu2O) have been mainly focused as the photoelectrode materials. The amount of progress that we showed with these materials have further strengthened our device concept based on metal oxides. The fact that we have obtained almost all deliverable targets within PECDEMO validates the concept for further development. Having mentioned that, we of course are not limited to the possibility that novel metal oxides may be developed in the future that have much better performance than the metal oxides studied by PECDEMO. Scaling up of these metal oxide photoelectrodes are proven to be challenging, but we have demonstrated the capability to do so within our consortium, and we expect this to be continuously improved within the community. As for the PV cells, both silicon-based and the new inorganic-organic perovskite cells have been extensively studied within PECDEMO. Both types did not show any major limitation; they instead provided flexibility in the resulting design of the PEC-PV device. Scaling up of these PV cells is not expected to be an issue, as also shown within our consortium. Beyond the materials consideration, we have also attempted to address several issues in arriving at the objective of this milestone, which is to identify the most promising hybrid PEC-PV device concept. The considerations include power management, cell design, and plant design (including life cycle and technoeconomic considerations). 1. Power management The most straightforward electrical coupling between the PV and PEC cells in PV-PEC tandem cells is connecting them in series. In this design, the same current flows in both cells and the photovoltages generated in each cell adds up together as in series connection of multijunction PV cells. The operation point of such tandem cells is at the cross point of the current voltage (J-V) curves of the PV and PEC cells. Page 9

12 The main problem in this design is that the photocurrent is limited by the worse cell, and the operation point is fixed and it cannot track the maximum power point of the system. This leads to significant power loss and generation of excess heat. These problems can be rectified by connecting another electric load in parallel to the PEC cell. This allows dividing the overall power generated by the PV cell between the PEC cell and the parallel load, thereby leading to co-production of both hydrogen and electrical power. This solution significantly increases the overall efficiency of the tandem device. Furthermore, since the parallel load acts as a sink for excess power generated by the PV cell, it reduces potentially harmful heating. This power management approach enables tracking the maximum power point by using a DC/DC convertor. Figure 5 compares the overall conversion efficiency of different device configurations that co-generated both hydrogen and electrical power, normalized by the power conversion efficiency of the PV cells that drive the system. Thus, the figure of merit, FOM = Ptotal / PPV, is plotted as a function of the fraction of chemical power (i.e., hydrogen production) out of the total power (electricity and hydrogen) generated by the system, Xchem = Pchem / Ptotal. The figure shows the potential advantages of this power management approach to enhance the efficiency of PV- PEC tandem cells. For more details see: A. Rothschild and H. Dotan, Beating the Efficiency of Photovoltaics- Powered Electrolysis with Tandem Cell Photoelectrolysis, ACS Energy Letters (2017), 2, Page 10

13 Figure 5. Calculated figure of merit (i.e., the overall power produced by the system devided by the power produced by the PV cells driving the system) as a function of the fraction of the total power production that goes toward chemical power generation, Xchem = Pchem/Ptotal, for hypothetical PVelectrolysis (black curve) and PV PEC tandem cells (color curves) that co-generate both chemical (i.e., hydrogen) and electrical power. The calculations assume PV cells with a power conversion efficiency of 20%, electrolysis efficiency of 68.3% for the electrolyzers, and PEC cells operating at the bias voltages and current densities as indicated by the color code in the legend. Solid lines represent a scenario in which all of the PV cells in the array are always covered by PEC cells in front of them that reduce 30% of their power due to light absorption in the PEC cells. Broken lines represent another scenario wherein the reduction in PV power production scales with the chemical power production by the PEC cells. The filled circles on the right-hand side of the figure represent conventional PV PEC tandem cells that produce only hydrogen and no electrical power. Their colors correspond to their respective current densities, as indicated in the legend. 2. Cell design PECDEMO proposed and assessed different PEC cell designs targeting maximizing the energy conversion efficiency, versatility and cost-effectiveness. A sub-modular prototype composed by four individual PEC cells, each with an active area of 50 cm 2, connected in parallel was designed, optimized, built and tested under artificial solar conditions and concentrated solar irradiation in field tests Figure 6. Covering all the PECEMO project requirements, this device has an open light path from the front to the back window allowing an easy incorporation of Page 11

14 PEC/PV tandem photoelectrodes, either arranged as a single photoelectrode or in a split configuration; cells within the module can operate separately or combined, where different combinations are possible. In addition, the metal counter-electrodes are located side-by-side to the photoelectrode physically separated by an ion exchange membrane. Computational Fluid Dynamics (CFD) simulations were used to develop the best geometry with a fully optimized electrolyte feed flowpath, improved heat dissipation, and efficient collection of evolved gas bubbles Figure 7. The cell design for the 10 x 5 cm 2 photoelectrodes weighted the ion transport resistance and the electric transport resistance at the transparent conductive oxide (TCO) layer of the photoelectrode. In the design of Figure 7, photoelectrode serves simultaneously as back window of the cell; this arrangement allows reducing the size of the device and also bringing several benefits in terms of material use and overall construction costs. This embodiment requires printing and sintering a metal frame around the window for decreasing the electrical resistance at the TCO layer and for allowing the electric contact between the photoelectrode and the external circuit. The versatility of the developed prototype allows operating either with a n- or p- type semiconductor as photoelectrode, in acid or alkaline media, indoor or outdoor and under standard or concentrated solar illumination conditions. Gathering together the knowledge gained during PECDEMO project and aiming the future commercialization of this technology, a 1 1 m 2 PEC-PV panel was designed as shown Figure 8. This innovative approach is now under development and comprises the use of tubular shape counter-electrodes supported in an ion exchange membrane where H2 or O2 can be collected separately from the electrolyte and a photoelectrode with incorporated metallic current collectors. Page 12

15 Figure 6. Sub-modular prototype assembled in the solar concentrator SoCRatus with a rectangular flat focus, built by DLR. Figure 7. CFD simulations of the PEC cell: a) Non optimized and b) optimized design in terms of temperature profile [electrolyte: water; total flow rate: 500 ml min -1 ; external temperature: 25 C under 10-SUN irradiance] and electrolyte flow distribution [electrolyte: water; total flow rate: 500 ml min -1 ; external temperature: 25 C]. Page 13

16 Figure 8. Envisioned 1 1 m 2 PEC-PV panel for future large scale application in the field. 3. Plant design The PEC-PV units have to be implemented in processes which allow an economically viable and an industrially applicable operation. A promising scenario is for example a hydrogen refueling station providing hydrogen at high pressure for vehicles. Numerous criteria are important for the choice of an appropriate location for hydrogen production plants (e. g., weather conditions, politics, terrain, and infrastructure). However, the most crucial criterion is a high level of solar irradiance as can be found for instance in Seville (Spain) or Negev (Israel). The solar plant should be operated at temperatures as high as reasonable in terms of stability and efficiency. Higher temperatures clearly decrease the required active cooling capacity because of enhanced passive cooling and correspondingly reduces the electricity demand of the plant. Active cooling by an air cooling system featuring blowers and suitable air / electrolyte heat exchangers is favored due to the secured availability of the cooling medium. Solar concentrators should be installed allowing a more compact design of the PEC-PV system including piping. Operation under concentrated solar irradiation offers a significant potential to reduce the hydrogen Page 14

17 production costs and the environmental impact of PEC-PV water splitting. Additionally, supply of all electric plant components (pumps, air blowers etc.) with electricity produced by solar cells of PEC-PV modules can further lower the environmental impact (global warming potential) of PEC-PV water splitting system. Novel plant design One of the greatest challenges in large-scale solar water splitting plants is the separation of the hydrogen from the oxygen and the collection and transport of the hydrogen from millions of PEC cells distributed in the solar field to a central hydrogen distribution facility. This involves an immense sealing and piping constructions that puts a heavy burden on the hydrogen production economy. To overcome this challenge we invented a new PEC cell design with separated oxygen and hydrogen cells. According to this design, the PEC solar cells produce oxygen that is simply discharged to the atmosphere, whereas the hydrogen is produced elsewhere in another cell. The ion exchange between the two cells is mediated with a set of auxiliary electrodes made of NiOOH and Ni(OH)2 that undergo a reversible redox reaction exchanging OH - ions with the primary electrodes (i.e., the photoanode and cathode) in the oxygen and hydrogen cells. This enables centralized hydrogen production far away from the solar field, as illustrated in Figure 9. For more details see: A. Landman et al., Photoelectrochemical Water Splitting in Separate Oxygen and Hydrogen Cells, Nature Materials (2017, DOI: /nmat4876). 7 Page 15

18 Figure 9. Conceptual illustration of a solar hydrogen refueling station with distributed PEC solar cells producing O2 and a centralized H2 generator. Page 16