Chapter 10
Electrolytes, Separators, and Membranes
Electrolytes, separators, and membranes (ESMs) constitute a class of components that serve important functions in electrochemical energy storage devices, and yet, their roles are often underestimated. Although they may not be directly involved in the main reduction and oxidation reactions of the chemical species that supply the electrochemical energy, the properties of the ESMs could significantly impact the performance and lifetime of the device.
ESMs are comprised of materials that are ion conductive but electron insulating. In other words, they allow the passage of ions such as Li+ and H+, but they must block the transport of the electrons. ESMs are generally placed between the two main electrodes (i.e., anode and cathode) inside the electrochemical cell. Although they have much in common, there are also subtle differences between “electrolytes”, “separators” and “membranes”. This is often influenced by the usage of the terms in the scientific literature, academia, and industry. The main function of an “electrolyte” in energy storage is considered to be ion transport and salt dissociation, and thus, an electrolyte generally can be a liquid or solid. On the other hand, “separators” and “membranes” physically separate the electrodes to prevent their contact (and electrical shorting), in addition to their basic function of ionic transport and permeation of species. Therefore, “separators” and “membranes” are generally made of a solid material like a thin polymeric sheet to serve all their necessary functions. In case of solid electrolytes, like solid polymers or ceramics, they may be considered as both an “electrolyte” and a “separator”.
Beyond the need for ion conduction and electrical insulation, ESMs can offer a host of other desirable functions in an electrochemical device. This generally depends on several factors including specific materials and phases, electrochemical cell design, and operational requirements of the device. For example, some ESMs allow the selective permeation of chemical species. In other words, they can block the transport of certain ions, but not all ions. Electrolytes also facilitate the dissociation of salt compounds. Solid polymer or ceramic electrolytes can also enhance the mechanical properties of the device. Solid or gel polymer electrolytes may provide adhesive properties that contribute to the mechanical integrity of a multilayer thin device.
One important aspect pertaining to electrolytes, in addition to their inherent desirable functional properties, is safety. As a component of an energy storage device, the electrolyte is required to be stable and safe, at least, within the operational ranges and conditions of the device, such as within the voltage window and the operational temperature range. In this chapter, we’ll discuss various aspects of electrolytes, membranes and separators in energy storage devices.
10.1 Electrolyte Classifications
We can classify all electrolytes into two main groups: (1) liquid electrolytes and (2) solid or quasi-solid electrolytes. Each group can be further divided into various electrolyte types as depicted in Figure 10.1. Liquid electrolytes can be either aqueous or non-aqueous. Solid electrolytes are usually either polymers or ceramics.
Figure 10.1 Electrolyte classification.
In this chapter, we will focus mainly on the electrolytes associated with three main electrochemical energy storage systems, specifically, batteries, supercapacitors and fuel cells. Although there are many common aspects about the ESMs used in these three devices, there are also several differences including their specific chemistry, ionic species, and morphology that will be addressed in later discussions. Before we start an in-depth discussion of electrolyte types for specific energy storage devices, let us take a look at two important and basic electrolyte properties, namely, ion conductivity and the transference number.
10.2 Ionic Conductivity
Ionic conductivity is arguably one of the most important properties of an electrolyte. In simple terms, it represents how fast and how much of the charged ions can pass through the material. Fundamentally, ion conductivity, σ, can be expressed as
where ni is the concentration of ionic species (m-3), zi is the charge of the ions (C), and mi is the mobility of the ions (m2 s–1 volt–1).
The ion conductivity in SI units is represented in Siemens per meter (S/m). Siemens (S) is the reciprocal of ohm, which is the unit of resistance or impedance. In the past, conductivity was represented in terms of “mho” (ohm spelled backward). A quick verification of the final unit from Equation 10.1, reveals the resulting units based on the individual parameters as coulomb sec-1 volt-1 meter-1, which simplifies to its equivalent, ampere volt-1 meter-1 or ohm-1 meter-1, and finally, represented as S/m.
10.2.1 Measurement Techniques
Ionic conductivity can be determined through several methods. If the mobility and concentration of the specific ions are known or can be directly measured, then the ionic conductivity can be determined through Equation 10.1. Often, the specific parameters in Equation 10.1 may not be known or are relatively difficult to directly measure. Therefore, another method of determining ion conductivity that is more practical is commonly employed based on the following equation
where R is the bulk impedance, d is the thickness, and S is the surface area of the material.
Equation 10.2 is especially suitable for solid electrolytes. The experimental procedure to obtain ion conductivity through Equation 10.3 is as following: First, the solid electrolyte layer (e.g., polymer film) is sandwiched between two electrically conductive metal electrodes (e.g., stainless steel), as shown in Figure 10.2. Care must be taken to keep the electrode internal surfaces clean and smooth to ensure good contact with the solid electrolyte film at the interface. The electrodes are then connected externally to a frequency response analyzer, and the solid electrolyte film is subjected to impedance spectroscopy (ranging from high to low frequencies). The plot of Z” vs. Z’ (imaginary vs. real impedance) is shown in Figure 10.2; this is a typical plot obtained for a solid polymer film, consisting of a semicircle impedance profile followed by a slanted line to the right.
Figure 10.2 Procedure to obtain bulk impedance for ion conductivity calculation of polymer electrolyte through impedance spectroscopy.
To calculate ion conductivity in Equation 10.2, we must determine the bulk Rb from the impedance spectra. We can accomplish this by two methods: (1) the resulting impedance spectra obtained is modeled through an electrical representative circuit model and the bulk resistance R is extracted, (2) the tangential slanted line from the right side of the impedance spectra is projected onto the real impedance axis (horizontal axis), and R is estimated. The former method (1) is a more precise and systematic way to determine bulk R, and furthermore, the circuit element modeling can provide useful insights on the characteristics and overall behavior of the electrolyte film.
10.2.2 Nyquist Plot Circuit Fitting
Example of Nyquist plots obtained from the impedance spectra on liquids and polymer electrolyte are shown in Figures 10.3 and 10.4, respectively.
The solution resistance Rs of the liquid electrolyte shown in Figure 10.3 and the bulk resistance Rb for the polymer in Figure 10.4 can be used to calculate their respective electrolyte ion conductivities. The contact resistance Rc can include all metallic, and lower impedance components. The element Q in the circuit model of Figure 10.4 is a nonlinear capacitor. The capacitor in the circuit model can be either linear or non-linear depending on the materials and mechanisms involved. Warburg diffusion impedance is represented by the model element W.
Figure 10.3 Typical nyquist plot of liquid or gel electrolyte.
Figure 10.4 Typical impedance spectra for a solid polymer electrolyte and its associated circuit model fitting.
To appreciate the advantages of the impedance spectroscopy accompanied by the circuit modeling, as well as some inherent challenges, let us take a more in-depth look at the fundamentals of this method.
In an ideal resistor, the impedance R is related to the voltage E and current I, and is independent of the frequency, and thus, it is expressed as
If we send an alternating current (AC signals) through an ideal resistor, it will be in-phase with its voltage. However, many electrolytes, electrochemical materials, and cells show complex impedance behavior and are therefore considered non-ideal. The impedance obtained, say from a polymer electrolyte film, is the resistance of the material to an applied current, that exhibits a complex behavior, and therefore, may not follow an ideal resistor. To show this behavior mathematically, the voltage and current are shown as frequency dependent functions as following
(10.4)
(10.5)
where Et is the potential (voltage) at time t, E0 is the amplitude of the signal, and ω is the radial frequency (radians/second). Note that It the current signal at time t as a function of amplitude I0 and ωt is shifted in phase by φ from the input potential signal.
We can obtain the expression for impedance Zt (similar to Ohm’s law) as
(10.6)
10.3 Ion Conduction Theory
The ion conduction mechanism in electrolytes can vary widely from one electrolyte type to another. Figure 10.5 shows a comparison of ionic conductivity in various types of electrolytes. As it can be seen, liquid electrolytes have higher ion conductivity compared to gels, and gels are more conductive than solid polymer electrolytes.
Figure 10.5 Comparison of electrolytes
(courtesy of Kamaya et al., Nature Materials 10, 2011).
Perhaps the slightly counterintuitive observation is that hard solid (inorganic) electrolytes like ceramic and glassy ion conductors show very high ion conductivity, even higher than liquids. Also, it can be seen that with higher temperatures, the ion conductivity increases, which is expected due to the relation between conductivity and diffusion coefficient, and the Arrhenius dependency of diffusion coefficient on temperature. Various additives such as nanometer sized fillers and liquid plasticizers may be added to the electrolyte to enhance its properties.
The fundamental process of the ion conduction strongly depends on a host of factors including the electrolyte phase, chemistry, structure, ionic species, and environmental conditions. For example, the mechanism of ion conduction in liquid electrolytes can be considerably different from that in polymer or ceramic electrolytes. Here, we discuss some of the mechanisms pertinent to specific electrolyte phases and types.
10.3.1 Ion Conduction in Liquid Electrolytes
Let us begin with the concept of ion diffusion. We can refer to a simple diffusion model known as Fick’s first law that relates the flux of the diffusing species to diffusivity and concentration gradient:
(10.7)
where J is the diffusion flux (m–2 s–1) which represents the amount of diffusing substance passing through a unit area at a unit time interval, D is the diffusion coefficient (m2/s) and c is the concentration of the diffusing species (mol/m3). In concentration dependent diffusion, the ions move from the highest to the lowest concentration regions and the ion flow (flux) is in the opposite direction of the concentration gradient, hence the negative sign in Equation 10.8.
Fick’s first law in one dimension can be expressed as
Fick’s second law of diffusion is related to the rate of concentration change:
(10.9)
We can show Fick’s second law in one-dimensional diffusion as
In many cases, it’s reasonable to assume D is constant due to the negligible local effects on diffusivity. Therefore, Equation 10.10 becomes
(10.11)
Assuming specific boundary and initial conditions, we can obtain the concentration curves as a function of dimension x and time t. A graphical depiction of the concentration profiles for one-dimensional Fickian diffusion is shown in Figure 10.6 for thickness L with concentration C1 at the outer surface (C=C1 at x=0 and x=L), and zero concentration initially inside the material (C=0 at t=0 and 0< x <L).
Figure 10.6 Concentration profile C(x, t) for one-dimensional Fickian diffusion in a thin material of thickness L.
In general, the Fickian model is adequate for many material systems to predict diffusion and time dependent concentration of the diffusing species. There are a few exceptions. For example, in cross-linked polymer materials that swell (expand volumetrically) during the diffusion process, non-Fickian diffusion models may be more suitable.
Let us now explore the effect of temperature on diffusion. The diffusion of ions in a liquid electrolyte can be expressed using the Einstein-Stokes relation:
where D is the ion diffusivity, k is Boltzmann’s constant, T is the absolute temperature (K), µ is the solvent liquid viscosity, ro is the ion radius.
As it can be seen, the diffusion coefficient for liquid electrolytes as shown in Einstein-Stokes expression (Equation 10.12) is proportional to temperature. The coefficient b is related to the size of the diffusing species A compared to that of the base solvent liquid molecules B. If the sizes of the two molecules A and B are the same, b=4. If A is larger than B, then b=6, and if A<B, b takes a value lower than 4. For ions diffusing through a liquid solvent electrolyte in batteries, a value of b equal to 6 is generally appropriate. It is important to note that the Einstein-Stokes equation is most suitable for low-viscosity liquids.
To relate the ionic conductivity, σ (S/cm) to diffusion coefficient D we can use Nernst-Einstein equation:
(10.13)
where q is ionic charge, n is ionic concentration, and k is Boltzmann’s constant. As noted, ion conductivity in liquids is relatively less influenced by temperature as compared to that in solid electrolytes where an exponential relation is often observed.
In an energy storage cell of diffusing species, where concentration gradient is the only driving force for diffusion, and Fickian diffusion is observed, the concentration profiles in Figure 10.5 are applicable. Depending on the initial and boundary conditions of the concentrations in the cell, we can apply either half-thickness (0<x<L/2 or L/2<x<L) or full thickness profiles from Figure 10.5.
In electrochemical cells, such as lithium ion batteries, other driving forces for diffusion are present that need to be considered. In other words, Fick’s first law in one dimension may not be adequate and Equation 10.8 should be expanded. An appropriate and commonly used ion diffusion model for electrolytes in Li-ion batteries and other similar electrochemical systems including supercapacitors is the Nernst-Plank equation:
(10.14)
where J is the flux of the ion species i at a distance x from the electrode/electrolyte interface (mol/m2s), D is the diffusion coefficient of the ion (m2/s), ∂c/∂x is the concentration gradient at a distance x (mol/m4), zi is the charge state of the ion (dimensionless), F is the Faraday’s constant (96485 C/mol), R is the universal gas constant (8.314 J/Kmol), T is the absolute temperature (K), ci is the concentration of ion i (mol/m3), ∂φ/∂x is the potential gradient (V/m), and v(x) is the velocity of the volume element (m/s).
The concentration gradients in lithium ion batteries can be generally linear at steady state during galvanostatic charging and discharging (Figure 10.7). This has been demonstrated both theoretically and experimentally, and more discussion of concentration gradients in batteries can be found in the literature including [Zhou et al., 2006].
Figure 10.7 Linear concentration gradient at steady state in the electrolyte during charging and discharging.
Another model, known as Maxwell-Stefan equation, can account for interactions between the ions and non-idealities in the electrolyte. The rate of lithium ion concentration change (∂cL/∂x) in the electrolyte based on Maxwell-Stefan equation is expressed as
(10.15)
where εL is the electrolyte volume fraction, βL is Bruggeman’s constant for the electrolyte filled pores, VmLiPF6 is the molar volume of the electrolyte salt LiPF6, DL is the diffusion coefficient of the electrolyte salt corresponding to the thermodynamic driving force, 1 + ∂ ln f ±/∂ ln cL is the thermodynamic enhancement factor, cL, tot is the charged species total concentration, csolv is the concentration of the solvent tsolvLi+ is the transport number for the solvent, jL is the current density in the electrolyte, and F is Faraday’s constant.
The potential distribution in the electrolyte can be determined from the Maxwell-Stefan equation as
(10.16)
where κ is the concentration-dependent ionic conductivity. A thorough analysis of ion transport and polarization in a lithium ion battery based on liquid electrolyte (LiPF6 dissolved in EC and EMC, 3:7) using Maxwell-Stefan model is provided in [Nyman et al., 2010].
10.3.2 Ion Conduction in Polymer Electrolytes
The ion conduction mechanism in polymer electrolytes and membranes can vary depending on many factors including the polymer structure and ion species. Li+ ion conduction in polyethylene oxide (a common solid electrolyte used for batteries) can be different from H+ proton conduction in polymer membrane (i.e., Nafion) used in hydrogen fuel cells. However, there are also some similarities. In this section, we’ll explore various aspects of the ion conduction mechanisms in polymer electrolytes.
To understand ion conduction in polymers, let us start with the concept of ‘free volume’. The irregular packing of the polymer chains in an amorphous (non-crystalline) polymer leads to the existence of free volume. Figure 10.8 illustrates the free and occupied volumes related to the polymer chains. The amount of the free volume in a polymer electrolyte can greatly influence its chain mobility and flexibility, and subsequently, its ion conductivity.
Figure 10.8 Illustration of free volume and occupied volume.
The free volume is the unoccupied volume in the polymer material. In other words, the total volume minus the occupied volume is the free volume as shown below
(10.17)
The occupied volume in a polymer material is the space taken by the actual polymer molecules and the thermally induced vibrations of the molecule:
(10.18)
Basically, the free volume (VFree) is the volume due to very small holes or voids (generally at nanoscale) that are created due to packing irregularities of the polymer chains.
It is important to note that the free volume can change with temperature, that is if the polymer chain has sufficient space to undergo conformational changes (e.g., chains unfolding, twisting) leading to free volume changes.
The model for lithium ion transport in a common solid polymer electrolyte, polyethylene oxide (PEO) with oxygen atoms on the chain, is drawn in Figure 10.9. The positive lithium ion can be attracted to the negative polarity created by the oxygen atoms on the polymer chain. As the polymer chain undergoes thermally induced configuration change, the lithium ion hops to another site in the vicinity with a stronger negative polarity. It can be noted that the free volume, polymer chain configuration, and flexibility play critical roles in the facilitation of lithium ion hopping and transport mechanism (Figure 10.10).
Figure 10.9 Schematics of lithium ion transport mechanism in polyethylene oxide electrolyte.
Figure 10.10 Schematics of polyethylene oxide with free volume in the presence of lithium cations and perchlorate anions.
Figure 10.11 shows the concept of glass transition and free volume based on the thermal expansion behavior of polymers. At the glass transition temperature Tg, the free volume reaches a minimum limit, where it is too small for any chain conformation movement; therefore, the free volume is effectively “frozen” at Tg.
Figure 10.11 Free volume VFree, coefficient of thermal expansion (CTE) and glass transition temperature (Tg).
The free volume model for ion diffusivity D (cm2/s) at temperatures above Tg can be expressed based on Cohen and Turnbull equation as
where D0 is a constant, γ is overlap correction (between 0.5 and 1), V* is the critical volume large enough to allow another molecule to jump after displacement.
The free volume as a function of temperature is expressed as
where fg is the free volume fraction at Tg, Vg is the volume of polymer, and α is the expansion coefficient of the free volume.
Substituting Equation 10.20 into Equation 10.19, we obtain the Williams-Landel-Ferry (WLF) equation:
(10.21)
where C1 = γ V*/Vgfg and C2 = fg/α.
Ionic conductivity, σ (S/cm), can be related to diffusion coefficient by Nernst-Einstein equation:
(10.22)
where q is ionic charge, n is ionic concentration, k is Boltzmann constant. Note, that in addition to the temperature T in the denominator, there can be another temperature dependency that is embedded inside the diffusivity term D.
The free volume (FV) model for ionic conductivity is generally the combination of Equations 10.16 and 10.19, and expressed as:
(10.23)
FV models are sometimes found to be inadequate for modeling ionic conductivity in solid polymer electrolytes. A semi-empirical model, namely, the Vogel–Tammann–Fulcher (VTF) model can better predict ionic conductivity in solid polymer electrolytes, when the FV models are not suitable.
Ionic conductivity, σ, as a function of temperature can be expressed using the VTF model as
(10.24)
where A and B are fitting constants, R is the universal gas constant, and T0 is referred to as the Vogel temperature (the temperature at which the extrapolated mobility of ions disappears).
10.3.3 Ion Conduction in Ceramic Electrolytes
Another class of solid electrolytes is inorganic compounds including ceramics and glasses. An interesting and rather surprising aspect of these electrolytes is that they can exhibit ion conductivity values as high as that of liquid electrolytes (or even higher). This is somewhat non-intuitive. We often view liquids as a medium for faster diffusion and better conduction as compared to the more mechanically restrictive structure in solid and hard ceramics. Many studies have looked into the reasons and mechanisms of high ion conduction in ceramic electrolytes. Point defects in ceramic lattices create “fast transport tunnels” for ions to pass through, and therefore, provide an effective high ion conduction medium.
Ceramic electrolytes can be ion conductive because of the basic imperfections and defects in their crystal structures. Point defects in crystals include Schottky defect and Frenkel defect. Schottky defect exists due to vacant crystal sites. Frenkel defect is formed when an ion (generally smaller sized ion) leaves its site vacant, and moves to an interstitial site.
Ion transport in crystals is possible due to hopping mechanism through continuous pathway of vacancies or interstitial sites. We can classify the ion conduction mechanism in ceramic electrolytes into three types as shown in Figure 10.12: (a) vacancy mechanism, (b) interstitial mechanism, and (c) interstitialcy mechanism.
Figure 10.12 Ion conduction in ceramic electrolytes by (a) vacancy mechanism, (b) interstitial mechanism, (c) interstitialcy mechanism.
In the vacancy mechanism, the ion hops onto an extrinsic or intrinsic vacancy. In an interstitial mechanism, an interstitial ion hops onto an adjacent vacant interstitial site. In the interstitialcy mechanism, an interstitial ion pushes a neighboring ion in a normal lattice site onto an interstitial vacancy, and subsequently, the ion takes its place in the regular lattice site. This process is also called “cooperative motion”. There are many material and structural factors that can affect ion conductivity in ceramic electrolytes including the ion size, activation energy of ion to move to a new site, availability of continuous transport pathways, number of crystal defects or vacancies compared to the number of ions. These factors may be considered during the design and fabrication of the ceramic material to enhance its ion conductivity and the overall performance of an electrochemical device.
A common ion conducting ceramic electrolyte considered for lithium ion batteries is lithium lanthanum titanate (LLTO). The process of ion conduction in this compound is by vacancy mechanism. Lithium ions move through a high concentration of A-site vacancies in the crystal, leading to high ion conductivity. More discussion of inorganic solid lithium ion conductors can be found in the literature including the review by Philippe Knauth [Knauth, P., 2009]. We’ll revisit the types and properties of ceramic and glassy electrolytes for Li ion batteries in a later section (10.6.2.2).
10.4 Factors Affecting Ion Conductivity
Many factors can influence the ion conductivity of an electrolyte. Let us look at one example, namely, the ion conductivity of a solid polymer electrolyte. We can organize all the factors in the form of “cause and effect” or “fish bone” diagram as shown in Figure 10.13.
Figure 10.13 A fish bone diagram of factors that influence ion conductivity measurements.
The factors can be grouped into material, fabrication and testing. Let’s evaluate each one. Perhaps the most important material factor is the polymer host. The polymer host factors include the chain backbone chemistry, side groups, molecular weight, the state of free volume, and percentage crystallinity (for semi-crystalline polymers). All these factors can affect ion conduction.
The second group of factors is related to the lithium salt including salt concentration, chemistry, and salt/solvent interaction. In case of solid polymer electrolytes, due to their low ion conductivity, often, solid fillers are added to boost the ion conductivity. Factors related to the fillers include filler material, size (e.g., nanoscale versus microscale), shape, and surface coating/functionality. We can also add liquid plasticizer to increase ion conductivity of the polymer electrolyte, and factors include the glass transition temperature, type of plasticizer, and concentration.
The fabrication process also plays a critical role. How we make the polymer can directly affect its ion conductivity. Factors include quality of dispersion and mixing, method of fabrication (e.g., casting versus coating), and cooling and heating.
Perhaps the most underestimated factor that we must consider is the testing condition. Ion conductivity must be assigned to the material based on measurements, and the measurement technique and conditions can cause variation in the conductivity assignment. For example, how accurately were the sample area and thickness measured? Were the samples exposed to room temperature and humidity, and for how long? How was the bulk impedance R from Equation 10.2 determined? Was there a circuit model fitting involved, and how precisely fitted? Similar to any type of property measured and assigned to a material, the measurement technique, and testing conditions must be carefully controlled to minimize variability in the ion conductivity values.
10.5 Transference Number
When the ions in the electrolyte carry the charge from one electrode to another, not all charges are carried by one type of ions. For example, in a lithium perchlorate (LiClO4) salt-based electrolyte, charges may be carried by both the lithium ion (Li+) (or cation) and the perchlorate ion (ClO4–) (or anion). This is undesirable for the battery operation as it causes concentration polarization in the electrolyte.
The ideal scenario is to have only the main ion type (that is involved in intercalation of electrodes) carry the charge, and in the case of lithium ion battery, this ion is Li+. Therefore, another property of importance when evaluating electrolytes is their transference number. The transference number (also referred to as the transport number) is defined as the fraction of total charge in the battery carried by the main ion type. If only half of the total charge is carried by the lithium ions, then the transference number is 0.5. In other words, the other half is carried by a different ion. This means that two types of ions are mobile in the electrolyte. If all the charge is carried by lithium ions, then the transference number is 1. That indicates that other ions are not mobile (or exhibit much lower mobility compared to the main ion).
Therefore, transference number (t) of an electrolyte can be expressed in terms of the mobility µ of the charge carrying ions as
(10.25)
The transference number can be measured in several different ways and has consequently led to some variation in the estimated values for electrolytes. Therefore, when comparing different transference numbers, care must be taken to use the values obtained from the same measurement techniques.
10.6 Electrolytes for Lithium Ion Batteries
In this section, we will discuss the electrolytes for lithium ion batteries including non-aqueous and aqueous liquid electrolytes, and polymer and ceramic solid electrolytes.
10.6.1 Liquid Electrolytes
Liquid electrolytes can be divided into two main groups: Aqueous (water based) and non-aqueous. An aqueous electrolyte can be neutral, acidic or alkaline. The non-aqueous electrolytes consist of organic liquids and ionic liquids. Organic liquid electrolytes (non-aqueous) are the most common electrolytes used in conventional batteries. They have one of the highest ion conductivities among the different electrolytes, and their use in electrochemical cells has been extensively investigated. More in-depth discussion of different electrolytes will be presented in the following sections.
10.6.1.1 Non-Aqueous Electrolytes
Organic liquid electrolytes offer several advantages including high ion conductivity and a relatively high transference number, however, they also come with several drawbacks. First, the electrolyte may undergo undesirable chemical reactions during the battery charge/discharge operation. Specifically, the electrolyte may react and form unintended (and undesirable) chemical by-products at the graphite-based anode (negative electrode). This chemical reaction leads to the creation of a thin solid layer on the surface of the electrode called solid electrolyte interphase (SEI). Although the SEI layer still allows the passage of ions, it can slow down their diffusion. The formation and growth of the SEI layer on the electrode surface can cause a noticeable degradation in the battery performance, causing capacity fading, especially in the first cycle (or more) until the SEI layer is fully stabilized. The presence of a stable SEI layer prevents further chemical reactions in the electrolyte, and thereafter, the battery can show a relatively steady capacity performance.
The SEI is characterized as a bilayer formation. The inner and outer layers of the SEI are shown in Figure 10.14. The inner SEI layer is located adjacent to the electrode (anode) surface and is inorganic. It is a relatively dense layer made of mainly Li2CO3 and LiF. The outer SEI layer that interfaces with the liquid electrolyte is porous and composed of organic components. The total thickness of the SEI layers can range from 2 nm to tens of nanometers. The nature of the SEI materials and the mechanisms of SEI formation in batteries are not yet fully understood and still under investigation. This is mainly due to the complex structure of the SEI and the challenges in precise in situ experiments. A review of the SEI modeling for lithium ion batteries is provided by Wang et al., 2018.
Figure 10.14 Solid electrolyte interphase (SEI) layers.
Common examples of organic liquid electrolytes used in lithium ion batteries include propylene carbonate (PC), ethylene carbonate (EC), di-methyl carbonate (DMC), and diethyl carbonate (DE). Often, a mixture of two (or more) electrolytes may be used in a battery (e.g., PC/DMC or EC/DMC) to achieve optimum properties. Table 10.1 shows the properties of a selection of organic (non-aqueous) electrolytes.
Table 10.1 Selection of non-aqueous electrolytes for Li-ion batteries.
| Solvent | MW | Tm (°C) | Tb (°C) | Tf (°C) | η (cP) | ε |
| EC | 88 | 36.4 | 248 | 160 | 1.9* | 89.78 |
| PC | 102 | –48.8 | 242 | 132 | 2.53 | 64.92 |
| NMO | 101 | 15 | 270 | 110 | 2.5 | 78 |
| DMC | 90 | 4.6 | 91 | 18 | 0.59** | 3.107 |
| DEC | 118 | –74.3 | 126 | 31 | 0.75 | 2.805 |
| DMM | 76 | –105 | 41 | -17 | 0.33 | 2.7 |
| DME | 90 | –58 | 84 | 0 | 0.46 | 7.2 |
η and ε at 25°C
* 40°C
** 20°C
Lithium salt is added to the electrolyte to provide a source for lithium ions as well as increase the ion conductivity of the electrolyte. A selection of commonly used lithium salts and their respective properties are shown in Table 10.2. The common Li salts include lithium perchlorate (LiClO4), lithium hexafluorophosphate (LiPF6), lithium hexafluoroarsenate (LiAsF6), lithium tetrafluoroborate (LiBF4), and lithium trifluoromethanesulfonate (LiCF3SO3), also known as lithium triflate (LiTf). More expanded discussion of the lithium salts can be found in the review by Xu, 2004 and other similar reviews.
Table 10.2 Selection of most commonly used lithium salts in lithium ion batteries.
| Salt | MW | Tm (°C) | Tdecomp. (°C) in solution | σ (S/cm) × 10–3 (1.0 M, 25 °C) | |
| in PC | in EC/DMC | ||||
| LiBF4 | 93.9 | 293 | >100 | 3.4 | 4.9 |
| LiPF6 | 151.9 | 200 | ~80 (EC/DMC) | 5.8 | 10.7 |
| LiAsF6 | 195.9 | 340 | >100 | 5.7 | 11.1 |
| LiClO4 | 106.4 | 236 | >100 | 5.6 | 8.4 |
| LiTf | 155.9 | >300 | >100 | 1.7 | – |
Another group of liquid electrolytes are known as ionic liquids (ILs). Ionic liquids are essentially molten salts. If their melting temperature is below the room temperature, then they are referred to as room temperature ionic liquids (RTILs). A few examples of ionic liquids for anode/cathode electrochemical systems are shown in Table 10.3.
Table 10.3 Examples of ionic liquids used in electrochemical systems.
| Electrochemical system | Ionic liquid |
| LixTiyOz/LiCoO2 | [EtMeIm+] [BF4–] |
| Li/LiCoO2 | [EtMeIm+] [NTf2–],
[BuMeIm+] [BF4–] |
| C/LiCoO2 | [EtMeIm+] [NTf2–], [BuMe2Im+] [PF6] |
| Li/LiMn2O4 | [Me3HexN+] [NTf2–], [MePrPyrrol+] [NTf2–] |
| Li/V2O5 | [MePrPyrrol+] [NTf2–] |
| Li4Ti5O12/LiFePO4 | [MePrPyrrol+] [NTf2–] |
The RTILs are considered for use in batteries because of several advantages including better safety and stability compared to their organic liquid counterparts. To create a suitable electrolyte for lithium ion batteries, lithium salt [Li+][X–] is generally added and dissolved in the RTIL [A+] [X–]. The resulting ionic liquid system, therefore, becomes [Li+]m [A+] n [X–]m+n which contains two types of cations, namely [Li+] and [A+]. Although we can achieve relatively high ion conductivity and safety with RTILs, a few drawbacks still remain. One functional drawback is the lower transference number.
In RTILs, the transference number of lithium ion t (Li+) is determined as
(10.26)
However, in the traditional organic electrolyte it is
(10.27)
Therefore, RTILs can have an inherent disadvantage of lower transference number due to the added mobility of the other species in the electrolyte system. Another drawback associated with ionic liquids is that they can be more expensive compared to the other types of electrolytes. More extensive information about the ionic liquids can be found in the review paper by Lewandowski et al., 2009 and other similar reviews.
10.6.1.2 Aqueous Electrolytes
Aqueous electrolytes offer several advantages over non-aqueous organic liquid electrolytes including higher ion conductivity (e.g., 2 orders of magnitude), safety, and less expensive battery assembly and fabrication process. However, there are several drawbacks that need to be addressed. One of the main problems with an aqueous electrolyte based battery is the instability in charge/discharge cycling. Past studies have investigated and reported poor cycling life of lithium ion batteries based on aqueous electrolyte including the original study conducted by Dahn’s group in 1994 [Li et al., 1994].
Recent efforts have shown some degree of success on improving cycle life [Luo et al., 2010]. Through a comprehensive investigation of suitable negative electrode materials in aqueous electrolytes, it was found that pH level of the aqueous electrolyte alone has less of a central role as previously thought. The main culprit in the case of cycling instability is the presence of oxygen. By removing oxygen from the electrochemical system during battery assembly (e.g., through vacuum sealing) a significant enhancement of cycling life can be achieved.
Figure 10.15 shows electrode materials that can be potentially used with aqueous based electrolyte in lithium ion batteries [Luo et al., 2010]. Both positive and negative electrode materials are included. The hydrogen and oxygen evolution at various pH levels are also shown.
Figure 10.15 Electrode materials that can be explored for use in aqueous electrolyte based lithium ion batteries and their respective intercalation potential at various pH levels of the aqueous electrolyte
(adapted from Luo et al, 2010).
We can investigate if the battery electrode is stable in the presence of both water and oxygen. We first identify the chemical reaction and the elements involved:
(10.28)
We can then calculate the potential for an intercalated electrode material using the following expression:
(10.29)
where 
is the chemical potential of intercalated Li in the electrode compound, 
is the chemical potential of Li in Li metal and e is the magnitude of the electron charge.
Plugging the respective values into 10.29 as a function of pH, we obtain the following:
From Equation 10.30, it can be seen that no electrode material can avoid reaction in the presence of oxygen. Now, let us investigate what would happen if we remove the oxygen. In other words, how does the electrode react with water in the absence of oxygen? The chemical reaction becomes:
(10.31)
We can then recalculate potential V(x) for the chemical reactions and obtain
According to Equation 10.32, it can be seen that chemical stability is possible with aqueous electrolytes that have pH of greater than 10. For example, a pH of 13 results in potential of 2.27 V. Therefore, a suitable negative electrode material can be selected with an adjusted electrolyte pH. For example, we can select LiTi2(PO4)3 which at pH of 13 can remain stable. Experimental results confirm the prediction of adjusted electrolyte pH in the absence of oxygen. Specifically, the reported capacity plots of cycled LiTi2(PO4)3/LiFePO4 aqueous lithium-ion battery show impressive enhancements in a battery’s cycle life [Luo et al., 2010]. More discussion of properties and requirements of electrochemical cells with aqueous electrolyte can be accessed in related studies including [Luo et al., 2010].
10.6.2 Solid and Quasi-Solid Electrolytes
Solid and quasi-solid electrolytes are commonly used in modern fuel cells, and non-traditional batteries and supercapacitors. Solid or quasi-solid electrolytes can also be further divided into two types: Polymer electrolytes (organic) and ceramic/glass electrolytes (inorganic).
10.6.2.1 Polymer Electrolytes
An important category of electrolytes is the polymer electrolytes. They are considered soft solid organic electrolytes. The term ‘soft’ describes their physical features; they provide more mechanical compliance and flexibility compared to ‘hard’ inorganic electrolytes. They are highly suitable for thin-film flexible or stretchable batteries or supercapacitors.
Polymer electrolytes can be fabricated as solid or gel (quasi-solid). Solid polymer electrolytes (SPEs) are referred to the polymers that are generally dry, and consist of one phase of polymer mixed with salt.
Gel polymer electrolytes (GPEs), on the other hand, are referred to the cross-linked polymers that contain liquid electrolyte at a relatively higher weight percentage (e.g., 50% or above). Gels can have a wide range of properties depending on the amount and type of liquid, and properties of the polymer matrix.
Let us focus on “gel” electrolytes. The gel polymer electrolytes (GPEs) constitute two distinct phases of liquid salt/plasticizer solution inside swollen cross-linked solid polymer matrix. Although commonly considered as a viscous paste, the density and viscosity of GPEs can vary widely ranging from liquid-like electrolytes to solid-like films, depending on the fraction of the liquid phase present in the polymer matrix, and the configuration, and characteristics of the material constituents.
Dry solid polymers have the lowest ion conductivity among all the electrolytes, and therefore, various techniques are used to improve their conductivity. For example, a small amount of liquid may be added (e.g., 10% wt.) to plasticize the polymer. By plasticizing, we are improving polymer chain flexibility and mobility (lower glass transition temperature) and thereby, we improve its ion conductivity. Another technique is to add a small amount of nanosized fillers (e.g., 1% or 5%) such as silica nanoparticles, or graphene oxide (GO) nanosheets. This will be discussed further in a later section.
10.6.2.2 Ceramic Electrolytes
Ceramic or glassy electrolytes are hard solid inorganic electrolytes that offer relatively high ion conductivity, despite being solid. In the case of solid polymers, liquids exhibit higher conductivities than solid or gel polymers. However, in solid inorganic (ceramic and glassy) electrolytes the mechanism of ion conduction is different. We discussed the three different types of ion hopping mechanisms in ceramics in an earlier section (10.3.3), which are vacancy mechanism, interstitial mechanism, and interstitialcy mechanism.
Here, we will mention the common inorganic solid electrolytes used in lithium ion batteries such as LiPON, LLTO, and sulfide glass. Table 10.4 provides a selection of ceramic and glassy materials commonly used as solid electrolytes for Li ion batteries and few corresponding properties [Knauth, 2009]. The ion conductivity of inorganic electrolytes at room temperature (RT) can generally range from 10-6 to 10-3 (S/cm). Fast tunnel ion transport through continuous pathways of crystal lattice defects is mainly responsible for the relatively high conductivity observed in some solid inorganic electrolytes.
Table 10.4 Inorganic solid electrolytes.
| Name | Composition | Type | Ion conductivity at RT (S/cm) | Activation energy (EV) |
| LLTO |  |
Crystalline | 10–3 | 0.3–0.4 |
| NASICON | Li1.3Al0.3Ti1.7(PO4)3 | Crystalline | 3 × 10–3 | 0.3–0.5 |
| LISICON | Li14ZnGe4O16 | Crystalline | 10–6 | 0.4–0.6 |
| Thio-LISICON | Li3.4Si0.4P0.6S4 | Crystalline | 6.4 × 10–4 | 0.5–0.6 |
| Garnet | Li6La2BaTa2O12 | Crystalline | 4 × 10–5 | 0.4–0.6 |
| Li ion conductor-mesoporous oxide | Lil–Al2O3 | Composite | 2.6 × 10–4 | 0.4–0.5 |
| Sulfide glass | GeS2+ Li2 + Lil + Ga2S3 and La2S3 | Amorphous | 10–3 | 0.4–0.5 |
| LiPON | Li2.88PO3.73N0.14 | Amorphous | 3.3 × 10–6 | 0.45–0.55 |
10.7 Electrolytes for Supercapacitors
Supercapacitors are most frequently mentioned synonymously with double-layer capacitors. An electrical double-layer capacitor (EDLC) consists of mainly two electrodes and an electrolyte/separator. The electrolyte is an ion conductor and a source of positive and negative ions. As shown in Figure 10.16 the ions in the electrolyte reach the pore walls in the electrode, and line up next to their opposite charges from the electrode, effectively forming two layers or a “double-layer” capacitor. In between these two layers is the electrolyte solution.
Figure 10.16 Schematics of an electrical double-layer mechanism at the negative porous electrode.
One may ask, why is a double-layer capacitor, a “super” capacitor? Two main factors contribute to the “super” performance of double-layer capacitors. The first factor is the distance between the capacitor electrode and the second layer formed by the ions (effectively, the second ionic electrode) which is dramatically small. Essentially, the ions can reach as close as possible to the solid electrode minimizing the distance. This very small distance between the solid electrode and the ionic electrode is called the “Helmholtz layer” and it’s about 0.3 to 0.6 nm (3 to 6 Angstroms), as shown in Figures 10.16 and 10.17. Recalling the equation of capacitance (Chapter 2, Equation 2.20), one can see that such a small distance d, in the denominator, can lead to orders of magnitude increase in capacitance, hence creating a “super” performance capacitor.
A second reason for the “super” performance is related to the area in the capacitor. The double-layer capacitor is generally made of porous electrodes with much higher specific surface area or area density (surface area per mass or volume) and much better accessibility of the ions. As such, higher area, A, in the numerator of the capacitance equation can also result in significant capacitance increase and contribute to the “super” performance.
Note that in a supercapacitor there are two electrodes, and at each electrode/electrolyte interface, a double-layer capacitor is formed. Thus, the total capacitance CTotal of a supercapacitor is the result of two double-layer capacitors with capacitance CD in series:
(10.33)
Supercapacitors or double-layer capacitors may also be referred to as electrochemical supercapacitors (ESs). However, no major electrochemical reaction occurs in the double-layer electrode as compared to the chemical reactions (oxidation and reduction) in the batteries and fuel cells. In some cases, the disassociation of the salt and the formation of ions in the supercapacitor electrolyte is considered as “electrochemical” phenomena. Thus, a double-layer capacitor may be considered as an “electrical” or “electrochemical” energy storage device.
In addition to double-layer capacitors, there are two other devices that more closely belong to the class of electrochemical supercapacitors (ESs). These two devices are pseudocapacitors and hybrid capacitors. Pseudocapacitors allow electrochemical reactions in the form of reversible faradaic redox reactions. On the other hand, hybrid capacitors consist of both faradaic redox reactions and electrical double layer storage, effectively part supercapacitor and part battery. Here, we will focus mainly on the double-layer supercapacitors.
For the double-layer capacitor, several different models have been proposed to predict and understand the behavior of the supercapacitor, especially at the electrode-electrolyte interface. Figure 10.17 shows the graphical depictions pertaining to three common models, namely, Helmholtz model, Guoy-Chapman point charge model and Stern model. In Helmholtz model a high potential gradient is proposed at the Helmholtz layer distance and the effect of the ions at larger distances are assumed negligible. In Guoy point charge model, the ions are treated as point charges and their locations and total charge are considered for the potential gradient prediction and thus a non-linear potential is assumed between the electrode and the ions in the electrolyte. The Stern model is basically a combination of Helmholtz and Stern models, where a linear potential gradient is assumed at Helmholtz layer, and a non-linear potential gradient due to the “diffuse layer” is assumed at larger distances in the electrolyte.
Figure 10.17 Models of double-layer supercapacitors (a) Helmholtz model, (b) Guoy-Chapman model (point charge model), (c) Stern model.
The electrolyte can play a very important role in the performance and functionality of a supercapacitor. The effects of various electrolyte factors on the performance of supercapacitors are presented in Figure 10.18 [Zhong et al., 2015]. A selection of aqueous and organic electrolyte based supercapacitors and their performances are also provided in Tables 10.5 and 10.6, respectively.
Figure 10.18 Effect of electrolyte on supercapacitor performance
adapted from [Zhong et al. 2015].
Table 10.5 Selection of aqueous electrolyte based supercapacitors and their performance at room temperature [Zhong et al., 2015].
| Aqueaous electrolyte | Electrode | Specific capacitance (Fg–1) | Cell voltage (V) | Energy density (Wlikg–1) | Power density (Wkg–1) |
| Strong acid | |||||
| H2SO4 (2M) | MMPGC | 105 (at 4 mVs–1) | 0.8 | 4 | 20 |
| H2SO4(1 M) | Microporous carbon | 100 (at 0.2 Ag–1) | 1 | 3.8 | 100 |
| H2SO4 (2 M) | Carbon nanofiber networks* | 204.9 (at 1 Ag–1) | 1 | 7.76 | 100 |
| H2SO4(1 M) | ANS–rGO | 375 (at 1.3 Ag–1) | 2.0 | 213 | 1328 |
| Strong alkaline | |||||
| KOH (6 M) | p-CNTn/CGBs | 202 (at 0.3 Ag–1) | 0.9 | 4.9 | 150 |
| Neutral | |||||
| Na2SO4 (1 M) | Mesoporous MnO2 | 278.8 (at 1 mVs–1) | 1 | 28.4 | 70 |
| KCl (1 M) | MnCl2-doped PANI/SWCNTs** | 546 (at 0.5 Ag–1) | 1.6 | 194.13 | 550 |
MMPGC: macro/mesoporous partially graphitized carbon; ANS: 6-amino-4-hydroxy-2-naphthalenesulfonic acid; rGO: reduced graphene oxide; p-CNTn/CGBs: porous CNT networks decorated crumpled graphne balls; PANI: polyaniline; SWCNTs: single-walled carbon nanotuibes.
*3D heteroatom and doiped
** Nanocomposite.
Table 10.6 Selection of organic electrolyte-based supercapacitors and their performance at room temperature [Zhong et al., 2015].
| Electrolyte | Electrode | Specific capacitance (Fg–1) | Cell voltage (V) | Energy density (Whkg–1) | Power density (Wkg–1) |
Carbon-based symmetric ES |
|||||
| 0.7 M TEABF4/ADN | AC | 25 (at 20 mV s–1) | 3.75 | 28 | – |
| 1 M LiPF6/(EC-DEC 1:1) | Heteroatom doped porous carbon flakes | 126 (at 1 Ag–1) | 3 | 29 | 2243 |
Pseudo-capacitative electrode material-based symmetric ES |
|||||
| 0.5 M Bu4NBF4/ACN | Heterostructured poly(3,6-dithien-2-yl-9H-carbazol-9-yl acetic acid)/TiO2 nanoparticles composites | 462.88 (at 2.5 mA cm–2) | 1.2 | 89.98 | – |
Asymmetric ES |
|||||
| 1.5 M TEMABF4/PC | Non-porous activated mesophase carbon microbeads/graphitized carbon | – | 4 | 60 | 30 |
| 1 M LiTFSI/ACN | MnO2 nanorodes–rGO/V2O5 NWs–rGO | 36.9 | 2 | 15.4 | 436.5 |
It should be noted that this diagram mainly shows the primary (direct) effects of the electrolyte. There are also secondary (indirect) effects that may need to be considered during the design and selection of the electrolyte. For example, the size of the ions in the electrolyte may affect the capacitance of the super capacitor, but it also influences the energy and power densities as they are both related to capacitance.
In general, electrolyte factors affecting the supercapacitor consist of electrochemical stability of the electrolyte, potential window, ionic conductivity, ion mobility, concentration, viscosity, ion size, and ion-electrode interaction. The electrolyte factors can directly and indirectly affect the power density, capacitance, energy density, thermal stability, cycle life and equivalent series resistance (ESR). A review of published modeling studies and strategies for the supercapacitor electrodes and electrolytes is provided by Aderyani et al., 2019.
10.8 Electrolytes for Fuel Cells
Fuel cells are another useful electrochemical energy storage system. Their input fuel material is oxygen, hydrogen, or air. The fuel undergoes reduction or oxidation similar to batteries. The main difference between fuel cells and batteries is that fuel cells are thermodynamically open systems while batteries are closed systems. Batteries are generally self-contained and do not require external materials, while in fuel cells, hydrogen or oxygen need to be re-supplied continuously. An exception is the oxygen battery which allows oxygen input. The electrolyte is generally a polymer membrane or solid inorganic depending on the type of fuel cell.
Figure 10.19 shows five common types of fuel cells and their respective chemical reactions at the anode and cathode. The main ion that transports in the electrolyte in PEMFC and PAFC is the hydrogen ion (proton). The main ions that transport in SOFC, AFC and MCFC are O2, OH-, and CO32– respectively.
Figure 10.19 Common types of fuel cells.
The electrolytes used vary depending on the fuel cell type. Table 10.7 shows the common electrolytes used for different fuel cells. It’s important to note that the electrolyte is designed and fabricated for optimum performance in a specific fuel cell.
Table 10.7 Electrolytes for common fuel cell types.
| Fuel cell | Alkaline (AFC) | Molten carbonate (MCFC) | Phosphoric acid (PAFC) | Proton exchange membrane (PEMFC) | Solid oxide (SOFC) |
| Aspect | |||||
| Electrolyte | Potassium Hydroxide (KOH) | Liquid Molten Carbonate (LiCO3 inLiAlO2matrix | Liquid Phosphoric Acid (in SIC matrix) | Sulfonated Organic Polymer (Nafion) | Ceramic (Yttria-stabilized) |
| Charge Carrier | OH– | CO32– | H+ | H+ | O2– |
| Operating Temperature | 90–100°C | 600–700°C | 150–200°C | 50–100°C | 700–1000°C |
| Fuel Efficiency | 60% | 50% | 40% | 60% | 60% |
| Power Density (mW/cm2) | 100–200 | 100 | 200 | 350 | 240 |
| Possible Applications | Military, space | Electric utility, distributed generation | Distributed generation | Transportation, portable power | Electric utility, distributed generation |
A very common solid polymer electrolyte used in proton exchange fuel cells (PEMFC) is the Nafion membrane. Nafion is also called a proton exchange membrane (PEM). Here, we’ll focus on PEMs and the proton conduction mechanism.
The proton conducting mechanism of Nafion can be described using a water channel model. This model is based on simulations obtained through the small-angle X-ray scattering data and the solid-state nuclear magnetic resonance studies. In this model, the hydrophilic functional groups (made of sulfonic acid groups) can organize themselves into channels of about 2.5 nm in diameter and 50 nm long. These channels allow the passage of hydrogen ions (protons). Since the channels can easily absorb water due to their hydrophilic nature, they are referred to as water channels (Figure 10.20).
Figure 10.20 Nafion water channel model for proton conduction: (a) top view of the hydrophilic water channels (b) side view of typically 50 nm long water channels.
Protons can transport well through the Nafion water channels, but if the Nafion is dried up, proton transport can become significantly suppressed. While the inner region of the Nafion channels are hydrophilic and can transport protons, the outer molecules (polymer backbone) surrounding the channels are hydrophobic crystallites as depicted in Figure 10.20. These outer region crystallites can provide good mechanical support and stability for the water channels. The figure shows the top and side views of the water channel model of Nafion.
The protons on the sulfonic acid groups (SO3H) in the Nafion membrane can also hop around from one site to another. While the cations (H+) pass through the porous pathways of Nafion through the hopping mechanism, the anions and electrons are effectively blocked and not allowed to transport through the membrane. This property of Nafion and other similar membranes is referred to as “permeation selectivity” or “permselectivity”.
Important Nafion properties include high ion conductivity (ion-exchange), thermal and chemical resistance, permselectivity, mechanical strength and insolubility in water. The Nafion structure can greatly affect the properties and functionality of the membrane including ion conductivity, water management, hydration stability especially at high temperatures, mechanical, thermal, and oxidative stability and electro-osmotic drag.
As mentioned earlier, the presence of water is essential in ensuring good proton transport in the Nafion membrane. Therefore, a humidifying (water management) system must be included with the hydrogen fuel cell, to keep the Nafion membrane continuously hydrated. Another strategy that can potentially reduce the cost of the water management system, and ensure higher material effectiveness is using nanofillers inside the Nafion membrane. The nanofillers can trap the moisture on their surface (interface between the nanofillers and the surrounding Nafion polymer), and thus, allow certain degree of self-humidification of the membrane [Kammoun et al., 2015].
Thorough reviews of fuel cells are provided in several published papers and books including [Kim, 2015] listed in the bibliography section.
10.9 Fillers and Additives
Fillers can be added to the electrolyte to enhance its properties. For example, nanoparticles can be added to the electrolyte to suppress dendrite growth, or make the electrolyte electrochemically more stable. Fillers in organic liquid electrolyte can especially affect the properties at the interface of the electrode with the electrolyte where unwanted chemical reactions may occur and the solid electrolyte interphase (SEI) layer forms. In general, fillers and additives in liquid electrolytes are designed to enhance the electrochemical performance, stability and safety of the liquid electrolyte and the respective energy storage device.
Fillers can also be added to solid polymer electrolytes. The main objective is to increase the ion conductivity of the polymer electrolyte. Inorganic non-ion conductive fillers with high aspect ratio (e.g., nanosheets or nanorods) have been shown to increase the ion conductivity of semi crystalline polymer electroylytes by increasing their amorphous and high conductivity regions. Fillers can have multiple benefits. For example, graphene oxide nanosheets added to polymer electrolytes for lithium ion batteries have been shown to i) increase ion conductivity ii) improve mechanical strength and iii) improve interfacial properties of the electrolyte and electrode inside the battery during operation, leading to both electrochemical and mechanical enhancements [Yuan et al., 2014; Kammoun et al., 2015].
Ion conductive ceramic fillers can also be combined with polymer electrolytes to form hybrid solid electrolyte systems. These materials have the potential to offer highly optimum properties due to the contributions of their constituents.