Alkali ion batteries (AIBs): Operating mechanism and key parameters
Batteries are electrochemical cells that store electrical energy through thermodynamically unfavorable chemical reaction, in which the energy could be spontaneously extracted when connected to an external load.1, 2 Alkali ion batteries (AIBs) belong to a category called “secondary batteries” and rely on the reversible conversion of electrical and chemical energy. A single cell comprises of 4 key components, namely the cathode, anode, electrolyte and separator.
The cathode and anode materials are typically host materials that can accommodate alkali ions through electrochemical redox reactions. They are the most crucial components that determine the overall performance (i.e. energy density, power density, cyclability and etc.) of the battery. The active cathode and anode materials are often intermixed with carbon additives (e.g. carbon black, Super P) and binder solution (e.g. PVDF in NMP) for a prolonged period to obtain coalescence slurry before coating onto battery-grade aluminum and copper foils respectively. The electrolyte is electrically insulating but provides an ionic conductive medium for Li or Na ions to shuttle between the cathode and anode during charge/discharge processes. Carbonate-based organic compounds (e.g. PC, EC, DEC) are the most commonly utilized electrolyte solvent, while fluorophosphate (e.g. LiPF6, NaPF6) and perchlorates (e.g. LiClO4, NaClO4) are the most widely used form of alkali salts in both LIBs and NIBs. The separator is a microporous polymeric membrane that is wetted with the electrolyte before use. The key function of the separator helps prevent a direct short between the cathode and anode while allowing alkali ions to permeate and shuttle between electrodes.
During the charging process in a full cell, alkali ions de-intercalate from the cathode and shuttle through the electrolyte before electrochemically reducing the anode. This process is thermodynamically unfavorable and requires an external power source.3 Conversely, the reverse occurs in the discharge process. The discharge reaction, on the other hand, is thermodynamically favorable and the driving force is essentially determined by the variation of Gibbs free energy between the reactants and products according to the equation: ΔG = Σ ΔG ( ) – Σ ΔG ( ) where ΔG is the change in Gibbs free energy from a reaction at standard conditions and ΔG corresponds to the Gibbs free energy of formation at standard conditions. The equilibrium potential or open-circuit potential could be calculated as: o −ΔG where Eo is the Open Circuit Potential (O.C.P) of the battery, z is the number of electrons generated from the redox reactions, and F corresponds to Faraday constant (96 485 C mol-1). Electrons produced through redox reactions travel through an external circuit where it does work and power devices. The overall voltage is restricted by the effective voltage limit of the electrolyte, where the upper limit is decided by electrolyte decomposition/oxygen evolution4, 5 and the lower limit defined by alkali metal plating potential6. The heart of batteries lies in the electrode materials, which determines how much charge could be stored and how long it could be cycled. Specific capacity is the key parameter that determines the quantity of charge that could be accumulated per unit mass of material. Theoretical specific capacity can be calculated using: -1 1000 Specific capacity (mAh g ) = x 3600
Where z is the electrons generated from the redox reactions, F corresponds to the Faraday constant (96 485 C mol-1) and M is the Molar mass of the active material (g mol-1). As could be deduced, the specific capacity is materials dependent and is also reliant on the materials’ reaction mechanism. The coulombic efficiency (C.E) of a particular material is a ratio between the discharging capacity and the charging capacity. It determines the reversibility and charge loss in each charging/discharging cycle.
A C.E value of 100% would be ideal as it simply means that all of the stored Li or Na ions could be reversibly extracted out. In practical devices, the C.E values are much often less than 100%, which is due to parasitic side reactions that irreversibly consume Li or Na ions from the electrolyte. Coulombic efficiency is not to be confused with energy efficiency, where voltage hysteresis influences the latter. The time required to charge/discharge a cell is dependent on the current density applied and the specific capacity of the material. Rate capability is one of the key performance indicators that determine how fast a material could be charged. C-rate defines how fast a cell is being cycled. An nC cycling rate could be defined (with respect to the theoretical capacity of the material) as the amount of current required for the cell completely charge/discharge in 1/n hours. The specific capacity drops when C rate increases, which is attributed to diffusion/kinetic limitations and is dependent on morphology, structure and the materials’ redox reaction mechanism.7
Current progress of anode materials
As mentioned in the previous segment, the parameter affecting cell performance depends largely on the choice of material. Anode materials can be divided into three categories as illustrated in Figure 2.7 The traditional graphite anode in LIB undergoes an insertion reaction where 1 Li+ is stored within the basal plane of 6 C atoms giving rise to a capacity of 372 mAh g-1. While intercalation materials often boast of long cyclability due to minimal size expansion, they are largely limited by their low specific capacities (170-400 mAh g-1).8, 9 Conversely, conversion materials are capable of delivering higher reversible capacity (often in the range 500-1100 mAh g-1) but is often accompanied by large polarization and a sloping voltage profile during charge-discharge processes which prevents practical usage in commercial applications.10 The last group is the alloying materials, which possesses the highest reversible capacity (800-1800 mAh g-1) among the three groups, along with low working potential and relatively good power capabilities.11
Among the three groups of materials, tremendous efforts have been made on the research of conversion and (or) alloying based transition metal oxides and sulfides. These metal oxides and sulfides possess much higher capacity compared to the traditional graphite anode.12 Moreover, they demonstrate relatively better rate capability and cycling performance. Nevertheless, both conversion and (or) alloying materials are often accompanied by substantial volume expansion that contributes to material fracture and is identified as the main factor for its rapid capacity fading. In 2005, Sony announced the use of Nexelion battery comprising of a Sn-Co-C anode, which makes it the first time since the commercialization of LIB that a conversion/alloying anode is utilized in commercial batteries. 13
While both Li and Na are Group I alkali metals, they have reacted differently towards to various materials. For instance, due to the higher reducing potential of Na (E = -2.70 V vs. S.H.E) as compared to Li (E = -3.04 V vs. S.H.E), anode materials such as Si (~0.1 V vs. Li+/Li) that operates at potential close to plating in LIB would be rendered inactive in NIB due to thermodynamic limitations.14-16 Additionally, the ionic size difference between Li+ and Na+ has a large influence on the reversibility during lithiation/sodiation.17 The bigger ionic size of sodium ions (0.97 Å) compared to lithium ions (0.68 Å) has resulted in a much larger volume change for all conversion and alloying materials during sodiation as compared to lithiation.14, 18 This implication could theoretically lead to poorer cyclability and lower volumetric energy in NIB systems.
Intercalation materials are characterized by their low working potential with a small, flat working voltage plateau. They require an open structure to host and accommodate the presence of Li+/Na+. During cation insertion, relatively small/no volume change occurs within the structure, thus translating to longer cyclability.19, 20 Graphite (~372 mAh g-1) stands as the most successful intercalation anode material ever used in commercial LIBs. Up to 1 Li+ can be inserted at a low operating voltage (~0.5 V vs. Li+/Li) within the hexagonal C6 ring, giving rise to a theoretical capacity of 372 mA h g-1. On the other hand, sodiation of graphite faces thermodynamic restrictions (small interlayer spacing) in forming binary Na-C compounds under carbonate-based electrolytes.21, 22 One of the ways to make insertion of Na+ into graphite possible was through the substitution of traditional EC: DEC electrolyte solvents with diglyme-based electrolytes.23 Asides from that, using expanded graphite that possesses a wider spacing of 0.43 nm as compared to pristine graphite (~0.34 nm) showed relatively good cyclability with capacity retention of 74 % after 2000 cycles.24 Several other notable insertion materials include graphene, which is a one-atom-thick sheet of sp2-bonded C atoms arranged in a hexagonal crystal matrix. It has allured plentiful of interest for application in various fields since its discovery owing to its superior electrical conductivity, excellent mechanical strength, high specific surface area and good thermal/chemical stability. TiO2 is another widely explored intercalation based material that exists in several different polymorphs (e.g. anatase,25 TiO2-(B),26 brookite27 and rutile25) depending on the synthesis condition. Nb2O5 exists in several polymorphs and the most commonly studied form is orthorhombic structure.28, 29 It is attractive as anode material owing to the rapid intercalation-pseudocapacitance property.
Asides from traditional insertion electrode materials, compounds that undergo different reaction mechanisms are also of large interests. The main advantage of these materials is their capability to store more than 1 Li+ or Na+ per unit compound, which would then translate to higher specific capacity and energy density. However, the ability to accommodate a large amount of alkali cations would be a double-edged sword as it brings along great size expansion and structural damage, thus often resulting in poor cyclability.
Materials that are capable of undergoing conversion reaction are often transition metal compounds with the formula McXd (where M = 3d transition metal, X = O, S, P, N, etc.). The preliminary work on utilization of such metal oxides was carried out by Tarascon’s group, focusing on the effect of particle size and the corresponding cell’s performance.30 During lithiation/sodiation, the compound converts into elemental metallic nanoparticles that are embedded within a matrix of LinX/NanX. The electrochemical reaction could be summarized with the equation: McXd + (d·n)Li+/Na+ <-> cM + dLinX/ dNanX where n is the redox number of the anionic compounds Xn-. Conversion materials typically generate capacities in the range between 700 – 1500 mAh g-1. A table with most of the widely studied metal oxides could be found in Appendix A1.
The iconicity of the M-X bonds within the compound determines the redox potential, which falls within a range between 0.5 – 1.6 V vs. Li+/Li or Na+/Na. The main downside of employing conversion based material lies with the large voltage hysteresis between charging and discharging processes, which translates into poor energy efficiency. Furthermore, conversion type materials often operate below 0.5 V vs. Li+/Li or Na+/Na, which results in the production of a polymeric like SEI layer that hinders ion transport at the interface between the electrode and electrolyte. This formation is known to be the main culprit for the irreversible consumption of alkali ions that leads to low C.E. value in the initial cycle.
The last type of alkali storage mechanism is the alloying reaction. It involves the direct formation of a Li-A/Na-A alloy compound, where A is a Li/Na electrochemically active element/compound (e.g. Group 13, 14, 15 and 16 elements). The alloying compatibility between 2 components can easily be predicted from their corresponding phase diagrams. Alloying based materials have much lower operating potential (<0.7 V vs. Li+/Li or Na+/Na) and higher specific capacity (900 – 3000 mAh g-1) as compared to conversion materials. The capability to store such large amount of ion/charge causes more extensive size expansions (Appendix A1), thus resulting in poorer cyclability compared to both insertion and conversion materials.
Group 14 elements (e.g. Si, Ge, Sn) are some of the most extensively explored alloying materials in both Li and Na systems. Si is known to have the highest gravimetric capacity in the Li system, because it can store up to 4.4 Li per unit Si. Ge is also widely explored due to its high Li diffusion coefficient value that permits fast charging/discharging of the cell. However, both Si and Ge were found to be electrochemically inactive in Na system due to thermodynamics restriction.14
Metal Oxides as Next Generation Anode
Despite attempts by companies who have tried to utilize different anodes such as Sn-Co-C and Si, the most widely used anode material in commercial LIBs is currently still graphite. As for NIBs, the road to commercialization is still a long shot and requires intensive research to look for better-performing electrodes and electrolytes. Metal oxides itself is a widely explored family of material. TiO2,31- 33 Li4Ti5O12,19, 34-36 Na2Ti3O737 and Nb2O528, 38-40 are known to possess intercalation mechanism, similar to that of traditional graphite. They have fast reaction kinetics owing to the “open” structure that facilitates ion diffusion into the insertion sites. However, they possess low practical capacity (<~250 mAh g-1) and high reaction potential besides graphite. Conversion based oxides such as Fe2O3,41-43 Co3O4,44 MoO245 and CuO46 have way much higher capacity (~600 – 1000 mAh g-1). The performances of these oxides are dependent on their morphology and size, which critically determines their diffusion kinetics and cyclability. Table A1 illustrates the electrochemical properties of different metal oxides in both LIB and NIB.
Amongst the various metal oxides, SnO2 is one of the more widely explored options due to its chemical reactivity in both LIB and NIB. It is a transparent, wide band gap semiconductor (band gap value of ~3.6 eV) and is widely explored in various applications such as catalysis,47 electrochemical sensors48 as well as alkali ion battery electrodes.49 SnO2 has a tetragonal unit cell with a rutile crystal structure (also known as Casserite). It possesses a space-group symmetry of P42/mnm and the lattice parameters are a = b = 4.737 Å and c = 3.185 Å. The electrochemical performance of pristine SnO2, in general, undergoes a conversion reaction to form Sn that is embedded within a Li2O/Na2O matrix. Subsequently, Sn alloys with Li+/Na+ to form Li4.4Sn/Na3.75Sn. The overall theoretical capacities of SnO2 in LIB and NIB assuming fully reversible conversion and alloying reactions are 1493 and 1378 mAh g-1 respectively.50-52 While the alloying reaction is well known to be highly reversible, the reaction corresponding to a reversible conversion of Sn back into SnO2 in the Li system has been comprehensively debated as of late. In-situ XRD was utilized by Dahn et al. which showed only the presence of crystalline Sn during charging up to 3.00 V vs.Li+/Li .53 The reversible conversion was deemed inactive due to the absence of SnO2. More recently, Kim et al. utilized ex-situ TEM and SAED to investigate on the reversibility.54 Similar to Dahn’s work, they arrived in conclusion that conversion reaction has no reversibility. The reversible high capacity (~1100 mAh g-1) achieved in the early cycles was claimed to be attributed to reversible electrolyte decomposition, forming LiOH. It is important to note that firstly, both works utilized micron sized SnO2 nanoparticles, which may hinder Li diffusivity for reversible conversion. Secondly, both XRD and TEM are excellent for the detection of crystalline materials but are poor in the analysis of amorphous phases. Ex-situ XAS was also employed to study the fundamental mechanism during lithiation.55 Kim et al. illustrated through XANES that the redox state of Sn rises above 0, but smaller than 4+. Also, EXAFS demonstrated the reversibility the Sn-O bond during de-alloying reaction up to 2.00 V vs. Li+/Li. However, while Sn-O bond forms, the reversibility lasts only in the first coordination shell, indicating poor long range order. This evidence hints at presence of short-order arrangement (in other words amorphous) SnOx that clearly could not be detected by XRD or TEM. The results were also backed by recent in-situ XAS results from Pelliccione et al.56 Conversely, the reaction mechanism of SnO2 in NIB is less controversial. The reversible conversion of SnO2 in NIB has been shown to be more feasible as compared to in LIB. Both Ding57 and Lu58 explored the sodiation mechanism of SnO2 through ex-situ XRD which revealed the presence of poorly crystalline SnO2 at the end of charge (3.00 V vs. Na+/Na). Nevertheless, the electrochemical performance of SnO2 in NIB has been relatively poor. Gu et al. investigated the failure mechanism of SnO2 nanowires in NIB through in-situ TEM. Due to the size expansion during sodiation, the reverse desodiation process results in the formation of pores within the material itself. These pores have shown to contribute largely towards electrical impedance and thus resulting in poor cyclability.59
Strategies for Improving Tin (IV) Oxide as Anode
Therefore, it is of paramount importance to design and fabricate SnO2 with advanced functionalities to overcome the aforementioned intrinsic shortcomings. Several strategies have been widely deployed over the years in an attempt to improve the overall electrochemical performance of SnO2. These examples (Figure 2.6) include designing intrinsic hollow structure, coating an amorphous carbon layer on SnO2 and the template synthesis of 0D, 1D and 2D nanostructures. Asides from SnO2, these methods could also be extended to other metal oxides.
Table of contents :
Chapter 1 Problem Statement
1.1.2 Hypothesis and objective
1.2 Dissertation Overview
1.3 Findings and Outcomes/Originality
Chapter 2 Alkali ion batteries (AIBs): Operating mechanism and key parameters
2.2 Current progress of anode materials
2.2.1 Insertion electrodes
2.2.2 Conversion electrodes
2.2.3 Alloying electrodes
2.3 Metal Oxides as Next Generation Anode
Chapter 3 Selection of methods and materials
3.2 Principle behind synthesis methods
3.2.2 Laser Pyrolysis
3.3 Principle behind characterization techniques
3.3.1 X-Ray Diffraction (XRD)
3.3.2 Electron Microscopy (SEM & TEM)
3.3.3 Energy Dispersive X-ray Spectroscopy (EDX)
3.3.4 X-ray Photoelectron Spectroscopy (XPS)
3.3.5 Raman spectroscopy
3.3.6 Thermogravimetric analysis
3.3.7 Synchrotron X-ray Absorption Spectroscopy (XAS)
3.4 Principle behind electrochemical methods
3.4.1 Coin cell fabrication
3.4.2 Ex-situ measurements
3.4.3 Galvanostatic cycling
3.4.4 Cyclic voltammetry
Chapter 4 Synthesis
4.2.2 Laser pyrolysis for synthesis of SnO2 and rGO- SnO2
4.3 Physical characterization
4.4 Electrochemical measurements
4.4.1 Cyclic voltammetry
5.4.2 Galvanostatic cycling
5.2.2 Laser pyrolysis
5.3 Physical characterization
5.4 Electrochemical measurements
5.4.1 Cyclic Voltammetry
5.4.2 Galvanostatic cycling
5.4.3 Power law analysis
5.5 Ex-situ synchrotron studies
6.2.2 Co-precipitation synthesis
6.3 Physical characterization
6.4 Electrochemical measurements
6.4.1 Cyclic Voltammetry
6.4.2 Galvanostatic cycling
7.1 Impact of findings
7.1.1 Outcome of hypothesis 1:
7.1.2 Outcome of hypothesis 2:
7.1.3 Outcome of hypothesis 3:
7.2 Outstanding questions and future work
7.2.1 Investigation onto the gradual capacity decay during initial lithiation cycles
7.2.2 Further explore the different types of substitution metal for ASnO3 with other cations in the A site
7.2.3 Laser pyrolysis of bimetallic tin oxides for LIB and NIB
7.2.4 Assembly of full cell