Novel proton and metal-ion conducting polymers and block copolymers
Viviani, Marco
DOI:
10.33612/diss.156496098
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Viviani, M. (2021). Novel proton and metal-ion conducting polymers and block copolymers. University of Groningen. https://doi.org/10.33612/diss.156496098
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Chapter 1
Climate change and the related need to reduce dependence on fossil fuels has brought increasing attention on alternative and renewable energy sources. International efforts to keep the increase in global average temperature below 2 °C above pre-industrial levels and, preferably, 1.5 °C by 2050 is one of the main goals of the “Paris Agreement”.1 This ambitious
target requires adequate policies and investments in favor of more sustainable energy resources and utilization. As an example, the EU recently signed the European Green Deal which paved the way to policies aimed at reaching carbon neutrality by 2050 and the reduction of the EU’s greenhouse gases emission by 55% by 2030 compared to 1990 levels.23
In general, to achieve a reduction in global warming, a substantial shift to “greener” energy sources and technologies is necessary to limit greenhouse gases (GHG) emissions. This involves important transformations, among which an energy transition toward electrification, renewable sources and hydrogen production/usage is expected. Harvesting renewable energy has several implications related to natural sources, making power generation less controllable and more difficult to manage in terms of energy distribution.4 Electrochemical energy devices
(EEDs) play a pivotal role in the energy transition, in particular, overcoming renewable energy instability issues and providing greener alternatives for power generation.
Figure 1.1a depicts different possible scenarios for future energy production/utilization
including some of the possible technologies involved. There are four main categories of electrochemical power sources: (I) batteries (primary (single use), secondary (rechargeable) and redox flow (RF)), (II) fuel cells, (III) supercapacitors (or electrochemical capacitors) and (IV) photovoltaics. The former two convert chemical energy generated by redox reactions into electrical energy, supercapacitors exchange electrical energy via electrostatic charge accumulation on the electrodes (capacitive interaction) and, finally, photovoltaics convert radiation energy into electrical power.
Figure 1.1 a) Classification of different types of energy production and storage technologies. Reprinted with permission from Ref. 8. b) Ragone plot illustrating the performances of specific power vs. specific energy for different electrical energy-storage technologies. Times reported in the plot are the discharge time, obtained by dividing the energy density by the power density. Reprinted with permission from Ref. 9. Copyright 2018 American Chemical Society.
A technological optimization of electrochemical devices is required to support future energy scenarios and to maintain the current standards in energy production/consumption. To better understand the last statement, the Ragone plot (Figure 1.1b) provides a useful and straightforward comparison between available technologies, in terms of their energetic performances. Power density (W L-1 or W kg-1) is plotted against energy density (Wh L-1 or
Wh kg-1). The Ragone plot clearly shows that, despite their environmental impact,
combustion engines still represent a benchmark in terms of combined energy and power output, without a single electrochemical device counterpart.5 However, a combination of
different technologies and technological advancements would shrink this gap.6,7
Ion conducting materials are an essential component in all EEDs as they represent one of the constituting components, i.e. the electrolyte (Figure 1.2).
Figure 1.2 Scheme of the four main electrochemical energy devices (EEDs). Adapted from Ref. 12 with
permission from The Royal Society of Chemistry and with permission from Ref. 5. Copyright 2004, American Chemical Society.
Depending on the EED in question, different types of electrolytes are needed. However, two common requirements hold in all cases, namely, complete electrical insulation and high ion conductivity.5 Ion conductivity requires dissociation and transport of ions. State of the art
technologies are mainly based on liquid electrolytes due to their higher conductivity. However, some applications such as fuel cells and capacitors already switched to solid electrolytes due to their lower safety and environmental risks. In general, switching from liquid to solid electrolytes comes at the expense of a substantial reduction in ion conductivity of the system, mainly due to the reduced mobility of the ions in the solid phase.
Nevertheless, ion conducting polymers are promising materials due to their flexibility in final applications, facile production, mechanical strength and ability to make intimate contact with the electrodes.10,11 Their solid-state enables applications not available for liquid electrolytes
(e.g. proton exchange fuel cells). Also, the substitution of the latter in batteries, provides substantial improvements in terms of safety and toxicity.12,13 Four main categories of ion
conducting polymers can be distinguished:14
- Category I: gel electrolytes. Polymer networks swollen with a classical electrolyte
- Category II: solid polymer electrolytes (SPEs) or dry polymer electrolytes. A mixture of
a polymer and an inorganic salt. The polymer is responsible for the solvation and transport of ions, e.g. PEO–LiTFSI in Li-ion batteries.
- Category III: dry single-ion conductors. One of the ions is covalently bond to the
polymer chain and the polymer must participate in ion solvation, e.g. PEO-b-P(STFSI). - Category IV: solvated single-ion conductors. Mixture of single-ion conductors and low
molecular weight solvents, e.g water in Nafion™.
In this thesis, we focus on solid polymer electrolytes (SPEs) (Cat. II) and solvated single-ion conductors (Cat. IV) for high energy applications, i.e. batteries and fuel cells.
Depending on the electrolyte category, for a given salt or solvent content, two main factors define the final ion conductivity: namely the chemical nature and the nanomorphology of the polymer.14–19
Considering neutral polymeric matrices, the overall ion conductivity is proportional to the number of charge carriers and their mobility:
𝜎 = ∑ 𝑛𝑖𝑞𝑖𝑢𝑖 Eq.1.1
where 𝑞𝑖 is the ionic charge, 𝑢𝑖 is the mobility of each charge carrier and 𝑛𝑖 is the number of free charged species (dissolved species) or charged mobile clusters. The number of charge carriers 𝑛𝑖, relies mainly on both dissociation energy (U) and dielectric permittivity (𝜀′) of the host material which governs the shielding of charges from each other and thereby influences ion pairing. Although the dielectric constant tends to be relatively low in polymer materials compared to liquids, salt dissolution is also related to the favorable complexation of the ions.20 This is particularly evident in systems like PEO, where structural factors
(chelating effects) play a crucial role.21,22
According to Eq. 1.1, the measured conductivity is not dependent on the nature of the ion. However, in most applications, only one ion is functional while the counterion is not. The transference number (ti) is a measure of the effective charge transported by the i-ion in the
absence of a concentration gradient:
𝑡+= (1 − 𝑡−) = 𝑢+
𝑢++𝑢− Eq. 1.2
The relevance of Eq. 1.2 is mostly limited to Cat. I and Cat. II, as for single-ion conductor
(Cat. III and Cat. IV) ti is unitary due to the immobilization of the counterion.
From these early considerations, it is evident that the chemical structure defines the crucial properties of the resulting electrolyte. In SPEs, high ion transport is usually ascribed to segmental mobility of the amorphous phase. Above the glass transition temperature (Tg), the
mobility of the polymer chains becomes significant and two mechanisms are responsible for ion conduction, namely, segmental motion and diffusion of the whole chains with the coordinated ions (relevant only for low molecular weight polymers).14 According to this,
fully amorphous polymers, with Tg as low as possible, are desired. On the other hand,
excessively low Tgs result in liquid-like behavior and reduced mechanical stability.
In this respect, the block copolymer approach represents an elegant strategy to decouple the mechanical properties and the segmental motion (ion conductivity) by combining rigid “inert” and soft conducting blocks.
Block copolymers (BCPs) are macromolecules consisting of two or more chemically distinct polymers covalently bound by their chain ends. When the constituent polymers are chemically incompatible, the enthalpic contribution is positive, overcomes the entropic term, and phase separation is observed.23,24 The presence of chemical bonds in between blocks
prevents macrophase separation, enabling the so-called self-assembly of block copolymers, i.e. the formation of nanoscale phase-separated domains with sizes in the order of the chain lengths (5 – 100 nm).
For a given temperature, phase separation depends on the Flory-Huggins parameter (), the degree of polymerization (N) of the blocks and the volume fraction of the block (f).25 The
product N determines the degree of phase separation defining three different segregation regimes:
a) Weak segregation regime N < 10,5
b) Intermediate segregation regime 10 < N < 100 c) Strong segregation regime N > 100
In the weak segregation regime, temperature affects the mixing enthalpy. Hence, with increasing temperature, the N product decreases to a critical value, where an order-to-disorder transition is observed (𝑇𝑂𝐷𝑇). By varying the volume fractions of an AB diblock copolymer, minimization of the interfacial energy between the blocks of the polymers leads to different morphologies: closed packed spheres (CPS), body centred spherical (BCC), hexagonally packed cylindrical (HPC), bicontinuous gyroid (GYR) and lamellar (LAM) (Figure 1.3).26 More complex geometries can be obtained by introducing one or more
chemically different blocks in the structure (ABC triblock, ABCD tetrablock, etc.) or by changing the architecture of the block copolymers (star and branched). The validity of the theories found experimental confirmations even though “real” phase diagrams differ from theoretical ones eventually exhibiting new morphologies (e.g. hexagonally perforated lamellae (HPL))27 or deviating from expected compositions when dispersity or charge effects
Figure 1.3 Calculated phase diagram of a neutral apolar symmetric block copolymer with related self-assembled structures. Reprinted from Ref. 28, copyright 2015 with permission from Elsevier. In the case of solid polymer electrolytes, appropriate choice of blocks’ chemistries may also result in selective distribution of the salts in the nanophase separated polymer. The relevance of self-assembled morphologies over conductivity for a system with a selective allocation of ions in one single phase can be estimated by the effective medium theory (EMT)29:
𝜎 = 𝑓𝜙𝑐𝜎0 Eq. 1.3
where 𝜎0 and 𝜙𝑐 are the intrinsic conductivity and the volume fraction of the conducting phase and f is the tortuosity or morphological factor. An ideal morphology value 𝑓𝑖𝑑𝑒𝑎𝑙 has been assigned based on the EMT29 for all four morphologies gyroid; lamellar; cylinders; and
spheres (Table 1.1).20
Table 1.1 Ideal morphology factors for 1D transport in phase-separated block copolymers with 𝝓𝒄<0.5
Morphology 𝒇𝒊𝒅𝒆𝒂𝒍
Gyroid 1
Lamellar 2/3
Cylinders 1/3
Spheres 0
From Table 1.1, it is clear how the percolation of the gyroid structure is expected to have the highest 𝑓𝑖𝑑𝑒𝑎𝑙 while the spherical morphology is ideally insulating. It is evident that nanomorphology plays a key role in the ion conductivity of a system.
In the case of dry and solvated single-ion conductor polymer electrolyte membranes, the chemistry of the polymer chains determines the acidity/basicity strength of the ionizable
groups and the interactions with the eventual solvent or other polymers. For solvated single-ion conducting polymers, the solvent is mainly responsible for the transport properties while the polymer must provide adequate percolation and mechano-chemical stability. The anticipated concepts related to morphology have also been widely applied to tune the nanostructure of polymer electrolyte membranes by a selective distribution of ionic groups in the polymer architecture or using a block copolymer approach. Major efforts have been made to recreate percolated structures where the tortuosity factor is maximized (Table 1.1). In the case of solvated single-ion conductors, nanophase separation also occurs in random copolymers if there is a strong incompatibility between the ionic groups and the polymer backbone. This is the case of perfluorosulfonic ionomers (PFSI) (or perfluorosulfonic acids, PFSA) where the phase separation and the nanomorphology is dictated by both the incompatibility and the swelling of the hydrophilic domains which defines their percolated structure.30
Although most EEDs already have benchmark electrolytes and commercial solutions, they have several drawbacks related to practical operations, safety, environmental impact and availability. Alternative solutions are needed among which novel polymer electrolytes have a primary role. The prevalence of emerging mobile and stationary energy devices strongly rely on novel ion conducting polymers capable of reducing EEDs cost while improving safety and performance. For these reasons, the field of polymers for energy applications has attracted much research interest. 9,11,16,20,31–38
To better understand the relevance of polymeric electrolytes in both fuel cell and battery applications, we will first describe the principles behind proton and metal-ion transport in such matrices, limiting the overview to the cases of interest for the present thesis.
1.1. PROTON-CONDUCTING POLYMERS AND SOLID POLYMER ELECTROLYTES Proton exchange membrane (PEM)
Proton exchange membranes are cation exchange materials bearing strong acid groups (i.e. sulfonic acid) capable of exchanging protons with the media in which they are immersed or in contact, without appreciable dissolution.39
PEMs are ionic polymers, macromolecules with ionizable sites covalently bound to the main chain or side chains of the polymer backbone. Depending on the charge content, a general distinction can be made between ionomers (≤ 10% charge content) and polyelectrolytes (> 10% charge content).39 PEMs are also known as hydrated acid polymers due to the necessity
of water molecules to ensure proton dissociation and transport phenomena. To maximize the proton conductivity, complete dissociation and high content of charges are desired but, on the other hand, sulfonic groups endow the polymers with high hygroscopicity and eventually solubility in water. To prevent this, an accurate choice of the chemistry and molecular design is necessary in order to maximize the acidity of sulfonic groups and exploit the intrinsic polymer phase separation which has a key role in the final performance of membranes.36,40,41
Acid doped membranes
Another kind of proton-conducting membrane is represented by acid doped polymers, polymeric matrixes doped with high boiling strong acids (H2SO4, H3PO4), able to conduct
under anhydrous conditions.42 Among others, polybenzimidazole-phosphoric acid (PBI-PA)
membranes represent the benchmark system of this category in terms of conductivity and physicochemical properties.43,44 This system strongly differs from the hydrated acids because
the ion exchange mechanism occurs in phosphoric acid instead of water. The anhydrous proton conductivity enables the use of acid doped membranes well above the water boiling point in applications such as high-temperature polymer electrolyte membrane fuel cells.43 Solid polymer electrolyte (SPE)
Solid polymer electrolytes are electrically conducting solutions of salt in a polymer39
meaning that SPEs are at least two-component systems in which both the polymer and the salt properties should be taken into account.
The polymers should maximize solvation and transport of the salt ions, which requires them to: (i) have a high dielectric constant; (ii) have high electron-donor characteristics, i.e. high concentration of sequential polar groups along their backbone ( e.g. –O–, –S–, –N–, –P–, C=O and C≡N)22; (iii) have an appropriate distance between coordinating centers; (iv) a
flexible backbone with low steric hindrance for bond rotation; (v) be easy to synthesize and (vi) have good interfacial contact with the electrodes.12 On the other hand, for a given
polymer, the nature of the salt determines the solubility and the relative mobility of the ions in the SPE. Salts with low lattice energy are those capable of forming complexes with a polymer host.22 Details regarding the proton-conducting mechanism in PEMs and metal-ion
transport in SPEs will be discussed in Section 1.3.
1.2. PROTON EXCHANGE MEMBRANE FUEL CELLS (PEMFC)
Fuel cells are electrochemical devices that generate electric power and heat by direct electrochemical oxidation of a fuel using oxygen from the air.45 Hydrogen fuel cells consume
hydrogen and oxygen to produce water, electricity and heat (Figure 1.4) resulting in a practically “zero-emission” power supply system. When alternative fuels (e.g. methanol, ethanol, 2-propanol, hydrazine, trimethoxymethane, propane46) are used carbon or nitrogen
oxides are also produced. However, their high efficiency (up to 60%), power density and (potential) pollution-free utilization47 make fuel cells extremely attractive power devices. A
Table 1.2 Comparison of Fuel Cell technologies 47 Fuel Cell Type Common Electrolyte
Operating Temperature (°C). Typical Stack Size (kW) Electrical Efficiency (LHV) Alkaline (AFC) KOH(aq) soaked in a porous matrix,
or alkaline polymer membrane <100 1–100 60% Polymer Electrolyte
Membrane (PEMFC) Perfluorosulfonic acid (PFSA) <120 <1–100
60% H2
40% ref. fuel Phosphoric Acid
(PAFC)
H3PO4 soaked in a porous matrix
or imbibed in a polymer membrane (PM) 150–200 5–100 (liq. PAFC) <10 (PM) 40% Molten Carbonate (MCFC) Molten Li2CO3, Na2CO3 and/or
K2CO3, soaked in a porous matrix 600–700 300–3000 50%
Solid Oxide (SOFC) Yttria-stabilized zirconia 500–1,000 1–2000 60%
Other possible classifications can be made based on the employed fuel (Direct Methanol Fuel Cell or DMFC) or the type of ion-exchanged by the electrolyte (Proton Exchange Membrane Fuel Cells or PEMFCs and Anion Exchange Membrane Fuel Cells or AEMFCs). In this thesis, we will focus largely on the polymer electrolyte membrane responsible for the proton transport in PEMFC.
The core of the PEMFC is the membrane electrode assembly (MEA) made of two catalytic electrodes pressed on the opposite sides of an electrolyte membrane and then placed in between two gas diffusion layers (GDL) (Figure 1.4, left).
𝑯𝟐
+
𝟏
𝟐
𝑶𝟐
→ 𝑯𝟐𝑶 + 𝑬𝒏𝒆𝒓𝒈𝒚
H298 = -286 kJ mol-1 S298 = 163,2 J mol-1k-1 G298 = -237,4 kJ mol-1 Etheo298= 1,23VFigure 1.4 Representation of the PEMFC single unit (left, reprinted with permission from Ref. 48. Copyright 2007 American Chemical Society) with the related hydrogen fuel cell reaction and thermodynamic variables (right).
The MEA constitutes the electrochemical reactor of the fuel cell. The electrodes allow gas permeation and catalyze the electrochemical reaction, namely the hydrogen oxidation
reaction at the anode and the oxygen reduction reaction at the cathode. Conductive bipolar plates are used to ensure electrical contact and collect electric power.
Since the voltage of each cell is relatively low, the elementary unit can be replicated and coupled with several other identical ones to create a stack which constitutes the real fuel cell.48
In the field of PEMFC, a sub-classification is possible based on the operative temperature:49,50
- Low-temperature PEMFC (< 70 °C)
- Intermediate temperature PEMFC (70 °C – 120 °C) - High-temperature PEMFC (120 °C – 200 °C)
To avoid ambiguity, it should be considered that these sub-categories only belong to the PEMFC which are all considered low-temperature fuel cells compared to the SOFC or MCFC (high-temperature fuel cells). PEMFC have been developed for three main applications: portable devices, transportation (fuel cell electric vehicle, FCEV) and stationary power generation (combined heat and power units, CHP). The PEM enables the transport of protons (H+) from the anode to the cathode where they recombine with reduced oxygen atoms (O2-)
to form water. Incomplete reduction of oxygen or crossover of this element on the anode side, lead to the formation of hydrogen peroxide (H2O2), which consequently decomposes to
aggressive hydroperoxyl (·OOH) and hydroxyl (·OH) radicals.
These radicals combined with the temperature and pressure inside the fuel cells, accelerate the degradation of the polymeric materials.51 Due to the aforementioned circumstances, the
PEM must satisfy stringent requirements to ensure high efficiency and durability of the device:34,52–54
✓ High ion conductivity at temperatures above 100 °C and below 0 °C; ✓ Electric insulation in all conditions;
✓ Low oxygen and fuel permeability;
✓ Adequate hydrolytic, thermal and oxidative stability; ✓ Low water transport via diffusion and electro-osmotic drag;
✓ Good water uptake, even above 100 °C or good ion conductivity at low RH;
✓ Good mechanical properties in all hydration states for reliable operation of MEA and to favor the fuel cell manufacturing processes;
✓ Satisfactorily low cost.
Besides the determination of the polymeric physicochemical properties, PEM characterization usually includes additional specific parameters that are listed in Table 1.3.
Table 1.3 Specific PEM characterization parameters
Parameter Symbol U.o.M. Description
Ion Exchange
Capacity IEC eq g
-1 The number of ionic groups on the ionomer backbone expressed
in terms of equivalent per gram of dry polymer Equivalent
Weight EW g eq
-1 Is the number of grams of dry polymer that contain one equivalent
of ionic groups and it is the inverse of the IEC
Water Uptake WU %
Is the weight per cent incremental ratio referred to the dry weight of the ionomer membrane after equilibration at a specific relative humidity (RH)
Hydration
number Number
The number of molecules of water that surround each sulfonic group at a certain RH
Degree of
sulfonation DS %
Is the number of sulfonic groups for each aromatic ring in the repeating unit of the polymer expressed as a percentage based on the unit value
Proton
conductivity S cm
-1
Is the ionic conductivity obtained by impedance measurements and corresponds to the ratio between the cell constant of the instrument and the resistance of the membrane
1.2.1. PROTON CONDUCTION MECHANISM IN HYDRATED ACID MEMBRANES
Understanding the transport phenomena behind ion mobility is of paramount relevance for the optimization of polymer electrolyte membranes. The peculiar behavior of charged polymers cannot be treated by extending the models valid for neutral macromolecules.17
In this work, we will focus our attention on the proton conductivity in a confined space, such as pores and nanochannels, within polymeric membranes.
For most proton exchange membrane applications, the solvents involved are water or hydroalcoholic solutions. Considering the hydrated acid polymer electrolytes, the final behavior and mobility of the ions are the result of the interplay between ions, solvent, membrane and, in the case of fuel cells, gases.17 The confinement effect generated by the
nanomorphology of channels as well as the high charge density inside the pores, lead to strong deviation from the prediction based on continuum theories valid for bulk solution and gases. Both theoretical55,56 and experimental approaches have been used to describe and
understand the exact transport mechanism in confined volumes with high charge density.57,58
Nafion™ and sulfonated poly(ether ether ketone) (sPEEK)59 have been extensively studied
as reference systems for PFSI membrane and sulfonated aromatic polymers, respectively. Even if the nature of the chain strongly affects the hydration behavior and the morphology of the pores, common features are recognizable for phase-separated polymers.17 In these
proton excess to follow a certain path inside the membrane produced by physical modification of the polymer structure depending on the water content. This is usually expressed as the hydration number () ([𝐻2𝑂]: [𝑆𝑂3𝐻]).
The physical dimensions of the nanochannels and the available volume for the transport mechanism depend on the water content () which in turns depends on the relative humidity (RH). For high values (> 10), the channels are broad enough to allow the proton excess to be transported via structure diffusion similar to the bulk water. For lower values ( < 6), the concentration of the protons is higher giving hyper-coordination. Also, the distance between the charged walls is smaller favoring the vehicle diffusion mechanism (Figure 1.5). Below = 3, the protons are considered undissociated and tend to remain in the proximity of the anionic sulfonic group.57
Classical continuum theories are inadequate to explain the morphological and transport features of proton exchange membranes. This is mainly because any inhomogeneity close to the charged interface is neglected.17 The dielectric constant () in the water phase shows a
gradual increase moving from the proximity of sulfonic groups to the middle of the channel where it reaches the conventional bulk value of 𝜀𝐻2𝑂 = 80,1 for 10 (Figure 1.5c).
60
Changes in the dielectric constant imply a different solvation property for the same ion depending on its location inside the pore and, subsequently, different coordination.
In liquid water, changes in temperature, pressure and concentration of H3O+ strongly affects
the ratio between the coexisting mechanisms. Structure diffusion is favored only by pressure (up to 0,6 GPa), while high temperature and high proton excess concentration inhibit this kind of mechanism by weakening the hydrogen bonding and providing hyper-coordination, respectively.17
Figure 1.5 Proton conduction in hydrated PFSA: (a) chemical structure and key structural factors influencing proton transport; (b) illustration of vehicular and hopping mechanisms; (c) distribution of proton concentration and dielectric constant across a hydrated domain for three water contents, λ. Adapted with permission from Ref. 40. Copyright 2017 American Chemical Society.
Inside PEM, the role of pressure is opposite to the one observed in homogeneous media. High pressures squeeze the membranes, increasing the water penetration and diffusion but reducing the volume of the pores and, consequently, inhibiting the structure diffusion mechanism. In the case of hydrocarbon aromatic polymers, the rigidity and the stiffness of
the molecular structures ensure less deformation under pressure effects, giving an advantage compared to PFSA.17
Apart from these general contributions, additional parasitic transports affect the overall ion conductivity of the material. Self-diffusion, chemical diffusion, permeation and
electroosmotic drag are phenomena related to the indirect spontaneous physicochemical
mechanisms that take place inside the nano-channels and interfere with the ideal expected behavior.17
1.2.2. LOW-TEMPERATURE PROTON EXCHANGE MEMBRANE (LT-PEM)
Nowadays, LT-PEM technology mostly relies on perfluorosulfonic ionomer (PFSI) electrolytes whose characteristics are illustrated below.
Perfluorosulfonic ionomers (PFSI)
In 1966, a major breakthrough in fuel cell technology was the discovery of PFSI membranes by E. I. DuPont de Nemours.61,62 Nafion™ was the name given to this family of ionomers.
Nafion™ is synthesized by copolymerizing tetrafluoroethylene (TFE) and an alkyl vinyl ether derivative with a sulfonyl fluoride group (Figure 1.6a)62 which is then converted into
sulfonic acid by subsequent hydrolysis.34 Envisioning the potential of this class of polymers,
other PFSIs similar to Nafion™ were developed by other companies such as Solvay Specialty Polymers (Hyflon™ or Aquivion™), FumaTech (Fumapem™) Asahi Chemical Company (Flemion™), Asahi Glass Company (Aciplex™), 3M™, Gore (Gore Membrane) and Dow (Dow Membrane™) (see Figure 1.6b). Nafion™ is a semicrystalline polymer insoluble in all solvents whose processability and characterization are limited using conventional methods. The molecular weight of Nafion™has been estimated at 105 -106 Da.30 However,
because of the uncertainty in the determination, the equivalent weight (EW) is used instead. The type of Nafion™ is defined by three digits that represents the EW divided by 100 (the first two) and the thickness expressed in thousandths of an inch (the third). For example, the most common Nafion 117 has an EW of 1100 meq g-1 and a thickness of 0.007 in. (≈ 175
𝜇m). The chemical structure consists of a PTFE backbone with perfluorosulfonic side chains separated by a variable number of repeating units depending on the EW. The TFE segments are mainly responsible for the partial crystallinity and mechanical properties of these ionomers. Concerning the side-chains, depending on the spacing of the sulfonic group from the backbone, it is possible to define long side chains (LSC) and short side chains (SSC) PFSI (Figure 1.6b).
The strength of the C-F bond (460 kJ mol-1) compared to the C-H bond (410 kJ mol-1)
provides exceptional chemical stability making PFSIs ideal candidates that can withstand the harsh conditions of fuel cells including low pH, radical oxidation, high temperature and high pressure.
Figure 1.6 a) Synthetic scheme of Nafion™ co-monomer synthesis62 and b) different perfluorosulfonic
ionomers (PFSI) molecular structures. Adapted with permission from Ref. 40. Copyright 2017 American Chemical Society.
The perfluorinated structure of PFSIs has a key role in their electrochemical stability and high performance in fuel cells that can be ascribed to: i) morphological features induced by the strong phase separation between the hydrophobic TFE backbone and the polar side chains bearing the sulfonic acid groups and ii) “super acidity” of the sulfonic group whose pKa value has been estimated around -6,59 three orders of magnitude below the first dissociation
of sulfuric acid (pKa = -3).
These observations make it clear that tuning the performance of these materials requires advanced knowledge of the nanomorphology and its relationship with the chemical structure and the proton conductivity.30
Understanding PFSI morphology
In order to rationalize the relationship between ionomer structure and its final properties, a wealth of studies have been dedicated to deepening the knowledge about PFSI membrane morphology.30,40 In particular, the dependence of the nanostructure on the hydration level
(i.e. ) is crucial to understanding the transport properties in PFSI and sulfonated polymers in general. Defining their nanostructure is a multi-scale problem which requires the complementary use of different microscopies including Atomic Force Microscopy (AFM), Transmission Electron Microscopy (TEM) and scattering techniques (Small and Wide Angle X-ray Scattering (SAXS, WAXS), Small Angle Neutron Scattering (SANS) (Figure
1.7a).30,40,63 Each technique has limitations and cannot provide comprehensive results by
itself. AFM offers the possibility to study in detail the surface of the sample neglecting the bulk morphology; whereas TEM has limitations regarding the size of the analyzed sample, dimensional resolution in the range 1-10 nm and restriction in investigating the effect of RH.41 On the other hand, X-ray and neutron scattering techniques offer the possibility to
analyze the bulk nanomorphology and its variation in time and at different RH. However, the necessity of models for the interpretation of scattering data in Fourier space is far from providing unambiguous results. The lack of a unique model able to define the nanomorphology of PFSI over the entire range of hydration explains the numerous studies on the topic and the “evolutionary” definition of the known interconnected structure, which is still the subject of scientific debate.64,65 X-ray scattering patterns of Nafion™ present three
main features in three separate regions. In order from low to high q range: the matrix knee, the ionomer peak, and the crystalline peak (Figure 1.7b). The matrix knee is related to the intercrystallite domain distance; the ionomer peak represents the correlation length for the hydrophilic domains inside the polymer defining the size of water domains; the crystalline peak appears in the WAXS region and correlates to the inter- and intra-crystalline spacing of the fluorocarbon chains in the crystalline domains. The presence and the shape of each peak is strongly dependent on the hydration (i.e. ), the polymer history (e.g. casting, annealing, hygrothermal ageing, etc.) and the EW.30 Considering the relationship between the scattering
vector (q), the wavelength (λ), the domain size (d) ( 𝑞 =4𝜋 sin (𝜃) 𝜆𝑥−𝑟𝑎𝑦 =
2𝜋
𝑑) and knowing the assignment of the peaks, it is possible to define the characteristic dimensions of the different domains and regions, but the major challenge is the definition of their structure, shape and connectivity.
Studies of the structural evolution of PFSI, from the dry state to solution (dispersion) state, suggest different morphologies based on the water content and the degree of swelling.
Figure 1.7 a) Schematic of Nafion™ at different length scales: 1) centimeter scale, 2) micrometer scale, observed using AFM, 3) nanometer scale, depicted using an artistic representation. The arrow reports a series of probing techniques for each length scale split into two categories: static with optical, electron and atomic force microscopies (OM, SEM/TEM and AFM respectively) and SAS and WAS techniques (top arrow) and dynamic with the quasi-elastic neutron scattering (QENS), NMR spectroscopy and diffusion techniques (bottom arrow). Adapted by permission from Ref. 63. Copyright 2008 Springer Nature. b) SAXS pattern of three different forms of Nafion™ and cartoons depicting the related domain size and nanomorphology. Adapted with permission from Ref. 40. Copyright 2017 American Chemical Society.
In the early 1980’s, Gierke66, Hsu67,68 and co-workers, proposed an interparticle cluster-network model to interpret SAXS and WAXS data of hydrated Nafion™ membranes. In this
nm) interconnected by narrow (1 nm) hydrophilic channels generating a percolated network (Figure .1.8a).66,67 Alternatively, Fujimura et al.69,70 compared a two-phase hard sphere and intraparticle core-shell models (Figure .1.8b) asserting that the latter was more likely to fit
the experimental data. However, the validity of Fujimura’s interpretation was questioned by the same group69 and later by electro-spin resonance (ESR) studies.71
On the other hand, the existence of hydrophilic channels between ionic clusters lacked experimental evidence 30 and the paracrystalline organization of the ionic clusters proposed
by Gierke’s model66 did not find large consensus.72–74 Alternatively, Roche72,73 et al.
suggested an inhomogeneous distribution of ionic domains upon hydration proposing an
intraparticle instead of a interparticle interpretation of the ionomer peak shifts.72,73 Kumar
and Pineri74 proposed a hard-sphere model where the sulfonic groups, thanks to the increased
elasticity of Nafion™ upon hydration, were also able to redistribute outside the clusters forming a new inhomogeneous domain distribution.74 The same authors also observed that
the small-angle upturn of the scattering spectra could be attributed to the electrostatic interactions between the ionic clusters.74
Figure 1.8 a) Graphical representation of interparticle cluster-network model. Reprinted from Ref. 67,
copyright 1983 with permission from Elsevier. b) Two-phase interparticle (left) and core-shell
intraparticle (right) models graphical representation. Adapted with permission from Ref. 69. Copyright
1982 American Chemical Society. c) Graphical representation of sandwich model. Reprinted from Ref.
75, copyright 2001 with permission from Elsevier, d) log-log representation of a standard small-angle X-ray scattering curve from a swollen Nafion 117 film with indications of the characteristic regions and slopes. Reprinted with permission from Ref. 76. Copyright 2002 American Chemical Society.
Signs of correlation between the upturn and the cluster spatial organization were also confirmed by the local order model elaborated by Dreyfus et al.77 Alternative models
resembling lamellar78 or “sandwich”75 (Figure .1.8c) morphologies were proposed to explain
the reversibility in water absorption/desorption and the linear scaling of both ionomer peak and cluster size upon hydration. However, the lamellar model did not consider the secondary correlation between the ionic domains74 which actually resulted in an anisotropic swelling
behavior.79,80 The sandwich model proposed by Haubold et al.75 did not provide an
unambiguous interpretation of the tridimensional hydrophilic-hydrophobic network.30
It was only in the early 2000’s that Gebel 81 proposed a conceptual description of the
transformation of Nafion™ morphologies upon swelling until dissolution (suspension). SAXS and SANS patterns of PFSI membranes and dispersion82,83 were analyzed using
different models depending on the hydration state. From the analysis, two regimes were defined based on the water volume fraction (w) and characterized by an inversion of the micellar structures: water-in-polymer (w < 0,5), and polymer-in-water (w > 0,5). For w << 0,5 the dry membrane is considered to contain isolated, spherical ionic clusters. With the absorption of water, the clusters swell to hold pools of water surrounded by ionic groups at the polymer-water interface. Between 0,3 < w< 0,5, a percolated structure of spherical ionic domains connected by cylinders of water dispersed in the polymer matrix is formed.
As the water content increases (w > 0,5), an inversion of the morphology occurs and the membranes transform into a connected network of rod-like polymer aggregates surrounded by water, as also suggested before by Lee et al.84 The spherical domain cannot be obtained
at w > 0,5 because the polymer volume associated with each ionic group is too large compared to the average distance between ionic groups along the polymer chain. The transition and the shapes assumed by the hydrated ionic network inside the PFSI resulted in being regulated by the polymer-water interfacial energy which imposes a specific surface of 55 Å281 (Figure 1.9). In an attempt to strengthen the hypothesis of the phase inversion at w
= 0,5 in absence of thermodynamic justification, Rubatat et al.76 extended the investigation
of the q range using Ultra Small Angle Scattering (USAS) techniques. Two additional trends in the intensity decay in the Porod region were observed. The log-log plot of the intensity presented a linear decay according to a power law of q-4 in between the SAS and WAS
regime, while according to q-1 power law at very low q. The former is characteristic for a
sharp interface while the second trend is associated with elongated rod-like objects (Figure
.1.8d).76 These observations brought the definition of a ribbon-like model for high dilution
of the PFSI in which locally flat ionic domains are interconnected through nodes and embedded in semicrystalline backbone domains. (Figure 1.9).
In 2008, Schmidt-Rohr and Chen85 proposed a new interpretation of Gebel’s data according
to a parallel cylinder morphology. According to this interpretation, also supported by NMR data, the nanostructure was organized in inversed micelle cylinders with large diameters (2.4 nm) even at water content as low as 20% (RH 80%). This model was criticized by Kreuer and Portale64 who found an erroneous assumption of the water fraction with a related
as proposed by Rubatat.76,86 A linear shift in the ionomer peak was observed instead, but on
the scale of 30-100 nm, Nafion™ morphology shows tortuosities strongly affected by pre-treatment associated with the crystalline domains.64 A layered polymer-water structure was
proposed for the local scale and w < 0,5. The water films act as a glue between negatively charged sulfonic groups of the PFSI stabilizing layered structures against the cylindrical model.
Figure 1.9 Evolution of Nafion™ nanomorphology upon hydration from dry membrane to dispersion in water. Adapted with permission from Ref. 40. Copyright 2017 American Chemical Society.
The validity of this assumption was limited to 0.2 < w < 0,6 suggesting that the progressive buckling followed by a phase inversion81 is expected only at w > 0,5 where the water
completely screens the charges removing the “glue” effect of the water film.64
The locally flat model found good consensus and, recently, 3-D direct imaging via TEM tomography of a fully hydrated Nafion™ membrane revealed the presence of the elongated flat ionic domains (Figure 1.10).65 Although a detailed description of PFSI nanomorphology
is still under debate, the main features have found reasonable explanations for their exceptional transport properties.
The copresence of sulfonic acid groups and the extremely apolar semicrystalline PTFE backbone generates unique nanostructures whose dimensions and shapes are strongly dependent on the hydration level. From the interpretation of USAS, SAS and WAS experiments suggest that a sharp polymer-water interphase delimits irregularly shaped and interconnected ionic domains having locally flattened geometries (especially at low hydration levels).
Figure 1.10 Bright-field cryo-TEM 3D reconstruction with two perpendicular slices through the tomogram shown; yellow marks the spatial distribution of the central region of the dark (hydrophilic) phase using isosurface rendering. Reprinted with permission from Ref. 65. Copyright 2015 American Chemical Society.
The retention of semicrystalline domains is crucial for mechanical properties and ion conductivity. Moreover, for a given EW, SSC PFSIs have longer TFE segments compared to Nafion™which lead to improved mechanical properties and higher Tg compared to LSC
enabling their use at T > 80 °C.87,88 PFSIs, although the benchmark technology for LT
PEMFC, have specific drawbacks that need to be addressed in order to facilitate the diffusion of PEMFC technology on a large scale. As described above, the ion conduction mechanism in hydrated acid membranes such as PFSI is strongly dependent on the Φ𝑤. This leads to restrictive conditions:
i. operating temperature in the range of 0 to 100 °C to avoid freezing or boiling of water and consequently membrane damage or dehydration.
ii. the conductivity of PFSI in most cases drops above 90 °C due to progressive dehydration and loss of mechanical properties of the polymer determining the collapse of the interconnected ionic channels.30 SSC PFSI showed better performances due to
iii. Besides these aspects, the performance of PFSI membranes are affected by radical attack (·OH, ·OOH)89 under operating conditions (Figure 1.11). Methanol crossover
is far from being negligible while their hard processability due to their insolubility constitutes further drawbacks along with their high price and environmental impact.
Figure 1.11 Schemes and reaction of PFSI chemical degradation occurring under oxidative attack in a PEMFC environment. Adapted with permission from Ref. 40. Copyright 2017 American Chemical Society.
Why higher temperatures?
In addition to the aforementioned electrolyte-related issues, the choice of a higher temperature for fuel cells offers several operative advantages compared to LT-PEMFC: 1) Lower CO poisoning of the Pt catalyst90
2) Limited H2 purification for integrated reformers91
3) Higher kinetics at the electrodes50,92
4) Better water management50
5) Improved heat management50,93
6) Lower catalyst loading94
7) Direct hydrogen95
8) Alternative non-noble metal catalyst systems50
All the benefits of higher temperature come at the expense of a harsher environment in the fuel cell with additional challenges:50,92,95
a) Faster degradation of the polymer electrolyte and loss of the catalyst resulting in an overall reduced lifetime of the system;
b) Long start-up time (up to 40 minutes for HT-PEMFC) that limit or exclude their use in certain sectors (i.e. automotive);50
c) Dehydration of the electrolyte requires humidification to aid proton conductivity to limit ohmic losses, alternatively different proton conductor media (i.e. phosphoric acid) can be used.
Despite their unique physicochemical properties, PFSI-based technology limitations are evident and hinder the diffusion of LT-PEMFC. In this work, we will focus our attention on the synthesis of alternative polymeric materials suitable for hydrogen intermediate and high-temperature PEM.
1.2.3. INTERMEDIATE TEMPERATURE PROTON EXCHANGE MEMBRANE (IT-PEM)
Fundamental studies on the structure of PFSI demonstrate that the excellent ion conductivity of sulfonated ionomers is a result of a complex interplay between nanostructure and hydration of the membrane. This enables the formation of a percolated structure, where ionic domains form an interconnected network of hydrated nanochannels filled with water. These structural features have been recognized as the benchmark nanomorphology for ion conduction. Over the past two decades, extensive studies have focused on the development of reliable, high-performance polymer electrolyte membranes able to address the PFSI-related downsides and especially their low performance at T > 80 °C.16,18,33,34,36,41,49,54,59,96–99 One of
the most promising alternatives is the use of fluorine-free sulfonated aromatic polymers (sAP). Different classes of polymers were studied as possible sulfonatable polymers including poly(p-phenylene), poly(p-xylylene), poly(1,4-oxyphenylene), poly(ether ether ketone) (PEEK), poly(arylene ether sulfone) (PAES), poly(phenylene sulfide) (PPS), polyimides (PI), polybenzimidazole (PBI).
Aromatic polymers present some advantages compared to PFSI: i) they are less expensive; ii) they possess higher thermal and mechanical stability; iii) sAPs containing polar groups have higher water uptake than PFSI over a wide temperature range; iii) decomposition of sAPs can be suppressed to a great extent by proper molecular design; iv) sAPs have better fuel barrier properties than PFSI.54 On the other hand, sAPs show weaker phase separation
compared to PFSIs because of the higher hydrophilicity of the backbones which has also lower oxidative stability due to weaker bond strength and substituent effects. The exceptional acidity of PFSI (pKa ≈ -6) compared to sAP (pKa ≈ -1)59 causes the former to have higher
proton conductivity for a given IEC. This requires the sAPs to have higher IEC to be competitive, which may imply excessive water uptake and swelling with consequent loss of mechanical stability of the membranes.36 Looking at the proton conduction mechanism, sAPs
have higher activation energy (Ea) compared to PFSI which makes them particularly suitable
for the IT-range where higher temperatures facilitate proton dissociation, under sufficient hydration.16
Although some of the aforementioned limitations appear to be intrinsic to the hydrocarbon nature of the sAP, appropriate chemical design and morphological effects have been demonstrated to effectively overcome several of the sPA drawbacks. Polymer architecture and chemistry allow tuning stiffness, filmability, acidity, hydrophobicity, nanomorphology and chemical resistance of the final material. In addition, nanostructural features define the state of water inside the membrane that tunes the proton transport properties (e.g. IEC, ) and the degree of swelling (i.e. mechanical properties) of the final membrane.16,18,100
Chemical stability issues
Sulfonated aromatic polymers are potentially susceptible to radical or hydrolytic degradation,101–103 which affect their durability.16 The main degradation reactions are oxo and
hydroxo radical attack and desulfonation via hydrolytic cleavage of the C-S bond (Figure
1.12a and 1.12b). The formation of radicals depends on the gas (fuel) crossover, which is
greatly suppressed in hydrocarbon membranes compared to PFSI. In particular, membranes with locally high IECs generally have a high dispersion of water and therefore a low dissolution and transport of gases.104,105
Figure 1.12 a) Different sulfonated polymer structures with black arrows indicating the most probable sites for radical attack. Adapted with permission from Ref. 101. Copyright 2011 John Wiley & Sons Inc. b) Reaction coordinates of the sulfonation/desulfonation reaction showing the stabilization of sAPs against hydrolysis by placing the sulfonic groups in ortho position relative to the EWG and increasing the Weland intermediate free energy. Adapted with permission from Ref. 106. Copyright 2007 American Chemical Society, and c) proposed mechanism for imide hydrolysis.
For what concerns the desulfonation (hydrolysis) reaction, the free energy of the transition state (Wheland intermediate or -complex) is mainly responsible for the stability of the
toward electrophilic aromatic substitution (EAS) by adding electron-withdrawing groups (EWG) is a valuable solution to stabilize sAPs against desulfonation, albeit being harder to sulfonate. The EWG (i.e. −SO2−) in ortho position to the sulfonic acid group increases the
hydrolytic and oxidative stability (Figure 1.12b). Another inconvenience of electron-donating groups in ortho position to sulfonic groups is their susceptibility to nucleophilic attack (hydrolysis) as a consequence of reduced electron density due to a resonance effect.101
In this regard, the EWG and electron-donating diverging effect call for an accurate balance also accounting for the hydrophilicity of each linker. In addition to radical attack and desulfonation, polyimide-based PEMs also suffer from high hydrolytic instability of the imide ring103 which undermines their use under fuel cell conditions (Figure 1.12c).
Another issue that must be taken into account for the synthesis and the treatment of PEM is the presence of metals and alien ions different from H+ which have detrimental effects for
the lifetime of PEM and its performance.107,108
1.2.3.1. SULFONATED POLY(PHENYLENE SULFIDE SULFONES) (SPSS) AND SULFONATED POLYPHENYLENE SULFONE (SPSO2) PROMISING IONOMER CANDIDATES
Sulfonated poly(phenylene sulfide sulfones) (sPSS) or sulfonated poly(arylene thioether sulfones (sPATS) have been proposed for the first time by McGrath et al.109,110 and lately
investigated by other groups.106,111–113 Since early studies, this class of compound appeared
to be suitable for fuel cell applications.111 High hydrolytic, thermal and chemical stability
characterize this class of materials and the corresponding sulfones (named as sPSO2). Moreover, the good fuel barrier properties and high proton conductivity even at low RH and high temperatures, make them ideal candidates for the intermediate temperature regime. sPSS have the additional advantages of being fluorine-free, relatively easy to synthesize, potentially cheaper and more environmentally friendly than PFSI. The latter two statements are supported by the fact that sulfur is considered a by-product of the petrochemical industry and produced in large excess compared to actual demand, making it cheap and with environmental concerns related to its massive storage.114 Recently, the direct use of elemental
sulfur as a greener building-block for polymer synthesis has received increasing attention.115,116 The catalyst-free synthesis of sPSS represents an additional advantage
compared to polymers and block copolymers obtained via metal-catalyzed processes (Ni-mediated oxidation, copper(I)-alkyne azide coupling (CuAAC) click reaction, ATRP, etc.) which might bring unforeseen consequences in the fuel cell environment.107,108 This implies
that the ideal synthetic strategy of PEM for fuel cell applications should avoid the use of metal-catalyzed reactions.
Synthesis
The synthesis of sPSS polymers could be performed using dichlorodiphenylsulfone (SDCDPS) or the fluorinated analogous
3,3’-disulfonate-4,4’-difluorodiphenylsulfone (SDFDPS) with different types of aromatic dithiols or even metal sulfides (i.e NaS2 or LiS2). The use of tetrasulfonated monomers with sulfonic groups
selectively placed in ortho position to the sulfone linkers have also been extended to the synthesis of sPSS polymers.117 The use of fluorine-terminated monomers is preferred due to
their higher reactivity allowing reduced reaction temperature and time.118 The purity of the
monomer is crucial for stoichiometric balance and control of the molecular weight in step-growth polymerization. The synthesis and handling of highly hygroscopic sulfonated monomers119, as well as the sensitivity to moisture and oxygen of the sulfurated organic and
inorganic compounds, constitute additional hurdles in the synthesis of sPSS that requires special care and precautions. Solution polymerization in polar aprotic solvents such as N-methyl-1-pyrrolidone (NMP),118 N,N-dimethylacetamide (DMAc)120 or sulfolane121 using
excess potassium carbonate (K2CO3) as a weak base have been proposed. A dehydration step
via azeotropic distillation with toluene is usually performed in the former cases while distillation-free polymerization has been reported in sulfolane. High reaction temperatures (160 °C < T < 200 °C) and different reaction times (5 – 36 h) are needed depending on the type of sPSS and the desired molecular weight.106,118,121 Purification of the polymer is also a
critical step, especially at high IEC where the water solubility of the polymer limits the possibility of separation from residual salt by precipitation. Depending on the necessity and the type of polymer, different purification techniques have been proposed among which dialysis is the best method, enabling the purification also from traces of reaction solvents.106
Varying the nature of the sulfide and the stoichiometry of the reagents, different copolymers can be synthesized. McGrath et al. firstly introduced the new sPSS (Figure 1.13a)109,118,122
which was lately reconsidered by Schuster et al. who expanded the family of sPSS to their analogous sPSO2 by simple oxidation of the sPSS precursor.106 (Figure 1.13a). Kreuer and
coworkers further broadened the sPSS scenario by copolymerizing SDFDPS with metal sulfides obtaining a new polyelectrolyte sPSS-204 and eventually its fully sulfonated analogue sPSO2-220 (Figure 1.13b).123,124 Jannasch et al, using tetra sulfonated monomers
with all the sulfonic groups in ortho position to sulfone linkages, obtained SPATSx which were then transformed and compared with their sulfone analogues (Figure 1.13c).117 The
same research group also synthesized an hypersulfonated sPSS/SO2 with octasulfonated repeating units (S125 and S142) (Figure 1.13d).125 A copolymer approach was instead
attempted by Lee et al. who explored the copolymerization of nitrile containing monomers to improve the water management and adhesion to the catalyst layer of the resulting polymers (Figure 1.13e).126
Figure 1.13 Chemical structures of different sulfonated poly(phenylene sulfide sulfones) and the corresponding oxidized sulfone analogues. a) sPSS-x and sPSO2-y (x,y = 312- 781) b) sPSS-204 and sPSO2-220 c) SPATSx and SPASx ( x= 41 – 100), d) SPAS1, SPAS2 and S142, S125, e) sPPSSfN.
Properties
The exclusive presence of sulfide and sulfone linking units alternatively to the ether or ketone groups of conventional sPAES or SPEEK bestows the resulting polymers with unique hydrolytic and oxidative stability.127 Water uptake and swelling of sPSS membranes
dramatically depend on the IEC (i.e. EW). However, WU of sPSS with 50% molar content of SDFDPS is lower than sPAES and sPEEK with an equivalent degree of sulfonation. But the proton conductivity is as high as 0.25 S cm-1 under fully hydrated conditions.127 Higher
DS adversely affects the swelling of the membrane under hydration, eventually causing its dissolution.111 Bai et al. 111,121 reported record conductivity of 0.36 S cm-1 and 0.215 S cm-1
measured at 65 °C and 85% RH for a sPSS with 100% and 80% SDFDPS content. By changing the sulfonic acid group placement and modifying the structure of the starting sulfonated monomer, Jannasch et al. obtained stable sulfonated membranes with IEC as high as 2.5 meq g-1 capable to retain proton conductivity of 10-2 mS cm-1 at 30% RH and 80 °C.128
A controversial trend in the conductivity was obtained in the case of hypersulfonated S125 (IEC ≈ 7 meq g-1) whose proton conductivity was lower than sPSO2-220 with IEC of 4.5
meq g-1.125 This unexpected trend of σ with respect to EW was explained in terms of
counterion condensation, responsible for reduced ion mobility at extreme DS due to the proximity of the sulfonic groups.129 This relevant feature suggested the existence of a
trade-off between IEC and physical spacing between the charged groups inside the polymer. However, above IEC of 1.8 meq g-1,the WU and swelling of the membrane increases until
complete dissolution in water for the highest possible content of SDFDPS. Oxidation of the thioether linkage to sulfone (sPSO2) brings serious advantages in terms of oxidative and dimensional stability under humidified conditions.106 However, excessive swelling or
dissolution of the membrane at high IEC are not prevented by simple oxidation of the thioether units. The rigidity and high degree of orientation130 of the -SO
2- linkage provides
additional mechanical strength but also exacerbates the brittleness of the polymers. On the other hand, the shallow rotational barrier of sulfide linkage promotes the phase separation of the sPSS. The sulfur atom is readily oxidized to sulfoxide and eventually sulfone in hydrogen peroxide, acting as an efficient radical scavenger in a fuel cell environment.131
Nanostructural studies via AFM revealed that in sPSS, with a molar content of SDFDPS above 40%, an inversion between the hydrophobic and hydrophilic domain exists resulting in the detrimental effect of hydration on the mechanical properties above that DS.118 WAXS,
SAXS and SANS analysis of various types of sPSS and sPSO2 revealed the formation of narrow ionic nanochannels with sizes comparable to Nafion™ (3.3 nm) for the TBBT-based sPSS in the dry condition (3.7 nm).132 In the “fully” sulfonated sPSO2-form, a progressive
narrowing of the domain size with increasing IEC is observed, passing from 1.91 nm of the sPSO2-360 to the 1.44 nm of the sPSO2-220 at 75% RH.105 Further increase in the IEC
brought to an opposite broadening of the ionic domains under the same hydration level, with S125 and S142 of 2.49 nm and 3.04 nm, respectively.125 Ion clustering and domain size are
fundamental properties that must be considered at different hydration levels because they define water management inside the membrane and proton conductivity.
Limitations and possible solutions
Similar to other sAPs, low EW sPSS and sPSO2 suffer from poor mechanical properties and poor water solubility, irrespective of the oxidation of the thioether linkages. In order to overcome these downsides while maintaining high proton conductivities, several approaches have been attempted including copolymerization104,112,113,117,133 and cross-linking (ionic or
covalent).127,134,135 Block copolymerization of sPSS or sPSO2 segment appeared to be the
most promising solution in terms of control over the properties of the PEM. The possibility to exploit the self-assembly of the resulting membranes and tune the nanomorphology allows further improvement in the proton-conducting properties. However, the presence of charges and the characteristic polydispersity of step-growth polymers complicate the predictability of the phase separation and the obtainable morphologies. Below, we report the main concepts related to charged and dispersed block copolymer self-assembly, together with successful applications of this approach for the synthesis of promising PEM materials.
Charged block copolymers and self-assembly
Among the most common copolymer architectures (Figure 1.14) the capability to give ordered phase-separated structure scales according to the following order: block copolymer
> graft copolymer > random.96 This explains the great interest in the block copolymer
approach in designing novel sAPs.20
Figure 1.14 Different possible copolymer architectures: random (I), alternating (II), block (III – VII) and graft (VIII) copolymers. Adapted with permission from Ref. 97. Copyright 2005 John Wiley & Sons Inc.
Most of the theories and experimental observations for block copolymer self-assembly relate to neutral and low disperse blocks. From a theoretical point of view, these two assumptions limit the prediction and the understanding of the phase behavior of polydisperse and/or charged block copolymers, which are particularly relevant in the field of sulfonated aromatic block copolymers for PEMFC.24,136
In the case of polydisperse systems, entropic contributions play major roles in the deviation from the predicted behavior for monodisperse systems. Firstly, the broadening of the dispersity index (Ð) reduces the entropic penalty associated with the chain stretching induced
to fill voids; this is because of the possibility to distribute long and short chains in the gaps and at the interfaces. In the second place, the configurational entropy of the block copolymers increases because short segments of the disperse block can be entirely pulled in the other block domain to relieve the chain stretching (Figure 1.15).24 As a result, block polydispersity
causes shifts in the composition windows associated with the well-known morphologies as compared to the corresponding monodisperse block copolymers.137–139
Figure 1.15 Schematic illustration of polydispersity in the center B blocks of ABA triblock copolymers results in: a) frustrated packing due to entropically unfavorable chain stretching; b) entropic relaxation with consequential interfacial crowding and desorption of short B chains from the interface and “swelling” of A domains; c) domain interfaces buckling toward the polydisperse domain, relieving chain stretching, thereby entropically stabilizing the microphase separated state. Reprinted with permission from Ref. 137. Copyright 2012 American Chemical Society.
In addition to polydispersity, the effect of charges on polymer self-assembly is relatively understood and is subject of great interest from both experimental and theoretical points of view.136,140 Charged block copolymers show strong deviation from the theory of neutral block
copolymers and self-assembly is dramatically affected by their charge content and distribution.140 Common theories cannot provide exhaustive explanations or predictions and
new parameters need to be computed such as the electrostatic interaction parameter (Г) that is dependent on other fundamental parameters such as the charge density, distance, ion valence and the dielectric constant of the polymer.136 Simulated phase diagrams based on
Self Consistent Field Theory (SCFT) and adapted strong stretching theory (SST)141 have been
proposed and in most cases, a distinctive “chimney” shape characterize these systems (Figure
1.16a-d). This anomalous self-assembly is characteristic of a “phase inversion” of the
charged block whose coalescence due to electrostatic forces inducing the formation of homogeneous phase segregation of the non-charged block even when the charged fraction (fa) is the minority. The shape and delimiting composition mainly depend on the charge
content and Г values. In this regime, percolated structures like inversed hexagonally packed cylinders with the charged domain surrounding the neutral cylinders are expected. 136 The
predicted phase diagrams allow the main geometries (sphere, cylinder, lamellae) and their inversed analogues. However, they do not exclude the presence of discontinuous phases in the chimney regime where interfacial tensions could be very low.141 The phase segregation
limit depends mainly on the charge fraction (density) and the electrostatic cohesion parameter that determine the binodal limits of miscibility of the two blocks. The higher the Г, the lower the and the Ncrit, above which phase separation is observed. That has also been predicted
