UPTEC K 20019
Examensarbete 30 hp
Juni 2020
Conducting redox polymers for
battery applications
Teknisk- naturvetenskaplig fakultet UTH-enheten Besöksadress: Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0 Postadress: Box 536 751 21 Uppsala Telefon: 018 – 471 30 03 Telefax: 018 – 471 30 00 Hemsida: http://www.teknat.uu.se/student
Abstract
Conducting redox polymers for battery applications
Mikael Svensson
The near future will put a lot of demand on the increasing need for energy production and storage. Issues regarding the modern-day battery technology’s environmental benignity, safety and cost to sustain such demands thus serve as a huge bottleneck, necessitating the research into alternative electrochemical energy storage solutions. Conducting redox polymers are a class of materials which combines the concepts of conducting polymers and redox active molecules to work as fully organic electrode materials. In this work three conducting redox polymers based on 3,4-ethylenedioxythiopene and 3,4-propylenedioxythiopene (EPE) with hydroquinone, catechol and quinizarin pendant groups were investigated. The polymers were electrochemically characterized with regards to their ability to cycle protons (aqueous electrolyte) and cations (non-aqueous electrolyte), their kinetics and charge transport and as cathodes in a battery. In non-aqueous electrolyte, hydroquinone and catechol did not exhibit redox activity in a potential region where the backbone was conducting as they were not matched. Quinizarin showed redox-matching as concluded by in situ conductance and UV-vis measurements when cycling Na+, Li+, Ca2+ and Mg2+-ions in acetonitrile. Comparison of the kinetics revealed that the rate constant for Ca2+-ion cycling was one magnitude larger than the rest, and galvanostatic charge/discharge showed that 90% of the polymer capacity was attainable at 5C. An EPE-Quinizarin cathode and metallic calcium anode coin cell assembly displayed output voltages of 2.4 V, and the presented material thus shows promising and exciting properties for future sustainable battery chemistries.
Populärvetenskaplig sammanfattning
Batterier är idag en sån självklar del av vår vardag att vi inte stannar upp och tänker på precis hur nödvändiga de är för ett modernt samhälle. Jorden och dess befolkning fortsätter snabbt att öka, och med det ökar efterfrågan på den energi som ska förse vår elektronik, våra fordon och hushåll med ström. Dagens batteriteknik är främst baserad på giftiga, sällsynta, dyra och i många fall oetiskt erhållna material som lämnar stora koldioxidavtryck längst hela deras livscykel. Att istället kunna basera batterier på de grundämnen som återfinns i naturen så som kol, syre och kväve, skulle bidra till en mer hållbar situation generationer framöver.
En typ av batterimaterial helt baserad på just dessa grundämnen är ledande redox polymerer. Polymerer är långa kedjor av repeterande molekylära enheter som de flesta känner igen i vardagen som plast. Plast är inte känt för att leda ström, men med rätt molekylstruktur kan detta uträttas, vilket skapar en elektriskt ledande ryggrad. Till denna kan sedan molekyler bindas in vilka kan lagra laddning genom att ges eller tas ifrån elektroner. Detta arbete syftar till att karaktärisera just denna typ av material för användning i framtida, mer hållbara batterier. Tre ledande redoxpolymerer med tre olika inbundna molekyler har utvärderats och jämförts med avseende på vilka joner som kan användas, hur snabbt elektroner överförs mellan dessa och deras potential i riktiga batteriapplikationer.
Contents
1 Introduction 1
2 Background 2
2.1 Battery basics . . . . 2
2.2 Conventional battery materials . . . . 2
2.3 Organic battery materials . . . . 2
2.4 Conducting redox polymers . . . .4
2.5 Cycling chemistries . . . . 5
3 Experimental 7 3.1 Chemicals and reagents . . . . 7
3.2 Polymerization and electrode preparation . . . 7
3.3 Electrochemical characterization . . . 7
4 Results and discussion 9 4.1 Polymerization and electrode preparation . . . 9
4.2 Proton cycling . . . 11
4.3 Metal cation cycling . . . 12
4.4 Kinetics . . . . 14
4.6 Effects on the formal potential and kinetics . . . 16
1
1 |
Introduction
The near future will put a lot of demand on the ever-increasing need for energy and energy storage solutions. This applies to everything from consumer electronics and electric vehicles to large-scale renewable energy production. The successful lithium-ion battery technology, however, may not be enough to satisfy these requirements as they rely on metals which are non-renewably sourced, fairly scarce and leaves a large carbon footprint across their entire manufacturing process. Not to mention the cost, these realizations have spiked an interest towards analogous battery technologies based on alternative electrode materials utilizing other cycling chemistries such as sodium and potassium,1–3 magnesium, zinc, calcium and
aluminium4–7 to name a few and even fully organic proton batteries.8
Organic materials can provide several advantages over their inorganic counterparts: they are made up of basic building blocks found in nature, which could potentially be easily accessible from biomass. Conversely, they could more easily be recycled and disposed of even by natural means through processes such as photodegradation.9 They can be more light weight than
inorganics and metals and could also offer smart, novel applications owing to their mechanical properties.10 One should however keep in mind that when demands such as energy density,
cyclability, sustainability, safety and most of all costs are being considered, one single battery chemistry might not satisfy all of these.11,12 However, organic materials could provide a rich
diversity of combinations to satisfy these demands in various ways.
The aim of this master thesis is to explore a novel class of organic cathode materials, conducting redox polymers (CRPs), for battery applications. CRPs are polymers based on a conducting backbone coupled with redox-active pendant groups. These type of novel materials combines the concepts of conducting polymers and conventional redox-active species which mitigates many of their individual weaknesses. Currently, the limited work on these materials has mostly been focused on and tailored towards proton8,13,14 or lithium3,15 cycling. In a bid to provide truly
sustainable electrode materials on-par with their inorganic counterparts, other chemistries than lithium needs to be explored. Herein, CRPs based on conducting EPE backbones with different quinone pendant groups were thus investigated with regards to their ability to cycle protons (aqueous electrolyte) and various mono– and multi–valent metal ions such as Li+, Na+, K+,
Mg2+ and Ca2+ (non-aqueous electrolyte). This work will serve to characterize CRPs for novel
2
2 |
Background
2.1 Battery basicsA basic battery consists of two electrodes separated by a porous membrane and a conductive electrolyte which can transport ions between the electrodes. An external circuit allows electrons to move between the electrodes. When a battery is charged, the positive electrode is oxidized while the opposite negative electrode is reduced. This allows a battery to store charge electrochemically, as energy is required to push electrons to the negative side and charge the battery. When the battery is being used, the energy is regained as electrical energy as the electrons spontaneously move to the positive side. The processes occurring at the respective electrodes are redox processes, and each electrode material has a redox potential at which their reactions occur. Thus, their difference in redox potential governs how much voltage we can get out of the cell. Another important battery characteristic is the specific capacity, which governs how much charge we can store per amount of material. The specific capacity therefore depends on how many electrons the material can transfer in a single reaction compared to its molecular weight.
2.2 Conventional battery materials
Cathode materials found in e.g. commercial Li-ion batteries are based on crystalline solid host structures which can store guest ions. Currently, commercialized cathode materials are most often based on polyanionic transition metal compounds or layered oxides such as LiNi0.33Mn0.33Co0.33O2, LiFePO4 and LiCoO2 to name a few, with capacities and average cell
voltages ranging between 100–200 mAhg–1 and 3–5 V vs Li/Li+ respectively.12 Lithium is an
excellent ion for these types of materials as it is small, mobile and has the lowest reduction potential of any element. Cathode materials for ions other than lithium such as Prussian blue analogues have also been explored for alternative sodium and potassium cycling chemistries with good progress,1,16,17 showing capacities around 150 mAhg–1 and discharge voltages around
3 V. Like most intercalation materials, however, the strain put upon the framework due to larger ion sizes does in many cases leave the capacity retention over many cycles lacking.
2.3 Organic battery materials
Some examples that have seen success in these types of applications include organosulfur compounds, radicals, conducting polymers (CPs), redox-active molecules (with carbonyls being the most successful in this regard)9,11,18 and various others. Herein, the materials of
interest will be those of CPs and redox-active species, both with different pros and cons related to them.
3
losing an electron, which dopes it either negatively (n-type) or positively (p-type). These charges are delocalized over a number of monomeric units which grows with increasing doping. Over a certain threshold, the polymer then becomes conducting.19 The inherent mechanism of
doping limits the CP to a certain amount of charge per polymeric unit usually in the order of 0.3–0.5, effectively limiting the gravimetric capacity of most CPs. Furthermore, the lack of localized charges results in redox potentials that rely on- and strongly varies with the degree of doping, resulting in varying cell voltages as compared to the plateau seen with discrete redox-active molecules.18
Figure 1. Positive doping (p-doping) of the conductive polymer EPE resulting in a conjugated structure of the
backbone which is conductive, counterions (X–) are present to counter the charge. A singly charged unit is called
a polaron while a doubly positive unit becomes a bipolaron.
Redox active molecules mainly based on carbonyls on the other hand exhibit known and well defined redox behaviors as seen in Figure 2, which can be easily tailored through organic synthesis. For example, it has been shown that substitution with either electron-withdrawing or -donating groups on the adjacent structure can shift the redox potential either higher or lower. This gives them great versatility as it allows them to act as both cathodes and anodes depending on the desired application. These types of compounds, however, exhibit low to no conductivity. Furthermore, their low molecular weight poses problems as dissolution into the electrolyte results in capacity fading over continuous cycling.
Figure 2. Redox conversion involving a 2-electron 2-proton reaction of benzoquinone, a carbonyl-class compound.
4
2.4 Conducting redox polymers
Combining a conductive polymer backbone with redox-active pendant groups is a strategy which can mitigate many of the issues discussed above. The possibility to covalently attach pendant groups of choice could open up their range of applications far beyond just battery materials. These CRPs could be tailored towards electrochemical storage, electrochromism, photosenzitation and various other sensory applications.19
In a CRP the redox active pendant groups provide the well-defined redox-active behavior and charge storage capacity. Meanwhile, the conductive polymer backbone provides an electronically conductive path through the material to make use of any generated charge (Figure 3). In this way, preferably, the individual characteristic properties of each component is self-contained and does not affect one another.14 For effectively working CRP materials, however,
certain interactions must be accounted for and conditions be met. As CPs are only conducting in their oxidized or reduced doped states, the redox reaction of the pendant group must occur in a potential region where the backbone is conducting, i.e. doped to benefit from any conductivity. This is known as redox matching.20 Furthermore, the choice of linker between the
pendant group and conducting backbone can have several effects on the functionality of the CRP. A conjugated linker, for example, can alter either the conductivity of the backbone or the behavior of the pendant group, effectively diminishing their individual properties.19
Investigations into a polypyrrole–quinone pendant CRP revealed that the quinone redox conversion caused twisting of the backbone due to the stiffness of the linker, negatively affecting the conductivity of the backbone.14,21 The choice of linker moiety can also affect the
redox potential of the pendant group as any substitution on the system can alter its energy levels.19
Figure 3. a) A schematic of a CRP. b) the structure of the CRP EP(1,4-QH2)E consisting of an EPE conducting
backbone and a hydroquinone redox-active pendant group bonded via a thio linker. The redox-active pendant group is responsible for the charge storage and well-defined electrochemical reactions while the conducting backbone transports the current to a current collector.
A very well-characterized class of pendant groups are the quinone family. Quinones are a class of carbonyl compounds with the simplest one being a benzoquinone (BQ). In aqueous, protic
5
electrolyte it undergoes a 2-electron, 2-proton redox reaction (Figure 3). Quinones are an attractive class of carbonyl compounds owing to their fast and reversible kinetics.19,22,23
The specific capacity which is related to the number of electrons transferred and the molecular weight allows quinones a great variety of energy densities. For the example of BQ above, the transfer of two electrons by reduction of the two carbonyl groups results in a theoretical specific capacity of 496 mAhg–1. Comparatively, it can be seen that the theoretical specific capacities
of 1,4-napthaquinone (1,4-NQ) and 9,10-anthraquinone (9,10-AQ) decrease to 339 mAhg–1 and
258 mAhg–1, respectively, as a direct consequence of the increasing molecular weight22 (Figure
4). While the smaller molecules theoretically can provide extremely high specific capacities, higher-molecular-weight materials or linkage to a polymer backbone can inhibit dissolution into the electrolyte. Furthermore, as was previously mentioned, the redox behavior of quinone compounds can be finely tuned by the introduction of electron withdrawing-groups (EWGs) or electron-donating groups (EDGs) to the aromatic core. An EWG has the effect of increasing the redox potential of the quinone as it makes the oxygen groups electron poor, thus making them more prone to accepting rather than donating electrons. Conversely, EDG groups will lower the redox potential.19 It should also be mentioned that the structure of the quinone
skeleton has similar effects. It is therefore reasonable that a balance between energy density and desired properties has to be balanced.
Figure 4.The structures of some common quinones and the effect of increasing molecular weight on the specific
capacity.
2.5 Cycling chemistries
As was mentioned previously, when energy density, cyclability, sustainability, safety and costs are being considered, no single battery chemistry will satisfy all of these requirements. For example, Emanuelsson and co-workers assembled an all-organic aqueous proton battery based on PEDOT-AQ and PEDOT-BQ conductive redox polymers.24 While definitely a leap forward
in the area of truly sustainable battery technology, the average cell voltage of 0.5 V limits the applications. Therefore, lithium-ion technology is still being investigated as it shows great promise in transitioning to more sustainable electrochemical energy storage systems. Various substituted benzoquinones have been shown to cycle lithium in a wide potential window, with electron-deficient perfluoroalkyl-substitued quinones demonstrating discharge potentials as high as 3.1 V vs. Li/Li+.22 The cycling stability, however, has shown mixed behaviors, ranging
6
from barely hundreds to thousands of cycles until major capacity decay.22,25 This is in most
cases attributed to a dissolution problem and not much work has been done on CRPs. Moreover, the substituting moieties in the example of BQ brings the theoretical capacity down from 496 mAhg–1 to 99 mAhg–1. Utilizing higher-molecular-weight quinones or CRPs could therefore
circumvent these issues.
Secondary batteries based on metal anodes ions such as Mg2+, Ca2+ and Al3+ could also offer
very attractive features for energy dense, cheap and safe batteries. Calcium and magnesium are extremely abundant and can offer electrodes with theoretical capacities of 1340 and 2000 mAhg–1 with standard redox potentials of –2.76 and –2.38 V vs. SHE, respectively.22,26
However, because of the often larger size and higher charge densities of multivalent ions, conventional cathodes based on intercalation materials can exhibit low performance because the species in question can strongly interact with the host lattice, resulting in slow kinetics and rapid capacity fading.4,5,12,27 Calcium, however, has been touted to offer better kinetics in these
regards due to its softer character.28 Because organic materials rely on coordination reactions
rather than intercalation and solid-state diffusion, they show great promise for these types of applications. For example, Gheytani and co-workers reported a 1.2 V, stable aqueous calcium battery based on a copper hexacyanoferrate cathode and a poly[N,N′-(ethane-1,2-diyl)-1,4,5,8-naphthalenetetracarboxiimide] (PNDIE) anode.5 This is no less true for quinones as well, as
their different redox properties and relative placements of coordinating groups could more or less effectively coordinate mono- and multi-valent cations as illustrated in Figure 5.6 While
organic electrolytes in these examples, however, would be preferred to avoid the stability window of water, issues with finding adequate organic electrolytes which allow a high salt solubility and conductivity has long hampered the progress.26,29
Figure 5.Universality of quinone chemistries, which could allow for aqueous proton batteries or non–aqueous
mono– or multi–valent metal ion battery applications.
7
3 |
Experimental
3.1 MaterialsFor the investigation of cation cycling, bis(trifluoromethanesulfonimide) (TFSI) salts of Li+,
Na+ and K+ (Aldrich), Mg2+ and Ca2+ (TCI) were used. The CRP monomers used in this work
were EPE trimers with a hydroquinone, catechol and quinizarin pendant group. All monomers were previously synthesized in-house as has been described elsewhere21.
3.2 Polymerization and electrode preparation
The compounds under study were polymerized by post-deposition polymerization. The respective monomers were dissolved in N-methyl-2-pyrrolidone (NMP) to a concentration of 0.1 mg/µL and ~0.1 µL was deposited with a pipette onto a glassy carbon (GC) electrode. The solvent was then evaporated under vacuum to leave behind a insoluble film, which could be polymerized by cyclic voltammetry. For aqueous cycling, polymerization was carried out by cyclic voltammetry at ambient temperature in a three-electrode beaker cell with a working electrode, a platinum wire counter electrode and an Ag+/Ag (AgCl) reference electrode. The
electrolyte was 0.5 M sulfuric acid. For non-aqueous cycling, polymerization was carried out in a similar way with a working electrode, a platinum wire counter electrode and an Ag+/Ag
(AgNO3) reference electrode in a separate compartment. The electrolyte was TFSI-salt in
acetonitrile (MeCN) and water in a 3:1 ratio. Values for each electrolyte against ferrocene can be found in Table 1. In a subsequent similar step the quinone pendant groups were deprotonated and the protons exchanged with the metal ion of choice. This was achieved by cyclic voltammetry in an electrolyte of TFSI-salt in MeCN together with 0.1 M pyridine. Electrodes were always thoroughly rinsed with MeCN between every step.
For interdigitated-array (IDA) measurements the monomers were polymerized onto an IDA-electrode with 90 pairs of Au bands on glass substrate (10 µm between bands, 10 µm wide, 150 nm high, MicruX Technologies, Spain). For UV-vis absorption measurements the monomers were polymerized onto an conducting indium tin oxide (ITO) glass slide. For all other experiments a glassy carbon (GC) electrode (Ø 3.0 mm) was used as the working electrode.
3.3 Electrochemical characterization
For proton cycling all polymers were characterized by cyclic voltammetry in 0.5 M sulfuric acid in a similar setup as described for polymerization. To investigate the influence of pH on the formal potential, cyclic voltammetry was conducted at a scan rate of 100 mVs–1 between
pH 2–14 at intervals of one pH unit. The pH was monitored using a pH-meter and varied using concentrated sulfuric acid and sodium hydroxide.
8
electrode and an Ag+/Ag (AgNO
3) reference electrode in a separate compartment was used as
reference.
To investigate the kinetics of the polymers, cyclic voltammetry was performed at various scan rates, between 10 mVs–1 and 3 Vs–1 for each TFSI electrolyte. Similar scan rate studies were
also performed between room temperature and 120 oC to investigate the kinetics dependence
on temperature. Conductance measurements were performed on polymers in situ during cyclic voltammetry using a bipotentiostat with a voltage bias between the two working electrodes of 10 mV. The conductance could in turn be determined from the current passing between the two electrodes. UV-vis absorption measurements were conducted with a three electrode setup like described above in a cuvette. An absorption measurement was taken after performing chronoamperometry and letting the current reach zero every 0.1 V between –0.6 V to 0.9 V. For characterization vs. metallic calcium all work was performed in a nitrogen-filled glove box. Calcium pellets were processed and folded onto a platinum wire to act as both reference and counter electrode. Polymers on GC working electrodes were characterized vs. metallic calcium in 0.1 M CaTFSI/MeCN in a beaker cell setup. For battery testing, coin cells were manufactured. 0.3 µL of EPE-Qz was polymerized and deprotonated as previously described in CaTFSI/MeCN onto graphite paper to work as a cathode while a thinly flattened piece of calcium metal served as the anode. The electrolyte used in the cells was 0.1 M Ca-TFSI in MeCN.
Table 1. Calculated potentials vs. Ferrocene for different electrolyte compositions.
Ion Electrolyte V vs Ferrocene
Li+ 0.1 M LiTFSI 0.08
Na+ 0.05 M NaTFSI 0.11
K+ 0.05 M KTFSI 0.077
Mg2+ 0.1 M MgTFSI2 0.017
9
4
|
Results and Discussion
4.1 Polymerization and electrode preparationThe CRPs were made up of two 3,4-ethylenedioxythiophene (EDOT, E) units coupled to a propylenedioxythiophene (ProDOT, P) central unit which forms the EPE conducting backbone. The central unit in turn accounts for the bonding via a thio linker to the pendant group. Three different quinone pendant groups were used in this work: a hydroquinone (1), a catechol (2) and a quinizarin (3) (Figure 6), respectively.
Figure 6. Structures of the three CRPs investigated in this work: an EPE–hydroquinone (1), an EPE–catechol (2)
and an EPE–quinizarin (EPE-Qz) (3).
The post-deposition polymerization process for all compounds gave a similar response due to having the same EPE backbone. Hydroquinone and catechol were polymerized in aqueous electrolyte for proton cycling studies as well as in non-aqueous electrolyte for metal cation cycling studies to avoid the stability window of water as seen in Figure 7. Their polymerization behaviors were identical, and thus only figures for EPE–hydroquinone is shown.
From Figure 8a, successful polymerization in 0.5 M sulfuric acid was interpreted from the appearance of two peaks centered around 0.55 V, attributed to the pendant group redox reactions. No more than three cycles were necessary to fully polymerize the material, as the current increased marginally with every additional cycle and a completely black film was produced on the electrode. Figure 7b shows the polymerization process in MeCN. The redox peaks are here located around 0.25 and –0.25 V with far less current being produced than for polymerization in aqueous electrolyte. This is a possible indication that the change in electrolyte solvent disturbs the redox matching of the pendant groups and the EPE–backbone.
O O S S O O S O O S HO OH O O S S O O S O O S OH OH O O S S O O S O O S O O HO OH
10
Figure 7. General polymerization behavior of EPE-hydroquinone and -catechol in 0.5 M H2SO4 (a), and in
MeCN (b) by CV at a scan rate of 100 mVs–1. (c) After deprotonation in 0.1 M pyridine/MeCN.
For cation cycling studies, polymerized electrodes had to be prepared by having the protons forcibly exchanged for any cation as protons provide a naturally stable reduced state for the quinones. This could be achieved by cycling the polymers in a good proton acceptor such as pyridine. From Figure 7c however, this seems to have had no effect on either EPE-hydroquinone or -catechol as the only remaining signal came from the backbone, evident by the low current.
From Figure 8a, successful polymerization of EPE-Qz in non-aqueous electrolyte could be interpreted from the irreversible build-up of capacitance seen between –0.45 V and 0.2 V during scanning which marginally increases as the material became fully polymerized. This is concomitant with increased doping currents as more material is being polymerized onto the electrode, and only a few cycles were necessary to form a black homogenous film. For cation cycling studies, EPE-Qz also had to be deprotonated beforehand. As seen in Figure 8b, a large peak centered at 0.4 V is seen on the anodic sweep of the 1st cycle when cycling quinizarin with
11
Figure 8. (a) General polymerization behavior of EPE-Qz by CV at a scan rate of 50 mVs–1 in 1:3 water/MeCn..
(b) Deprotonation and cation exchange of EPE-Qz in MeCN and 0.1 M pyridine at a scan rate of 20 mVs–1. The
large peak at 0.4 V disappears after the initial cycle and is attributed to the removal of protons.
Depending on the compound, a number of factors were observed to contribute to a successful polymerization: electrolyte composition, monomer concentration and layer thickness. The smaller and more polar hydroquinone and catechol were easily polymerized in sulfuric acid, while in the case of the quinizarin a 1:3 ratio of water/MeCN provided for the best polymerization conditions. This can be rationalized by the bigger and less polar Qz pendant group requiring a certain degree of solvation for the entire material to be reachable and the resulting network not to dense. However, at too high concentrations of MeCN in the electrolyte the film rapidly dissolved. For all monomers a high concentration of at least 0.1 mg/µL of NMP was optimal. At lower concentrations the films either dissolved or exhibited low currents as a result of the disperse and less cohesive polymer network. Deposition of too much material on the other hand led to large amounts of material not being polymerized as it could not be reached by the electrode surface. This could be observed visually as some areas did not turn black.
4.2 Proton cycling
Figure 9a shows the anodic and cathodic response of hydroquinone (red) and EPE-catechol (blue) in sulfuric acid during cyclic voltammetry, which correspond to a two–electron two–proton transfer reaction. The formal potentials could be calculated to be 0.47 V for EPE-hydroquinone and 0.57 V for EPE-catechol, a 100 mV potential difference. With the only structural difference being the position of the hydroxyl groups (para position for hydroquinone and ortho position for catechol) with a similar backbone and linker, the catechol with its hydrogens in ortho position exhibits a more stable reduced state. This is likely the effect of the ortho-positioned hydrogens ability to in a greater extent pull electron density away from the aromatic core, which has been shown with other electron-withdrawing substituents to shift the redox potential higher.30 This observation has an important implication, in that it can help
12
coordinated metals such as Ca2+, Mg2+ or Zn2+ would most likely also result in the pendant
group exhibiting a more stable reduced state as compared to being singly coordinated.6
The formal potential E0 of quinones depends heavily on the pH (Figure 9b) The differences in
formal potentials here are attributed to the relative positions of the hydrogens as discussed previously. Both quinones show a pH dependency, which when extrapolated across the entire pH range exhibit an ideal slope close to 59 mV/pH (–62 mV/pH for hydroquinone and –57 mV/pH for catechol). This is indicative of a solely proton-coupled reaction, with a one-to-one stoichiometry between protons and electrons. From the extrapolation to pH 1, which means a proton activity of 1, the formal potential could be calculated to be 0.50 V for hydroquinone and 0.59 V for catechol. For catechol, polymer degradation most likely occurred beyond pH 7, which could be the of result of the higher reactivity of a ortho-substituted ring. Between pH 7 and 10 a break in the trend is observed for hydroquinone before it is regained again after pH 10.
Figure 9. (a) CV of the hydroquinone (red) and the catechol (blue) performed at a scan rate of 100 mVs–1 in 0.5
M sulfuric acid. Catechol exhibits a higher reduced state, with E0’ 100 mV above hydroquinone. (b) The
pH-dependency of E0’ with pH for both polymers. By extrapolation, the formal potentials E0 could be calculated to be
0.50 V and 0.59 V for the hydroquinone and the catechol respectively.
4.3 Metal cation cycling
13
Figure 10. CV of hydroquinone (a) and catechol (b) in MeCN containing 0.1 M CaTFSI, performed at a scan rate
of 100 mVs–1 in a wide potential window.
EPE–Qz showed redox activity in MeCN and the formal potential could be calculated from the anodic and cathodic peak responses seen, which varied largely with the cation being cycled (Figure A1a-d, Appendix). Figure 11 shows the calculated formal potentials of the cycled cations in relation to each other. The read area represents the hard limit at which the EPE– backbone becomes completely insulating. For quinones in general, the complete conversion between oxidized (Q) and fully reduced state (QM2) occurs in two steps with an intermediary
semi-quinone radical state (QM•). The two peaks, or sometimes shoulders present on the anodic and cathodic sweeps can therefore be attributed to a QM2/QM• conversion, while the second
peak is attributed to the QM•/Q conversion. The formal potential E0’ could be estimated for the
QM•/Q conversions, but not for each of the QM2/QM• conversions. The formal potential and
degree of splitting between the two conversions depend on the degree to which the coordinating cation stabilizes the reduced and radical states; e.g. the higher electron-withdrawing capability of Li+ on the aromatic core compared to that of Na+ shifts the redox potential of the second
conversion to higher potentials. Larger cation sizes also show similar effects, which was observed by the larger splitting of Ca2+ with its larger ionic radius but relatively low charge
density, and Mg2+ which is somewhat larger than Li+ but has a much higher charge density. As
observed for the proton-cycling quinones, the two redox conversions cannot be discerned as H+
14
Figure 11. Formal potential of EPE-Qz cycling K+-ions, Na+-ions, Li+-ions, Ca2+-ions and Mg2+-ions. Performed
at a scan rate of 20 mVs–1 in MeCN containing 0.05 M KTFSI, 0.05 M NaTFSI, 0.1 M LiTFSI, 0.1 M CaTFSI
and 0.1 M MgTFSI respectively.
4.4 Kinetics
To investigate the kinetics and nature of charge transfer of EPE-Qz cycling different cations, cyclic voltammetry experiments were conducted at scan rates v between 20 mVs–1 and 3 Vs–1,
as seen in Figure 12a–d. At sufficiently high scan rates, the anodic and cathodic peaks starts to largely drift apart; however, the logarithmic dependence of the peak currents with scan rate reveals a slope of 1, indicating that the reactions are not diffusion-limited by the polymeric film at these scan rates. At scan rates where the peak-to-peak distances drift apart more than 200 mV, a logarithmic dependence of the peak potentials with scan rate appears as shown in the insets in Figure 13a-d. From these linear regions, an apparent rate constant, k, could be obtained31 for each cation being cycled according to Eq (1), thus the kinetics could qualitatively
be compared. The result of the calculated rate constants for oxidation and reduction are summarized in Table 2.
log k = 𝛼 log(1 − 𝛼) + (1 − 𝛼)log𝛼 − log ,/01-.2 − 𝛼(1 − 𝛼)/0∆45
6.8-. (1)
Table 2. Calculated rate constants for EPE-Qz cycling different cations.
15
Figure 12. CV of EPE-Qz cycling Li+ (a), Ca2+ (b), Mg2+ (c) and Na+-ions (d). Performed at scan rates between
10 mVs-1 and 3 Vs-1 in MeCN containing 0.1 M Li-TFSI, 0.1 M Ca-TFSI, 0.1 M Mg-TFSI and 0.5 M Na-TFSI
respectively. The insets show the peak potentials (Ep) dependence on scan rate. From the linear regions the transfer
coefficients 𝛼 could be estimated and the rate constants 𝑘; calculated. (e) dependence of the peak currents (i
p) on
scan rate (v). For all cations a linear trend with a slope of 1 appeared, indicating no diffusion contributions to the kinetics.
The scan rate study reveals that Ca2+ exhibits rate constants for the oxidation and reduction
about one hundred times larger compared to the rest of the ions. Furthermore, the separation of the two redox conversions are still detectable while the two peaks merge for other cations. Meanwhile, the other metal ions exhibits similar rate constants, with Mg2+ exhibiting the lowest.
This means that for e.g. Ca2+–cycling, the entire material is converted in 40 seconds, while it
would take almost 9 hours for the Mg2+–cycling. No values could be determined for K+–cycling,
as most of the signal came from capacitance. The decreasing values for the rate constants follow a similar trend as observed for the increasing values of ∆E which relates to the reversibility of the reaction. For comparison, other studied quinones cycling protons report rate constants from 6 s–1 up to 27 s–1, which varied with the linker being used.14,19 However, here an identical thio–
16
4.5 Effects on the formal potential and kinetics
It has previously been shown for catechols cycling metal cations, that an increasing valency results in stronger interactions between the chatechol and metal, with the formal potential shifting higher according to M+ < M2+ < M3+. Concomitantly, the calculated rate constants saw
a reverse trend, owing to the increasingly sluggish kinetics of more strongly bonded metal cations.6 This gives a plausible explanation for EPE-Qz cycling different metal cations, as the
calculated formal potential increases roughly according to an increasing cation charge density: K+ < Na+ < Li+ < Ca2+ < Mg2+. However, for the case of the Qz pendant-group, it is not known
whether it has the possibility coordinate multivalent ions as can a catechol via its two adjacent oxygen groups. Furthermore, the calculated rate constants show that EPE-Qz cycling Ca2+
strongly deviates from the trend, having a rate constant a thousand and hundred times greater than Mg2+ and Li+ respectively.
Scan rate studies were performed between room temperature and 100 oC to relate the rate
constants (and formal potential) to an energy barrier ∆Ea. Figure 13 shows the fitted Arrhenius
plots of Li+, Ca2+ and Mg2+. From the slopes, activation energies of 0.28, 0.84 and 0.91 eV
could be calculated for Ca2+, Li+ and Mg2+ respectively. The lower activation energy of Ca2+–
cycling can be thought to be mirrored in the higher calculated rate constant.
Figure 13. Temperature dependence of the rate constant of EPE-Qz cycling Li+, Ca2+ and Mg2+–ions. All data
points were collected at a scan speed of 1 Vs–1 in MeCN containing 0.1 M LiTFSI, 0.1 M CaTFSI and 0.1 M
MgTFSI respectively. The negative slopes were used to calculate the activation energies using the Arrhenius equation, which were calculated to be 0.28 eV for Ca2+, 0.84 eV for Li+ and 0.91 eV for Mg2+.
17
While these results do not take into account complete solute-solvent interactions, it is here thought to not be a major contribution to the formal potential since the cation with the greatest value of ∆G should also be more prone to stay in solution, i.e. exhibit the lowest reduced state. However, neither the activation energies calculated for Li+ and Mg2+ nor the Born solvation
energy provides an satisfactorily explanation for the differences in either the kinetics or formal potential.
Nonetheless, with these encouraging results, Ca2+ was chosen for further investigation as it is
thought to be an ideal candidate for future sustainable battery chemistries.
4.6 In situ conductance
Conductance measurements were performed by polymerizing and deprotonating EPE-Qz in CaTFSI, as well as LiTFSI for reference onto an IDA electrode. Measurements were conducted
in situ during cyclic voltammetry to confirm a good redox-matching between the quinizarin
pendant group and the EPE backbone, as well as to acquire insights into any pendant group– backbone interplay during cycling (Figure 14).
Figure 14. Conductance and CV data of EPE-Qz cycling Li+-ions (a) and Ca2+-ions (b). Both were performed at
a scan rate of 20 mVs–1 in MeCN containing 0.1 M LiTFSI and 0.1 M CaTFSI respectively, using an IDA-electrode.
Conductance data during cyclic voltammetry reveals that for both of the widely differing cations, there is indeed redox matching since the redox reactions occur in a region where the polymer backbone is conducting. The similarity in polymer doping onset potentials where the conductance starts to increase, which lies at –0.4 V vs Fc+/0 also reveals that the polymer
18
going towards higher potentials, a large decrease in conductance is observed in both cases, which could be attributed to polymer degradation as the hysteresis is quite large and the conductance maximum is not regained.
UV-vis spectroscopy was therefore employed to more closely examine the doping process of EPE-Qz during Ca2+ cycling. By performing an absorbance measurement between –0.6 V and
0.9 V every 0.1 V, the evolution of the different energy transitions corresponding to the bandgap and polaron/bipolaron states could be closely followed. At each potential step, chronoamperometry was performed until no more current was generated, indicating that the doping process was complete. Figure 15a shows the absolute absorbance spectrum between 400 nm and 1100 nm. In figure 15b the absorbance data points can be seen overlayed with the previously conducted in situ conductance measurement.
Figure 15. (a) In situ UV-vis absorbance between 400 nm and 1100 nm of EPE-Qz cycling Ca2+ during redox
conversion, coated onto an ITO glass slide. A UV-vis measurement was taken at each potential step after performing chronoamperometry until no current was observed. The absorbance at 2.15 eV corresponds to the band gap energy of the un-doped backbone, while the 1.23 eV absorbance corresponds to the bipolaron energy states.
(b) Absorbance and conductance data showing decreasing band gap absorbance and increasing bipolaron
absorbance going towards higher potentials.
As seen in figure 15a and 15b, UV-vis indicates a steady decrease in the absorbance of the 2.15 eV transition related to the bandgap of the un-doped backbone at increasing potentials, which indicates a process of continuous doping. Meanwhile for the 1.23 eV transition, related to bipolaron states,30 the absorbance sees a small increase near the onset doping potential, with a
19
practically make the backbone insulating.32 The conductance response can therefore be
attributed to an interband/intraband hopping mechanism at high levels of doping rather than along the backbone, and polymer degradation can thus be excluded. Still, the observed conductance behavior could pose problems in any real battery application as the uneven anodic and cathodic current distributions would affect the achievable capacity over continuous cycles. Increasing the mean polymer chain-length by either improved polymerization conditions or larger monomeric units could likely abate these effects.
4.7 Battery characterization
Figure 16 shows the galvanostatic charge/discharge conducted on a EPE-Qz electrode cycled in CaTFSI, which shows two plateaus upon charge and discharge which relates to an anodic and cathodic QM2/QM• and QM•/Q conversion as seen during cyclic voltammetry. The first
plateau/conversion upon charging is poorly defined compared to the rest, which is in agreement with the smaller peak seen during cyclic voltammetry. At a current rate of 5C it can be seen that >90% of the theoretical capacity, calculated by the amount of material put onto the GC electrode, could be obtained upon charge, with ~85% being achievable upon discharge. Thus, the material could be effectively prepared as a cathode.
Figure 16. Galvanostatic charge/discharge curve of EPE-Qz electrode cycling Ca2+-ions, performed at a current
rate of 5C.
EPE-Qz electrodes prepared onto GC electrodes were cycled inside a nitrogen filled glove box against metallic calcium at 20 mVs–1. The cyclic voltammogram in Figure 17a shows the anodic
and cathodic responses with similar current contributions, and a large peak split, ∆Ep of 2.36 V
20
reveals two plateaus upon charge and discharge at potentials corresponding to the quinizarin conversions also seen during cyclic voltammetry.
Figure 17. (a) CV of EPE-Qz against metallic calcium in 0.1 M Ca-TFSI in MeCN at a scan rate of 20 mVs–1, the
formal potential was calculated to be 2.35 V (b) Galvanostatic charge/discharge of EPE-Qz/calcium coin cells at 0.5C and 5C.
As seen in Figure 17b, galvanostatic cycling of the coin cells showed the initial capacity to be about 15% of the theoretical capacity of EPE-Qz (71 mAhg–1) at 5C, which slowly decreases
over 25 cycles. The capacity fading is here attributed to the poor choice of electrolyte as well as the limiting anode side, where the kinetics of the metallic calcium anode is simply too slow for the cathode side to fully utilize its full capacity. The calculated capacity was, however, calculated to be 69 mAhg–1 for both charge and discharge for all cycles at 0.5C, implying that
almost the entire material could be active. Unfortunately, not enough time was available to properly conduct a C-rate study and the results are thus not entirely trustworthy.
The large resistance present in the system can be attributed to a wide range of factors. In general, the use of calcium metal anodes has been plagued with optimization issues regarding suitable electrolytes which allow a high salt solubility, reasonable desolvation energy and conductivity. The inability of the Ca2+-ion to diffuse through the passivation layer or SEI due to its large size
21
5 |
Conclusions
With the ever increasing need for energy globally, the resource scarcity, cost and large carbon footprint of currently metal–based battery technology serves as a bottleneck. Furthermore, the transition to alternative cycling chemistries based on Na+, Mg2+, Ca2+, Zn2+ and others is
plagued with insufficiently functioning intercalation cathode materials. In this work a novel class of organic compounds for sustainable electrode materials, conducting redox polymers, have been investigated. Three polymers based on a EPE conducting backbone with hydroquinone, catechol and quinizarin pendant groups were evaluated for optimal polymerization conditions and electrode preparation, their cycling chemistries as well as prototype battery testing. The different pendant groups have capabilities of cycling either protons or various metal cations. However, it is evident that the requirement of a backbone-pendant redox matching is a hard condition that limits these systems and their flexibility, evident by the hydroquinone’s and catechol’s inability for cation cycling going from aqueous to non-aqueous electrolyte. For the redox matched quinizarin, Ca2+-cycling revealed superior
kinetics to Na+, Li+ and Mg2+. From batteries assembled using an EPE-Qz cathode with metallic
calcium functioning as anode, it is showed that EPE-Qz have the potential to work in a full cell with a output voltage of around 2.4 V and a theoretical capacity of 71 mAhg-1, although much
work is lacking on the anode side and electrolyte optimization. Nonetheless, this work has shown the promising future prospects of conducting redox polymers for various sustainable battery chemistries.
6 |
Future Outlook
From the presented material it is evident that the material show great promise to work in a full cell with a reasonable output voltage and capacity. However it is also evident that much progress is lacking on the anode side of pure metals. To further build upon this proof of concept, it is suggested that these materials and chemistries be studied in various other electrolytes such as PC/EC which would be more optimal for the voltage window, while still retaining redox-matching.
7 |
Acknowledgments
I want to start by thanking my supervisor Martin Sjödin, who’s expertise and apparent love for the field of electrochemistry is least to say contagious. Thanks to everyone in the battery group at NFM, who at several occasions have assisted me in my lab work or otherwise provided fruitful discussions which has helped form this thesis. A special thanks also to my colleague
Ronnie Mogensen from Structural Chemistry, who due as a consequence of his never ending
22
8 |
Bibliography
(1) Qian, J.; Wu, C.; Cao, Y.; Ma, Z.; Huang, Y.; Ai, X.; Yang, H. Sodium-Ion Batteries: Prussian Blue Cathode Materials for Sodium-Ion Batteries and Other Ion Batteries (Adv. Energy Mater. 17/2018). Adv. Energy Mater. 2018, 8 (17), 1870079. https://doi.org/10.1002/aenm.201870079. (2) Liang, Y.; Jing, Y.; Gheytani, S.; Lee, K.-Y.; Liu, P.; Facchetti, A.; Yao, Y. Universal Quinone Electrodes for Long Cycle Life Aqueous Rechargeable Batteries. Nat. Mater. 2017, 16 (8), 841–848. https://doi.org/10.1038/nmat4919.
(3) Åkerlund, L.; Emanuelsson, R.; Renault, S.; Huang, H.; Brandell, D.; Strømme, M.; Sjödin, M. The Proton Trap Technology—Toward High Potential Quinone-Based Organic Energy Storage. Adv. Energy Mater. 2017, 7 (20), 1700259. https://doi.org/10.1002/aenm.201700259. (4) Bitenc, J.; Lindahl, N.; Vizintin, A.; Abdelhamid, M. E.; Dominko, R.; Johansson, P. Concept and Electrochemical Mechanism of an Al Metal Anode ‒ Organic Cathode Battery. Energy
Storage Mater. 2020, 24, 379–383. https://doi.org/10.1016/j.ensm.2019.07.033.
(5) Gheytani, S.; Liang, Y.; Wu, F.; Jing, Y.; Dong, H.; Rao, K. K.; Chi, X.; Fang, F.; Yao, Y. An Aqueous Ca-Ion Battery. Adv. Sci. 2017, 4 (12), 1700465.
https://doi.org/10.1002/advs.201700465.
(6) Patil, N.; Mavrandonakis, A.; Jérôme, C.; Detrembleur, C.; Palma, J.; Marcilla, R. Polymers Bearing Catechol Pendants as Universal Hosts for Aqueous Rechargeable H+, Li-Ion, and Post-Li-Ion (Mono-, Di-, and Trivalent) Batteries. ACS Appl. Energy Mater. 2019, 2 (5), 3035–3041. https://doi.org/10.1021/acsaem.9b00443.
(7) Bančič, T.; Bitenc, J.; Pirnat, K.; Kopač Lautar, A.; Grdadolnik, J.; Randon Vitanova, A.; Dominko, R. Electrochemical Performance and Redox Mechanism of Naphthalene-Hydrazine Diimide Polymer as a Cathode in Magnesium Battery. J. Power Sources 2018, 395, 25–30. https://doi.org/10.1016/j.jpowsour.2018.05.051.
(8) Emanuelsson, R.; Sterby, M.; Strømme, M.; Sjödin, M. An All-Organic Proton Battery.
J. Am. Chem. Soc. 2017, 139 (13), 4828–4834. https://doi.org/10.1021/jacs.7b00159.
(9) Shea, J. J.; Luo, C. Organic Electrode Materials for Metal Ion Batteries. ACS Appl.
Mater. Interfaces 2020, 12 (5), 5361–5380. https://doi.org/10.1021/acsami.9b20384.
(10) Liao, C.; Zhang, M.; Yao, M. Y.; Hua, T.; Li, L.; Yan, F. Flexible Organic Electronics in Biology: Materials and Devices. Adv. Mater. 2015, 27 (46), 7493–7527.
https://doi.org/10.1002/adma.201402625.
(11) Heiska, J.; Nisula, M.; Karppinen, M. Organic Electrode Materials with Solid-State Battery Technology. J. Mater. Chem. A 2019, 7 (32), 18735–18758.
https://doi.org/10.1039/C9TA04328D.
(12) Nitta, N.; Wu, F.; Lee, J. T.; Yushin, G. Li-Ion Battery Materials: Present and Future.
Mater. Today 2015, 18 (5), 252–264. https://doi.org/10.1016/j.mattod.2014.10.040.
23
Aqueous Conducting Redox-Polymer-Based Proton Battery That Can Withstand Rapid Constant-Voltage Charging and Sub-Zero Temperatures. Angew. Chem. Int. Ed. 2020, 59 (24), 9631–9638. https://doi.org/10.1002/anie.202001191.
(14) Sterby, M.; Emanuelsson, R.; Huang, X.; Gogoll, A.; Strømme, M.; Sjödin, M. Characterization of PEDOT-Quinone Conducting Redox Polymers for Water Based Secondary Batteries. Electrochimica Acta 2017, 235, 356–364. https://doi.org/10.1016/j.electacta.2017.03.068. (15) Oka, K.; Strietzel, C.; Emanuelsson, R.; Nishide, H.; Oyaizu, K.; Strømme, M.; Sjödin, M. Characterization of PEDOT-Quinone Conducting Redox Polymers in Water-in-Salt Electrolytes for Safe and High-Energy Li-Ion Batteries. Electrochem. Commun. 2019, 105, 106489.
https://doi.org/10.1016/j.elecom.2019.106489.
(16) Wang, L.; Song, J.; Qiao, R.; Wray, L. A.; Hossain, M. A.; Chuang, Y.-D.; Yang, W.; Lu, Y.; Evans, D.; Lee, J.-J.; Vail, S.; Zhao, X.; Nishijima, M.; Kakimoto, S.; Goodenough, J. B. Rhombohedral Prussian White as Cathode for Rechargeable Sodium-Ion Batteries. J. Am. Chem. Soc.
2015, 137 (7), 2548–2554. https://doi.org/10.1021/ja510347s.
(17) Brant, W. R.; Mogensen, R.; Colbin, S.; Ojwang, D. O.; Schmid, S.; Häggström, L.; Ericsson, T.; Jaworski, A.; Pell, A. J.; Younesi, R. Selective Control of Composition in Prussian White for Enhanced Material Properties. Chem. Mater. 2019, 31 (18), 7203–7211.
https://doi.org/10.1021/acs.chemmater.9b01494.
(18) Muench, S.; Wild, A.; Friebe, C.; Häupler, B.; Janoschka, T.; Schubert, U. S. Polymer-Based Organic Batteries. Chem. Rev. 2016, 116 (16), 9438–9484.
https://doi.org/10.1021/acs.chemrev.6b00070.
(19) Sterby, M. Electrochemical Characterizations of Conducting Redox Polymers: Electron
Transport in PEDOT/Quinone Systems; Digital Comprehensive Summaries of Uppsala Dissertations
from the Faculty of Science and Technology; Acta Universitatis Upsaliensis: Uppsala, 2019. (20) Sterby, M.; Emanuelsson, R.; Mamedov, F.; Strømme, M.; Sjödin, M. Investigating Electron Transport in a PEDOT/Quinone Conducting Redox Polymer with in Situ Methods.
Electrochimica Acta 2019, 308, 277–284. https://doi.org/10.1016/j.electacta.2019.03.207.
(21) Huang, H.; Karlsson, C.; Strømme, M.; Gogoll, A.; Sjödin, M. Synthesis and Characterization of Poly-3-((2,5-Hydroquinone)Vinyl)-1H-Pyrrole: Investigation on
Backbone/Pendant Interactions in a Conducting Redox Polymer. Phys. Chem. Chem. Phys. 2017, 19 (16), 10427–10435. https://doi.org/10.1039/C6CP08736A.
(22) Han, C.; Li, H.; Shi, R.; Zhang, T.; Tong, J.; Li, J.; Li, B. Organic Quinones towards Advanced Electrochemical Energy Storage: Recent Advances and Challenges. J. Mater. Chem. A
2019, 7 (41), 23378–23415. https://doi.org/10.1039/C9TA05252F.
(23) Tong, L.; Jing, Y.; Gordon, R. G.; Aziz, M. J. Symmetric All-Quinone Aqueous Battery. ACS Appl. Energy Mater. 2019, 2 (6), 4016–4021. https://doi.org/10.1021/acsaem.9b00691. (25) Lu, Y.; Zhang, Q.; Li, L.; Niu, Z.; Chen, J. Design Strategies toward Enhancing the Performance of Organic Electrode Materials in Metal-Ion Batteries. Chem 2018, 4 (12), 2786–2813. https://doi.org/10.1016/j.chempr.2018.09.005.
24
Ponrouch, A. Cation Solvation and Physicochemical Properties of Ca Battery Electrolytes. J. Phys.
Chem. C 2019, 123 (49), 29524–29532. https://doi.org/10.1021/acs.jpcc.9b07308.
(27) Selvakumaran, D.; Pan, A.; Liang, S.; Cao, G. A Review on Recent Developments and Challenges of Cathode Materials for Rechargeable Aqueous Zn-Ion Batteries. J. Mater. Chem. A 2019,
7 (31), 18209–18236. https://doi.org/10.1039/C9TA05053A.
(28) Ponrouch, A.; Frontera, C.; Bardé, F.; Palacín, M. R. Towards a Calcium-Based Rechargeable Battery. Nat. Mater. 2016, 15 (2), 169–172. https://doi.org/10.1038/nmat4462.
(29) Wang, M.; Jiang, C.; Zhang, S.; Song, X.; Tang, Y.; Cheng, H.-M. Reversible Calcium Alloying Enables a Practical Room-Temperature Rechargeable Calcium-Ion Battery with a High Discharge Voltage. Nat. Chem. 2018, 10 (6), 667–672. https://doi.org/10.1038/s41557-018-0045-4. (30) Wang, H.; Emanuelsson, R.; Liu, H.; Edström, K.; Mamedov, F.; Strømme, M.; Sjödin, M. Redox-State-Dependent Interplay between Pendant Group and Conducting Polymer Backbone in Quinone-Based Conducting Redox Polymers for Lithium Ion Batteries. ACS Appl. Energy Mater.
2019, 2 (10), 7162–7170. https://doi.org/10.1021/acsaem.9b01130.
(31) Laviron, E. General Expression of the Linear Potential Sweep Voltammogram in the Case of Diffusionless Electrochemical Systems. J. Electroanal. Chem. Interfacial Electrochem. 1979,
101 (1), 19–28. https://doi.org/10.1016/S0022-0728(79)80075-3.
(32) Skompska, M.; Mieczkowski, J.; Holze, R.; Heinze, J. In Situ Conductance Studies of P- and n-Doping of Poly(3,4-Dialkoxythiophenes). J. Electroanal. Chem. 2005, 577 (1), 9–17. https://doi.org/10.1016/j.jelechem.2004.11.008.
25
9 |
Appendix
Electrochemical characterization
Figure A1 shows the bare cyclic voltammograms of EPE-Qz and the different redox potentials depending on the cation being cycled. Figure A2 shows the galvanostatic cycling performed on the assembled coin cell batteries at 5C and 0.5C.
Figure A1. CV of EPE-Qz cycling K+ (a), Na+ (b), Li+ (c), Ca2+ (d) and Mg2+-ions (e). Performed at a scan rate
of 20 mVs–1 in MeCN containing 0.05 M KTFSI, 0.05 M NaTFSI, 0.1 M LiTFSI, 0.1 M CaTFSI and 0.1 M
MgTFSI respectively.
26
Solvation energy calculations
Solvation energies for Na+, Li+, Ca2+ and Mg2+ were calculated from Born’s equation according
to Eq (2) below. Figure A2 shows the calculated formal potentials vs. the calculated solvation energies.
∆𝐺 = −=>?@A@ BCDEFE (1 −
G
DH) (2)
Figure A3. Formal potentials vs. calculated solvation energies in MeCN. A lower solvation energy implies a less
stable Qz-cation reduced state. Here the trend shows the opposite, meaning that the solvation energy does not play a major part in influencing the formal potential.
