Rechargeable Aqueous Batteries Based on Available Resources Investigation and Development towards Efficient Battery Performance Mylad Chamoun Academic dissertation for the Degree of Doctor of Philosophy in Inorganic Chemistry at Stockholm University to be publicly defended on Friday 15 February 2019 at 13.00 in Magnélisalen, Kemiska övningslaboratoriet, Svante Arrhenius väg 16 B.
Abstract Batteries employing water based electrolytes enable extremely low manufacturing costs and are inherently safer than Li-ion batteries. Batteries based on zinc, manganese dioxide, iron, and air have high energy relevancy, are not resource restricted, and can contribute to large scale energy storage solutions. Zinc has a rich history as electrode material for primary alkaline Zn–MnO2 batteries. Historically, its use in secondary batteries has been limited because of morphological uncertainties and passivation effects that may lead to cell failure. Manganese dioxide electrodes are ineffective as rechargeable electrodes because of failure mechanisms associated with phase transformations during cycling. The irreversibility of manganese dioxide is strongly correlated to the formation of the electrochemically inactive spinel, Mn3O4/ZnMn2O4. The development of the iron electrode for Fe–air batteries was initiated in late the 1960s and these batteries still suffer from charging inefficiency, due to the unwanted hydrogen evolution reaction. Meanwhile, the air electrode is limited in long-term operation because of the sluggish oxygen evolution and reduction kinetics. These limitations of the Fe–air battery yield poor overall efficiencies, which bring vast energy losses upon cycling. Herein, the limitations described above were countered for rechargeable Zn–MnO2 and Fe–air batteries by synthesizing electrode materials and modifying electrolyte compositions. The electrolyte mixture of 1 M KOH + 3 M LiOH for rechargeable alkaline Zn–MnO2 batteries limited the formation of the inactive spinels and improved their cycle life significantly. Further, the formation of the inactive spinels was overcome in mildly acidic electrolytes containing 2 M ZnSO4, enabling the cells to cycle reversibly at lower pH via a distinctive reaction mechanism. The iron electrodes were improved with the addition of stannate, which suppressed hydrogen evolution. Furthermore, optimal charge protocols of the iron electrodes were identified to minimize the hydrogen evolution rate. On the air electrode, the synthesized NiCo2O4 showed excellent bifunctional catalytic activity for oxygen evolution and reduction, and was incorporated to a flow assisted rechargeable Fe–air battery, in order to prove the practicability of this technology. Studies of the electrode materials on the micro, macro, nano, and atomic scales were carried out to increase the understanding of the nature of and interactions between of these materials. This included both in operando and ex situ characterization. X-ray and neutron radiation, and analytical- and electrochemical methods provided insight to improve the performance and cycle life of the batteries.
Keywords: rechargeable aqueous batteries, alkaline electrolytes, aqueous sulfate electrolytes, zinc electrodes, manganese dioxide electrodes, iron electrodes, air electrodes, oxygen electrocatalysts.
Stockholm 2019 http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-163154
ISBN 978-91-7797-552-6 ISBN 978-91-7797-553-3
Department of Materials and Environmental Chemistry (MMK)
Stockholm University, 106 91 Stockholm
RECHARGEABLE AQUEOUS BATTERIES BASED ON AVAILABLE RESOURCES
Mylad Chamoun
Rechargeable Aqueous Batteries Based on Available Resources
Investigation and Development towards Efficient Battery Performance
Mylad Chamoun ©Mylad Chamoun, Stockholm University 2019
ISBN print 978-91-7797-552-6 ISBN PDF 978-91-7797-553-3
Cover: Vector images adapted from Vecteezy.com & Freepik.com
Printed in Sweden by Universitetsservice US-AB, Stockholm 2019 Doctoral Thesis 2019 Department of Materials and Environmental Chemistry Arrhenius Laboratory, Stockholm University SE-10691 Stockholm, Sweden
Faculty opponent:
Prof. Ann Mari Svensson Department of Materials Science and Engineering Norwegian University of Science and Technology (NTNU)
Evaluation committee:
Prof. Göran Lindbergh Department of Chemical Engineering and Technology The Royal Institute of Technology (KTH), Sweden
Dr. Helena Berg CEO & Owner, AB Libergreen
Prof. Daniel Brandell Department of Chemistry - Ångström Laboratory Uppsala University
Substitute:
Prof. Jiayin Yuan Department of Materials and Environmental Chemistry Stockholm University
Cover: Investigated rechargeable aqueous battery chemistries for electrical power systems with renewable energy sources installed such as solar and wind power.
List of publications
This thesis is based on the following publications:
Paper I:
Effect of Multiple Cation Electrolyte Mixtures on Rechargeable Zn-MnO2 Alkaline Battery B. Hertzberg, A. Huang, A. Hsieh, M. Chamoun, G. Davies, K. J. Seo, Z. Zhong, M. Croft, C. Erdonmez, S. Meng, D. Steingart. Chemistry of Materials, 2016, 28 (13), 4536-4545 My contribution: Synthesized the MBDB material, contributed to the collection, and processing of the operando EDXRD data, conducted parts of the electrochemical characterization, and wrote parts of the manuscript.
Paper II: Stannate Increases Hydrogen Evolution Overpotential on Rechargeable Alkaline Iron Electrodes M. Chamoun, B. Skårman, H. Vidarsson, R. I. Smith, S. Hull, M. Lelis, D. Milcius, D. Noréus. Journal of The Electrochemical Society, 2017, 164 (6), 1251-1257 My contribution: Conducted the electrochemical and structural characterization (SEM, XRD and EDS), contributed to the collection and processing of the operando neutron diffraction data, and wrote most of the manuscript except the XPS part.
Paper III: 2+ Rechargeability of Aqueous Sulfate Zn/MnO2 Batteries Enhanced by Accessible Mn ions M. Chamoun, W. R. Brant, CW. Tai, G. Karlsson, D. Noréus. Energy Storage Materials, 2018, 15, 351-360 My contribution: Conducted the electrochemical and structural characterization (SEM, XRD, EDS and ICP-AES), contributed to data collection and the processing of the operando XRD data, performed the operando pH measurements and quantification of hydrogen on zinc electrodes, and wrote most of the manuscript except the TEM/EELS parts.
Paper IV: Electrochemical Performance and in Operando Charge Efficiency Measurements of Cu/Sn-Doped Nano Iron Electrodes A. R. Paulraj, Y. Kiros, M. Chamoun, H. Svengren, D. Noréus, M. Göthelid, B. Skårman, H. Vidarsson, M. B. Johansson. Batteries, 2019, 5, 1-15 My contribution: Contributed to data collection and processing of quantifying hydrogen on iron electrodes and wrote parts of the manuscript.
i
Paper V: Bifunctional Performance of Flow Assisted Rechargeable Iron-Air Alkaline Batteries M. Chamoun, A. R. Paulraj, B. Skårman, H. Vidarsson, Y. Kiros, D. Noréus. In manuscript My contribution: Synthesized the oxygen electrocatalysts (except LCMO), conducted the electrochemical and structural characterization (SEM and XRD), developed the Fe–air cell setup, and wrote most of the manuscript.
Publications not included in this thesis:
Paper VI: Water Splitting Catalysis Studied by using Real-Time Faradaic Efficiency Obtained through Coupled Electrolysis and Mass Spectrometry Svengren H., Chamoun M., Grins J., Johnsson M. ChemElectroChem, 2017, 5 (1), 44-50
Reprints were made with permission from the publishers.
ii
Contents
List of publications ...... i Abbreviations ...... v 1. Introduction ...... 1 1.1. Large-scale energy storage systems ...... 1 1.2. Electrochemical energy storage ...... 2 1.3. Rechargeable aqueous batteries based on available resources ...... 2 1.4. Investigated electrode materials for rechargeable aqueous batteries ...... 3 1.4.1. Zinc ...... 3 1.4.2. Electrochemical challenges of zinc ...... 4 1.4.3. Manganese dioxide ...... 5 1.4.4. Electrochemical challenges of manganese dioxide ...... 7 1.4.5. Iron ...... 8 1.4.6. Electrochemical challenges of iron ...... 9 1.4.7. Oxygen electrocatalysts ...... 9 1.4.8. Electrochemical challenges of oxygen electrocatalysts ...... 10 1.5. The aim of the thesis ...... 11 2. Experimental ...... 13 2.1. Synthesis ...... 13
2.1.1. Bismuth-doped β-MnO2 ...... 13 2.1.2. Oxygen electrocatalysts ...... 13 2.2. Cell preparation and electrochemical characterization ...... 13
2.2.1. Alkaline Zn–MnO2 ...... 13
2.2.2. Aqueous sulfate Zn–MnO2 ...... 14 2.2.3. Iron electrode ...... 14 2.2.4. Air electrode...... 15 2.2.5. Fe–air prototype ...... 15 2.2.6. Coulombic efficiency ...... 16 2.3. Operando techniques ...... 17 2.3.1. Energy-dispersive X-ray diffraction (EDXRD) ...... 17 2.3.2. Neutron diffraction ...... 17 2.3.3. X-ray diffraction (XRD) ...... 17
iii
2.4. Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) ...... 18 2.5. XRD ...... 18 2.6. Electron energy loss spectroscopy (EELS) ...... 19 2.7. X-ray photoelectron spectroscopy (XPS) ...... 19 3. Results and discussion ...... 20
3.1. Cation electrolyte mixtures for rechargeable Zn–MnO2 alkaline batteries (Paper I) ...... 20
3.1.1. Structural characterization of bismuth-doped β-MnO2 (MBDB) ...... 20 3.1.2. Electrochemical performance of KOH:LiOH electrolytes ...... 21 3.1.3. Phase evolution investigation of MBDB electrodes ...... 21
2+ 3.2. Reversible aqueous sulfate Zn–MnO2 batteries with Mn (Paper III) ...... 23
3.2.1. Electrochemical characterization of MnO2 electrodes ...... 23
3.2.2. Characterization of cycled MnO2 electrodes ...... 24
3.2.3. Progression of the MnO2 charge and discharge mechanism ...... 27 3.3. Effect of stannate on rechargeable iron electrodes (Paper II) ...... 30 3.3.1. Electrochemical and structural characterization of iron electrodes ...... 30 3.3.2. Phase evolution characterization and structure refinement ...... 33 3.4. Hydrogen evolution on iron and zinc electrodes (Papers III and IV) ...... 35 3.5. Flow-assisted rechargeable Fe–air batteries (Paper V) ...... 38 3.5.1. Structural characterization of oxygen electrocatalysts ...... 38 3.5.2. Electrochemical performance of air electrodes ...... 40 3.5.3. Electrochemical performance of the Fe–air prototype ...... 42 4. Conclusions ...... 46 5. Future perspectives ...... 48 6. Sammanfattning ...... 49 Acknowledgements ...... 51 References ...... 53
iv
Abbreviations
CAES Compressed-air energy storage CV Cyclic voltammetry DOD Depth-of-discharge EDS Energy-dispersive X-ray spectroscopy EDXRD Energy-dispersive X-ray diffraction EELS Electron energy loss spectroscopy EMD Electrolytic manganese dioxide EPDM Ethylene propylene diene monomer ESS Energy-storage systems GDL Gas diffusion layer HER Hydrogen-evolution reaction IHP Intermediate hydride phase
LCMO La1–xCaxMnO3 LDH Layered double hydroxide NMP N-Methyl-2-pyrrolidone
MBDB Bismuth doped β-MnO2 MS Mass spectrometer PDF Powder diffraction file PE Polyethylene PHS Pumped hydroelectric storage PTFE Polytetrafluoroethylene PVC Polyvinyl chloride PVP Polyvinylpyrrolidone OER Oxygen-evolution reaction ORR Oxygen-reduction reaction SEM Scanning electron microscopy SHE Standard hydrogen electrode SMES Superconducting magnetic energy storage SOC State-of-charge SS Stainless steel XPS X-ray photoelectron spectroscopy XRD X-ray diffraction
v
1. Introduction 1.1. Large-scale energy storage systems
We are now shifting our electrical power systems from using fossil fuels to renewable energy sources. The incentive is to replace large traditional power plants with smaller non-dispatchable renewable energy sources. The paradigm shift to renewable energy needs, however, to cope with the increasing global energy demand, estimated to double by 2050, and without increasing carbon dioxide emission levels or reliance on restricted fossil fuel resources1. Shifting the entire electrical power system to renewable energy sources such as wind, solar, and hydro is a complex task. Their inherently intermittent character is dependent on time and weather and results in unpredictable generation profiles, which may not match with the energy-consumption profile. Imbalance between power generation and demand already exists in current power systems. With increasing renewable energy sources installed, the uncertainty in predicting power generation adequacy will increase2. Thus, technologies that improve the resiliency of the power system are necessary. To make the best use of a power system with renewable energy installed, we need efficient energy-storage systems (ESS) to handle load fluctuations and ensure reliable power delivery whenever needed. In spite of the advantages with energy storage, only 1% of the global energy presently used had been stored, mostly through pumped hydroelectric storage, which accounts for 98% of the total installed storage systems1. ESS that are available on a large scale can be divided into four groups: 1) mechanical, 2) electrical, 3) electrochemical and 4) chemical. These consist of technologies such as 1) pumped hydroelectric storage (PHS), compressed-air energy storage (CAES) and flywheels, 2) superconducting magnetic energy storage (SMES) and electrical supercapacitors, 3) batteries and electrochemical supercapacitors, and 4) power-to-hydrogen or synthetic natural-gas production3,4. In Table 1.1, the characteristics of these ESS are compared with regard to power output and discharge time. Depending on the time scale of service, these technologies can support electrical power systems by facilitating frequency regulation and load balancing, enhancing power quality, and providing an uninterruptible power supply. These assets will improve power systems quality, stability and reliability2,5.
Table 1.1. Characteristics of different energy storage technologies6. Technology Power output (MW) Discharge time Efficiency (%) Start time
PHS < 5000 1 – 24 h 65 – 85 s – min CAES Depending on storage size 1 – 24 h 42 – 70 min Flywheels 0.002 – 20 s – min 95 s – min SMES 0.001 – 10 s 90 ms Supercapacitors 0.01 – 1 ms – s 95 ms Lead-acid batteries 0.001 – 50 s – 3 h 60 – 95 < ms Lithium-ion batteries 0.001 – 2 min – h 85 – 99 < ms Vanadium redox flow batteries 0.03 – 20 s – 10 h 85 ms Sodium sulfur batteries 0.5 – 50 s – h 85 – 90 < ms Power to H2 gas kW – GW s – months 62 – 82 s – min Power to CH4 gas kW – GW s – months 49 – 56 min – h
1
1.2. Electrochemical energy storage
Electrochemical energy-storage technologies include batteries, redox flow batteries, electrochemical supercapacitors, and fuel cells. The technologies are distinguished by their energy storage mechanisms. Batteries store energy through electron transfer reactions, wherein the oxidation states of reactants change. Redox flow batteries follow the same mechanism except that the redox species circulate. By comparison, supercapacitors undergo capacitive charging from the electric double layer at the interface between the electrode and the electrolyte. The technologies all function as closed systems except for fuel cells, which store energy from external reactants fed to the cells7. Batteries are of high interest and contain electrochemical cells with two electrodes where the redox reactions ensue. An electrolyte separates the electrodes and contains dissolved ions that can be transferred freely from one electrode to the other. The electrodes are connected externally and, as the redox reactions proceed, electrons transferred via the outer circuit create a current. These cells can then be connected in series and/or in parallel to provide the desired voltage and/or capacity, respectively8. Important characteristics of batteries are their energy density (Wh L–1), specific energy (Wh kg–1) and specific power (W kg–1), i.e. how much energy they can store per unit mass or volume, and how quickly they can deliver this energy. There is a trade-off in batteries between short-term (power density) and long-term storage (energy density), as described by Ragone9, and these can be customized to the application. Table 1.2 lists energy and power characteristics of mature battery technologies for large-scale applications. In large-scale energy storage the primary factors are not energy and power density, but rather low installation cost, long cycle life, high energy efficiency and the ease of scaling up the storage capacity10.
Table 1.2. Energy and power characteristics of mature battery technologies for large-scale applications5. Battery Energy density Specific energy Specific power Cycle life (Wh L–1) (Wh kg–1) (W kg–1)
Lead-acid 60 – 75 30 – 40 60 – 110 100 – 500 Nickel-Cadmium 130 – 150 40 – 60 40 – 100 2000 Nickel-Metal hydride 250 – 330 70 – 100 70 – 200 1000 Lithium-ion - Li(Ni,Co,Mn)O2 – C 200 – 250 120 – 160 200 – 300 300 –1000 Lithium-ion - LiFePO4 – C 120 – 150 80 – 90 200 – 300 1500 – 2000 Sodium-Sulfur 70 – 150 60 – 120 15 – 70 4000 Vanadium Redox Flow 10 – 20 10 – 20 1 – 4 5000
1.3. Rechargeable aqueous batteries based on available resources
The most important aspect of manufacturing batteries for large-scale energy storage is the price set by the market. The availability of the materials and the processes used to manufacture the devices drive the cost. These two factors must coincide with a sustainable life cycle for large-volume markets. Thus, scrutinizing feasible materials to develop sustainable and efficient batteries is critical and not an easy task. Forecasts of the availability of materials are inaccurate and vary depending on the state of the art in industrial sectors1. Among the various types of batteries available today, non- aqueous lithium-ion batteries are the most prominent choice because of their high energy density and versatile design capabilities that allow them to meet energy and power demands11. However,
2 cost12, safety13 and lifetime will limit their full-scale implementation in electrical power systems, for which low-cost and long service life are the main concerns. For instance, the use of cobalt in the layered LiCoO2 electrode material has been the benchmark in lithium-ion batteries despite that the availability of cobalt is low14. The European Commission identified cobalt as a critical raw material in 2017 because of its significant supply risk15. The supply risk originates from geopolitical issues with the Democratic Republic of Congo, which is the dominant global producer of cobalt (64%). The unsustainable cobalt supply has encouraged research into other battery chemistries16–20. Batteries using aqueous electrolytes are inherently safer and less expensive than their non-aqueous counterparts. Aqueous electrolytes have significantly higher ionic conductivities (up to 1 S cm–1), than non-aqueous ones (typically around 1 – 10 mS cm–1 21). This favors aqueous electrolytes for high-rate operations e.g. when sudden energy deliveries are needed, in particular in quick-response balancing systems in the electrical grid2. At present, lead-acid batteries dominate the aqueous battery market because of their high rate capability and low system-installation price. Lead-acid batteries find its use as start battery or backup battery with low demand for cycle life10. Proposed potential electrode materials for large-scale energy storage are zinc, manganese dioxide, iron, and non-precious-metal-based catalysts for the air electrode. These electrode materials are 22,23 24 adopted in rechargeable aqueous Zn–MnO2 and Fe–air batteries, both having high energy relevancy and unrestricted availability25. Their challenges concerning irreversibility and inefficiency are in this thesis investigated and alleviated in order to make viable systems. The studied electrode materials are not exclusively applicable for rechargeable aqueous batteries and the work is intended to shed light on the importance of using available materials in the development of future electrical power system. The batteries designed in this thesis are relevant for large scale energy storage with assured safe operation and low total cost.
1.4. Investigated electrode materials for rechargeable aqueous batteries 1.4.1. Zinc Zinc has a rich history in alkaline26–28, mildly acidic29 and redox-flow rechargeable batteries30–32. It possesses attractive attributes for an electrode material, such as abundancy, low toxicity and a high specific theoretical capacity of 820 mAh g–1. Furthermore, it is the most electropositive metal that does not have noteworthy corrosion issues in aqueous electrolytes between pH 4 and 1433, making it an outstanding electrode material34. The alkaline battery has been the working horse in the primary battery market for over 60 years. This battery contains zinc and manganese dioxide and delivers a specific energy density of 150 Wh kg–1, comparable to some lithium-ion chemistries35. In recent years, zinc has been coupled with several electrolyte and electrode combinations for high- performance rechargeable batteries. Electrodes used in alkaline electrolytes include Ni–Zn36,37, Zn– 38–40 23,41,42 air and Zn–MnO2 . For mildly acidic electrolytes, electrode materials with open crystal 20,43–46 47,48 structures that are capable of hosting zinc ions are used, and in redox-flow cells, Zn–Br2 , 49 50,51 Zn–I2 and Zn–Fe have been used as electrodes. In static cells, zinc is found as composites, pastes, or powders, whereas flow cells use dissolved zinc ions sourced from various salts. 2– In alkaline electrolytes operating above pH 14, zinc exists in equilibrium with zincate ions, Zn(OH)4 , and zinc oxide, ZnO, precipitates when zincate exceeds its supersaturated concentration limit34:
3
2− − − 푍푛(푂퐻)4 + 2푒 ⇌ 푍푛 + 4푂퐻 E° = –1.12 V vs. Standard hydrogen electrode (SHE) (1.1)
2− − 푍푛(푂퐻)4 ⇌ 푍푛푂 + 퐻2푂 + 2푂퐻 (1.2) where E° is the standard potential relative to the SHE at 25 °C. In mildly acidic electrolytes, at pH 4–6, zinc dissolves to Zn2+ during discharge and is electrodeposited as zinc metal during charge22:
푍푛2+ + 2푒− ⇌ 푍푛 E° = -0.76 V vs. SHE (1.3)
In Figure 1.1, the Pourbiax diagram shows that the redox potential of zinc is below that at which the hydrogen evolution reaction (HER) occurs. Pourbaix diagrams depict possible stable phases of an electrochemical system at equilibrium and do not consider kinetic effects. Based on thermodynamics, the HER should dominate at zinc redox potentials, but luckily that reaction is sluggish, which enables zinc to be used as electrode material.
mildly acidic
alkaline
Figure 1.1. Pourbaix diagram of 10–5 M Zn2+(aq) at 25 °C, created by the software Medusa®. Highlighted arrows in red and blue depict the mildly acidic and alkaline pH regions, respectively, and the dashed green lines correspond to the oxygen evolution and hydrogen evolution reactions.
1.4.2. Electrochemical challenges of zinc Zinc faces several challenges when adopted in aqueous rechargeable alkaline batteries. Shape changes and morphological uncertainties affect the deposition of the metal during charge. Zinc tends to plate anisotropically and this induces localized mass-transport-limited regions52. The anisotropic growth ramifies, and dendrite formation increases as the mass-transport limitation increase. These dendrites eventually cause short-circuiting if they penetrate the separator53. Figure 1.2 illustrates the uneven deposition of zinc and the formation of dendrites that may short-circuit the Zn–MnO2 cell.
4
- e
Electrolyte
2
Zn
MnO
Short-circuit
Figure 1.2. Illustration of uneven zinc deposition and the formation of dendrites that short-circuit the cell.
Engineering zinc structures can extend the cycle life but shape changes upon cycling are not easily avoided. Reported fine three-dimensional structures of zinc have provided longer cycle life while sacrificing energy density54–56. Nevertheless, these three-dimensional structures have limited utilization due to rapid dissolution of the metal57. Low zinc utilization during dissolution is mainly associated with corrosion and passivation effects33. Hence, utilization is limited to 60% or less58. Limiting passivation and corrosion effects is a major challenge related to the particle size of zinc and the solubility of zincate. The smaller the zinc particles are (or the higher their surface area is), the more aggravated is the corrosion. The corrosion reaction is caused by the HER and is parasitic because it consumes water without contributing any useful charge capacity to the battery:
− − 2퐻2푂 + 2푒 → 퐻2 + 2푂퐻 E°= –0.83 V vs. SHE (1.4)
Using zinc particles with low surface area limits corrosion at the expense of passivation35. Passivation is dependent on the solubility of zincate and is caused by precipitation of ZnO when the solution becomes saturated. Initially, porous ZnO forms, but this densifies overtime and eventually passivates the zinc59. Mildly acidic electrolytes are more forgiving than alkaline electrolytes. Passivation from ZnO is prevented because the pH is lower. Instead, zinc dissolves to Zn2+ during discharge and electroplates back during charge (1.3). The reversibility has been extensively studied and is good60–64. Concerns of the HER lowering the Coulombic efficiency and potential dendrite formation remain. Other work has focused on substituting sulfate with other anions, adding surfactants or adjusting the concentration of the salts used43. The pH of the electrolyte must be maintained at 4–6 during battery operation to avoid severe corrosion and passivation33.
1.4.3. Manganese dioxide Manganese dioxide is used in primary alkaline batteries for a wide range of power electronics65, as well as in secondary batteries such as lithium-66, sodium-67, magnesium-68, and zinc-ion batteries43.
5
Inexpensive production, high theoretical energy capacity, high redox potential, and low environmental impact make manganese dioxide an excellent electrode material69. Manganese dioxide comes in a variety of polymorphs depending on the synthesis conditions: tetragonal pyrolusite β-MnO2, orthorhombic ramsdellite, tetragonal hollandite ɑ-MnO2, hexagonal birnessite δ-MnO2, monoclinic romanèchite and cubic λ-MnO2. These structures can be described by 4+ different distributions of Mn cations over octahedral sites in the oxygen atom arrangement. MnO6 octahedra sharing opposite octahedral edges form MnO6 chains parallel to the c-axis. The chains further connect to each other in different ways and tunnels form along the c-axis. The tunnels can be 70 classified by the number of MnO6 units and chains between two basal planes . The stable form of manganese dioxide β-MnO2 is built by MnO6 units forming a 1 1 tunnel structure. Among the polymorphs, β-MnO2 is the least electrochemically active. Electrolytic manganese dioxide (EMD), also known as γ-MnO2, is composed of intergrown β-MnO2 and ramsdellite. In modern alkaline, lithium or other types of batteries, EMD is used because of its high electrochemical activity, high manganese content, and purity. EMD is produced via electrochemical deposition from acidic sulfate baths containing Mn2+ ions and undergoes a two-electron oxidation65:
2+ + − 푀푛 + 2퐻2푂 → 푀푛푂2 + 4퐻 + 2푒 (1.5)
The redox reactions of manganese dioxide vary with pH. Two pH regions are of interest: one above pH 14 (alkaline electrolytes) and the other at pH 4–6 (mildly acidic electrolytes). Figure 1.3 shows the Pourbaix diagram of these two pH regions and corresponding redox reactions.
mildly acidic
alkaline
Figure 1.3. Pourbaix diagram of 10–5 M Mn2+(aq) at 25 °C, created by the software Medusa®. Highlighted arrows in red and blue depict mildly acidic and alkaline pH regions, respectively, and the dashed green lines correspond to the oxygen evolution and hydrogen evolution reactions. The discharge reaction of EMD in alkaline electrolytes includes a homogeneous one-electron reduction via proton insertion to form MnOOH:
− − 푀푛푂2 + 퐻2푂 + 푒 → 푀푛푂푂퐻 + 푂퐻 E° = +0.36 V vs. SHE (1.6)
6
Further one-electron reduction proceeds through heterogeneous dissolution to a give soluble 71 hydroxymanganese complex, which then precipitates as Mn(OH)2 :
− − 푀푛푂푂퐻 + 퐻2푂 + 푒 → 푀푛(푂퐻)2 + 푂퐻 E° = –0.28 V vs. SHE (1.7)
In mildly acidic electrolytes, EMD initially dissolves to trivalent manganese (푀푛4+ + 푒− → 푀푛3+). This Jahn–Teller Mn3+ cation is unstable because of its high-spin electronic configuration and disproportionates to Mn4+ and Mn2+ ions (2푀푛3+ → 푀푛4+ + 푀푛2+)61. The total discharge reaction mechanism may be simplified to72:
+ − 2+ 푀푛푂2 + 4퐻 + 2푒 → 푀푛 + 2퐻2푂 E° = +1.23 V vs. SHE (1.8)
The two-electron transfer in both electrolytes corresponds to a theoretical capacity of 617 mAh g–1.
1.4.4. Electrochemical challenges of manganese dioxide Manganese dioxide in alkaline electrolytes can evolve several failure mechanisms upon battery cycling. These mechanisms are strongly correlated to the irreversible formation of inactive phases over time. Figure 1.4 shows the reactions during charge-discharge cycling of EMD. The figure highlights the significant phase transformations leading to irreversibility.
Figure 1.4. Reaction mechanism of EMD (γ-MnO2) upon charge-discharge cycling in alkaline electrolyte. Figure reproduced from Paper I69 with permission from American Chemical Society.
The first discharge involves a phase transformation of γ-MnO2 into ɑ-MnOOH (2 1 tunnel structure) and γ-MnOOH (1 1 tunnel structure) via proton insertion, with a change of the manganese valence state from 4+ to 3+73. The reduction proceeds through the partial formation of the spinel phase
Mn3O4, or ZnMn2O4 if zinc is present. The reduction continues from Mn3O4 to the final discharge 73,74 product Mn(OH)2, with the manganese valence state of 2+ . The partial formation of
7
Mn3O4/ZnMn2O4 does not involve the exchange of electrons but is based on whichever complex ion is available to fill the tetrahedral lattice73:
2− 2− − 2푀푛푂푂퐻 + 푀푛(푂퐻)4 /푍푛(푂퐻)4 → 푀푛3푂4/푍푛푀푛2푂4 + 2퐻2푂 + 2푂퐻 (1.9)
The following charge step includes the oxidation of Mn(OH)2 to β-MnOOH (layered structure), γ-
MnOOH and γ-Mn2O3 (spinel structure) during the first electron transfer. Charging continues with the formation of δ-MnO2 upon the second electron transfer. The layered δ-MnO2, with large interlayer spacing, hosts cations and structural water to stabilize its crystal structure75. The electrochemically inactive Mn3O4/ZnMn2O4 can only be partially reduced to Mn(OH)2 and the rest remains inert in the electrode, resulting in a significant loss of capacity. This does progress over multiple cycles, generating more of the inactive spinel phase and eventually leading to failure. The presence of zinc accelerates this process, as ZnMn2O4 is less electrochemically active than Mn3O4. A 74,76,77 similar reaction mechanism has been reported elsewhere . The formation of Mn3O4/ZnMn2O4 limits the rechargeability of alkaline manganese dioxide based batteries78. In mildly acidic electrolytes, a different reaction mechanism takes place. The electrochemically inactive spinel phase is suppressed because the pH is buffered below 6 by basic salts precipitating. These electrolytes containing zinc sulfate have been reported to deliver long cycle life and excellent battery performance60,64,79. The proposed discharge reaction mechanism of EMD in zinc sulfate electrolytes can be described as the co-insertion of protons and zinc ions72,79,80. Upon the first stage of discharge, protons are inserted into the EMD structure, leading to an increased local pH at the electrode surface. With continuous pH increase, the second discharge regime proceeds where zinc is being inserted, precipitating zinc hydroxide sulfate pentahydrate, Zn4SO4(OH)6·5H2O:
2+ 2− − 4푍푛 + 푆푂4 + 6푂퐻 + 5퐻2푂 → 푍푛4푆푂4(푂퐻)6 ⋅ 5퐻2푂 (1.10)
This does not involve any electron exchange and occurs at pH 561. The formation of the precipitate buffers the pH and thus prevents the formation of Mn3O4/ZnMn2O4, which would be formed at higher pH. However, it is electrochemically inactive and forms large crystalline flakes on active particles. These flakes block active sites, plug pores, and impede mass transport. Thus, preventing the precipitate from insulating the surface is critical for maintaining reversibility. Another challenge is manganese dissolution, which generates manganese vacancies where zinc ions can be inserted. Excessive zinc-ion insertion into EMD leads to structural collapse and to cell failure72.
1.4.5. Iron Iron is the second most abundant metal on Earth25 and is the electrode material for large-scale aqueous rechargeable batteries. Iron is cheap and energy dense—both enviable properties for batteries. Thomas Edison first developed the iron electrode in the early 20th century for the Ni–Fe battery27. Later, significant interest in developing Fe–air batteries for fossil-free traction arose in the late 1960s at NASA81. Fe–air batteries gained serious interest because of their remarkably high theoretical energy densities, up to 9700 Wh L–1, and specific energies of more than 1200 Wh kg–1 24. Later pioneering work by the Swedish National Development Company in the 1970s demonstrated the feasibility of this battery, with a lifetime over 1000 cycles and an energy density of 80 Wh kg–1 82,83. Renewed interest in the iron electrode has led to recent advances in nanostructured materials and has improved their performance further, resulting in higher energy densities84,85.
8
Iron electrodes use aqueous alkaline electrolytes because of their high ionic conductivity and the 24 availability of reversible redox reactions . The main discharge product, iron (II) hydroxide Fe(OH)2, is insoluble. This favors the iron electrode because a solid-state reaction ensues and prevents the diffusion of dissolved species86. Conventional iron-based batteries limit deep discharging and operate upon the first two-electron reduction of iron to iron (II) hydroxide. This reaction offers a theoretical specific capacity of 960 mAh g–1. The deep discharge regime involves further reduction to – the less electrochemically active magnetite, Fe3O4, with a theoretical specific capacity of 199 mAh g 1. When recharged, the iron (II) hydroxide or magnetite transforms back to metallic iron. The charge and discharge reactions can be described as87,88:
− − 퐹푒(푂퐻)2 + 2푒 ⇌ 퐹푒 + 2푂퐻 E° = –0.88 V vs. SHE (1.11)
− − 퐹푒3푂4 + 4퐻2푂 + 2푒 ⇌ 3퐹푒(푂퐻)2 + 2푂퐻 E° = –0.76 V vs. SHE (1.12)
1.4.6. Electrochemical challenges of iron Even though iron electrodes are robust and have been used for over a century27,82, long-term inefficiency and unwanted side reactions have limited their large-scale use. The primary limitation is, as for the zinc electrode, the HER (1.4)89. Unfortunately, iron is a good hydrogen evolution catalyst leading to inadequate charging efficiencies in the range of 55–70%27. As such, significant amount of water is lost, and superfluous hydrogen gas is evolved. Sulfide compounds are added to suppress the HER and uphold efficient operation. Adsorbed sulfide on iron poisons the HER and is incorporated 89–91 into the electrode as FeS with Bi2O3, or Bi2S3 by itself, or in the electrolyte as Na2S or K2S . Moreover, it is critical not to operate at deep discharge regimes and encourage formation of the less active phases Fe3O4 and Fe2O3. Utter attention is required to limit the depth-of-discharge (DOD) and avoid passivation.
1.4.7. Oxygen electrocatalysts Rechargeable aqueous metal–air batteries such as Zn–air92 and Fe–air24 attract research interest because of their exceptionally high energy density. The air electrode uses bifunctional catalysts, i.e. substances that can catalyze both the oxygen-evolution reaction (OER) and the oxygen-reduction reaction (ORR). Metal–air batteries need an open cell design to deliver oxygen to the catalytic sites. Oxygen is fed from an external source of air from the outer atmosphere, hence the name “air electrode”. The air is not stored in the cell, which therefore exhibits notably high theoretical specific energy density93. Air electrodes require a three-phase boundary between the solid electrode in contact with the ion conducting electrolyte and gas phase. To satisfy the three-phase interface, air electrodes are designed with an open structure to facilitate gas diffusion while combining hydrophilic and hydrophobic properties in separate layers. The hydrophilic layer ensures proper wetting and contact with the aqueous electrolyte. The hydrophobic counterpart, commonly with a wet-proofing agent such as polytetrafluoroethylene (PTFE), prevents electrolyte penetration and facilitates oxygen diffusion to the catalytic sites92. Air electrodes can be adopted in different electrolytes, but alkaline ones are favored because of rapid kinetics94. Alkaline electrolytes render possible the use of non- precious-metal-based catalysts, most of which dissolve in acid. In these electrolytes, oxygen is reduced to solvated hydroxide ions during discharge, and is regenerated upon charge. The ORR
9 mechanism is complex because it involves a multistep electron transfer and follows either a four electron- or a two electron pathway95,96. The reaction pathway varies with the catalytic material and electronic structure97. For the direct four-electron pathway on metals, the reaction follows:
− − 푂2 + 2퐻2푂 + 2푒 → 2푂퐻푎푑푠 + 2푂퐻 (1.13)
− − 2푂퐻푎푑푠 + 2푒 → 2푂퐻 (1.14)
Giving the overall reaction:
− − 푂2 + 2퐻2푂 + 4푒 → 4푂퐻 E° = +0.40 V vs. SHE (1.15)
The alternate two-electron pathway proceeds through intermediate peroxide formation:
− − − 푂2 + 퐻2푂 + 2푒 → 퐻푂2 + 푂퐻 (1.16)
− − − 퐻푂2 + 퐻2푂 + 2푒 → 3푂퐻 (1.17)
Metal oxides follow the same pathways but with a different surface charge distribution. The metal oxide cations on the surface are not completely coordinated with oxygen atoms but instead with the oxygen of a water molecule98. The four-electron pathway is favored on precious metals, silver and particular metal-oxide structures such as spinels and perovskites95. The two-electron peroxide pathway dominates on carbon-based catalysts, gold and other metal-oxide structures98. The OER mechanism is also complex. To simplify, oxygen is generally evolved from the oxide phase instead of the metal and is followed by a release of two coordinated oxygen atoms to a metal ion on the catalyst surface99. Transition metal oxides based on Ni, Co and Mn spinels100–103 and perovskites97,104,105 have proven to be active bifunctional catalysts with good corrosion resistance in alkaline electrolytes.
1.4.8. Electrochemical challenges of oxygen electrocatalysts Bifunctional oxygen electrocatalysts are the main bottleneck of batteries that use air electrodes because of slow kinetics and corrosion issues93. The sluggish kinetics are due to the ORR step. In 2– – alkaline electrolyte, the competing displacement of O2 /OH as well as hydroxide-ion conversion are reported as the rate-limiting steps in the ORR97. The redox reaction upon operation yields large polarization losses from high overpotentials. Overpotential is the magnitude of deviation from the equilibrium potential and is constituted of activation, concentration, and resistance losses106:
휂푡표푡푎푙 = 퐸푐푒푙푙 − 퐸푒푞 = 휂푎푐푡 + 휂푐표푛푐 + 푖푅 (1.18) where 퐸푐푒푙푙 is the measured cell potential, 퐸푒푞 is the equilibrium potential, 휂푎푐푡 is the activation overpotential and defined as the required activation energy to proceed with the redox reaction,
휂푐표푛푐 is the concentration overpotential and describes mass transport limitation by depletion of charge carriers in the electrolyte at the electrode surface, and 푖푅 is the ohmic drop losses caused by resistance in the hardware and electrolyte. Figure 1.5 shows the total OER and ORR overpotential of an air electrode in alkaline electrolyte using
NiCo2O4 as bifunctional catalyst. Upon a full charge-discharge cycle, the total overpotential achieved was 693 mV at a moderate current density rate of ±10 mA cm–2 and under a flow of air. This overpotential is significant and prompts low voltaic and energy efficiencies. For instance, the energy
10 efficiency of an alkaline Fe–air battery is 50% and the remaining part is unmitigated losses24. Active oxygen electrocatalysts based on Pt, Pd, Ru and Ir encounter high total overpotentials as well; these are typically above 500 mV92,93,107–109.
ƞOER-ORR = 693 mV
Figure 1.5. A full charge-discharge cycle at ±10 mA cm–2 of an air electrode using the bifunctional catalyst –3 NiCo2O4 in 6 mol dm KOH. The figure highlights the total overpotential of OER and ORR. Air was used as the oxygen feed to the air electrode. Figure reproduced from Paper V.
Polarization losses are a major challenge with oxygen electrocatalysts. The catalyst must sustain oxidizing environments under high overpotentials. Furthermore, competition between the four- and two-electron pathways of the ORR deteriorates the electrode, if the reaction mechanism favors the latter. The two-electron pathway generates corrosive peroxide species, harming the electrode upon battery operation110. Another concern with air electrodes in alkaline electrolyte is carbonate formation when carbon dioxide reacts with hydroxide ions92:
− 2− 퐶푂2 + 2푂퐻 → 퐶푂3 + 퐻2푂 (1.19)
The poorly soluble carbonates clog electrode pores and block the electrolyte channels, retarding the electrochemical activity. To circumvent carbonate formation, it is important to purify the airflow or use pure oxygen. Another option is to circulate the electrolyte in order to prevent the carbonate from reaching supersaturation93.
1.5. The aim of the thesis
This thesis describes the investigation and development of rechargeable aqueous Zn–MnO2 and Fe– air batteries to overcome hurdles in their performance. Mixed cation electrolytes containing KOH and LiOH enhanced the cycle life and proved its potential as a drop-in electrolyte replacement for traditional alkaline Zn–MnO2 batteries. The structural evolution and failure mechanisms were investigated using electron microscopy and operando energy-dispersive X-ray diffraction (EDXRD) techniques. In the analogous battery with mildly acidic electrolytes, the complex reaction mechanism was explained to answer why perpetual access of Mn2+ ions at the electrode/electrolyte interface enhanced the rechargeability. Differentiation of the phase evolution of cycled MnO2 electrodes used both ex situ and operando X-ray radiation analytical techniques.
11
The examined hurdles for rechargeable Fe–air batteries were countered by suppressing the evolution of hydrogen gas on the iron electrode and optimizing bifunctional catalysts for the air electrode, in order to improve performance and cycle life. The evolution of hydrogen gas was minimized using potassium stannate as an additive. The rationale behind the additive was that tin metal would deposit as the iron electrode charged and serve as a barrier to hydrogen gas. Operando neutron diffraction measurements described the phase evolution of the iron electrode with stannate during charge-discharge cycling. Hydrogen evolution on iron, evaluated as a function of charge current density, was studied by coupled electrochemistry and mass spectrometry. New bifunctional catalysts for the air electrode were characterized structurally and electrochemically. The findings presented an excellent catalyst candidate with superb activity for oxygen evolution and reduction with excellent long-term stability. The catalyst was used in a rechargeable alkaline Fe–air battery as a demonstration of this technology, and its effect on performance and cycle life is presented later.
This thesis aims to investigate the aforementioned battery systems and develop them as viable technologies for large-scale energy storage. The motivation of this study emphasized rechargeable batteries based on available resources to actualize the transition from fossil fuels to renewable energy sources.
This thesis presents the background and motivation of the work, explains how experiments were carried out and how analysis methods were applied, discusses the relevant results from Papers I–V, and finally summarizes these results.
12
2. Experimental 2.1. Synthesis
2.1.1. Bismuth-doped β-MnO2
The bismuth-doped β-MnO2 (MBDB) material was synthesized via the thermolysis of manganese and bismuth nitrates. Two solutions were prepared separately before heat treatment: 1) 50 g of
Mn(NO3)2·4H2O (Sigma-Aldrich, >97%) in 80 mL deionized water and 2) 4.27 g of Bi(NO3)3·5H2O
(Sigma–Aldrich, >98%) in 18.6 mL deionized water and 6.4 mL nitric acid (HNO3, Sigma–Aldrich, 70%). The solutions were mixed together and heated to 125 °C. Observable color changes of the solution upon heating indicated the oxidation of manganese from 2+ to 4+, and the final solution was black. The black solution was heated under vacuum at 125 °C for 12 h and the solid residue then baked at 325 °C for 5 h in air. Lastly, the solid material was ground into a fine powder.
2.1.2. Oxygen electrocatalysts
Three materials were synthesized as of oxygen electrocatalysts: 1) La1–xCaxMnO3 (LCMO), 2) Ni–Fe layered double hydroxide (LDH) and NiCo2O4. For the LCMO catalyst, a solution of La, Ca and Mn nitrates (VWR Chemicals, >99%), in the molar ratio 0.1:0.9:1, was first prepared. The solution was then added dropwise to a heated solution of 0.2 M Na2CO3 (Sigma-Aldrich, >99.5%) at 50 °C, resulting in a precipitate. The precipitate was washed with deionized water, filtered, and dried before being calcined at 700 °C for 6 h in air. After calcination, the solid was quickly quenched in a water-cooled zone of the furnace. Lastly, the solid material was washed with 10% aqueous acetic acid, rinsed with deionized water, dried overnight at 150 °C and then ground. Synthesis of the Ni–Fe LDH catalyst consisted of adding a reducing agent to a mother solution containing nickel and iron nitrates. The mother solution consisted of 8 mL of 0.5 M Ni(NO3)2·6H2O
(Merck, >97%), 2 mL of 0.5 M Fe(NO3)3·9H2O (Pro Analysi, >98%) and 0.50 g of polyvinylpyrrolidone (PVP, Fluka, K30, Mw = 40 000). This solution was mixed to ensure full dissolution of the compounds and then transferred to a beaker with 50 mL deionized water. The reducing agent solution contained
1 g NaBH4 (Sigma-Aldrich, >98%) in 20 mL deionized water, and was added dropwise to the mother solution. Mixing of the total solution continued overnight, resulting in a precipitate. The precipitate was collected, filtered, and washed with both deionized water and ethanol before drying at 60 °C for 12 h.
Preparation of the NiCo2O4 catalyst comprised thermal decomposition of 3.62 g Ni(NO3)2·6H2O
(Merck, >97%) and 7.26 g Co(NO3)2·6H2O (Sigma-Aldrich, >98%) in 200 mL deionized water at 375 °C for 2 h. After heating, the solid material was cooled to room temperature and then ground to a powder.
2.2. Cell preparation and electrochemical characterization
2.2.1. Alkaline Zn–MnO2 After synthesis of the MBDB active material, electrodes were prepared by mixing 45 wt% MBDB, 45 wt% graphite (Timcal, KS6) and 10 wt% PTFE (Sigma-Aldrich, 60 wt% dispersion in H2O). The electrodes were worked into pastes, dried in a vacuum oven at 125 °C for 1 h, and then pressed onto a perforated nickel wire screen at 98 MPa. The pressed MBDB electrodes were assembled into
13 planar cells against zinc foil as counter electrode and separated with one layer of polyvinyl chloride (PVC) and two layers of non-woven cellulose (Freudenberg LLC, FV-4304). Acrylic plates with screws held the cell under compression. After assembly, the cells were immersed in a beaker of 5 mL electrolyte. The cycling protocol used constant current rates of 205, 31 or 147 mA g–1, with cells connected to a battery analyzer instrument (MTI, BST8-3). The voltage range was kept between 0.4 and 1.8 V vs. Hg/HgO (+0.098 V vs. SHE) with a constant voltage step of 1.8 V during charge until the current dropped below 10% of the input value.
2.2.2. Aqueous sulfate Zn–MnO2
2+ Characterization of aqueous sulfate Zn–MnO2 cells used two electrolytes, 2 M ZnSO4 (Mn -free 2+ electrolyte) and 2 M ZnSO4 + 0.1 M MnSO4 (added-Mn electrolyte). Preparation of MnO2 electrodes consisted of mixing 0.80 g EMD (Tronox, Ultrafine), 0.02 g graphite (Timcal, BNB90), 0.09 g carbon black (Imerys, Super C65), and 0.09 g polyvinylidene fluoride (PVDF, Arkema, KynarFlex 2801) with 2 mL N-Methyl-2-pyrrolidone (NMP, VWR) solvent. This mixture was ball-milled (SPEX, 8000 Mill) for 20 min to form a slurry and then cast as a thin film on carbon paper (Freudenberg, H23) to a thickness of 0.18 mm. Afterwards, the electrodes were dried in two steps, first at 60 °C for 3 h and then at 120 °C for 12 h under vacuum. For the counter electrode, a slurry prepared in similar fashion was casted on zinc foil (Alfa Aesar, 0.25 mm thickness) before drying at 60 °C overnight. The zinc slurry composition consisted of 0.80 g zinc powder (Sigma–Aldrich, <10 µm, >98%), 0.10 g activated carbon (Merck, Activated charcoal for analysis), 0.05 g graphite (Timcal, BNB90), 0.05 g PVDF (Arkema, KynarFlex 2801) and 1.2 mL NMP. One layer of glass fiber paper (Whatman Grade GF/F) separated the electrodes and the cell was contained between two acrylic pieces before it was immersed in a container with 5 mL electrolyte. Cyclic voltammetry (CV) measurements were performed on a potentiostat (BioLogic, SP-50) at potentials between 1 and 1.8 V vs. Zn/Zn2+ using a sweep rate of 0.2 mV s–1. Charge-discharge cycling was done by connecting the cell to a current source (Wuhan LAND Electronics, CT2001A) at a constant current rate of 60 mA g–1. Voltage cut-offs were set at 1 and 1.8 V with a constant voltage step at 1.8 V during charge and ended when the current dropped below 20% of the input value.
2.2.3. Iron electrode The iron electrodes used in the stannate-additive study consisted of 80 wt% iron (Höganäs AB,
Nutrafine RS), 5 wt% Bi2S3 (Sigma–Aldrich, 99%), 8 wt% graphite (Imerys, KS6L), 2 wt% carbon black
(Imerys, Super C65) and 5 wt% PTFE (Sigma–Aldrich, 60 wt% dispersion in H2O). After mixing, the formed paste was rolled to a thickness of 0.1 mm, dried at 110 °C for 1 h and then pressed into a nickel wire screen (Dexmet, 100 mesh) at 30 MPa. The cell consisted of a commercial sintered nickel electrode (Gates Energy) as counter electrode and Ag/Ag2O as reference electrode (+0.242 V vs. Hg/HgO or +0.098 V vs. SHE)111. The measured half-cell potentials were converted to be relative to that of Hg/HgO. One layer of non-woven cellulose (Freudenberg, 700/18F) separated the electrodes and the cell was contained between acrylic plates before being submerged in 30 mL 6 M KOH + 1 M
LiOH electrolyte, with or without 0.1 M K2SnO3 included. The electrodes were cycled using a current source (Wuhan LAND Electronics, CT2001A) operated at a constant current rate of 192 mA g–1. Voltage cut offs were set to –0.458 and –1.158 V vs. Hg/HgO. Nanostructured copper- and/or tin-doped iron materials, denominated as CuSn and Sn, were provided by Höganäs AB. Paper IV details the structural and elemental composition analysis of these
14 powders. The two iron-electrode materials were combined with single-walled carbon nanotubes (SWCNT, OCSiAl, TUBALL™ BATT). Another sample investigated the effect of 0.65 M LiOH added to the 6 M KOH electrolyte. In total, four samples were evaluated: 1) CuSn, 2) CuSnCNT, 3) CuSnCNTLi and 4) SnCNT. Preparation of the electrodes involved mixing 80 wt% of the nanostructured iron materials with 5 wt% carbon black (AkzoNobel, Ketjenblack EC-300J), 5 wt% Bi2S3 (Sigma–Aldrich,
99%) and 10 wt% PTFE (Sigma–Aldrich, 60 wt% dispersion in H2O). The mixture was homogenized in laboratory blender (Waring, LB20ES) at 6000 RPM for 15 min in an aliphatic solvent (Shell Chemicals, ShellSol D70). Afterwards, the filtered wet mass was rolled on a nickel wire screen (100 mesh) to a thickness of 0.7 mm, pressed at 375 kg cm–2 and then sintered at 325 °C for 30 min. Paper IV shows the active mass loading of iron in each sample. The nanostructured metal-doped iron material containing copper and with SWCNT was adapted to the Fe–air prototype in Section 2.2.5.
2.2.4. Air electrode Preparation of the air electrodes covered three parts: a catalyst layer, a current collector, and a gas diffusion layer (GDL). Preparation of the catalyst layers were done for three sets of samples: 1) 65 wt% LCMO + 10 wt% Ni–Fe LDH, 2) 65 wt% LCMO + 10 wt% NiCo2O4 and 3) 75 wt% NiCo2O4, with 10 wt% carbon black (Imerys, Super C65) and 15 wt% PTFE (Sigma–Aldrich, 60 wt% dispersion in H2O), constituting the rest for all three samples. The mixtures were homogenized in 20 mL ethanol per 1 g of solids using an ultrasonic probe (Hielscher, UP200St) at 30% amplitude for 10 min. After mixing and filtration, ethanol was added to the collected cakes to form pastes that were rolled to a thickness of 0.4 mm. The rolled pastes were then pressed at 375 kg cm–2 onto a nickel wire screen (Dexmet, 100 mesh) that served as current collector. Lastly, the electrodes were sintered at 340 °C for 25 min before being cold-pressed onto the GDL, a porous PTFE foil (Guarniflon, TPF020), at 375 kg cm–2. Paper V provides the active mass loadings of the catalysts used in the prepared electrodes. Electrochemical characterization of the OER and ORR in 6 M KOH used a specially designed cell connected to a potentiostat (Bio-Logic, SP-50). Nickel mesh served as the counter electrode and
Ag/Ag2O as the reference electrode while the recorded half-cell potentials were converted to the Hg/HgO reference. The cells were submerged in 50 mL of electrolyte and wetted the catalyst layer, while the GDL layer was preserved dry with air flowing at a rate of 20 mL min–1. The active geometric surface area was 4 cm2 and reported current densities were based on this value. Air electrodes were preconditioned by CV over 20 cycles between –0.108 and 0.592 V at 5 mV s–1. Then, the cycle life of these electrodes was assessed by charging and discharging for 2 h in steps at ±10 mA cm–2.
2.2.5. Fe–air prototype Paper V details the requirements that must be considered when fabricating the Fe–air prototype to enable stable cell operation. In short, resistant cell materials must withstand corrosive environments, facilitate both electrolyte and gas flow, and secure the mechanical integrity and sealing of the cell. Figure 2.1 shows the cell breakdown of the Fe–air prototype. Both stainless steel (SS) and PVC end plates tighten the cell body, which has total dimensions of 12 10 3 cm. Inlets and outlets for gas and electrolyte were fitted on the front. Within the cell, one iron and one air electrode were aligned in parallel and separated by two polyethylene mesh spacers (PE, 1.35 mm thickness, 3.4 3.2 mm mesh size). Another PE mesh spacer was placed on the GDL side of the air electrode, facilitating gas transport to the catalytic sites. Ethylene propylene diene monomer rubber gaskets (EPDM, Kuntze, ESO2 425-010, 0.5 mm thick) sealed the cell, and contained holes to ensure
15 gas and electrolyte flows to the electrodes. The cell contained 10 mL electrolyte and had an active geometric surface area of 49.2 cm2. Spot-welded nickel-foil tabs served as external connections.
Figure 2.1. Cell breakdown of the Fe–air prototype components. Figure reproduced from Paper V.
The flow-assisted Fe–air prototype cell was operated using copper-doped iron for the iron electrode and NiCo2O4 as the bifunctional catalyst for the air electrode. The 6 M KOH electrolyte was circulated through a closed system at 1 mL min–1 by a peristaltic pump (LKB Bromma, 2132 MicroPerpex). Oxygen feed to the cell flowed at the rate of 35 mL min–1. Upon operation with electrolyte and gas flows, the cell was connected to a current source (Wuhan LAND Electronics, CT2001B) and the potential was recorded from the full cell. The cell operation protocol included three steps: 1) a formation step to activate both electrodes at ±4 mA cm–2, 2) a rate-capability step at current densities between ±5 and ±25 mA cm–2, and 3) a cycle-life-assessment step at ±10 mA cm–2. For the first two steps, the cell was charged to the theoretical specific capacity of iron, 960 mAh g–1, whereas in the last step, the charge capacity was optimized to maximize efficiency.
2.2.6. Coulombic efficiency Detection of gaseous products with a specially designed electrochemical cell coupled to a mass spectrometer (MS), as described in previous work112, enabled the quantification of hydrogen gas evolved at iron and zinc electrodes. The measurements assumed that the HER was the only cause of deviation from 100% Coulombic efficiency. Coulombic efficiency is defined as the ratio of total discharge capacity output from the cell to the total charge capacity input into the cell over a full cycle. The cell consisted of two separated chambers filled with electrolyte and with volumes of 48 cm3 each. Both chambers were purged continuously with argon to exhaust accumulated gas. From each chamber, the exhaust was collected into a sampling point for the MS (Pfeiffer, Thermostar GSD320-QMG220) to detect the gaseous products. Evaluated iron electrodes in alkaline electrolyte were cycled against an oversized commercial nickel electrode (Gates Energy). The iron electrodes were activated prior to the efficiency analysis, and then fitted into the setup to quantify the amount of hydrogen gas. The electrodes were cycled using a potentiostat (BioLogic, SP-50) at charge current densities of 5–15 mA cm–2 while the discharge was kept constant at 5 mA cm–2. In the case of zinc electrodes, symmetric cells were used in mildly acidic electrolytes at current densities of 1–100 mA cm–2. The geometric surface areas of the zinc working and counter electrodes were 1 and 9 cm2, respectively.
16
2.3. Operando techniques 2.3.1. Energy-dispersive X-ray diffraction (EDXRD)
Operando EDXRD enabled the MnO2 phase transformations that occurred during cell operation in alkaline electrolytes to be distinguished. A 3D printed cell holder (Formlabs, Form 1 3D printer) was prepared and from a resin that was stable in concentrated alkaline solution and highly transparent to X-rays. The cell had an open design with no compression. It was filled with 1 M KOH + 3 M LiOH as electrolyte and cycled against a zinc foil as the counter electrode. The cell was cycled at 147 mA g–1 for both charge and discharge, using a battery cycler (MTI, BST8-3). Data collection was conducted at the National Synchrotron Light Source on the beamline X17B1 at Brookhaven National Laboratory. Prior to running the cell, the incident X-ray beam and detection at a fixed angle of 2θ = 3° were aligned. The incident beam consisted of white-beam radiation in the energy range of <20 to 200 keV and diffracted X-rays were collected with a cryostat-cooled Ge detector (Canberra). The collected X- ray signals were digitally processed using an 8192-channel analyzer. Data collection points occurred every 1 min as the cell cycled. Calibration of the X-ray energy used LaB6 and CeO2 as standards. Paper I further details the positioning of the incident beam and detector as well as the complete scan procedure.
2.3.2. Neutron diffraction Operando neutron diffraction measurements were carried out to distinguish the phase evolution of an iron electrode with stannate, which used a cell design described elsewhere113. A 6 M KOD + 0.1 M
K2SnO3 electrolyte was used to minimize the incoherent scattering from H nuclei. Therefore, on the nickel-mesh counter electrode, O2 and D2 gas evolved upon cycling. The iron electrode was prepared in similar procedure as described in Section 2.2.3 but on a larger scale (1.25 g of iron) to achieve satisfactory intensities of diffracted neutrons. Before data collection, the iron electrode was activated by charge-discharge cycling until it reached a stable discharge capacity. Time-of-flight neutron diffraction data collection was performed at the ISIS pulsed spallation neutron source using the Polaris diffractometer114. Collection time per data point was set to 60 min to ensure good statistical quality. The cell discharged at 96 mA g–1 for 5 h and charged at 192 mA g–1 for 4 h, with a 30-min rest step in between. Rietveld analysis of the collected data used the GSAS software and data from the low-angle detector bank (40° < 2θ < 67°, dmax = 7 Å, Δd/d = 0.86%). The Bragg peak profiles were described by a pseudo-Voigt function although only the Gaussian width of the function was refined115.
2.3.3. X-ray diffraction (XRD) Modified pouch cells, based on cells used in other reported work116, were constructed for operando
XRD measurements to identify the phase evolution of MnO2 in aqueous sulfate Zn–MnO2 cells. The pouch cell had 13-mm holes punched on both sides to avoid contribution of the polymer-coated aluminum foil to the XRD patterns, and these were sealed with Kapton tape. The zinc counter electrode also contained a 5-mm hole to allow the incident X-ray beam window to be focused on the
MnO2 electrode. The incident beam passed through the following components: Kapton tape, electrolyte, glass fiber paper and the MnO2 electrode. Figure 2.2 shows a schematic of the pouch cell.
17
Figure 2.2. Schematic of the operando XRD pouch cell components with incident X-ray beam direction illustrated. Figure reproduced from the supplementary information for Paper III72 with permission from Elsevier.
Data collection upon phase transformations in the MnO2 electrode used an in-house single crystal diffractometer (Bruker AXS) with Mo source and a CCD detector (2D Apex II). XRD patterns were collected in transmission mode every 10 min using a 0.5 or 0.8-mm beam collimator. The cell was charged and discharged at 30 mA g–1, including a constant-voltage step at 1.8 V during charge, until the current dropped below 10% of the input value.
2.4. Scanning electron microscopy (SEM) and energy- dispersive X-ray spectroscopy (EDS)
SEM micrographs were captured from secondary electrons for visualizing the sample morphology and particle size, while EDS provided information on the sample surface composition. SEM and EDS investigations were performed for the MBDB, oxygen electrocatalysts, MnO2 electrodes after cycling in aqueous sulfate electrolytes and iron electrodes after cycling in the stannate-additive study. SEM micrographs of the synthesized MBDB were captured using an environmental scanning electron microscope (ESEM, Philips XL 30) operated at 20 kV. SEM images of the oxygen electrocatalyst and cycled MnO2 and iron electrodes were recorded using a field emission microscope (JEOL, JSM- 7001F), operated at 15 kV, and integrated with EDS capabilities to analyze the elemental composition of the materials. Paper IV includes SEM images of the nanostructured copper- and/or tin-doped iron materials.
2.5. XRD
XRD patterns were mainly used for phase identification and structural analysis. Ex situ XRD patterns were collected on MBDB and the oxygen electrocatalyst materials. The XRD pattern of MBDB was acquired using a Bruker D8 Advance diffractometer operated in Bragg–Brentano θ-2θ geometry and
18 using a Cu Kɑ1 source, λ = 1.54056 Å. The XRD patterns of the oxygen electrocatalysts were acquired using a PANalytical X´Pert Pro diffractometer, in Bragg–Brentano θ-θ geometry, and a Cu Kɑ source, λ = 1.54056 Å. Papers II, III and IV include XRD patterns of the iron electrodes used in the stannate additive study, MnO2 electrodes used in aqueous sulfate electrolyte, and nanostructured copper- and/or tin-doped iron materials, respectively.
2.6. Electron energy loss spectroscopy (EELS)
EELS provided structural information of cycled MnO2 electrodes in aqueous sulfate electrolytes. EELS data collection used a Schottky field emission electron microscope (JEOL, JEM-2100F) operated at 200 kV, with a probe size of 0.7 nm and a camera length of 2 cm. The microscope had an integrated CCD camera (Gatan, Ultrascan 1000), post-column energy filter (Gatan Imaging Filter, Tridiem 863), and a Gatan annular dark field detector. MnO2 electrode samples cycled in aqueous sulfate electrolyte were washed, dried, and then scratched off the carbon paper with a diamond scriber and collected on a holey carbon support film (SPI, 300 mesh Cu). Acquired EELS spectra covered the Mn
L2,3 edges (651 and 640 eV, respectively), the O K edge (530 eV) and the Zn L2,3 edges (1043 and 1020 eV, respectively). These spectra were recorded in the scanning transmission electron microscopy mode with applied spatial drift correction and subpixel scanning modality. All EELS spectra were acquired at inelastic mean free path regions between 0.3 and 0.8. Full width at half maximum of the zero-loss peak was 1.20 eV. Multiple scattering effects in the EELS spectra were removed by Fourier ratio deconvolution of low-loss features and the background fitted by a power-law model.
Estimation of the Mn oxidation states was based on integrated L3/L2 intensity ratios using a double arctan continuum model117,118 as detailed in other work119,120.
2.7. X-ray photoelectron spectroscopy (XPS)
Surface sensitive XPS analysis was used to quantify tin in the cycled iron electrode with stannate. XPS analysis used a ULVAC-PHI Versaprobe 5000 spectrometer with monochromatic Al Kɑ line (1486.6 eV) as the X-ray source. The XPS operated at a nominal power of 25 W with a beam size of 100 µm and at a 45° angle while the sample was charge-neutralized of both low-energy electrons and Ar+ ions. Calibration used two reference points for the spectrometer energy scale: Au 4f7 (84 eV) and Cu
2p3 (932.7 eV). Detection of tin was performed on iron electrodes that had been cycled with stannate by measuring the Sn 3d spectrum (484–498 eV) to a depth of 5 nm with an energy pass of 23.5 eV and step size of 0.1 eV. Paper II includes the deconvolution of the Bi, F, C, and O XPS spectra.
19
3. Results and discussion
The results and discussion in this thesis are a synopsis of those given in Papers I–V. Paper I covers the detailed synthesis, composition, structure characterization and electrochemical properties of bismuth doped β-MnO2 in electrolyte mixtures containing multiple cations for rechargeable Zn–
MnO2 alkaline batteries. Paper III elucidates the reaction mechanism of rechargeable aqueous sulfate Zn–MnO2 batteries by structural and electrochemical characterization. Paper II deals with the effect of stannate on rechargeable alkaline iron electrodes by electrochemical, morphological and phase evolution characterizations. Paper IV explores nanostructured doped iron materials for high- performance rechargeable iron electrodes and gives insight into optimal cell operation. Lastly, Paper V investigates proper bifunctional oxygen electrocatalysts for air electrodes and adapts one of them to an Fe–air battery in order to prove the feasibility of this technology.
3.1. Cation electrolyte mixtures for rechargeable Zn–MnO2 alkaline batteries (Paper I)
3.1.1. Structural characterization of bismuth-doped β-MnO2 (MBDB)
Figure 3.1 shows the XRD pattern and SEM image of the synthesized bismuth doped β-MnO2 (MBDB) material. The reflections at 2θ = 31–34° were from Bi2O3 (ICSD 417638). This 2θ region was excluded in the Rietveld refinement to improve the precision of the fit. The remaining reflections matched well with those from β-MnO2 (ICSD 73716). The stoichiometric ratio of Bi:Mn corresponding to 3.5% of the manganese atoms in the (0,0,0) position replaced with bismuth was used in the refinement, while oxygen atoms were located in the (0.3046,0.3046,0) position, with full occupancy, yielding Rp =
9.8% and Rwp = 15.4%. The synthesized MBDB material was produced as fine interconnected particles in the size range of 50–100 nm, as shown in Figure 3.1b.
Figure 3.1. a) Ex situ XRD pattern of MBDB including observed data (black), calculated pattern (red), ICSD peak positions for β-MnO2 (green), and the difference between raw data and fitted model (blue). b) SEM image of synthesized MBDB. Figure reproduced from Paper I69 with permission from American Chemical Society.
20
3.1.2. Electrochemical performance of KOH:LiOH electrolytes
Electrochemical performance of the alkaline Zn–MnO2 cells was analyzed for mixed-cation electrolytes with varied KOH and LiOH concentrations. Figure 3.2a shows discharge capacity results for a wide composition range of electrolyte mixtures measured at 205 mA g–1 to identify optimal mixture concentrations. The cells containing 4:0, 3:1 and 2:2 mixtures of KOH:LiOH (total concentration 4 M) had similar cycle trends. These cells attained high discharge capacities of 230– 370 mAh g–1 in the first cycle but these values quickly declined to 50–100 mAh g–1 in subsequent cycles. Meanwhile the 1:3 and 0:4 KOH:LiOH mixtures sustained higher capacities over 10 cycles. The 4 M LiOH cell failed after 10 cycles, however, because of passivation of the zinc electrode as ZnO is less soluble in the weakly basic pure LiOH electrolyte121. The 1 M KOH + 3 M LiOH cell cycled over 50 times and showed less capacity fading during these cycles. Figure 3.2b shows results obtained in electrolyte mixtures with concentration ratios close to 1:3 and tested at slower rates, with the first cycle at 31 mA g–1 and subsequent cycles at 147 mA g–1. Superior cell performance was achieved within the approximate range of 1:3 to 3:7 KOH:LiOH ratios. At the higher ratio of 1:2 or the lower ratio of 1:4, the capacity dropped significantly after the first cycle, to 110–80 mAh g–1. This quick decay in performance can be explained by a disfavored proton insertion mechanism when the
KOH:LiOH ratio was too low and the favored formation of ZnMn2O4 when the ratio was too high. These results indicate that an optimized KOH:LiOH ratio can improve the overall cycle life of Zn–
MnO2 cells. Paper I elucidates the electrochemical effect of bismuth doping in β-MnO2 of cells cycled in 1 M KOH + 3 M LiOH electrolyte. In short, bismuth affected the second-electron discharge region 3– 4– by complexing with dissolved manganese species of Mn(OH)2 or Mn(OH)2 and inhibited the 122 formation of Mn3O4/ZnMn2O4 . This improved the cyclability of Zn–MnO2 cells in 1 M KOH + 3 M
LiOH, where discharging occurred by simultaneous proton and lithium insertion into the MnO2 structure.
Figure 3.2. MBDB cells cycled against zinc metal in different KOH:LiOH electrolyte mixtures, at rates of a) 205 mA g–1 and b) 31 mA g–1 for the first cycle and 147 mA g–1 for subsequent ones. Figure reproduced from Paper I69 with permission from American Chemical Society.
3.1.3. Phase evolution investigation of MBDB electrodes The details of phase evolution in MBDB electrodes upon cycling in 1 M KOH + 3 M LiOH were studied using both ex situ and operando techniques. Figure 3.3 shows ex situ XRD patterns of MBDB
21 electrodes cycled once and stopped at the discharged (Figure 3.3c) and charged (Figure 3.3d) states.
The structural transformation that occurred upon the first full discharge reduced β-MnO2 into a mixture of LiMn2O4 and Mn(OH)2. Observed shifts in the (111), (113) and (222) peaks of LiMn2O4 at 19.1, 33.9 and 44.9° evidenced lithium insertion. Concurrent proton insertion formed the other product, Mn(OH)2. These reduced products demonstrated the plausible simultaneous lithium and proton insertion into the β-MnO2 structure over the first full discharge. During the reoxidation step, the original MBDB structure was not recovered. Instead, a new set of peaks emerged, especially the noteworthy peak at 12°. The strongest emerging peaks can be assigned to the layered birnessite 123 structure (δ-MnO2) . Bi2O3 peaks appearing at 31–34° for the pristine electrode disappeared during discharge due to reduction of Bi2O3 to amorphous bismuth metal. The other components, PTFE, carbon black and nickel, remained unchanged throughout cycling.
Figure 3.3. Ex situ XRD patterns of MBDB electrodes in various states of charge and discharge in 1 M KOH + 3 M LiOH: a) MBDB powder, b) pristine MBDB electrode, c) after the first full discharge and d) after the first full recharge. Figure reproduced from Paper I69 with permission from American Chemical Society.
Figure 3.4 shows the operando EDXRD phase evolution investigation of MBDB, which revealed the same reaction mechanism seen from the ex situ XRD measurements. MBDB was fully reduced to
Mn(OH)2 and LiMn2O4, and never recovered upon recharge. The Mn(OH)2 peaks appeared at 81.42, 95.2, 128.5 and 141.4 keV during discharge, while an increased intensity at 77.8 keV, close to the graphite peak, corresponded to LiMn2O4. Interestingly, peaks stemming from metallic bismuth developed during discharge and then faded away when bismuth dissolved into the electrolyte during charge. The birnessite phase was not detected after the first cycle because of the limited energy range of the EDXRD scan. The absence of the intermediate trivalent MnOOH phase, usually formed during normal proton insertion71,73, may be explained by the favored lithium insertion that formed
LiMn2O4 instead. This may also be the key to the improved cyclability, as the irreversible
Mn3O4/ZnMn2O4 phases form through the trivalent intermediate. These irreversible phases cannot be oxidized back to birnessite, but Mn(OH)2 can.
22
discharged
charged
discharged
charged
Figure 3.4. Operando EDXRD measurements of a MBDB cell cycled against zinc metal in 1 M KOH + 3 M LiOH. The colored vertical arrows indicate peaks associated with particular phases and the black horizontal arrows indicate the end of charge and discharge of the MBDB electrode. Figure reproduced from Paper I69 with permission from American Chemical Society.
Ultimately, mixed cation electrolytes containing KOH and LiOH enabled a competing proton- and lithium-insertion mechanism in rechargeable alkaline Zn–MnO2 batteries. This improved the use of
MnO2 and upheld reversibility, while suppressing the inactive ZnMn2O4 phase from proton insertion in favor for LiMn2O4.
2+ 3.2. Reversible aqueous sulfate Zn–MnO2 batteries with Mn (Paper III)
3.2.1. Electrochemical characterization of MnO2 electrodes
2+ Figure 3.5 depicts the electrochemical impact of Mn in aqueous sulfate Zn–MnO2 cells by 2+ 2+ comparing results for cells with 2 M ZnSO4 (Mn -free) and 2 M ZnSO4 + 0.1 M MnSO4 (added-Mn ) electrolytes. The cells used overdimensioned zinc metal counter electrodes to allow full investigation of the MnO2 electrodes. The CV curves in a) and b) show the redox reactions that occur and the distinction between cells. In both cases, the first reduction entailed a one-electron transfer at 1.21 V while the subsequent ones had two peaks, at 1.25–1.28 and at 1.37–1.39 V. This shifting electrochemistry from the first cycle to following ones was caused by the conversion of ɣ-MnO2 to a layered hydrated MnO2 structure containing zinc. The two electron-transfer peaks after the first cycle correlated to the insertion mechanisms of protons and zinc. The electron transfer characteristics of both samples were analogous except that the added Mn2+ cell upheld reversibility over 50 cycles. Figure 3.5c validates the reversibility of charge-discharge cycling in the cell that had Mn2+ added to the electrolyte. Both samples showed increasing capacity over the first 15 cycles, which can be explained by poor wettability of the electrode. Eventually, maximum capacities were
23 reached and the added-Mn2+ cell maintained a stable capacity of 220 mAh g–1 over 100 cycles while the Mn2+-free cell deteriorated. The Coulombic efficiencies were unaffected by the declining capacity and remained over 98%. Figure 3.5d and e shows the charge and discharge profiles from these measurements. The figures depict a significant capacity fade when no Mn2+ is added to the electrolyte. The discharge over the last cycles in d) showed a declining capacity for the second plateau; this was related to zinc insertion and altered the reversibility. The added-Mn2+ cell however showed a stable performance that implied that Mn2+ restrained zinc insertion. The rate capability of the added-Mn2+ cell shown in Figure 3.5f achieved discharge capacities of 233, 216, 188, 141, 97 and –1 –1 76 mAh g at 30, 60, 120, 300, 750 and 1500 mA g , respectively. Overall, the Zn–MnO2 cell offered satisfactory rate capability.
2+ Figure 3.5. Electrochemical performance of aqueous sulfate Zn–MnO2 cells in 2 M ZnSO4 (Mn free) and 2 M 2+ –1 2+ ZnSO4 + 0.1 M MnSO4 (added Mn ). Cyclic voltammetry over 50 cycles at 0.2 mV s of a) Mn -free and b) added-Mn2+ electrolytes. c) Charge-discharge cycle summary of the two samples over 100 cycles at 60 mA g–1, where the left y-axis corresponds to specific discharge capacity (blue markers) and the right y-axis to the 2+ Coulombic efficiency (red markers). Charge and discharge profiles of the MnO2 electrodes in d) Mn -free and 2+ –1 2+ e) added-Mn electrolytes at 60 mA g . f) Rate capability analysis of the MnO2 electrode in the added Mn electrolyte. All cells were cycled against zinc metal and the current and capacity were normalized based on the 72 active mass of MnO2. Figure reproduced from Paper III with permission from Elsevier.
3.2.2. Characterization of cycled MnO2 electrodes
SEM images and elemental analysis of the surface in Figure 3.6 show structural changes to the MnO2 electrodes after cycling. The pristine electrode in Figure 3.6a consisted of EMD, carbon black, 2+ graphite and PVDF. After cycling in Mn -free electrolyte, and at a fully charged state, the MnO2 electrode shown in Figure 3.6b had a densified surface and contained a significant amount of zinc. This indicated that zinc insertion occurred during cycling, and in a weight amount close to that of manganese. Moreover, the dense surface decreased the porosity of the electrode structure and inhibited electrolyte contact with active sites. The electrode cycled in an electrolyte with added Mn2+ had a fibrous morphology and contained less zinc than manganese at its surface. As will be explained in a later section, zinc insertion into the MnO2 structure formed Zn1–xMnO2·nH2O. The availability of
24
Mn2+ in the electrolyte affected the amount of zinc inserted. In Figure 3.6c and d, the detected B and Si on the surface originate from the borosilicate glass paper. The glass paper tended to stick to the surface because of the precipitate, Zn4SO4(OH)6·5H2O, which formed during discharge. The precipitate adsorbed large amount of H2O and dehydrated the surface. Figure 3.6d shows the precipitate as large hexagonal layered flakes 10–30 µm in size. The precipitate completely covered the electrode surface and has low electrical conductivity. It was indeed surprising that the cell even managed to cycle with such coverage, indicating facile dissolution of the precipitate during charge via decreasing pH61.
Figure 3.6. SEM images and elemental analysis results (inset) for MnO2 electrodes in aqueous sulfate electrolytes. a) Pristine MnO2 electrode composed of EMD, carbon black, graphite and PVDF. The MnO2 electrodes were cycled 100 times and stopped at a charge state in b) Mn2+-free and c) added-Mn2+ 2+ electrolytes. d) MnO2 electrode cycled 100 times and stopped at a discharged state in added-Mn electrolyte. In c), the highlighted blue color area corresponds to glass fiber paper and the red color area to carbon paper. Figure reproduced from Paper III72 with permission from Elsevier.
EELS provided further structural analysis of cycled MnO2 electrodes at fully charged states. The intensities in EELS spectra reflect the density of states and enabled a determination of the oxidation state of manganese. Figure 3.7a shows representative EELS spectra of the O K edge, Mn L2,3 edges and Zn L2,3 edges from the pristine MnO2 electrode and electrodes cycled in electrolytes with and without Mn2+. All samples showed strong white line features at Mn L edges from the non-filled 3d
25
117,118,124 conduction band . The energy position and characteristics of Mn L2,3 edges were similar for the uncycled electrode and the electrode cycled in electrolyte with added Mn2+; the former had an
L3/L2 intensity ratio of 2.31 ± 0.15 and the latter 2.42 ± 0.19. The intensity ratio of the pristine MnO2 deviated from the theoretical value 2 ± 0.1117,125 but it has been reported elsewhere that high values 126 2+ are obtained for nanosized MnO2 . The slight increase in intensity ratio for the added Mn sample indicates a lower oxidation state of Mn, which correlates with the addition of zinc to the structure, as detected in the EELS spectra shown in Figure 3.7b. By comparison, the sample that was cycled in 2+ Mn -free electrolyte sample showed a small shift of the Mn L2,3 edges to lower energies. The intensity ratio of the Mn L2,3 edges increased here to 3.05 ± 0.56, indicating an even lower oxidation state of Mn than in the added Mn2+ sample. The O K edge characteristics of the samples were more distinct. The “a” and “b” peaks for the uncycled electrode and the one cycled in added Mn2+ had similar features. The “a” peak was located at 531.40 eV for both samples and the “b” peak was found at 556.60 and 543.8 eV for the respective samples. However, another peak appeared between these features after cycling in added Mn2+ electrolyte and is attributed to an additional state given by zinc in the structure127. In the sample cycled Mn2+-free electrolyte, the features are in contrast with those of the other samples and are labeled as “a*” and “c*”. These two peaks were located at 538.4 and 555.8 eV, respectively, and did not match with typical fine-structure features of manganese oxides. Instead, the broad energy width and range covered by the “a*” peak suggested that it may originate from mixed manganese oxides with different valence states. This sample contained a significant amount of zinc, consistent with the EDS results shown above.
26
Figure 3.7. a) EELS spectra covering the Mn L2,3 and O K-edges of the pristine and cycled MnO2 electrodes at charge states in both electrolytes. The intensity of Mn L2,3 and O K-edges are normalized to the maximum intensity of Mn L3. b) EELS spectra covering the Zn L2,3 spectra for the cycled MnO2 electrodes at charge states 2+ in both electrolytes. The intensity of Zn L2,3 edge from the sample cycled in added-Mn electrolyte is amplified for visualization and not relative to the Mn2+-free sample. Figure reproduced from Paper III72 with permission from Elsevier.
3.2.3. Progression of the MnO2 charge and discharge mechanism
Operando measurements of the phase evolution of MnO2 electrodes used a suitable designed beam window in the modified pouch cell to collect XRD patterns in transmission mode every 10 min. Reference patterns were measured for the individual cell components, EMD, carbon paper, glass fiber paper and electrolyte, before running the cell. Figure 3.8 shows the reference patterns of these components. The electrolyte had the highest background contribution and shifted during cycling because of varying H2O volumes. The background was set to be the same when evaluating different state-of-charge (SOC) patterns upon cycling.
27
Figure 3.8. Reference XRD patterns of the electrolyte, MnO2 electrode, carbon paper, and glass fiber paper from operando XRD data collection. The background was set to be the same throughout different SOC patterns upon cycling. The patterns of the three given SOC at 0, 50 and 100% show the significant background shift. Figure reproduced from the supplementary information for Paper III72 with permission from Elsevier.
2+ Figure 3.9 shows the XRD patterns and charge and discharge profiles of MnO2 cells in the Mn -free and added-Mn2+ electrolytes. The significant peak arising at 2θ below 10° during charge can be assigned to the (001) reflection of the hydrated phyllomanganate, Zn-buserite. Zn-buserite, a layered structure similar to that of birnessite (δ-MnO2), but with an interlayer spacing of 10 Å containing two 2+ 128 H2O molecules instead of one and Zn balancing the negative charge in the octahedral layers . The intensities of the (001)b and (002)b reflections belonging to the Zn-buserite phase and their peak positions differed in the cells containing electrolyte without and with Mn2+. The former had 2θ values of 8.65 and 18.16° (d = 10.22 and 4.88 Å, respectively) while the latter had values of 9.61 and 18.43° (d = 9.20 and 4.81 Å, respectively). The larger d-spacing values can be explained by the increased quantity of Zn2+ inserted into the interlayer region, which expands the unit cell22. The remaining reflections highlighted correspond to Zn1–xMnO2·nH2O, which is present in two intergrown phases: hexagonal pyrolusite (β-MnO2) and orthorhombic ramsdellite. In both cells, the evolved phases during charge represented Zn1–xMnO2·nH2O and Zn-buserite, whereas ZnMnO2·nH2O and
Zn4SO4(OH)6·5H2O formed during discharge. In Figure 3.9a, the unit cell of Zn1–xMnO2·nH2O expanded and contracted during cycling, as evidenced from the peak shifts to lower and higher 2θ, 2+ respectively. The peak shifts suggested Zn insertion to form ZnMnO2·nH2O and extraction back to
Zn1–xMnO2·nH2O. Similar operando phase evolution of zinc insertion into MnO2 has been reported elsewhere45. In contrast, the cell containing added Mn2+ had smaller peak shifts, which indicated that the crystal structure was more stable to minor unit cell volume changes than the cell without Mn2+.
For comparison, the (102)h and (110)h peak positions of Zn1–xMnO2·nH2O from fully discharged to charged states of the Mn2+-free cell increased by 1.8 and 2.7%, respectively, whereas for the added- Mn2+ cell they increased by 0.4 and 1.1%, respectively. The Zn-buserite peaks faded upon discharge, suggesting dissolution of the compound. From these observations, the impact of added Mn2+ is correlated to the quantity of Zn2+ insertion. The accessible Mn2+ suppressed dissolution and inhibited Mn vacancies for Zn2+ insertion. Higher Zn2+ insertion may destabilize the crystal structure and inevitably result in deteriorating capacity.
28
The peaks that arose during discharge included those of the crystalline Zn4SO4(OH)6·5H2O. This compound precipitated due to increased pH (lowered concentration of H+ during proton insertion) at pH around 5.561,62 without any exchange of electrons. The initial pH values of the Mn2+-free and added-Mn2+ electrolytes were 4.2 and 3.9, respectively. The precipitate directly influenced the cell 129 reversibility by buffering the pH around 5.5 and inhibited formation of inactive ZnMn2O4 at pH > 7 . However, the precipitate has low electrical conductivity and could impede diffusion of ions to active sites by blocking the surface and plugging the pores. Limiting the amount of this precipitate is important to maintain the rechargeability of the aqueous sulfate Zn–MnO2 cell.
Figure 3.9. Operando phase evolution analysis by XRD of MnO2 electrodes upon charge and discharge. Acquired XRD patterns in samples a) Mn2+-free and c) added-Mn2+ electrolyte at 30 mA g–1 with their corresponding charge and discharge profiles in b) and d), respectively. Assigned diffraction indices with subscripts correspond to the following phases: b = Zn-buserite, o = orthorhombic and h = hexagonal part of
Zn1–xMnO2·nH2O. In the XRD patterns, the blue line represents the fully charged state and red the fully discharged state. Straight dashed orange lines display peak position shifts of the hexagonal reflections. Figure reproduced from Paper III72 with permission from Elsevier.
To summarize, the MnO2 reaction mechanism in aqueous sulfate Zn–MnO2 cells can be described as:
+ − 2+ 푀푛푂2 + 4퐻 + 2푒 ⇌ 푀푛 + 2퐻2푂 (3.1)
2+ − 푍푛1−푥푀푛푂2 ⋅ 푛퐻2푂 + 푥푍푛 + 2푥푒 ⇌ 푍푛푀푛푂2 ⋅ 푛퐻2푂 (3.2)
29
2+ 2− − 4푍푛 + 푆푂4 + 6푂퐻 + 5퐻2푂 ⇌ 푍푛4푆푂4(푂퐻)6 ⋅ 5퐻2푂 (3.3)
The reaction during the first discharge proceeded via proton insertion into EMD and dissolution to Mn2+ (3.1). In subsequent cycles, the same mechanism took place, but with Zn-buserite dissolving instead. From these reactions, the concentration of H+ decreased upon proton insertion and lead to 2+ the precipitation of Zn4SO4(OH)6·5H2O (3.3). Simultaneously, Zn insertion occurred where x denotes the number of reversibly inserted cations (0 < x < 1) and formed ZnMnO2·nH2O (3.2). Paper III describes a hypothesis for the intermediate Mn3+ reaction that is excluded here.
3.3. Effect of stannate on rechargeable iron electrodes (Paper II) 3.3.1. Electrochemical and structural characterization of iron electrodes Charge and discharge profiles of iron electrodes cycled in alkaline electrolytes, shown in Figure 3.10, illustrate the impact of stannate in the electrolyte. The cells were cycled over Fe0/Fe2.67+ and included deep discharge regimes and the formation of Fe3O4. This regime is avoided in commercial alkaline iron electrodes, where the discharge reaction is limited to Fe(OH)2 with an Fe oxidation state of 2+. Both electrodes contained Bi2S3 as an additive to suppress the HER; this compound reduced during charging130:
− 2− 퐵푖2푆3 + 6푒 ⇌ 2퐵푖 + 3푆 E° = –0.92 V vs. Hg/HgO (3.4)
Metallic bismuth is deposited on the electrode surface while sulfide is adsorbed on iron and poisons its activity towards hydrogen90. The discharge profile of the iron electrode in an electrolyte without stannate (Figure 3.10a) had two plateaus. The first involved a two-electron oxidation to Fe(OH)2 at –
0.85 V and the second further oxidation to Fe3O4 at –0.70 V. The capacity increased successively over 30 cycles and stabilized thereafter around 400 mAh g–1. This increase in discharge capacity is a known formation step that occurs as the electrolyte gradually penetrates to the core of the iron particles89. During charge, metallic iron recovered and overcharged to 500 mAh g–1; the remaining capacity was attributed to the HER. In the cell with stannate added, different charge and discharge 2– trends evolved as shown in Figure 3.10b. K2SnO3 dissolved to give stannate ions, Sn(OH)6 , that were reduced to tin metal, which increased the HER overpotential during charge131:
2− − − 푆푛(푂퐻)6 + 4푒 ⇌ 푆푛 + 6푂퐻 E° = –1.02 V vs. Hg/HgO (3.5)
The redox potentials with stannate shifted, and after formation, the discharge profile had one main plateau at –0.84 V that stabilized around 250 mAh g–1. This shift in redox potentials is attributable to the formation of an alloy of iron and tin. The voltage cut-off during charge was set to –1.158 V to avoid the excessive reduction of stannate ions in the electrolyte to metallic tin. Interestingly, the HER overpotential shifted to more negative potentials, which allowed the Coulombic efficiency to be determined. The efficiency obtained after formation was 85 ± 5%. If the two cells were compared over the first discharge plateau, where commercial iron electrodes typically operate, the cell with stannate (250 mAh g–1) outperformed the one without (150 mAh g–1).
30
A
B
Figure 3.10. Charge and discharge profiles of iron electrodes cycled against nickel electrodes over 150 cycles at –1 111 192 mA g in a) 6 M KOH + 1 M LiOH and b) + 0.1 M K2SnO3. Figure reproduced from Paper II with permission from Journal of the Electrochemical Society.
Figure 3.11 shows SEM images of charged iron electrodes after cycling in alkaline electrolytes, as well as of a pristine electrode. The pristine electrode in Figure 3.11a consisted of large graphite flakes, together with smaller carbon black particles in the size range 40–60 nm, and iron powder with an average particle size of 10 µm. After cycling the electrode, the particle size of iron dropped to 0.2–1 µm, as shown in Figure 3.11b. This decrease in particle size occurred as iron dissolving and was redeposited during cycling. In the electrode that was used in an electrolyte with stannate added, octahedron-like particles 0.6–1 µm in size were deposited on the surface, as shown in Figure 3.11c. Spot elemental analysis with EDS showed an accumulation of tin in these particles. The surface in Figure 3.11b consisted of 45wt% Fe and 4 wt% Bi, whereas that in Figure 3.11c was 22 wt% Fe, 2 wt% Bi, and 3 wt% Sn.
31
Figure 3.11. SEM images of a) pristine iron electrode, b) charged iron electrode cycled in 6 M KOH + 1 M LiOH 111 and c) charged iron electrode cycled in 6 M KOH + 1 M LiOH + 0.1 M K2SnO3. Figure reproduced from Paper II with permission from Journal of the Electrochemical Society.
XPS analysis confirmed the presence of tin in the iron electrode cycled in the presence of stannate. Figure 3.12 depicts the Sn 3d spectrum of the electrode to a depth of 5 nm. The fitted data matched that of SnO2 formed from tin oxidized under exposure to air. Detection of tin here agrees with the EDS results described earlier. Other elements detected on the surface included C, O, Bi, F and Fe. However, the signal of Fe was weak and indicated only a minor amount on the surface. Paper II includes the XPS spectra of the other elements.
32
Figure 3.12. Deconvolution of the Sn 3d XPS spectrum measured to a depth of 5 nm depth on the iron electrode that was cycled 150 times in 6 M KOH + 1 M LiOH + 0.1 M K2SnO3, at charge state. Figure reproduced from Paper II111 with permission from Journal of the Electrochemical Society.
3.3.2. Phase evolution characterization and structure refinement Operando neutron diffraction measurements used a cell design reported in other work for the Polaris instrument at ISIS113,132. Prior to the measurement, the iron electrode was activated by cycling in deuterated 6 M KOD + 1 M K2SnO3 electrolyte. After activation, the discharge capacity reached 420 mAh g–1. Figure 3.13 depicts a representative neutron diffraction pattern and Rietveld refinement of the activated iron electrode in the cell, which also contained nickel (current collector and counter electrode). The background from the amorphous and liquid components was fitted by a shifted Chebyschev function with 12 coefficients. Initially, the refinement included the following phases: Ni(s) (PDF 00-004-0850), Fe(s) (PDF 00-006-0696), Fe3O4(s) (both nuclear and magnetic structures, PDF 00-019-0629), graphite (“C”, PDF 00-056-0159), and Bi(s) (PDF 00-044-1246). Ni dominated the pattern and Bi was found to have a very small contribution to the refinement and was hence excluded. A set of unidentified peaks whose intensity changed during cycling indicated a new phase taking part in the electrochemical reaction. The peaks of this new phase did not shift in position and were therefore not the result of any intercalation reaction, which would induce changes in the volume of the unit cell. The peaks could be indexed to a tetragonal unit cell with the dimensions a = 3.4049(7) and c = 3.255(2) Å. This unit cell was reminiscent of a high-pressure body-centered Sn modification with a = 3.70(1) and c = 3.37(1) Å133. Fe, Bi and C have also been reported to form isostructural solid solutions with Sn134–136. The observations that Sn metal was not detected in Figure 3.13, but that the presence of Sn was confirmed ex situ by EDS and XPS, suggest the presence of a new solid solution phase, composed of Sn in combination with other available elements in the electrode. Indeed, Fe, Bi and C have been found to form related structures. A good fit to the neutron diffraction data was accomplished using the space group P4/mmm with Fe, Sn, Bi and C located at (0,0,0) and (½,½,½) and D atoms located in (½,½,0). Thus, this structure model contains an octahedral interstitial site in the metal atom framework, similar to ones found for metal hydride structures. The presence of D obstructs a determination of the element ratios in the solid solution, in particular considering the correlation between site occupancies and thermal parameters, and the similar neutron scattering cross-sections of the elements present. D and C have almost equal scattering lengths, but the 1.63 Å distance between the interstitial sites and the metal atoms is in better agreement with D being the
33 interstitial atom. Thus, the new phase is identified as an intermediate hydride phase (IHP). A good agreement between observed and calculated data was achieved, shown in Figure 3.13, with the refinement converging with a goodness of fit value χ2 = 2.56. The corresponding weight fractions of phases were: Ni (59.2 wt% and RF = 1.1%), C (1.1 wt% and RF = 4.2%), (Fe,Sn,Bi,C)2D (15.3 wt% and RF
= 7.3%), Fe (14.4 wt% and RF = 4.7%), Fe3O4 (9.0 wt% and RF = 4.5%), and Bi (1.0 wt% and RF = 8.8%).
Figure 3.13. Fitted neutron diffraction pattern at a charge state of the iron electrode in 6 M KOD + 0.1 M
K2SnO3. All observed diffraction peaks could be assigned with their respective phase and diffraction indices. Figure reproduced from Paper II111 with permission from Journal of the Electrochemical Society.
Figure 3.14a shows a series of neutron diffraction patterns collected over a full charge-discharge cycle. Specific reflections of the Fe, Fe3O4, and IHP phases underwent the most noteworthy changes during cycling. Figure 3.14b compares the relative weight fractions of these phases over time. Refinement of each neutron diffraction pattern included all phases. Prior to the operando analysis, the cell was fully charged in fresh 6 M KOD + 0.1 M K2SnO3 electrolyte, to ensure a full reduction of 2– the iron electrode. This explained the high initial weight fraction of IHP as Sn(OD)6 is expected to undergo reduction upon D2 evolution. The discharge capacity achieved from this cell corresponded to 380 mAh g–1, consistent with the performance of half-cells. The weight fraction trends of Fe and
Fe3O4 followed the expected phase evolution during charge and discharge. The amount of IHP, however, correlated inversely with that of Fe. When Fe formed, IHP was depleted and vice versa.
Neither Fe(OH)2 nor Sn metal were detected operando. XRD analysis of the fully charged iron electrode recovered after these measurements showed the presence of Sn (see supplementary information of Paper II). This corroborated that Sn was part of the solid solution in the IHP and reacted during cycling. The IHP phase was metastable and decomposed to Sn among other species.
34
a
b Discharge Rest Charge
Figure 3.14. a) Operando neutron diffraction patterns depicting the phase evolution of Fe, Fe3O4 and the intermediate hydride phase (IHP) in the iron electrode during charge and discharge cycling in 6 M KOD + 0.1 M
K2SnO3. Each neutron diffraction pattern was collected over 1 h. b) Weight fraction evolution of representative phases (markers with error bars) and cell potential vs. O2/D2 (dashed green line) as a function of time. Figure reproduced from Paper II111 with permission from Journal of the Electrochemical Society.
In summary, the evolution of hydrogen gas was minimized using potassium stannate as an additive. Tin prolonged the discharge profile of iron electrodes and increased the overpotential required for hydrogen evolution, both desirable attributes for batteries. Operando neutron diffraction measurements showed a correlation of these attributes to the formation of a novel intermediate hydride phase during charge-discharge cycling.
3.4. Hydrogen evolution on iron and zinc electrodes (Papers III and IV)
The HER (1.4) contributes to deviation from 100% Coulombic efficiency in rechargeable alkaline iron electrodes and zinc electrodes in aqueous sulfate electrolytes. Hydrogen evolution on both
35 electrodes was quantified using electrochemistry coupled with mass spectrometry. The design of a special electrochemical cell connected to the MS allowed the detection of hydrogen gas accumulated in the system. The measurements assumed that the HER was the sole contributor to the deviation from 100% efficiency and were based on Faraday’s law:
퐼푡 푛 = (3.6) 푧퐹
Where n = number of moles (mol), I = current (A), t = time (s), z = electrons transferred per ion (z = 2 –1 for H2) and F = 96485 C mol . Calibration used SS as a standard to quantify hydrogen gas. This included linear sweep voltammetry over 0.3–0.55 V vs. Ag/AgCl at 0.2 mV s–1 upon accumulating hydrogen gas. The charge passed (Coulombs) was correlated to the recorded counts of H2 on the MS. From calibration, quantification factors were determined:
퐼 푡 푞푢푎푛푡. 푓푎푐푡표푟 = ∫ 퐸퐶 (3.7) ∫ 퐼푀푆푡
Where IEC = current recorded from the potentiostat (A) and IMS = extracted ion signal from the MS 6 (A). Calculated quantification factors corresponded to (12.1 ± 1) 10 . The number of moles of H2 detected per second based on (3.6) and (3.7) was:
푛 퐼 퐻2 = 푞푢푎푛푡. 푓푎푐푡표푟 ⋅ 푀푆 (3.8) 푡 푧퐹
This was further derived to give the cumulative capacity of H2:
푄퐻2 = 푞푢푎푛푡. 푓푎푐푡표푟 ⋅ 퐼푀푆푡 (3.9)
Where QH2 = charge passed of H2 (As). The charge passed values of H2 were converted to mAh for comparison with the charge capacity of the iron electrode. Finally, the charge efficiency was defined as:
푄푡표푡−푄퐻2 휂푐ℎ푎푟𝑔푒 = (3.10) 푄푡표푡
Where Qtot = total capacity obtained upon charging the iron electrode. Figure 3.15 shows the evolution of charge efficiency for the four iron electrodes: a) CuSn, b) CuSnCNT, c) CuSnCNTLi, and d) SnCNT. For each figure, the top plot shows the cell potential and moles of H2 formed per second (3.8) as a function of time, while the bottom plot shows the charging efficiency (3.10) and the cumulative H2 capacity (3.9) as a function of time. The extended H2 detection window from the MS accounted for all accumulated hydrogen gas. As such, the cumulative
H2 capacity and charge efficiency remained relatively constant close to the end of charge. The highest efficiencies were obtained for the CuSnCNT sample, followed by the CuSn sample. Charging did not improve with LiOH added to the electrolyte, as this decreased the efficiency. This can be explained by the increased dissolution of iron, which facilitates electrolyte access to active sites and stimulates hydrogen evolution137. The sample containing Sn as the sole dopant performed more poorly because of its lower electrical conductivity and had a lower charge acceptance of iron. The efficiency evolution showed three distinct regions during charge. In the first region, below 1.5 V, nucleation of iron occurred, and the hydrogen gas evolution was lowest. In the second region, between 1.5 and 1.6 V, iron reduction proceeded as hydrogen gas evolved at a constant rate. In the
36 last region, above 1.6 V, the iron became fully reduced, and the hydrogen evolution rate increased dramatically and overcharged the cell. From these results, the general trend showed that iron electrodes achieved higher charge efficiencies at current densities above 10 mA cm–2 (C/5 or 192 mA g–1 charge rates). The efficiency values were validated by comparing the calculated charge efficiencies (3.10) acquired from the MS with Coulombic efficiencies from the potentiostat, and only small deviations, ±3%, were observed.
Figure 3.15. Charge efficiency analysis at 5, 10 and 15 mA cm–2 of nanostructured copper and/or tin doped iron electrode materials: a) CuSn, b) CuSnCNT, c) CuSnCNTLi and d) SnCNT. In each figure, the top plot shows the cell potential (left y-axis) and number of H2 moles evolved per second (right y-axis) vs. time. The bottom plot shows the charge efficiency (left y-axis) and cumulative H2 capacity (right y-axis) vs. time. The total charge capacity of the iron electrode is annotated in the bottom plot in each sample. Figure reproduced from Paper IV138 with permission from MDPI.
Figure 3.16 shows hydrogen gas detected upon plating and stripping zinc in aqueous sulfate electrolytes. The setup used symmetric Zn vs. Zn cells coupled to the MS and operated at current density rates of 1–100 mA cm–2. Results showed no noteworthy overpotential difference between
37 the two cells, and neither evolved significant quantities of hydrogen gas; the detection of hydrogen remained close to zero. This was consistent with the Coulombic efficiency values obtained from the potentiostat, which were 99% and higher at all current densities in both cells. The zinc electrodes cycled performed particularly efficient and had excellent electrochemical reversibility.
Figure 3.16. Zinc electrodeposition and dissolution in Zn vs. Zn cells in a) Mn2+ free and b) added Mn2+ electrolytes at 1, 10, 25 and 100 mA cm–2. The blue line corresponds to the voltage profile and red to the ion current of H2 from the mass spectrometer. Figure reproduced from the supplementary information for Paper III72 with permission from Elsevier.
Conclusively, hydrogen evolution on iron and zinc, evaluated as a function of charge current density, was studied by coupled electrochemistry and mass spectrometry. From this, it was concluded that the iron electrode benefited from high-rate operation, which lowered the rate of hydrogen evolution, whereas hydrogen gas was not detected on zinc electrodes in aqueous sulfate electrolytes.
3.5. Flow-assisted rechargeable Fe–air batteries (Paper V) 3.5.1. Structural characterization of oxygen electrocatalysts New bifunctional catalysts for the air electrode were characterized structurally and electrochemically. The structures of the synthesized oxygen electrocatalysts were determined prior to the electrochemical performance analysis. Figure 3.17 shows the XRD patterns of the LCMO, Ni–
Fe LDH and NiCo2O4 catalysts. LCMO consisted of two phases with different La:Ca ratios: 1)
La0.1Ca0.9MnO3 and 2) La0.5Ca0.5MnO3. The peaks of the first phase can be indexed to an orthorhombic unit cell with a = 5.3063(3) Å, b = 7.4923(5) Å and c = 5.2988(3) Å in the space group Pnma139. The second phase, also orthorhombic, can be indexed with unit cell parameters a = 5.4164(5) Å, b = 5.4454(7) Å and c = 7.7099(8) Å in the space group Pbnm140. The structure of the Ni–
Fe LDH material is related to that of β-Ni(OH)2, but with larger interlayer spacing, and incorporated Fe3+ replacing of Ni2+ 141. The pattern for of this material showed poor crystallinity and matched with 142 the reported LDH with an 8:2 ratio of Ni:Fe . The NiCo2O4 XRD pattern is ascribed to a face- centered cubic structure with the unit cell parameter a = 8.1164(5) Å, space group Fd-3m143.
38
311
440 220 400 511 003 111 222 422 012 015 006 110113
0021 220 2 202 022 1 0202 2 2201 1011 2042
Figure 3.17. XRD patterns and assigned diffraction indices for synthesized oxygen electrocatalyst materials:
LCMO, Ni–Fe LDH and NiCo2O4. For the LCMO catalyst, the subscripts belong to the compounds: 1)
La0.1Ca0.9MnO3 and 2) La0.5Ca0.5MnO3. Figure reproduced from Paper V.
Figure 3.18 shows SEM micrographs of the a) LCMO, b) Ni–Fe LDH, and c) NiCo2O4 catalysts. The synthesis of LCMO included a quenching step to break up agglomerates and produced fine interconnected particles 100–300 nm in size. The Ni–Fe LDH material was composed of large amorphous nanosheets, with a needle-like structure, and showed poor crystallinity as seen in the
XRD pattern in Figure 3.17. The single-phase NiCo2O4 catalyst consisted of varied crystallite morphologies, with irregularly shaped spheres and rod-like structures. This material exhibited a broad particle size distribution ranging from 50 to 500 nm.
39
A
500 nm
B
500 nm
C
500 nm
Figure 3.18. SEM images of synthesized oxygen electrocatalysts: A) LCMO, B) Ni–Fe LDH and C) NiCo2O4. Figure reproduced from Paper V.
3.5.2. Electrochemical performance of air electrodes The electrochemical performance of air electrodes was analyzed first by CV, then by charge- discharge cycling until cell failure. The prepared air electrodes were highly hydrophobic and therefore activated with CV before cycling. This established a three-phase boundary between
40 electrolyte, oxygen, and catalyst. In each measurement, nickel served as the counter electrode to produce oxygen or hydrogen gas during operation, and an airflow supplied as oxygen to the electrode. Three air electrode combinations were examined: 1) LCMO + Ni–Fe LDH, 2) LCMO +
NiCo2O4 and 3) NiCo2O4. The rationale for the two first air electrodes was to use combinations of materials to merge active OER and ORR catalysts40,144. LCMO has excellent ORR performance but lacks OER activity, hence it was mixed with superior OER catalysts such as Ni–Fe LDH and 97,105 NiCo2O4 . On the contrary, Ni–Fe LDH catalysts perform poorly in reduction but are highly active 40 for oxidation . NiCo2O4 has been investigated both as an OER enhancer but also as a bifunctional catalyst39,145,146. Figure 3.19a shows the first CV cycle of the examined air electrodes during OER and ORR. The maximum current densities of OER/ORR for the three samples were: 1) +4/–8, 2) +4/–11 –2 and 3) +9/–16 mA cm . The air electrode with NiCo2O4 as single catalyst attained the highest current densities in both directions and also had earlier onset potentials for OER (+5 mA cm–2 and +479 mV) and ORR (–4.8 mA cm–2 and +80 mV). Figure 3.19b shows the assessed cycle life of the air electrodes subjected to 2 h charge and discharge steps until they failed. Mixtures of LCMO with Ni-Fe LDH or NiCo2O4, proved less activity and less stable than the single NiCo2O4 catalyst. The first sample had significant overpotentials in ORR and failed after 110 h of operation. This is ascribed to the limited ORR activity of the Ni–Fe LDH, which caused rapid cell failure147. In the second sample, a large current spike in ORR occurred after 205 h because of an unintended stop of airflow. The airflow was recovered shortly thereafter but the cell deteriorated in subsequent cycles. Nevertheless, the air electrode containing only NiCo2O4 exhibited lower total overpotentials of OER and ORR and sustained 440 h of operation. This electrode had a total overpotential of 693 mV after 175 h operation at ±10 mA cm–2 (shown in Figure 1.5). Based on these results, the NiCo2O4 air electrode was as the best candidate for an Fe–air prototype.
41
A -2 +9 mA cm
-2 -8 mA cm
-2 +4 mA cm
-2 -11 mA cm -2 -16 mA cm
B
Figure 3.19. A) Cyclic voltammetry analysis (1st cycle) at 5 mV s–1 for the three sample catalysts: 1) LCMO + Ni–
Fe LDH, 2) LCMO + NiCo2O4 and 3) NiCo2O4. The maximum current densities for the OER and ORR are highlighted. B) Charge and discharge cycle life performance of the three samples at ±10 mA cm–2 with 2 h per step. Figure reproduced from Paper V.
3.5.3. Electrochemical performance of the Fe–air prototype Electrochemical cell testing of the Fe–air prototype operated with assisted flow of both electrolyte and oxygen. The electrolyte flow dissipated heat and removed trapped bubbles, while oxygen gas inhibited carbonate formation from carbon dioxide in air. The cell-adapted electrodes were aligned in parallel with identical geometric surface areas of 49.2 cm2. The active mass of iron in the electrode was 1.41 g while the air electrode used 1.04 g NiCo2O4. Charge and discharge were based on the capacity of the iron electrode and the total cell had a theoretical capacity of 1355 mAhFe. For –2 –1 example, ±10 mA cm corresponded to a current rate of 349 mA g Fe or C/2.75 based on iron. The reactions taking place in the cell during charge involved Fe(OH)2 reduction to iron and hydroxide ions oxidized to oxygen and water (OER). Upon discharge, the opposite reactions followed. Running the cell incorporated a preconditioning step of both electrodes at ±4 mA cm–2 until the discharge capacity stabilized. Paper V includes the preconditioning data. During this step, the cell was overcharged to the theoretical capacity of iron. After 300 h the discharge capacity stabilized and –1 output 450 mAh g Fe with an energy efficiency of 51%. The low energy efficiency achieved is attributed to the sluggish OER and ORR kinetics. After preconditioning, the Fe–air rate capability testing, shown in Figure 3.20, used charge and discharge rates between ±5 to ±25 mA cm–2. In a), the discharge profiles show that specific capacities
42 of 510 to 333 mAh g–1 were obtained, depending on the varying rate used. Discharge potentials decreased with increased current densities, induced by a significant increase of ohmic and polarization losses. The limited kinetics of the cell reaction resulted in impractical cell potentials at current densities above ±15 mA cm–2 (below 0.6 V). Figure 3.20b summarizes the discharge –1 capacities and calculates the energy densities. The highest energy density of 377 Wh kg Fe, achieved at 10 mA cm–2, represented a deviation from the expected trend of decreasing with increasing current densities. The charge efficiency evolution of iron electrodes, seen in Figure 3.15, showed the highest charge acceptance at this point (10 mA cm–2) and might explain the trend here. When accounting for the total mass of the electrodes (4.22 g for the iron electrode and 3.78 g for the air –1 electrode), the energy density corresponded to 90 Wh kg electrodes. Figure 3.20c depicts the power density curves based on the geometric surface area and mass of iron. The highest power output –2 –1 reached 11.3 mW cm / 555 mW g Fe.
A
B
C
Figure 3.20. A) Discharge performance of the Fe–air prototype at 5–25 mA cm–2. B) Discharge capacities and calculated energy densities based on the nominal voltages of respective current densities. C) Power density curve of the prototype cell. Figure reproduced from Paper V.
43
The investigated durability of the Fe–air cell upon cycling at ±10 mA cm–2, limited the charge capacity –1 to 440 mAh g Fe and set the discharge cutoff at 0.58 V. Figure 3.21a shows the charge and discharge curves over 76 cycles with an average discharge potential of 0.73 V, discharge capacity of 355 mAh –1 g Fe, and energy efficiency of 47%. The charge and discharge characteristics of the cell followed the electron-exchange character of the iron electrode, with the OER/ORR plateaus remaining flat as shown in Figure 1.5. The cell cycled over 200 h and exhibited stable performance until failure. In the last charge cycles, another plateau evolved at 1.8 V, corresponding to the HER. The increased hydrogen evolution rate indicated poor charge acceptance of iron caused by passivation. Figure 3.21b summarizes the measured nominal cell discharge voltages, specific discharge capacities, and –1 calculated energy densities. The cell averaged an energy density of 259 Wh kg Fe and Coulombic efficiency of 79%. This compares to other reported work on Fe–air cells, which have achieved 290 –1 85 Wh kg Fe . Figure 3.21c shows the ohmic resistance (IR) over all cycles, with stable values around 0.37 Ω. This corresponds to a resistivity of 0.68 Ω m from a geometric surface area of 49.2 cm2 and interlayer spacing between the electrodes of 0.27 cm.
A
B
C
Figure 3.21. A) Charge and discharge profiles of Fe–air prototype cell over 76 cycles at ±10 mA cm–2 and with oxygen flow. B) Measured and calculated cell discharge potentials, specific discharge capacities, and energy densities over 76 cycles. C) Measured Coulombic efficiency and IR drop over 76 cycles. Figure reproduced from Paper V.
44
In summary, new bifunctional catalysts for the air electrode were characterized structurally and electrochemically. The findings presented an excellent catalyst candidate in the non-precious-metal phase NiCo2O4. This catalyst showed superb activity for oxygen evolution and reduction with excellent long-term stability. The catalyst was used in a rechargeable alkaline Fe–air battery as a demonstration of this technology, and its effect on performance and cycle life was presented. The results proved a robust bifunctional performance of the Fe–air cell and demonstrated the obstacles and opportunities of Fe–air batteries. The main limitations of poor rate capability and low energy and Coulombic efficiencies are hard to overcome. The outstanding advantage of this battery is the low cost of the materials it uses. Further development of this prototype should use lightweight materials, shorten the interlayering spacing between the electrodes, and eliminate the pump to flow electrolyte. Upon these improvements, the cell needs to show minimal unwanted side reactions such as the HER, which hurts the cell efficiency.
45
4. Conclusions
This thesis undertook to outline the limitations of rechargeable aqueous Zn–MnO2 and Fe–air batteries and address these issues using new and modified electrode and electrolyte materials. The impact of LiOH addition to alkaline electrolytes for Zn–MnO2 cells depend on the concentration of the KOH:LiOH mixtures due to the competing reaction mechanisms of proton and lithium insertion. When lithium insertion or lithium–proton insertion dominated, the reaction mechanism inhibited zinc poisoning i.e. by forming ZnMn2O4. This allowed the cell to cycle up to 60 cycles with minimal capacity fade, and average around 250 mAh g–1. The optimal electrolyte composition consisted of 1 M KOH and 3 M LiOH and enabled both lithium and proton insertion. This electrolyte composition, demonstrated to be optimal using operando EDXRD and ex situ XRD, suppressed the intermediate Mn3+ phases from proton insertion in favor of lithiated Mn spinel phases, and upheld reversibility.
Aqueous sulfate Zn–MnO2 cells offered better reversibility and performance than their alkaline equivalents. Cells with Mn2+ added to the zinc sulfate electrolyte cycled over 100 times and had stable capacities of 220 mAh g–1 at 1.35 V. The findings showed that accessible Mn2+ in the electrolyte limited Mn dissolution. This limited Mn vacancies for Zn2+ insertion, as demonstrated by EDS, which showed a lower Zn/Mn ratio than the cell cycled without Mn2+ in the electrolyte. Furthermore, operando XRD showed smaller unit cell volume expansion upon zinc insertion when Mn2+ was available. These observations correlated to the enhanced cell rechargeability. Operando phase evolution analysis revealed that layered Zn-buserite was formed during the first recharge. This explained the transition in electrochemical redox reaction characteristics from the first cycle to subsequent cycles. Zinc sulfate hydroxide precipitated during discharge due to increased pH upon proton insertion but did not involve the exchange of electrons. This precipitate showed significant coverage on the electrode surface after 100 cycles but still managed to dissolve easily during charge. Rechargeable iron electrodes with stannate added to the electrolyte presented stable performance over 150 cycles with 85 ± 5% in Coulombic efficiencies. The additive prolonged the first reduction plateau and increased this capacity by 100 mAh g–1. Post-mortem SEM and XPS analysis confirmed that tin was deposited on the surface of the iron electrode, while operando neutron diffraction analysis did not. Instead, operando phase analysis unveiled a suggested intermediate hydride phase composed of a solid solution of the participating elements. The joint effect of iron and the new phase both increased the hydrogen evolution overpotential and prolonged the discharge state for practical batteries. Quantification of hydrogen evolution on rechargeable iron electrodes showed three distinct rate domains for hydrogen evolution during charge. Low hydrogen rates developed during charge until the cell overcharged, which significantly increased the hydrogen evolution rate. The higher rates of hydrogen evolution deteriorated the electrodes and capacity declined over cycles. The highest charge acceptance of iron occurred at charge rates higher than 192 mA g–1. This finding should promote suitable cycling protocols for commercial iron electrodes. On zinc electrodes in aqueous sulfate electrolytes, hydrogen gas was not detected, Coulombic efficiencies above 99% were achieved for rates of 1–100 mA cm–2. The optimized charge protocol for iron electrodes was used to demonstrate a proof-of-concept flow- assisted Fe–air cell. The cell performed with long cycle life and output an energy density of 377 Wh –1 kg Fe. The cell operated over 588 h with oxygen flow while preventing the HER and Fe3O4 passivation on the iron electrode. The findings presented NiCo2O4 as a highly active bifunctional catalyst with total overpotential (OER/ORR) of 693 mV at ±10 mA cm–2. This performance is encouraging, and if
46 the weight of the cell can be lowered in future, it may increase the energy density while refining the cumbersome limitations of low efficiencies. This system shows potential for large-scale energy storage solutions because of the materials used.
These results are a stepping-stone toward the use of rechargeable Zn–MnO2 and Fe–air batteries for large-scale energy storage as they motivate low-cost and safe operation. These batteries show compelling potential for sustainable solutions in the electric grid.
47
5. Future perspectives
Rising concerns on the dependence of Li-ion batteries limited cobalt supply encourage research on alternative batteries using available materials. Current studies focused on improving the performance and cycle life of rechargeable aqueous Zn–MnO2 and Fe–air batteries, both of which are environmentally friendly and not resource restricted.
Improvements in the cyclability of alkaline Zn–MnO2 batteries were made by promoting the lithium insertion mechanism on the manganese dioxide electrode. This formed lithiated Mn spinels instead of the known irreversible phase, ZnMn2O4. Investigations to further prolong the cycle life of these batteries are necessary. One possibility is to limit the DOD of the Zn–MnO2 battery. Limiting the DOD to the one-electron exchange of manganese dioxide (Mn4+/Mn3+) may inhibit further reduction to the inactive spinels. In addition to validating the electrochemical reversibility, this would need accurate structural analysis to confirm that the desired redox reaction of the manganese dioxide electrode occurs. The downside with limiting the DOD is clearly that the battery would not fully utilize the capacity potential of the manganese dioxide electrodes. Thus, replacing alkaline electrolytes with aqueous sulfate ones is exceptionally promising for reversible and high- performance Zn–MnO2 batteries. These electrolytes, in which the redox reactions occur at lower pH, overcome the main obstacles of alkaline ones. However, actualizing aqueous sulfate Zn–MnO2 batteries for commercial use will require a transitioning to practical battery conditions, in order to scale up. This entails not using flooded cells, using low active mass loadings of the manganese dioxide and carbon paper as current collector. Focus on the manganese dioxide electrode, which is significantly less energy-dense than the zinc electrode, is essential to reach high energy densities in this battery. Moreover, morphological investigations of zinc plating are recommended to minimize dendrite formation. Overcoming the hurdles associated with Fe–air batteries is challenging considering the cumbersome nature of the redox reactions involved. The HER on the iron electrode and the large overpotentials of the air electrode impair the cell efficiency. Further development to suppress the HER on the iron electrode is required to reach higher Coulombic efficiencies. Reported efficiencies around 90% are not enough for practical batteries; research must aim for efficiencies above 99%. This may be possible with more efficient hydrogen-evolution suppressors than tin or sulfide. Another possibility is to deposit a few monolayers of tin or sulfide compounds on each individual iron particle to efficiently increase the hydrogen overpotential. Advanced engineering of active bifunctional catalysts that lower the OER/ORR overpotentials of the air electrode is vital to enhance efficient operation of the Fe–air battery. An interesting approach is to directly synthesize active catalysts on conductive metal backbones without carbon additives. This would minimize degradation of the air electrode performance due to carbon corrosion. Further opportunities should incorporate lightweight cell materials, optimize the cell design, and eliminate the use of a pump for electrolyte flow, in order to enhance the energy density of the cell.
Conclusively, the suggested studies would be of interest for rechargeable Fe–air and Zn–MnO2 batteries.
48
6. Sammanfattning
Globala energisystem övergår alltmer ifrån fossila bränslen till förnybara energikällor för att upprätthålla en hållbar lösning för människans framtida energibehov. Incitamentet är att ersätta stora traditionella kraftverk med mindre förnybara energikällor. Paradigmskiftet till förnybar energi behöver emellertid hantera den ökande globala energikonsumtionen som uppskattas att fördubblas år 2050. Att fullända övergången till förnybara energikällor som vindkraft, solkraft, vattenkraft och geotermisk energi är en komplex uppgift. Deras fluktuerande karaktär är beroende av årstid och tid på dygnet, som i sig resulterar i oförutsägbara energigenerationsprofiler. Därför behövs effektiva energilagringssystem för att hantera dessa fluktuationer och säkerställa tillförlitlig energiförsörjning vid behov. Inom energilagringteknologier är batterier den snabbast växande teknologin på grund av dess mångsidighet av att möta uppsatta energi- och effektkrav samt dess enkla förmåga att skala upp energilagringskapacitet från kWh till MWh. Dessutom är batterier inte beroende av tid, väder eller geografiska förutsättningar vilket möjliggör snabb installation och expansion. För storskalig implementering av batterier krävs energieffektiva teknologier som är miljövänliga, baserade på lättillgängliga råvarumaterial, har låga installationskostnader samt lång livslängd. Dagens dominerade batteriteknologier, exempelvis litiumjon och blysyra, uppfyller inte alla dessa krav och därför motiveras forskning av alternativa batterilösningar. Batterier består av elektrokemiska celler som innehåller två elektroder där de elektrokemiskareaktionerna sker. En elektrolyt separerar elektroderna och innehåller joner som vandrar fritt från en elektrod till den andra. Elektroderna ansluts externt för att påbörja sina reaktioner medan elektroner överförs i en yttre krets som genererar ström. Dessa elektrokemiska celler kan sedan kopplas i serie eller parallellt för att uppnå den önskade spänningen respektive energikapaciteten. Viktiga egenskaper hos batterier är energitäthet (Wh L–1), specifik energi (Wh kg– 1) och specifik effekt (W kg–1), dvs mängden energi de kan lagra per massa eller volym samt tidslängden av energitillförseln. I storskalig energilagring är de primära faktorerna inte energitäthet eller effekt, utan istället låga installationskostnader, lång livslängd, hög energieffektivitet och enkelheten med att skala upp lagringskapaciteten. Batterier som använder vattenbaserade elektrolyter möjliggör låga tillverkningskostnader och är i sig säkrare än litiumjonbatterier. Zink-, mangandioxid-, järn- och luftbaseradebatterier har hög energirelevans, de är inte resursbegränsade och kan bidra till storskaliga energilagringslösningar. Zink har länge använts som elektrodmaterial för primära alkaliska zink-mangandioxidbatterier. Historiskt har zinkelektroder begränsats för uppladdningsbara batterier på grund av morfologiska osäkerheter och passiveringseffekter. Mangandioxidelektroder är ineffektiva som uppladdningsbara elektroder på grund av felmekanismer associerade med fastransformationer under battericykling. Irreversibiliteten av mangandioxid är starkt korrelerad med bildningen av de elektrokemiskt inaktiva spinellfaserna, Mn3O4/ZnMn2O4. Utvecklingen av järnelektroden för järnluftbatterier initierades redan i slutet på 1960-talet men begränsas än idag av undermålig laddningseffektivitet. Dessutom är luftelektrodens livslängd begränsad vid långvarig drift på grund av den långsamma syrgasutvecklings- och reduktionskinetiken. Dessa begränsningar av järnluftbatterier medför stora energiförluster vid battericykling. I detta avhandlingsarbete motverkades begränsningarna för uppladdningsbara zink-mangandioxid- och järnluftbatterier genom att syntetisera elektrodmaterial och modifiera elektrolytkompositioner. Olika katjonelektrolytblandningar av kaliumhydroxid och litiumhydroxid för uppladdningsbara alkaliska zink-mangandioxidbatterier medförde konkurrerande proton- och
49 litiuminsättningsmekanismer. Med den optimala elektrolytkompositionen av 1 molar kaliumhydroxid och 3 molar litiumhydroxid, visade elektromikroskopologiska och röntgenkristallografiska resultaten att den inaktiva fasen ZnMn2O4 undertrycktes till förmån för litiuminnehållande manganspinellfaser, som i sig förbättrade livslängden av uppladdningsbara mangandioxidelektroder. I det analoga batteriet med samma elektrodmaterial men istället med ett milt surt zinksulfat elektrolytlösning undersöktes. Med hjälp av operando röntgendiffraktionsmätningar klargjordes både den komplexa reaktionsmekanismen och varför tillsättningen av Mn2+ joner i elektrolyten förbättrade uppladdningsförmågan. Den föreslagna reaktionsmekanismmodellen föreslog insättning av både protoner och zinkjoner i mangandioxiden. Vid urladdning detekterades en zinksulfathydroxid fällning, utlöst av en lokal pH-ökning. Denna fällning involverade inga elektronöverföringar och visade sig ha en signifikant påverkan på den elektrokemiska prestandan. Vidare visade studien att tillgängliga Mn2+ joner vid elektrod-elektrolytskiktet begränsade mangandioxidupplösningen. Detta i sig begränsande zinkinsättning och förbättrade uppladdningsbarheten av mangandioxidelektroden. De undersökta hindren för uppladdningsbara järnluftbatterier motverkades genom att undertrycka vätegasutveckling på järnelektroden och optimering av bifunktionella syrgaskatalysatorer för luftelektroden. Minimering av vätgasutveckling utnyttjade kaliumstannat som tillsatsmedel. Bakgrunden till tillsatsen var att tennmetall elektropläterades under uppladdning av järnelektroden och agerade som en vätgasutvecklingsbarriär. Tillsatsen av kaliumstannat förlängde urladdningsprofilen av järnelektroderna och ökade överpotentialen av vätgasutvecklingen, båda önskvärda attribut för batterier. Operando neutrondiffraktionsmätningar visade en korrelation av dessa attribut till en bildad intermediär fas vid uppladdning. Vidare utvärderades vätgasutvecklingen av nanostrukturerade järnmaterial dopade med koppar och tenn, genom kopplad elektrokemi och masspektrometri. Studien visade att järnelektroden gynnades av höga uppladdningsströmmar eftersom vätgasutvecklingen reducerades. För luftelektroden syntetiserades bifunktionella katalysatorer som karakteriserades strukturellt och elektrokemiskt. Resultaten visade en utmärkt katalysatorkandidat i NiCo2O4 med hög katalysatoraktivet för både syrgasutveckling och syrgasreduktion. Katalysatorn anpassades till ett uppladdningsbart alkaliskt järnluftbatteri koncept. Konceptet omfattade assisterande flöden av både elektrolyt och syrgas. Batteriets prestanda och livslängd påvisades samt diskuterades möjligheterna och begränsningarna av järnluftsbatterier. Mikro-, makro-, nano- och atomskala studier av elektrodmaterialen genomfördes för att öka förståelsen för dessa materials växelverkan och natur. Detta inkluderade både operando och ex situ karaktärisering. Röntgen- och neutronstrålning, analytiska och elektrokemiska metoder gav insikt på förbättringar av batteriernas prestanda och livslängd. Resultaten rapporterade i detta avhandlingsarbete är en språngbräda för att realisera uppladdningsbara zink-mangandioxid- och järnluftbatterier för storskalig energilagring. Dessa batterier motiverar låg kostnad och säker drift samt visar potential för hållbara lösningar i framtidens energisystem.
50
Acknowledgements
Foremost, I would like to express my sincere gratitude to my supervisor, Prof. Dag Noréus, for supporting me throughout my PhD studies, guidance, and invaluable knowledge. I highly value your trust, our discussions, and your encouragement to strive for perfection.
I want to thank my co-supervisors, Prof. Gunnar Svensson and Dr. Jekabs Grins, for taking interest in my work, for fruitful discussions, and supporting me at times when I needed it the most.
I want to thank my former PhD fellow, Dr. Yang Shen, for the support and amazing times shared together. Further thanks to Erik Grape, Mirva Eriksson, Mohammed Naeem Iqbal, and Ghislaine Robert-Nicoud for the enjoyable atmosphere in the office.
Special thanks to my collaborators in the FAIR project, Alagar Raj Paulraj (KTH), Prof. Yohannes Kiros (KTH), Dr. Björn Skårman (Höganäs AB) and Dr. Hilmar Vidarsson (Höganäs AB), as well as Dr. Gunder Karlsson (SiteTel Sweden AB). Thank you for excellent collaborations and for the achievements we accomplished together. Thanks to the Swedish Energy Agency for supporting this work.
Thanks to my collaborators, Dr. Cheuk-Wai Tai, Dr. Henrik Svengren, Prof. Mats Johnsson, Dr. William Brant (Uppsala University), Dr. Darius Milcius (LEI) and Dr. Martynas Lelis (LEI). Your engagement and contributions are much appreciated.
I would also like to thank my former supervisor, Prof. Dan Steingart (Princeton University), former colleagues, Dr. Benjamin Hertzberg (Princeton University), Dr. Andrew Hsieh (Princeton University), Dr. Greg Davies (Princeton University), Dr. Can Erdonmez (BNL), and collaborators, Dr. An Huang (UCSD), Dr. Joon Kyo Seo (UCSD), Dr. Zhong Zhong (BNL), Prof. Mark Croft (Rutgers University), Prof. Ying Shirley Meng (UCSD).
My appreciation goes to Jordi Jacas Biendicho for preparing the neutron beam time proposal provided by the UK Science and Technology Facilities Council, as well as Dr. Ron Smith and Dr. Stephen Hull for your technical support and measurements at ISIS, UK.
Many thanks to Dr. Lars Eriksson, Dr. Kjell Jansson and Dr. Zoltán Bascik for user training and instructions on various instruments.
I want to thank Dr. Tamara Church and Prof. Xiaodong Zou for proofreading my thesis and for your helpful feedback.
I offer my appreciation to the entire administrative, technical, safety, and workshop staff at MMK: Tatiana Bulavina, Helmi Frejman, Ann Loftsjö, Daniel Emanuelsson, Camilla Berg, Elisabeth Pernbom, Rolf Eriksson, Anne Ertan, Kadir Abdul Karim, Hans-Erik Ekström, Per Jansson and Jakob Paulin.
To all my friends and colleagues at MMK (past and present), I thank you for making my time as a PhD student enjoyable: Andrew Kentaro Inge, Stef Smeets, Yunchen Wang, Laura Samperisi, Ning Yuan, Viktor Bengtsson, Alexandra Neagu, Dickson Ojwang, Bin Wang, Jingjing Zhao, German Salazar
51
Alvarez, Thomas Thersleff, Hani Nasser Abdelhamid, Fei Peng, Eleni Mitoudi-Vagourdi, Ahmed Etman, Magdalena Cichocka, István-Zoltan Jenei , Tom Willhammar, Hongyi Xu, Elina Kapaca, Jonas Ångström, Dariusz Wardecki, Paulo Barros, Fredrik Björnerbäck, Aditya Dharanipragada, Daniel Eklöf, Chris Celania, Alisa Gordeeva, Korneliya Gordeyeva, Yulia Trushkina, Valentina Guccini, Zhehao Huang, Martin Kupuscinski, Konstantin Kriechbaum, Molly Lightowler, Yi Luo, Taimin Yang, Pierre Munier, Przemyslaw Rzepka, Jon Olsén, Hugo Voisin, Varvara Apostolopoulou-Kalkavoura, Sara Skoglund, Xia Wang, Anne Willert, Shun Yu and Yuan Zhong.
Last but not least, I am grateful to my wonderful family; my parents (Marin and Nabil), brothers (Ninus and Tony) and my life partner (Eloisa) for their constant support, love, and care. They have always been by my side throughout this journey, both in happy and difficult moments. I am also grateful to my friends and relatives for the encouragement that helped me reach this moment.
52
References
1. Larcher, D. & Tarascon, J.-M. Towards greener and more sustainable batteries for electrical
energy storage. Nat. Chem. 7, 19–29 (2015).
2. Castillo, A. & Gayme, D. F. Grid-scale energy storage applications in renewable energy
integration: A survey. Energy Convers. Manage. 87, 885–894 (2014).
3. Soloveichik, G. L. Battery technologies for large-scale stationary energy storage. Annu. Rev.
Chem. Biomol. Eng. 2, 503–527 (2011).
4. Ibrahim, H. & Ilinc, A. Techno-Economic Analysis of Different Energy Storage Technologies.
Energy Storage - Technologies and Applications (ed. Zobaa, A.) (InTech, 2013).
5. Dunn, B., Kamath, H. & Tarascon, J.-M. Electrical energy storage for the grid: a battery of
choices. Science 334, 928–935 (2011).
6. Energilagring - Teknik för lagring av el. (Kungliga Ingenjörsvetenskapsakademien (IVA), 2015).
7. Goodenough, J. B. Electrochemical energy storage in a sustainable modern society. Energy
Environ. Sci. 7, 14–18 (2014).
8. Tarascon, J. M. & Armand, M. Issues and challenges facing rechargeable lithium batteries.
Nature 414, 359–367 (2001).
9. Ragone, D. V. Review of battery systems for electrically powered vehicles. (SAE Technical Paper,
1968).
10. Posada, J. O. G., Rennie, A. J. R., Villar, S. P., Martins, V. L., Marinaccio, J., Barnes, A., Glover, C.
F., Worsley, D. A. & Hall, P. J. Aqueous batteries as grid scale energy storage solutions.
Renewable Sustainable Energy Rev. 68, 1174–1182 (2017).
11. Goodenough, J. B. & Kim, Y. Challenges for Rechargeable Li Batteries. Chem. Mater. 22, 587–
603 (2010).
12. Wanger, T. C. The Lithium future—resources, recycling, and the environment. Conservation
Letters 4, 202–206 (2011).
13. Roth, E. P. & Orendorff, C. J. How electrolytes influence battery safety. Electrochem. Soc.
53
Interface, 45–49 (2012).
14. Yabuuchi, N. & Ohzuku, T. Novel lithium insertion material of LiCo1/3Ni1/3Mn1/3O2 for advanced
lithium-ion batteries. J. Power Sources 119-121, 171–174 (2003).
15. European Commission. 2017 list of Critical Raw Materials for the EU. (2017).
16. Massé, R. C., Uchaker, E. & Cao, G. Beyond Li-ion: electrode materials for sodium- and
magnesium-ion batteries. Science China Materials 58, 715–766 (2015).
17. Thackeray, M. M., Wolverton, C. & Isaacs, E. D. Electrical energy storage for transportation—
approaching the limits of, and going beyond, lithium-ion batteries. Energy Environ. Sci. 5, 7854–
7863 (2012).
18. Goodenough, J. B. & Park, K.-S. The Li-ion rechargeable battery: a perspective. J. Am. Chem. Soc.
135, 1167–1176 (2013).
19. Luntz, A. Beyond Lithium Ion Batteries. J. Phys. Chem. Lett. 6, 300–301 (2015).
20. Kundu, D., Vajargah, S. H., Wan, L., Adams, B., Prendergast, D. & Nazar, L. F. Aqueous vs.
nonaqueous Zn-ion batteries: consequences of the desolvation penalty at the interface. Energy
Environ. Sci 11. (2018).
21. White, C. D. & Zhang, K. M. Using vehicle-to-grid technology for frequency regulation and peak-
load reduction. J. Power Sources 196, 3972–3980 (2011).
22. Zhang, N., Cheng, F., Liu, J., Wang, L., Long, X., Liu, X., Li, F. & Chen, J. Rechargeable aqueous
zinc-manganese dioxide batteries with high energy and power densities. Nat. Commun. 8, 405
(2017).
23. Ingale, N. D., Gallaway, J. W., Nyce, M., Couzis, A. & Banerjee, S. Rechargeability and economic
aspects of alkaline zinc–manganese dioxide cells for electrical storage and load leveling. J.
Power Sources 276, 7–18 (2015).
24. McKerracher, R. D., Ponce de Leon, C., Wills, R. G. A., Shah, A. A. & Walsh, F. C. A Review of the
Iron–Air Secondary Battery for Energy Storage. ChemPlusChem 80, 1-14 (2014).
25. Reich, M. & Vasconcelos, P. M. Geological and Economic Significance of Supergene Metal
54
Deposits. Elements 11, 305–310 (2015).
26. McBreen, J. Zinc Electrode Shape Change in Secondary Cells. J. Electrochem. Soc. 119, 1620–
1628 (1972).
27. Falk, S. U. & Salkind, A. J. Alkaline storage batteries. (John Wiley & Sons, 1969).
28. Chakkaravarthy, C., Waheed, A. K. A. & Udupa, H. V. K. Zinc—air alkaline batteries — A review.
J. Power Sources 6, 203–228 (1981).
29. Shoji, T., Hishinuma, M. & Yamamoto, T. Zinc-manganese dioxide galvanic cell using zinc
sulphate as electrolyte. Rechargeability of the cell. J. Appl. Electrochem. 18, 521–526 (1988).
30. Kalu, E. E. & White, R. E. Zn/Br2 cell: Effects of plated zinc and complexing organic phase. AIChE
J. 37, 1164–1174 (1991).
31. Tomazic, G. Zinc-bromine battery with circulating electrolytes. J. Power Sources 70, 168–168
(1998).
32. Lim, H. S., Lackner, A. M. & Knechtli, R. C. Zinc‐Bromine Secondary Battery. J. Electrochem. Soc.
124, 1154–1157 (1977).
33. Zhang, X. G. Corrosion and electrochemistry of zinc. (Springer Science & Business Media, 2013).
34. McLarnon, F. R. & Cairns, E. J. The Secondary Alkaline Zinc Electrode. J. Electrochem. Soc. 138,
645–656 (1991).
35. Reddy, T. Linden’s Handbook of Batteries, 4th Edition. (McGraw Hill Professional, 2010).
36. Parker, J. F., Chervin, C. N., Pala, I. R., Machler, M., Burz, M. F., Long, J. W. & Rolison, D. R.
Rechargeable nickel–3D zinc batteries: An energy-dense, safer alternative to lithium-ion.
Science 356, 415–418 (2017).
37. Parker, J. F., Pala, I. R., Chervin, C. N., Long, J. W. & Rolison, D. R. Minimizing Shape Change at
Zn Sponge Anodes in Rechargeable Ni–Zn Cells: Impact of Electrolyte Formulation. J.
Electrochem. Soc. 163, A351–A355 (2016).
38. Mainar, A. R., Iruin, E., Colmenares, L. C., Kvasha, A., de Meatza, I., Bengoechea, M., Leonet, O.,
Boyano, I., Zhang, Z. & Blazquez, J. An overview of progress in electrolytes for secondary zinc-air
55
batteries and other storage systems based on zinc. Journal of Energy Storage 15, 304–328
(2018).
39. Pichler, B., Weinberger, S., Reščec, L., Grimmer, I., Gebetsroither, F., Bitschnau, B. & Hacker, V.
Bifunctional electrode performance for zinc-air flow cells with pulse charging. Electrochim. Acta
251, 488–497 (2017).
40. Li, Y., Gong, M., Liang, Y., Feng, J., Kim, J.-E., Wang, H., Hong, G., Zhang, B. & Dai, H. Advanced
zinc-air batteries based on high-performance hybrid electrocatalysts. Nat. Commun. 4, 1805
(2013).
41. Mehta, S. Investigation of capacity fade in flat-plate rechargeable alkaline MnO₂/Zn cells.
(University of British Columbia, 2016).
42. Tang, L., Wang, J.-W., Meng, L.-R. & Jin, C.-C. Influence of the EMD on the high-rate alkaline Zn-
MnO2 battery. Chinese Battery Industry 4, 006 (2012).
43. Fang, G., Zhou, J., Pan, A. & Liang, S. Recent Advances in Aqueous Zinc-Ion Batteries. ACS Energy
Lett. 3, 2480–2501 (2018).
44. Xu, C., Li, B., Du, H. & Kang, F. Energetic zinc ion chemistry: the rechargeable zinc ion battery.
Angew. Chem. Int. Ed Engl. 51, 933–935 (2012).
45. Zhang, N., Cheng, F., Liu, Y., Zhao, Q., Lei, K., Chen, C., Liu, X. & Chen, J. Cation-Deficient Spinel
ZnMn2O4 Cathode in Zn(CF3SO3)2 Electrolyte for Rechargeable Aqueous Zn-Ion Battery. J. Am.
Chem. Soc. 138, 12894–12901 (2016).
46. Gupta, T., Kim, A., Phadke, S., Biswas, S., Luong, T., Hertzberg, B. J., Chamoun, M., Evans-
Lutterodt, K. & Steingart, D. A. Improving the cycle life of a high-rate, high-potential aqueous
dual-ion battery using hyper-dendritic zinc and copper hexacyanoferrate. J. Power Sources 305,
22–29 (2016).
47. Yang, H. S., Park, J. H., Ra, H. W., Jin, C.-S. & Yang, J. H. Critical rate of electrolyte circulation for
preventing zinc dendrite formation in a zinc–bromine redox flow battery. J. Power Sources 325,
446–452 (2016).
56
48. Wu, M. C., Zhao, T. S., Jiang, H. R., Zeng, Y. K. & Ren, Y. X. High-performance zinc bromine flow
battery via improved design of electrolyte and electrode. J. Power Sources 355, 62–68 (2017).
49. Li, B., Nie, Z., Vijayakumar, M., Li, G., Liu, J., Sprenkle, V. & Wang, W. Ambipolar zinc-polyiodide
electrolyte for a high-energy density aqueous redox flow battery. Nat. Commun. 6, 6303 (2015).
50. Xie, C., Duan, Y., Xu, W., Zhang, H. & Li, X. A Low-Cost Neutral Zinc-Iron Flow Battery with High
Energy Density for Stationary Energy Storage. Angew. Chem. Int. Ed Engl. 56, 14953–14957
(2017).
51. Gong, K., Ma, X., Conforti, K. M., Kuttler, K. J., Grunewald, J. B., Yeager, K. L., Bazant, M. Z., Gu,
S. & Yan, Y. A zinc–iron redox-flow battery under $100 per kW h of system capital cost. Energy
Environ. Sci. 8, 2941–2945 (2015).
52. Davies, G., Hsieh, A. G., Hultmark, M., Mueller, M. E. & Steingart, D. A. Utilization of Hyper-
Dendritic Zinc during High Rate Discharge in Alkaline Electrolytes. J. Electrochem. Soc. 163,
A1340–A1347 (2016).
53. Bass, K., Mitchell, P. J., Wilcox, G. D. & Smith, J. Methods for the reduction of shape change and
dendritic growth in zinc-based secondary cells. J. Power Sources 35, 333–351 (1991).
54. Parker, J. F., Nelson, E. S., Wattendorf, M. D., Chervin, C. N., Long, J. W. & Rolison, D. R.
Retaining the 3D framework of zinc sponge anodes upon deep discharge in Zn-air cells. ACS
Appl. Mater. Interfaces 6, pp. 19471-19476 (2014).
55. Ko, J. S., Sassin, M. B., Parker, J. F., Rolison, D. R. & Long, J. W. Combining battery-like and
pseudocapacitive charge storage in 3D MnOx@ carbon electrode architectures for zinc-ion cells.
Sustainable Energy & Fuels 2, 626-636 (2018).
56. Chamoun, M., Hertzberg, B. J., Gupta, T., Davies, D., Bhadra, S., Van Tassell, B., Erdonmez, C. &
Steingart, D. A. Hyper-dendritic nanoporous zinc foam anodes. NPG Asia Materials 7, e178
(2015).
57. Hilder, M., Winther-Jensen, O., Winther-Jensen, B. & MacFarlane, D. R. Graphene/zinc nano-
composites by electrochemical co-deposition. Phys. Chem. Chem. Phys. 14, 14034–14040
57
(2012).
58. Parker, J. F., Chervin, C. N., Nelson, E. S., Rolison, D. R. & Long, J. W. Wiring zinc in three
dimensions re-writes battery performance—dendrite-free cycling. Energy Environ. Sci. 7, 1117–
1124 (2014).
59. McKubre, M. C. H. & Macdonald, D. D. The Dissolution and Passivation of Zinc in Concentrated
Aqueous Hydroxide. J. Electrochem. Soc. 128, 524–530 (1981).
60. Pan, H., Shao, Y., Yan, P., Cheng, Y., Han, K. S., Nie, Z., Wang, C., Yang, J., Li, X., Bhattacharya, P.,
Mueller, K. T. & Liu, J. Reversible aqueous zinc/manganese oxide energy storage from
conversion reactions. Nature Energy 1, Article number: 16039 (2016).
61. Lee, B., Seo, H. R., Lee, H. R., Yoon, C. S., Kim, J. H., Chung, K. Y., Cho, B. W. & Oh, S. H. Critical
Role of pH Evolution of Electrolyte in the Reaction Mechanism for Rechargeable Zinc Batteries.
ChemSusChem 9, 2948–2956 (2016).
62. Kim, S. H. & Oh, S. M. Degradation mechanism of layered MnO2 cathodes in Zn/ZnSO4/MnO2
rechargeable cells. J. Power Sources 72, 150–158 (1998).
63. Xu, D., Li, B., Wei, C., He, Y.-B., Du, H., Chu, X., Qin, X., Yang, Q.-H. & Kang, F. Preparation and
Characterization of MnO2/acid-treated CNT Nanocomposites for Energy Storage with Zinc Ions.
Electrochim. Acta 133, 254–261 (2014).
64. Sun, W., Wang, F., Hou, S., Yang, C., Fan, X., Ma, Z., Gao, T., Han, F., Hu, R., Zhu, M. & Wang, C.
Zn/MnO2 Battery Chemistry With H(+) and Zn(2+) Coinsertion. J. Am. Chem. Soc. 139, 9775–
9778 (2017).
65. Biswal, A., Tripathy, B. C., Sanjay, K., Subbaiah, T. & Minakshi, M. Electrolytic manganese
dioxide (EMD): a perspective on worldwide production, reserves and its role in
electrochemistry. RSC Adv. 5, 58255–58283 (2015).
66. Dose, W. M. & Donne, S. W. Optimising heat treatment environment and atmosphere of
electrolytic manganese dioxide for primary Li/MnO2 batteries. J. Power Sources 247, 852–857
(2014).
58
67. Ortiz-Vitoriano, N., Drewett, N. E., Gonzalo, E. & Rojo, T. High performance manganese-based
layered oxide cathodes: overcoming the challenges of sodium ion batteries. Energy Environ. Sci.
10, 1051–1074 (2017).
68. Kim, J.-S., Chang, W.-S., Kim, R.-H., Kim, D.-Y., Han, D.-W., Lee, K.-H., Lee, S.-S. & Doo, S.-G. High-
capacity nanostructured manganese dioxide cathode for rechargeable magnesium ion
batteries. J. Power Sources 273, 210–215 (2015).
69. Hertzberg, B. J., Huang, A., Hsieh, A., Chamoun, M., Davies, G., Seo, J. K., Zhong, Z., Croft, M.,
Erdonmez, C., Meng, Y. S. & Steingart, D. Effect of Multiple Cation Electrolyte Mixtures on
Rechargeable Zn–MnO2 Alkaline Battery. Chem. Mater. 28, 4536–4545 (2016).
70. Julien, C. M., Massot, M. & Poinsignon, C. Lattice vibrations of manganese oxides. Part I.
Periodic structures. Spectrochim. Acta A Mol. Biomol. Spectrosc. 60, 689–700 (2004).
71. Gallaway, J. W., Menard, M., Hertzberg, B., Zhong, Z., Croft, M., Sviridov, L. A., Turney, D. E.,
Banerjee, S., Steingart, D. A. & Erdonmez, C. K. Hetaerolite Profiles in Alkaline Batteries
Measured by High Energy EDXRD. J. Electrochem. Soc. 162, A162–A168 (2015).
72. Chamoun, M., Brant, W. R., Tai, C.-W., Karlsson, G. & Noréus, D. Rechargeability of aqueous
2+ sulfate Zn/MnO2 batteries enhanced by accessible Mn ions. Energy Storage Materials 15, 351–
360 (2018).
73. Gallaway, J. W., Hertzberg, B. J., Zhong, Z., Croft, M., Turney, D. E., Yadav, G. G., Steingart, D. A.,
Erdonmez, C. K. & Banerjee, S. Operando identification of the point of [Mn2]O4 spinel formation
during γ-MnO2 discharge within batteries. J. Power Sources 321, 135–142 (2016).
74. Boden, D., Venuto, C. J., Wisler, D. & Wylie, R. B. The Alkaline Manganese Dioxide Electrode: I .
The Discharge Process. J. Electrochem. Soc. 114, 415–417 (1967).
75. Devaraj, S. & Munichandraiah, N. Effect of Crystallographic Structure of MnO2 on Its
Electrochemical Capacitance Properties. J. Phys. Chem. C 112, 4406–4417 (2008).
76. McBreen, J. The electrochemistry of β-MnO2 and γ-MnO2 in alkaline electrolyte. Electrochim.
Acta 20, 221–225 (1975).
59
77. Maskell, W. C., Shaw, J. E. A. & Tye, F. L. Manganese dioxide electrode—IV. Chemical and
electrochemical reduction of an electrolytic γ-MnO2. Electrochim. Acta 26, 1403–1410 (1981).
78. Mondoloni, C., Laborde, M., Rioux, J., Andoni, E. & Lévy‐Clément, C. Rechargeable Alkaline
Manganese Dioxide Batteries: I . In Situ X‐Ray Diffraction Investigation of the (EMD‐Type)
Insertion System. J. Electrochem. Soc. 139, 954–959 (1992).
79. Zhao, S., Han, B., Zhang, D., Huang, Q., Xiao, L., Chen, L., Ivey, D. G., Deng, Y. & Wei, W.
Unravelling the reaction chemistry and degradation mechanism in aqueous Zn/MnO2
rechargeable batteries. J. Mater. Chem. A Mater. Energy Sustain. 6, 5733–5739 (2018).
80. Huang, J., Wang, Z., Hou, M., Dong, X., Liu, Y., Wang, Y. & Xia, Y. Polyaniline-intercalated
manganese dioxide nanolayers as a high-performance cathode material for an aqueous zinc-ion
battery. Nat. Commun. 9, 2906 (2018).
81. Wagner, O. C. Secondary iron-air batteries. (1968).
82. Öjefors, L. Self-discharge of the alkaline iron electrode. Electrochim. Acta 21, 263–266 (1976).
83. Öjefors, L. & Carlsson, L. An iron—air vehicle battery. J. Power Sources 2, 287–296 (1978).
84. Paulraj, A. R., Kiros, Y., Skårman, B. & Vidarsson, H. Core/Shell Structure Nano-Iron/Iron Carbide
Electrodes for Rechargeable Alkaline Iron Batteries. J. Electrochem. Soc. 164, A1665–A1672
(2017).
85. Figueredo-Rodríguez, H. A., McKerracher, R. D., Insausti, M., Luis, A. G., de Leόn, C. P., Alegre,
C., Baglio, V., Aricò, A. S. & Walsh, F. C. A Rechargeable, Aqueous Iron Air Battery with
Nanostructured Electrodes Capable of High Energy Density Operation. J. Electrochem. Soc. 164,
A1148–A1157 (2017).
86. Hang, B. T., Eashira, M., Watanabe, I., Okada, S., Yamaki, J.-I., Yoon, S.-H. & Mochida, I. The
effect of carbon species on the properties of Fe/C composite for metal–air battery anode. J.
Power Sources 143, 256–264 (2005).
87. Hang, B. T., Watanabe, T., Egashira, M., Watanabe, I., Okada, S. & Yamaki, J.-I. The effect of
additives on the electrochemical properties of Fe/C composite for Fe/air battery anode. J.
60
Power Sources 155, 461–469 (2006).
88. Sundar Rajan, A., Ravikumar, M. K., Priolkar, K. R., Sampath, S. & Shukla, A. K. Carbonyl-Iron
Electrodes for Rechargeable-Iron Batteries. Electrochemical Energy Technology 1, (2014).
89. Manohar, A. K., Malkhandi, S., Yang, B., Yang, C., Surya Prakash, G. K. & Narayanan, S. R. A High-
Performance Rechargeable Iron Electrode for Large-Scale Battery-Based Energy Storage. J.
Electrochem. Soc. 159, A1209–A1214 (2012).
90. Manohar, A. K., Yang, C. & Narayanan, S. R. The Role of Sulfide Additives in Achieving Long Cycle
Life Rechargeable Iron Electrodes in Alkaline Batteries. J. Electrochem. Soc. 162, A1864–A1872
(2015).
91. Yang, B., Malkhandi, S., Manohar, A. K., Prakash, G. & Narayanan, S. R. Organo-sulfur molecules
enable iron-based battery electrodes to meet the challenges of large-scale electrical energy
storage. Energy Environ. Sci. 7, 2753-2763 (2014).
92. Li, Y. & Dai, H. Recent advances in zinc–air batteries. Chem. Soc. Rev. 43, 5257–5275 (2014).
93. Wang, Z.-L., Xu, D., Xu, J.-J. & Zhang, X.-B. Oxygen electrocatalysts in metal–air batteries: from
aqueous to nonaqueous electrolytes. Chem. Soc. Rev. 43, 7746–7786 (2014).
94. Kinoshita, K. Electrochemical oxygen technology. 30, (John Wiley & Sons, 1992).
95. Jörissen, L. Bifunctional oxygen/air electrodes. J. Power Sources 155, 23–32 (2006).
96. Christensen, P. A., Hamnett, A. & Linares-Moya, D. Oxygen reduction and fuel oxidation in
alkaline solution. Phys. Chem. Chem. Phys. 13, 5206–5214 (2011).
97. Suntivich, J., Gasteiger, H. A., Yabuuchi, N., Nakanishi, H., Goodenough, J. B. & Shao-Horn, Y.
Design principles for oxygen-reduction activity on perovskite oxide catalysts for fuel cells and
metal–air batteries. Nat. Chem. 3, 546–550 (2011).
98. Vielstich, W., Gasteiger, H. A. & Yokokawa, H. Handbook of Fuel Cells: Fundamentals Technology
and Applications: Advances in Electrocatalysis, Materials, Diagnostics and Durability. 5, (John
Wiley & Sons, 2009).
99. Mattioli, G., Giannozzi, P., Amore Bonapasta, A. & Guidoni, L. Reaction pathways for oxygen
61
evolution promoted by cobalt catalyst. J. Am. Chem. Soc. 135, 15353–15363 (2013).
100. Hamdani, M., Singh, R. N. & Chartier, P. Co3O4 and Co-based spinel oxides bifunctional oxygen
electrodes. Int. J. Electrochem. Sci. 5, 556–577 (2010).
101. Hu, L., Wu, L., Liao, M., Hu, X. & Fang, X. Electrical Transport Properties of Large, Individual
NiCo2O4 Nanoplates. Adv. Funct. Mater. 22, 998–1004 (2012).
102. Zhao, Q., Yan, Z., Chen, C. & Chen, J. Spinels: Controlled Preparation, Oxygen
Reduction/Evolution Reaction Application, and Beyond. Chem. Rev. 117, 10121–10211 (2017).
103. Yuan, C., Wu, H. B., Xie, Y. & Lou, X. W. Mixed transition-metal oxides: design, synthesis, and
energy-related applications. Angew. Chem. Int. Ed. 53, 1488–1504 (2014).
104. Karlsson, G. Perovskite catalysts for air electrodes. Electrochim. Acta 30, 1555–1561 (1985).
105. Suntivich, J., May, K. J., Gasteiger, H. A., Goodenough, J. B. & Shao-Horn, Y. A perovskite oxide
optimized for oxygen evolution catalysis from molecular orbital principles. Science 334, 1383–
1385 (2011).
106. Hamann, C. H., Hamnett, A. & Vielstich, W. Electrochemistry. 2nd. Completely Revised and
Updated Edition, New York (2007).
107. Chen, J., Lim, B., Lee, E. P. & Xia, Y. Shape-controlled synthesis of platinum nanocrystals for
catalytic and electrocatalytic applications. Nano Today 4, 81–95 (2009).
108. Lee, Y., Suntivich, J., May, K. J., Perry, E. E. & Shao-Horn, Y. Synthesis and Activities of Rutile IrO2
and RuO2 Nanoparticles for Oxygen Evolution in Acid and Alkaline Solutions. J. Phys. Chem. Lett.
3, 399–404 (2012).
109. McKerracher, R. D., Alegre, C., Baglio, V., Aricò, A. S., Ponce de León, C., Mornaghini, F., Rodlert,
M. & Walsh, F. C. A nanostructured bifunctional Pd/C gas-diffusion electrode for metal-air
batteries. Electrochim. Acta 174, 508–515 (2015).
110. Cheng, F. & Chen, J. Metal–air batteries: from oxygen reduction electrochemistry to cathode
catalysts. Chem. Soc. Rev. 41, 2172–2192 (2012).
111. Chamoun, M., Skårman, B., Vidarsson, H., Smith, R. I., Hull, S., Lelis, M., Milcius, D. & Noréus, D.
62
Stannate Increases Hydrogen Evolution Overpotential on Rechargeable Alkaline Iron Electrodes.
J. Electrochem. Soc. 164, A1251–A1257 (2017).
112. Svengren, H., Chamoun, M. & Grins, J. Water Splitting Catalysis Studied by using Real‐Time
Faradaic Efficiency Obtained through Coupled Electrolysis and Mass Spectrometry. (2018).
113. Biendicho, J. J., Roberts, M., Offer, C., Noréus, D., Widenkvist, E., Smith, R. I., Svensson, G.,
Edström, K., Norberg, S. T., Eriksson, S. G. & Hull, S. New in-situ neutron diffraction cell for
electrode materials. J. Power Sources 248, 900–904 (2014).
114. Hull, S., Smith, R. I., David, W. I. F., Hannon, A. C., Mayers, J. & Cywinski, R. The Polaris powder
diffractometer at ISIS. Physica B Condens. Matter 180, 1000–1002 (1992).
115. Larson, A. C. & Von Dreele, R. B. Gsas. General Structure Analysis System. LANSCE, MS-H805, Los
Alamos, New Mexico (1994).
116. Blidberg, A., Gustafsson, T., Tengstedt, C., Björefors, F. & Brant, W. R. Monitoring LixFeSO4F (x=
1, 0.5, 0) Phase Distributions In Operando to Determine Reaction Homogeneity in Porous
Battery Electrodes. Chemistry of Materials (2017).
117. Leapman, R. D., Grunes, L. A. & Fejes, P. L. Study of the L2,3 edges in the 3d transition metals and
their oxides by electron-energy-loss spectroscopy with comparisons to theory. Phys. Rev. B
Condens. Matter 26, 614–635 (1982).
118. van Aken, P. A. & Liebscher, B. Quantification of ferrous/ferric ratios in minerals: new
evaluation schemes of Fe L2,3 electron energy-loss near-edge spectra. Phys. Chem. Miner. 29,
188–200 (2002).
119. Mayence, A., Wéry, M., Tran, D. T., Wetterskog, E., Svedlindh, P., Tai, C.-W. & Bergström, L.
Interfacial strain and defects in asymmetric Fe–Mn oxide hybrid nanoparticles. Nanoscale 8,
14171–14177 (2016).
120. Ibrahem, I., Iqbal, M. N., Verho, O., Eivazihollagh, A., Olsén, P., Edlund, H., Tai, C.-W., Norgren,
M. & Johnston, E. V. Copper Nanoparticles on Controlled Pore Glass and TEMPO for the Aerobic
Oxidation of Alcohols. ChemNanoMat 1, 1–6 (2017).
63
121. Shen, Y. & Kordesch, K. The mechanism of capacity fade of rechargeable alkaline manganese
dioxide zinc cells. J. Power Sources 87, 162–166 (2000).
122. Im, D., Manthiram, A. & Coffey, B. Manganese(III) Chemistry in KOH Solutions in the Presence of
Bi-or Ba-Containing Compounds and Its Implications on the Rechargeability of γ-MnO2 in
Alkaline Cells. J. Electrochem. Soc. 150, A1651–A1659 (2003).
123. Musil, M., Choi, B. & Tsutsumi, A. Morphology and Electrochemical Properties of α-, β-, γ-, and
δ-MnO2 Synthesized by Redox Method. J. Electrochem. Soc. 162, A2058–A2065 (2015).
124. Garvie, L. A. J., Craven, A. J. & Brydson, R. Use of electron-energy loss near-edge fine structure
in the study of minerals. Am. Mineral. 79, 411–425 (1994).
125. Schmid, H. K. & Mader, W. Oxidation states of Mn and Fe in various compound oxide systems.
Micron 37, 426–432 (2006).
126. Laffont, L. & Gibot, P. High resolution electron energy loss spectroscopy of manganese oxides:
Application to Mn3O4 nanoparticles. Mater. Charact. 61, 1268–1273 (2010).
127. Tan, H., Verbeeck, J., Abakumov, A. & Van Tendeloo, G. Oxidation state and chemical shift
investigation in transition metal oxides by EELS. Ultramicroscopy 116, 24–33 (2012).
128. Lee, B., Lee, H. R., Kim, H., Chung, K. Y., Cho, B. W. & Oh, S. H. Elucidating the intercalation
mechanism of zinc ions into α-MnO2 for rechargeable zinc batteries. Chem. Commun. 51, 9265–
9268 (2015).
129. Massé, R. C. & Gerken, J. B. Assembly of a Robust and Economical MnO2-Based Reference
Electrode. J. Chem. Educ. 92, 110–115 (2015).
130. Rajan, A. S., Sampath, S. & Shukla, A. K. An in situ carbon-grafted alkaline iron electrode for
iron-based accumulators. Energy Environ. Sci. 7, 1110–1116 (2013).
131. Soler, L., Candela, A. M., Macanás, J., Muñoz, M. & Casado, J. Hydrogen generation from water
and aluminum promoted by sodium stannate. Int. J. Hydrogen Energy 35, 1038–1048 (2010).
132. Biendicho, J. J., Roberts, M., Noréus, D., Lagerqvist, U., Smith, R. I., Svensson, G., Norberg, S. T.,
Eriksson, S. G. & Hull, S. In situ investigation of commercial Ni (OH)2 and LaNi5-based electrodes
64
by neutron powder diffraction. J. Mater. Res. 30, 407–416 (2015).
133. Liu, M. & Liu, L.-G. Compressions and phase transitions of tin to half a megabar. High
Temperatures. High Pressures 18, 79–85 (1986).
134. Dobersek, M., Kosovinc, I. & Schubert, K. Metallographie und Konstitution der Fe-Ecke des
Systems Fe-C-Sn. Archiv für das Eisenhüttenwesen 263–266 (1984).
135. Degtyareva, V. F., Degtyareva, O. & Allan, D. R. Ordered Si-VI-type crystal structure in BiSn alloy
under high pressure. Phys. Rev. B Condens. Matter 67, 212105 (2003).
136. Savidan, J.-C., Joubert, J.-M. & Toffolon-Masclet, C. An experimental study of the Fe--Sn--Zr
ternary system at 900° C. Intermetallics 18, 2224–2228 (2010).
137. Lei, D., Lee, D.-C., Magasinski, A., Zhao, E., Steingart, D. A. & Yushin, G. Performance
Enhancement and Side Reactions in Rechargeable Nickel-Iron Batteries with Nanostructured
Electrodes. ACS Appl. Mater. Interfaces 8, 2088-2096 (2016).
138. Paulraj, A. R., Kiros, Y., Chamoun, M., Svengren, H., Noréus, D., Göthelid, M., Skårman, B.,
Vidarsson, H. & Johansson, M. Electrochemical Performance and in Operando Charge Efficiency
Measurements of Cu/Sn-Doped Nano Iron Electrodes, Batteries 5, 1-15 (2019).
139. Taguchi, H. Relationship between Crystal Structure and Electrical Properties of the Ca-Rich
Region in (La1- xCax) MnO2. 97. J. Solid State Chem. 124, 360–365 (1996).
140. Zhu, J. L., Yu, R. C., Li, F. Y., Jin, C. Q. & Zhang, Z. The structure and properties of the manganate
with nominal composition La1.0Ca2.0Mn2O7. Mater. Sci. Eng. B 95, 19–23 (2002).
141. Tang, D., Liu, J., Wu, X., Liu, R., Han, X., Han, Y., Huang, H., Liu, Y. & Kang, Z. Carbon quantum
dot/NiFe layered double-hydroxide composite as a highly efficient electrocatalyst for water
oxidation. ACS Appl. Mater. Interfaces 6, 7918–7925 (2014).
142. Zhang, K., Wang, W., Kuai, L. & Geng, B. A facile and efficient strategy to gram-scale preparation
of composition-controllable Ni-Fe LDHs nanosheets for superior OER catalysis. Electrochim. Acta
225, 303–309 (2017).
143. Lapham, D. P. & Tseung, A. C. C. The effect of firing temperature, preparation technique and
65
composition on the electrical properties of the nickel cobalt oxide series NixCo1 − xOy. J. Mater.
Sci. 39, 251–264 (2004).
144. Paulraj, A. R. & Kiros, Y. La0.1Ca0.9MnO3/Co3O4 for oxygen reduction and evolution reactions
(ORER) in alkaline electrolyte. J. Solid State Electrochem. 22, 1697–1710 (2018).
145. Flegler, A., Hartmann, S., Settelein, J., Mandel, K. & Sextl, G. Screen printed bifunctional gas
diffusion electrodes for aqueous metal-air batteries: Combining the best of the catalyst and
binder world. Electrochim. Acta 258, 495–503 (2017).
146. Price, S. W. T., Thompson, S. J., Li, X., Gorman, S. F., Pletcher, D., Russell, A. E., Walsh, F. C. &
Wills, R. G. A. The fabrication of a bifunctional oxygen electrode without carbon components
for alkaline secondary batteries. J. Power Sources 259, 43–49 (2014).
147. Paulraj, A., Kiros, Y., Göthelid, M. & Johansson, M. NiFeOx as a Bifunctional Electrocatalyst for
Oxygen Reduction (OR) and Evolution (OE) Reaction in Alkaline Media. Catalysts 8, 328 (2018).
66