- Email: [email protected]
Faradaic contributions in the supercapacitive charge storage mechanisms of manganese dioxides
Accepted Manuscript Title: Faradaic contributions in the supercapacitive charge storage mechanisms of manganese dioxides Author: Laura Coustan Pierre Lannelongue Paul Arcidiacono Fr´ed´eric Favier PII: DOI: Reference:
S0013-4686(16)30239-0 http://dx.doi.org/doi:10.1016/j.electacta.2016.01.212 EA 26594
To appear in:
Electrochimica Acta
Received date: Revised date: Accepted date:
1-11-2015 28-1-2016 29-1-2016
Please cite this article as: Laura Coustan, Pierre Lannelongue, Paul Arcidiacono, Fr´ed´eric Favier, Faradaic contributions in the supercapacitive charge storage mechanisms of manganese dioxides, Electrochimica Acta http://dx.doi.org/10.1016/j.electacta.2016.01.212 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Faradaic contributions in the supercapacitive charge storage mechanisms of manganese dioxides
Laura Coustan1,2 , Pierre Lannelongue1,2, Paul Arcidiacono1,2 and Frédéric Favier1,2*
1. Institut Charles Gerhardt Montpellier UMR 5235 CNRS, Université de Montpellier, Campus Triolet, cc1502, 34095 Montpellier cedex 05, France 2. Réseau sur le Stockage Electrochimique de l’Energie (RS2E), FR CNRS 3459, France * [email protected] (Frédéric Favier)
Abstract: Electrode materials based on four different manganese dioxides, amorphous, birnessite, crytpomelane and spinel were fabricated and their electrochemical behaviors compared in two electrolytes, Li2SO4 and (NMe4)2SO4. With respect to the structural characteristics of the various prepared MnO2, these electrolytes can be differentiated by their cation size. Voltammetric studies showed that these electrode materials presented distinct capacitive behaviors depending on the electrolyte used. Modeling of the electrode capacitances measured at various scan rates allowed to discriminate the surface and material bulk contributions to the overall specific capacitance of the fabricated electrodes.
Keywords: manganese dioxide, cation size, electrolyte, surface capacitance, bulk capacitance
1. Introduction
Thanks to Faradaic contributions to their storage mechanism, metal oxides are the most studied pseudo-capacitive materials [1–4]. Nevertheless, they should address many technological and process specifications before being produced at the industrial-scale and
introduced as commercial products. Among these, their synthesis must be easy and at low cost, and the balance between energy density delivered thanks to the Faradaic reactions and power density must be favorable. Manganese dioxide is known to meet most of the needed requirements while being somehow a good compromise in term of performances, between purely capacitive and cheap activated carbons and the costly but highly pseudocapacitive ruthenium oxide. Indeed, it has been demonstrated that MnO2, easily prepared by various routes, presents attractive electrochemical performances [5,6] The pseudocapacitive charge storage mechanism in MnO2, described by the equation 1, has been studied many times and established as involving fast Mn(III)/Mn(IV) redox reactions and cationic exchange to balance the changes in the manganese oxidation state [7]. MnO2 + (x + y) e- + x H+ + y C+
->
MnOOHxCy
(C: cation from electrolyte)
Even if the storage mechanism has been described as occurring at the surface of the electrode material, it has also been shown to be independent on the developed surface area [8]. Moreover, charge/discharge mechanism has also been shown by in-situ XRD, to depend on the considered MnO2 polymorph on the material bulk structure, beyond the surface [9,10]. This
pseudocapacitance
is
a
“core”
or
“bulk”
capacitance,
involving
cation
intercalation/insertion, eventually dissociated from the Faradaic surface contribution to the total capacitance of the material. There are many types of manganese dioxides built on MnO6 octahedra sharing corner, edges or faces. The structural complexity of the various polymorphs has been extensively investigated and can be illustrated by the large number of synthetic or natural compounds referring to an identical structure. Birnessite is a MnO2 phase with a 2 dimensional layered structure, with layers made of edge sharing MnO6 octahedra and separated by a gap of about 7 Å. This phase is also called δ-MnO2 or OL-MnO2 for octahedrallayer [11,12]. In the same way, the channel-based cryptomelane MnO2 can be identified as α-MnO2 or OMS-2 for octahedral molecular sieves 2, and the tri-dimensional compact spinel phase as λ-MnO2 [13,14]. Due to their different cationic exchange properties governed by their cavity size and connectivity, the various structures of these polymorphs are suspected to lead to different balanced contributions of the surface or bulk charge storage processes to the overall capacitance. But, contrarily to the electrochemical double layer capacitors (EDLCs), where the charge storage process is purely electrostatic, the simultaneous contributions of both
surface and bulk Faradaic phenomena make complex the unambiguous description of the charge/discharge mechanism [15–17]. If the electrode material structure has a critical role in the charge storage mechanism, it is the same for the electrolyte ion nature. Ion adsorption mechanisms onto carbon materials have been extensively investigated and correlations between material porosity and capacitance have highlighted the ion size influence [18]. In MnO2, the anion volume and size were shown to impact on the intercalation of the cations in the structure and the involved redox reactions. As the anion size, the cation size plays also a crucial role on the MnO2 electrochemical performances [19]. Indeed, depending on the structural characteristics of the polymorph, too large electrolytic cations may degrade the electrode material behavior by the limitation of the associated intercalation/insertion in too small crystallographic sites and lead to a strong discrepancy between surface and bulk contributions. In contrast, small cations should be able to “probe” either surface and bulk electroactive sites. Moreover, cation access to the bulk is controlled by the solid-state diffusion while much faster surface electrosorption kinetics are governed by the ion diffusion at the solid-liquid interface. As such, it is anticipated that the surface and bulk capacitive contributions should by highly dependent on the charge/discharge rate [20]. The goal of this study is to highlight the contributions of surface and bulk to the overall capacitance by using different electrolytes and different MnO2 polymorphs. These MnO2 polymorphs present different crystallographic structures and different ion exchange properties. The behavior in cyclic voltammetry of amorphous MnO2, birnessite, cryptmelane and spinel phases have been compared at various scan rates, in two aqueous electrolytes of distinct cation sizes, Li2SO4 and (NMe4)2SO4, used at the same 1 M concentration.
2. Experimental All chemical reagents were analytical grade and were used as purchased without further purification. Several techniques were used to prepare a set of MnO2 based materials with various microstructures. All aqueous solutions were prepared using deionized water (18 MΩ, obtained from an ELGA Lab Water deionized system). 2.1 Synthesis of amorphous MnO2 Amorphous phase of manganese dioxide was prepared by the co-precipitation route via the reaction between KMnO4 and MnSO4.H2O as detailed in the following: 0.013 moles KMnO4
(99%, Alfa Aesar) were dissolved in 100 mL of deionized water (resistivity > 18 MΩcm, from an ELGALab Water deionized system) kept at 20 °C (water bath) to avoid any solution overheating. Separately, 0.019 moles of MnSO4.H2O (99%, VWR) was dissolved in 100 mL of deionized water. This latter solution was added drop-wise to the former MnO4- solution under stirring. Stirring was kept for 1 h after complete mixing of the solutions. Obtained suspension was filtered on Buchner, and washed many times with distilled water. Final product was dried at 90 ° C in air overnight. 2.2 Synthesis of birnessite phase MnO2 Firstly, a solution (A) was prepared by the dissolution of 0.04 moles of KMnO4 (99% Alfa Aesar) ad 1.2 moles of NaOH (ACS grade, VWR) in 400 mL of distilled water. A solution (B), prepared with 0.112 moles de MnCl2.4H2O (99%, Aldrich) in 400 mL of water, was added to the solution (A), drop by drop, and under stirring in an ice bath. Obtained suspension has kept under stirring, at room temperature for 24 h and then was washed with water and dried at 90 °C overnight. 2.3 Synthesis of cryptomelane phase MnO2 The cryptomelane were obtained from the ignition in air of MnO2 birnessite at 400 °C for 60 h. 2.4 Synthesis of spinel phase MnO2 In a first time, LiMn2O4 was prepared by dissolution of 0.01 mole of LiNO3 (99%, Fisher Scientific) and 0.02 moles of Mn(NO3)2·2H2O (ACS grade, Fisher Scientific) in a 100 mL mixture of citric acid (ACS grade, Aldrich) and ethylene glycol (ACS grade, Aldrich) (molar ratio 1:4). Then, the solution was heated at 90 ◦C for 30 min and the temperature was subsequently increased to 140 °C in order to activate esterification and remove the ethylene glycol excess until a coloration of the solution. This solution was heated at 180 °C under vacuum and brown slurry was obtained. The resulting slurry was annealed at 250 °C in air leading to LiMn2O4. Finally, the MnO2 spinel was synthesized by hydrolysis of LiMn2O4 in a 0.5 M HCl (ACS grade, Aldrich) aqueous solution for 24 h. 2.5 Material characterization
The crystalline structure of the materials was investigated with X-ray diffraction using a PANanalytical X’Pert Pro powder X-ray diffractometer (Cu Kα1 radiation) in a BraggBrentano configuration. The Brunauer – Emmett – Teller (BET) surface area and pore volume were measured with N2 sorption at liquid nitrogen temperature (77 K) using a Micrometrics ASAP 2020 instrument. Samples were degassed at 200 °C for 5 h in flowing nitrogen prior to measurements to remove guest molecules. The total surface area was determined by the multipoint Brunauer-Emmett-Teller (BET) method.
2.6 Electrode preparation and electrochemical characterization Prepared MnO2 powders and PTFE (from 60% wt PTFE dispersion in water) in a 90/10 ratio, were manually mixed in acetone until a homogeneous slurry was obtained. This slurry was spread on a glass plate and rolled several times to obtain thick films. Square pieces of films, about 1 cm2 in surface area and 150 µm in thickness, were cut out and pressed at 10 tons for a few minutes in a stainless steel grid used as current collector. In the whole series of prepared electrodes, MnO2 loadings ranged from 9.1 to 12.3 mg.cm-2. Before electrochemical characterization, the electrodes were dried overnight under vacuum at room temperature and immersed, under vacuum, in the electrolyte solutions for 4 h. Electrochemical measurements were carried out using a Biologic VMP3 potentiostat controlled by EC-Lab V10.38 software. The electrochemical cell used in a conventional three-electrode setup was composed of a Pt counter electrode, an Ag/AgCl (3 M KCl) reference electrode and the composite electrode as the working electrode.
3. Results and discussion
Crystallographic structures of birnessite, cryptomelane and spinel forms of MnO2 are schematized in Figure 1 (a), (b) and (c), respectively. These structural arrangements are built on MnO6 octahedra building blocks sharing faces, edges or corners. Birnessite is obtained by slow reduction of MnO4+ by MnCl2 leading to a layered form of MnO2 made by stacking edge sharing MnO6. The presence of Na+ cations in the interlayer gap of about 7 Å accounts for the birnessite structure stability. Heating birnessite at 400 °C under air leads to the cryptomelane form of MnO2. During this heat treatment, part of Na+ is extracted from the layered pristine structure. Destabilized, the layered structure is converted as a channeled one. Resulting
channels, made of 2 x 2 arrangements of MnO6 octahedra sharing their edges, have side dimensions of 5 Å and 6.3 Å in diagonal. The spinel form of MnO2 is a tridimensional lithiated phase, formed by 1 x 1 channels built on octahedra sharing edges. The resulting tetrahedric site is about 2.9 Å in size.
Thanks to its layered structure, 2D-MnO2 birnessite allows a good mobility of the cations into the bulk. Because of more constrained structural arrangements, the progressive ionic conductivity loss from birnessite to cryptomelane and then spinel is expected to impact on the Faradaic bulk contribution to the capacitance and favors surface reactions [10]. XRD pattern of the various MnO2 phases synthesized during this study, amorphous MnO2 (a), birnessite (b), cryptomelane (c) and spinel (d) are shown in Figure 2. Amorphous behavior of MnO2 is confirmed by the presence of broad and low intensity diffraction peaks (a). Diffractogramm (b) shows similar features as (a) but with more defined diffraction peaks at 2θ angles of 12.2°, 24.7°, 36.9°. These peaks are unambiguously assigned to the monoclinic (C2/m) layered structure of the MnO2 birnessite polymorph (JCPDS 43-1456) [12,21]. In such a structure, layers are containing a mixture of Mn(IV) and Mn(III) [22–24]. The peak at 12.2° (2θ) corresponds to the (001) planes and to the interlayer distance at 7.0 Å. Its shape and intensity, as well as angular position, are highly sensitive to the phase crystallinity and, nature and content of the alkali cations intercalated in between the layers. The (002) peak at 24.5° (2θ) is also related to the interlayer spacing. Cryptomelane XRD pattern in Figure 2(c) is very different from those of the other MnO2 polymorphs in the series. Indeed, this diffractogramm present narrow and sharp peaks, indicating that the birnessite heating increases the crystallinity of the resulting cryptomelane. Main peaks are at 2θ angles of 12.7°, 18.1°, 28.7°, 37.2°, 41.9° and 49.5° (JCPDS 29-1020) corresponding to a 2 x 2 channeled structure in the I4/m group [25]. Spinel XRD pattern in Figure 2(d) is characteristic of a lower crystallinity. Nevertheless, main diffraction peaks at the 2θ angles of 19.3°, 37.3°, 38.9°, 45.2° and 49.5° (JCPDS 44-0992) confirm the tridimensional cubic structure, assigned to the Fd3m group [14]. The peaks at about 37° (2θ), shared by the diffractogramms in the series, is commonly observed in manganese oxide diffraction patterns since it is related to interatomic distances in materials based on edge-sharing MnO6 octahedra.
In this study, two different electrolytic salts were used, Li2SO4 and (NMe4)2SO4, at a 1 M concentration. Each salt obviously presents distinct characteristics. Some of them are related to the cation size which is given in Table 1. In aqueous solution, such monovalent cations are usually surrounded by one or more hydration shell. Li+ cation, the smallest of both, is the most polarizing and builds a hydration shell made of four water molecules in a tetrahedral symmetry. However, interactions are so strong that additional hydration shells surround the first one [26].
Actually, Li+ locally modifies the structure of the water molecules by breaking hydrogen bonds, and promotes the local water environment as structured hydration shells. This explains why its hydrated radius (3.82 Å) is very large before its radius as bare cation (0.68 Å). In contrast, NMe4+ is a destructuring cation, as it hardly forms any hydration shells, leading to a lower solvation than Li+ [26,27]. Table 2 shows the cavity sizes of prepared manganese oxides and a rapid comparison with cation size may stress some steric issues unfavorable to intercalation/insertion into the material bulk. It seems however important to keep in mind that the radius of hydrated cation is not the effective radius when intercalated/inserted/confined in the material structure [10,28,29]. Indeed, many authors have shown that ions could be desolvated in confined media [28,29]. As such, in the material bulk, Li+ is smaller (1.95 Å radius) than NMe4+ (3.22 Å radius). It is also obvious that, because of its large size, tetramethylammonium ion, even as desolvated cation, cannot access the bulk of any of the manganese dioxide crystallized polymorphs in the prepared series but birnessite for which a 7 Å interlayer space can a priori afford the intercalation of a 6.45 Å diameter NMe4+ bare cation. Cyclic voltammograms (CVs) for the various prepared MnO2 polymorph electrodes, amorphous, birnessite, cryptomelane and spinel, are shown in Figure 3A, B, C and D respectively. These were measured at various scan rates from 5 to 200 mV.s-1 in a three electrode configuration from 0 to 0.85 V (vs AgAgCl) in 1 M Li2SO4 (a) and (NMe4)2SO4 (b) aqueous electrolytes. The roughly rectangular shape of the CVs is characteristic of the pseudocapacitive behaviour of the prepared electrode materials. As electrode materials, amorphous and birnessite phases of MnO2 are presenting the most capacitive electrochemical signatures, either in Li2SO4 or (NMe4)2SO4 -based electrolytes. Some Faradaic contributions are visible, especially for birnessite (Figure 3B) and cryptomelane (Figure 3C) -based
electrodes in Li2SO4 electrolyte. These CVs highlight the influence of the electrolyte nature on the electrochemical behaviour of each electrode. Generally speaking, prepared materials show greater capacitances when operated in Li2SO4 electrolyte. However, even if the electrode capacitance is lower in (NMe4)2SO4 -based electrolyte, pseudocapacitive behaviour is confirmed in this electrolyte too. To highlight the effect of the electrolyte nature on the electrochemical behaviour of the prepared MnO2 phases, Figure 4 shows the 50th voltammetric cycle of the prepared electrode materials carried out at a scan rate of 5 mV.s-1 in 1M aqueous Li2SO4 (solid line) and (NMe4)2SO4 (dashed line). As already mentioned, depicted voltammetric cycles are more or less large but remain characteristic of a pseudocapacitive behaviour. Switching from Li2SO4 and (NMe4)2SO4 electrolyte, induces several changes in terms of shape and amplitude (current density) of the measured CVs, depending on the MnO2 phase considered. For example, the capacitive behaviour of the birnessite based electrode is greater in Li2SO4 than in (NMe4)2SO4 electrolyte (Figure 4b). In the former electrolyte, the cycle is more rectangular but especially larger, corresponding to higher current densities than in the latter. Similar current density fading can also be observed for the other electrode materials in the series. The electrolyte nature also impact on the eventual Faradaic contributions such as those observed for the cryptomelane electrode material. When operated in Li2SO4, the corresponding voltammogram in Figure 4c shows redox peaks, highlighting the contribution of Faradaic phenomena involved in the charge storage. These peaks, at about 0.5 and 0.2 V (vs AgAgCl), are hardly visible when the electrode is operated in 1 M aqueous (NMe4)2SO4 electrolyte. Table 3 lists the specific capacitances measured by CV at 5 mV.s-1 for the various prepared MnO2 based electrodes in both electrolytes. As suggested by the CV shapes in Figure 4, the capacitance losses are not the same for all the samples when switching from Li2SO4 to (NMe4)2SO4 -based electrolytes. For the spinel and amorphous MnO2, the capacitance losses are about 26 and 30 %, respectively. For birnessite and cryptomelane based electrodes, the capacitance losses are more marked. The capacitance of the birnessite based electrode at 200 F.g-1 in Li2SO4, drastically decreases when operated in (NMe4)2SO4 electrolyte (112 F.g-1 ). The corresponding capacitance loss is about 44 % and is accompanied by a strong deformation of the voltammogram shape. Cryptomelane electrode also presents a lower capacitance when used in (NMe4)2SO4 electrolyte. When operated in Li2SO4, the
corresponding capacitance is about 102 F.g-1 and falls down to 55 F.g-1 when in (NMe4)2SO4 electrolyte, corresponding to a 45 % loss.
In regards to the scan rate, a similar behavior is observed for the prepared electrodes in both electrolytes. Measured CVs remain characteristic of pseudocapacitive materials as scan rate is increased but the corresponding current densities are strongly affected. Figure 5 shows the changes in the relative capacitance with the scan rate. Whatever the considered MnO2 phases, capacitance is highly dependent on the scan rate and drastically decreases until 50 mV.s-1. For example, from 5 to 50 mV.s-1, up to 95 % capacitance is lost for amorphous and spinel MnO2 phases in Li2SO4 -based electrolyte. Except for the birnessite based electrode, the relative capacitances at high scan rate are higher in (NMe4)2SO4 electrolyte than in Li2SO4 electrolyte. This lower capacitance loss can be explained by (NMe4)+ being too large to be intercalated/inserted in the cryptomelane and spinel structures. The corresponding capacitance, in these cases, is originating from surface contributions, eventually limited by the ion diffusion at the solid/liquid interface.
Figure 6 is an alternative way to compare the capacitance fading of the prepared electrodes in Li2SO4 and (NMe4)2SO4 electrolytes. Using such a logarithmic scale, resulting curves are roughly linear and their slopes dependent on the material nature and electrolyte. Calculated slopes are given in the table 4. Electrodes made of amorphous (a) and birnessite (b) MnO2 are characterized by very similar slopes when cycled in Li2SO4 or (NMe4)2SO4 electrolytes, respectively at - 0.58 and - 0.53 for amorphous MnO2, and - 0.42 and - 0.46 for birnessite. These slight slope variations suggest that the behavior of Li+ and NMe4+ cations towards both amorphous and birnessite MnO2 is about the same when scan rate is increased. Slopes extracted from measurements done with cryptomelane and spinel -based electrodes (Figures 6c and d, respectively) present larger discrepancies when cycled in Li2SO4 and (NMe4)2SO4 electrolytes as a proof of charge storage mechanisms being different in these materials depending on the electrolyte. For these two electrode materials, the line slopes obtained from cycling in (NMe4)2SO4 electrolyte are lower than those in Li2SO4 electrolyte. This difference in scan rate capability comes from the charge storage in Li2SO4 involving the intercalation/insertion of Li+ in the structure of the material, which is drastically altered at high scan rate because of a limited solid state diffusion in the materials. In contrast, NMe4+ is
too large to diffuse into the material structure and its interaction, limited to the surface of the electrode material, is responsible for the lower capacitance than in Li2SO4 electrolyte at low scan rate but also less or un-affected by eventual diffusion limitations at high scan rate. In contrast, the slope from measurements with birnessite MnO2 based electrode in (NMe4)2SO4 is slightly greater than in Li2SO4 electrolyte. As already mentioned, the interlayer distance of about 7 Å is large enough for this material to host either Li+ or NMe4+ cations. Obviously, the larger size of the latter is less favourable to diffusion in the solid and the impact of the increased scan rate on the measured capacitance is consequently greater. Finally, despite a “structure” that can be described as a disordered birnessite, the electrochemical behavior of amorphous MnO2 is mostly led by surface interactions and the measured slope is slightly greater in Li2SO4 than in (NMe4)2SO4 electrolyte. In summary, in the case of (NMe4)2SO4 electrolyte, the charge storage mechanism is less dependent on diffusion properties of the cations into the bulk of the prepared electrode materials but in MnO2 birnessite, and the effect of the scan rate on the capacitance is lesser. In contrast,
because
of
storage
mechanisms
involving
solid
state
diffusion
and
intercalation/insertion phenomena of Li+, capacitance is more scan rate sensitive when using Li2SO4 electrolyte. Using such logarithmic scale, it is possible to express the specific capacitance C of the electrode materials as a linear function of the scan rate v by the following equation [30]: log C= a log v + logC 1
Where a is the line slope discussed above, and C1 the capacitance value at 1 mV.s-1 scan rate. As such, a linear relationship between the specific capacitance, C and the opposite of the scan rate square root, v-1/2, can be expressed by the following equation: C = constante × v
− 1 /2
+C ∞
In this equation, the intercept with the C axis, C ∞ , is the capacitance at infinite scan rate. At such infinite scan rate, diffusion of electrolytic species into the electrode material bulk cannot be considered and C ∞ corresponds to the contribution of the surface material only to the overall specific capacitance. The corresponding plots are depicted in figure 7. In a similar way, the capacitance inverse, C-1, can be expressed as a function of the scan rate square root to obtain, at the intercept, the material capacitance at infinitely low scan rate thanks to the following equation: −1
C = constante × v
1/ 2
−1
+ C0
−1 Here, C0 corresponds to the specific capacitance inverse at infinitely low scan rate. This
capacitance can be assigned to the total capacitance that can be developed by the electrode material, involving both surface and bulk contributions, without any diffusion limitations neither at liquid nor solid states. These two equations have already been used, in particular by Ardizzione et al, for the study of RuO2 capacitance contributions in the more or less accessible regions of the electrode material [31,32]. Figures 7 and 8 show the capacitance changes as a function of the inverse of the scan rate square root (C = f (v-1/2)) and the capacitance inverse as a function of the scan rate square root (C-1 = f (v1/2)), respectively. The surface contribution to the capacitance of the electrode material C∞ is extracted from figure 7 whereas figure 8 highlights the total capacitance of the electrode material, C0. Thanks to these values, the capacitance material core contribution Cbulk can be easily calculated by considering the total capacitance as the sum of the surface and bulk capacitances and the following: Cbulk = C0 - C∞. Table 5 gives both C∞ and C0 capacitance values with their uncertainty values (automatically generated by OriginLab software). From the graphs in Figure 7, the plot intercepts correspond to very small surface capacitances, ranging from 2 to 8 F.g-1, in the whole series of electrode materials and electrolytes. Spinel C∞, at about 2 F.g-1, is the lowest in the series. In a first approximation, C∞ seems to hardly depend on the electrolyte and polymorph natures.
The total capacitance developed by the electrodes is extracted from the graphs in Figure 8. At such an infinitely low scan rate, the kinetic transport and the ion diffusion into the electrolyte and the material do not alter the charge storage mechanism and both surface and bulk contribution maximize the overall capacitance of the considered electrode material. Total capacitance values are way greater than measured surface capacitances, in some cases, by more than two orders of magnitude. At such a very low scan rate, bulk contribution to the overall capacitance is the most important but seems to strongly depend on the considered electrode materials in the series. Specific capacitance for birnessite phase is calculated at about 1210 and 1159 F.g-1 in Li2SO4 and (NMe4)2SO4 respectively. These remarkable values close to the theoretical capacitance of MnO2 at about 1300 F.g-1, demonstrate that birnessite structure allows enhanced cationic exchanges regardless of the considered electrolyte. In this case, Faradaic phenomena are dominating and intercalation processes are favored. For other
prepared MnO2 phases, the total capacitances are lower (from 200 to 700 F.g-1) and drastically decrease when used in (NMe4)2SO4 instead of Li2SO4 electrolyte. This later point expresses the difficulty or impossibility for the large NMe4+ cations to be intercalated/inserted into the structure of either amorphous, cryptomelane or spinel MnO2 prepared materials.
On the other hand, considering this size effect and the surface to solely contribute to the capacitance, the total capacitances (C0) developed by cryptomelane and spinel in (NMe4)2SO4 electrolyte should be equal to the surface capacitance determined at infinitely high scan rate (C∞). This is obviously not true since for cryptomelane, C0= 196 F.g-1 and C∞= 6 F.g-1 while for spinel, C0= 204 F.g-1 and C∞= 2 F.g-1. A first hypothesis to account for these discrepancies could be that, contrarily to what was supposed earlier, C∞ capacitance, as calculated here, does not represent the total contribution of the material surface. The very low capacitance values C∞, listed in the Table 5, could actually be assigned to the unique double layer contribution to the capacitance, originating from electrostatic interactions, to differentiate from a Faradaic surface contribution. If these C∞ values correspond to the double layer capacitance developed by MnO2 electrode materials studied here, then it should be possible to compare them to the corresponding specific surface area values listed in the Table 6. Amorphous MnO2, which develops the higher specific surface area about 78 m².g-1, shows a specific capacitance at high scan rate similar to the other materials. Cryptomelane, with a specific surface area about 40 m².g-1, similar to that of spinel, but the half of that of amorphous phase, still presents a specific capacitance about 6 F.g-1 at infinite scan rate. At this stage, it does not seem possible to establish any correlation between these reported capacitance values and the specific surface area developed by these materials. It can also be interesting to compare these surface capacitances to the electronic conductivities of the various polymorphs given in the Table 6. Electronic conductivity values strongly depend on the material nature. While they present a similar C∞ capacitance, birnessite MnO2 appears as 103 less conductive than cryptomelane or spinel forms, while the latter present the lowest C∞ capacitance in the series. Again, the electronic conductivity characteristics of the prepared materials do not fit the values and their assignment to double layer contributions cannot be confirmed. A second hypothesis can be raised for the very low calculated C∞ capacitances. Whatever is the considered material and electrolyte, they all range from 2 to 8 F.g-1. These values do not
depend on the developed surface area of the electrode material nor on its electronic conductivity and are not correlated to the size of the electrolytic cation. They however express the impact of the scan rate on the surface contribution and at this infinitely strong regime, diffusion limitations of the electrolytic species at the liquid state within the electrolyte bulk and/or at the electrode/electrode interface may have to be considered showing that the scan rate increase does not only impact on the material bulk contribution but also on the surface contribution. It remains thus difficult, in regard to the different observations, to establish any correlation between C∞ and a double layer capacitance, and to conclude on the origin of this contribution, although limited, to the MnO2 capacitance. In contrast, since only the surface of cryptomelane or spinel MnO2 is “probed” (and not the bulk) when operated in (NMe4)2SO4 electrolyte, C0 values at about 200 F.g-1 correspond to the Faradaic contribution that can be provided at most by the material surface. Thanks to its “structure” deriving from birnessite, a mixed surface/bulk contribution to the C0 capacitance, calculated at 500 F.g-1, should probably be considered for the amorphous MnO2 based electrode.
4. Conclusion This study is based on the comparison of the electrochemical behavior of four MnO2 based electrode materials, amorphous, birnessite, cryptomelane and spinel, showing various structural arrangements and cation hosting capabilities. To promote cation size effects, two different electrolytic media, 1M aqueous Li2SO4 and (NMe4)2SO4 were used. Material specific capacitances were calculated and expressed as a function of the scan rate used in cyclic voltammetry experiments. Materials showed distinct capacitive behaviors depending on the used electrolyte and specific capacitances were shown to drastically decrease with the scan rate. To differentiate and assign the surface and bulk contributions to the overall capacitance, a diffusion -based modeling was used. The expression of the capacitance as a function of the inverse of the scan rate square root, and the capacitance inverse as a function of the scan rate square root, demonstrated that the total capacitance is balanced by both bulk and surface contributions. This balance strongly depends on the electrode material structure (and cationic exchange capabilities), on the electrolyte nature, and, consequently, on the charge/discharge rate. Although a contribution from the electrochemical double layer cannot be excluded, it is fairly limited to at most 1% of the overall capacitance. The limited
capacitance retention observed upon scan rate originates from the kinetics of Faradaic phenomena involved in their charge storage mechanisms. These phenomena are mostly influenced by the ion mobility into the material structure.
Acknowledgements The Centre National de la Recherche Scientifique is gratefully acknowledged for funding. This work has been performed with the support of the PROGELEC program of the Agence Nationale de la Recherche (ANR) through the FlexCap project (grant #073616). More information available at www.flexcap.fr.
References 1.
Ghodbane, O., Ataherian, F., Wu, N.-L. & Favier, F. In situ crystallographic investigations of charge storage mechanisms in MnO2-based electrochemical capacitors. J. Power Sources 206, 454–462 (2012).
2.
Augustyn, V., Simon, P. & Dunn, B. Pseudocapacitive oxide materials for high-rate electrochemical energy storage. Energy Environ. Sci. 7, 1597 (2014).
3.
Conway, B. E., Birss, V. & Wojtowicz, J. The role and utilization of pseudocapacitance for energy storage by supercapacitors. J. Power Sources 66, 1–14 (1997).
4.
Jo, C., Hwang, I., Lee, J., Lee, C. W. & Yoon, S. Investigation of pseudocapacitive charge-storage behavior in highly conductive ordered mesoporous tungsten oxide electrodes. J. Phys. Chem. C 115, 11880–11886 (2011).
5.
Coustan, L. & Favier, F. Microwave-Assisted Decoration of Carbon Substrates for Manganese Dioxide-Based Supercapacitors. J. Electrochem. Soc. 162, A5133–A5139 (2015).
6.
Cormie, A., Cross, A., Hollenkamp, A. F. & Donne, S. W. Cycle stability of birnessite manganese dioxide for electrochemical capacitors. Electrochim. Acta 55, 7470–7478 (2010).
7.
Toupin, M., Brousse, T., Belanger, D. & Be, D. Charge Storage Mechanism of MnO Electrode Used in Aqueous Electrochemical Capacitor Charge Storage Mechanism of MnO 2 Electrode Used in Aqueous Electrochemical Capacitor. Chem. Mater. 3184– 3190 (2004). doi:10.1021/cm049649j
8.
Brousse, T., Toupin, M., Dugas, R., Athouel, L., Crosnier, O. & Belanger, D. Crystalline MnO[sub 2] as Possible Alternatives to Amorphous Compounds in Electrochemical Supercapacitors. J. Electrochem. Soc. 153, A2171 (2006).
9.
Ghodbane, O., Louro, M., Coustan, L., Patru, A. & Favier, F. Microstructural and Morphological Effects on Charge Storage Properties in MnO2-Carbon Nanofibers Based Supercapacitors. J. Electrochem. Soc. 160, A2315–A2321 (2013).
10.
Ghodbane, O., Pascal, J.-L., Fraisse, B. & Favier, F. Structural in situ study of the thermal behavior of manganese dioxide materials: toward selected electrode materials for supercapacitors. ACS Appl. Mater. Interfaces 2, 3493–505 (2010).
11.
Julien, C. ., Massot, M. & Poinsignon, C. Lattice vibrations of manganese oxides. Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 60, 689–700 (2004).
12.
Athouël, L. Moser, F., Dugas, R.,Crosnier, O., Bélanger, D &Brousse, T. Birnessite as Possible Candidate for Hybrid Carbon/MnO2 Electrochemical Capacitor. ECS Trans 112, 7270–7277 (2008).
13.
Ghodbane, O., Pascal, J.-L. & Favier, F. Microstructural effects on charge-storage properties in MnO2-based electrochemical supercapacitors. ACS Appl. Mater. Interfaces 1, 1130–9 (2009).
14.
Lin, Y.-P., Tsai, C.-B., Ho, W.-H. & Wu, N.-L. Comparative study on nanostructured MnO2/carbon composites synthesized by spontaneous reduction for supercapacitor application. Mater. Chem. Phys. 130, 367–372 (2011).
15.
Simon, P., Gogotsi, Y. & Dunn, B. Where do batteries end and supercapacitors begin? Science 343, 1210–1211 (2014).
16.
Winter, M. & Brodd, R. J. What are batteries, fuel cells, and supercapacitors? Chem. Rev. 104, 4245–69 (2004).
17.
Brousse, T., Belanger, D. & Long, J. W. To Be or Not To Be Pseudocapacitive? J. Electrochem. Soc. 162, A5185–A5189 (2015).
18.
Simon, P. & Gogotsi, Y. Materials for electrochemical capacitors. Nat. Mater. 7, 845– 854 (2008).
19.
Chmiola, J. Yushin, G., Gogotsi, Y., Portet, C., Simon, P &Taberna, P L Anormalous increase in carbon capacitance at pore sizes less than 1 nanometer. Science (80-. ). 313, 1760–1763 (2006).
20.
Boisset, A. & Athouël, L. Comparative Performances of Birnessite and Cryptomelane MnO2 as Electrode Material in Neutral Aqueous Lithium Salt for Supercapacitor Application. J. 2, (2013).
21.
Moser, F., Athouël, L., Crosnier, O., Favier, F., Bélanger, D. & Brousse, T. Transparent electrochemical capacitor based on electrodeposited MnO2 thin film electrodes and gel-type electrolyte. Electrochem. commun. 11, 1259–1261 (2009).
22.
Drits, V. A., Silvester, E., Gorshkov, A. I. & Manceau, A. Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite .1. Results from X-ray diffraction and selected-area electron diffraction. Am. Miner. 82, (1997).
23.
Silvester, E., Manceau, A. & Drits, V. A. Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite .2. Results from chemical studies and EXAFS spectroscopy. Am. Miner. 82, (1997).
24.
Ramstedt, M., Shchukarev, A. V. & Sjoberg, S. Characterization of hydrous manganite (γ-MnOOH) surfaces- An XPS study. Surf. Interface Anal. 34, 632–636 (2002).
25.
Toupin, M., Brousse, T. & Be, D. Influence of Microstucture on the Charge Storage Properties of Chemically Synthesized Manganese Dioxide. Chem. Mater. 3946–3952 (2002).
26.
Jolivet, J. P. De la solution à l’oxyde. Inter Ed. Editions C, (1994).
27.
Ohtaki, H. & Radnai, T. Structure and dynamics of hydrated ions. Chem. Rev. 93, 1157–1204 (1993).
28.
Lynden-Bell, R. & Rasaiah, J. Using simulation to study solvation in water. J. Chem. Phys. 105, 9266 (1996).
29.
Tanimura, A., Kovalenko, A. & Hirata, F. Molecular theory of an electrochemical double layer in a nanoporous carbon supercapacitor. Chem. Phys. Lett. 378, 638 (2003).
30.
Fregonara, G., Fregonara, G., Trasatti, S. & Trasatti, S. ‘Inner’ and ‘Outer’ Active Surface Electrodes. 5–9 (1989).
31.
Athouel, L. Arcidiacono, P., Ramirez-Castro, C., Crosnier, O., Hamel, C. Investigation of cavity microelectrode technique for electrochemical study with manganese dioxides Electrochim. Acta 86, 268–276 (2012).
32.
Ardizzione, S., Fregonara, G. & Trasatti, S. “Inner” and “outer” active surface of RuO2 electrodes. Electrochim. Acta 35, 263 (1990).
Figure 1: Crystallographic structures of birnessite (a), cryptomelane (b) and spinel (c) MnO2 phases
Figure 2: XRD patterns of MnO2 polymorphs: amorphous (a), birnessite (b), cryptomelane (c) and spinel (d)
Figure 3: Cyclic voltammograms for the various prepared MnO2 electrodes: amorphous (A), birnessite (B), cryptomelane (C) and spinel (D) for Li2SO4 (a) and (NMe4)2SO4 (b) at scan rates from 5 to 200 mV.s-1
Figure 4: Cyclic voltammograms at the 50th cycle in both electrolytes at 5 mV.s-1 for amorphous (a), birnessite (b), cryptomelane (c) and spinel (d)
Figure 5: Relative capacitance upon scan rate for amorphous (a), birnessite (b), cryptomelane (c) and spinel (d) MnO2 based electrodes
Figure 6: Specific capacitance upon scan rate in log/log scale for amorphous (a), birnessite (b), cryptomelane (c) and spinel (d) in logarithmic scale
Figure 7: Specific capacitance as a function of the inverse of the square root scan rate for the various prepared electrode materials operated in Li2SO4 and (NMe4)2SO4 electrolytes
Figure 8: Inverse of the specific capacitance as a function of the scan rate square root for the various prepared electrode materials operated in Li2SO4 and (NMe4)2SO4 electrolytes
Table 1: Radii of the cations in used electrolytes Cation
Li+
Bare cation radius (Å) 0.68
Hydrated cation radius (Å) 3.82
1.95
3.22
3.22
3.22
(NMe4)+
Radius with first hydration shell (Å)
Table 2: Cavity sizes in MnO2 polymorphs MnO2 polymorph Cavity dimension (Å) Birnessite
7
Cryptomelane
5 x 6.2
Spinel
2.9
Table 3: MnO2 polymorph specific capacitances from measurements at 5 mV.s-1 in Li2SO4 and (NMe4)2SO4 1 M aqueous electrolytes Li2SO4 (NMe4)2SO4
Amorphous 139 F.g‐1 97 F.g‐1
Birnessite 200 F.g‐1 112 F.g‐1
Cryptomelane 102 F.g‐1 55 F.g‐1
Table 4: log C = f(log v) line slopes for MnO2 based electrodes Amorphous Birnessite Cryptomelane Spinel
Li2SO4 ‐0.58 ‐ 0.42 ‐ 0.52 ‐ 0.59
(NMe4)2SO4 ‐0.53 ‐ 0.46 ‐ 0.36 ‐ 0.39
Spinel 78 F.g‐1 56 F.g‐1
Table 5: C∞ and C0 capacitances extracted from graphs 7 and 8 for amorphous, birnessite, cryptomelane and spinel MnO2 phases MnO2 polymorph Amorphous Birnessite Cryptomelane Spinel
Li2SO4 C∞ (F.g‐1) 6.1 ± 1.4 7.7 ± 1.4 6.3 ± 1.1 2.1 ± 0.3
(NMe4)2SO4 C∞ (F.g‐1) C0 (F.g‐1) 5.8 ± 0.9 490 ± 87 6.4 ± 1.0 1159 ± 166 6.3 ± 1.1 196 ± 25 1.6 ± 0.3 204 ± 29
C0 (F.g‐1) 709 ± 95 1210 ± 181 476 ± 68 416 ± 64
Table 6 Specific surface areas and electronic conductivities of the studied MnO2 polymorphs MnO2 polymorph Specific surface area (m².g‐ 1 ) Electronic conductivity S.cm‐1
Amorphous
Birnessite
Spinel
40
Cryptomelan e 50
78 2 .10‐5
6 .10‐6
9 .10‐3
3 .10‐3
50
S0013-4686(16)30239-0 http://dx.doi.org/doi:10.1016/j.electacta.2016.01.212 EA 26594
To appear in:
Electrochimica Acta
Received date: Revised date: Accepted date:
1-11-2015 28-1-2016 29-1-2016
Please cite this article as: Laura Coustan, Pierre Lannelongue, Paul Arcidiacono, Fr´ed´eric Favier, Faradaic contributions in the supercapacitive charge storage mechanisms of manganese dioxides, Electrochimica Acta http://dx.doi.org/10.1016/j.electacta.2016.01.212 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Faradaic contributions in the supercapacitive charge storage mechanisms of manganese dioxides
Laura Coustan1,2 , Pierre Lannelongue1,2, Paul Arcidiacono1,2 and Frédéric Favier1,2*
1. Institut Charles Gerhardt Montpellier UMR 5235 CNRS, Université de Montpellier, Campus Triolet, cc1502, 34095 Montpellier cedex 05, France 2. Réseau sur le Stockage Electrochimique de l’Energie (RS2E), FR CNRS 3459, France * [email protected] (Frédéric Favier)
Abstract: Electrode materials based on four different manganese dioxides, amorphous, birnessite, crytpomelane and spinel were fabricated and their electrochemical behaviors compared in two electrolytes, Li2SO4 and (NMe4)2SO4. With respect to the structural characteristics of the various prepared MnO2, these electrolytes can be differentiated by their cation size. Voltammetric studies showed that these electrode materials presented distinct capacitive behaviors depending on the electrolyte used. Modeling of the electrode capacitances measured at various scan rates allowed to discriminate the surface and material bulk contributions to the overall specific capacitance of the fabricated electrodes.
Keywords: manganese dioxide, cation size, electrolyte, surface capacitance, bulk capacitance
1. Introduction
Thanks to Faradaic contributions to their storage mechanism, metal oxides are the most studied pseudo-capacitive materials [1–4]. Nevertheless, they should address many technological and process specifications before being produced at the industrial-scale and
introduced as commercial products. Among these, their synthesis must be easy and at low cost, and the balance between energy density delivered thanks to the Faradaic reactions and power density must be favorable. Manganese dioxide is known to meet most of the needed requirements while being somehow a good compromise in term of performances, between purely capacitive and cheap activated carbons and the costly but highly pseudocapacitive ruthenium oxide. Indeed, it has been demonstrated that MnO2, easily prepared by various routes, presents attractive electrochemical performances [5,6] The pseudocapacitive charge storage mechanism in MnO2, described by the equation 1, has been studied many times and established as involving fast Mn(III)/Mn(IV) redox reactions and cationic exchange to balance the changes in the manganese oxidation state [7]. MnO2 + (x + y) e- + x H+ + y C+
->
MnOOHxCy
(C: cation from electrolyte)
Even if the storage mechanism has been described as occurring at the surface of the electrode material, it has also been shown to be independent on the developed surface area [8]. Moreover, charge/discharge mechanism has also been shown by in-situ XRD, to depend on the considered MnO2 polymorph on the material bulk structure, beyond the surface [9,10]. This
pseudocapacitance
is
a
“core”
or
“bulk”
capacitance,
involving
cation
intercalation/insertion, eventually dissociated from the Faradaic surface contribution to the total capacitance of the material. There are many types of manganese dioxides built on MnO6 octahedra sharing corner, edges or faces. The structural complexity of the various polymorphs has been extensively investigated and can be illustrated by the large number of synthetic or natural compounds referring to an identical structure. Birnessite is a MnO2 phase with a 2 dimensional layered structure, with layers made of edge sharing MnO6 octahedra and separated by a gap of about 7 Å. This phase is also called δ-MnO2 or OL-MnO2 for octahedrallayer [11,12]. In the same way, the channel-based cryptomelane MnO2 can be identified as α-MnO2 or OMS-2 for octahedral molecular sieves 2, and the tri-dimensional compact spinel phase as λ-MnO2 [13,14]. Due to their different cationic exchange properties governed by their cavity size and connectivity, the various structures of these polymorphs are suspected to lead to different balanced contributions of the surface or bulk charge storage processes to the overall capacitance. But, contrarily to the electrochemical double layer capacitors (EDLCs), where the charge storage process is purely electrostatic, the simultaneous contributions of both
surface and bulk Faradaic phenomena make complex the unambiguous description of the charge/discharge mechanism [15–17]. If the electrode material structure has a critical role in the charge storage mechanism, it is the same for the electrolyte ion nature. Ion adsorption mechanisms onto carbon materials have been extensively investigated and correlations between material porosity and capacitance have highlighted the ion size influence [18]. In MnO2, the anion volume and size were shown to impact on the intercalation of the cations in the structure and the involved redox reactions. As the anion size, the cation size plays also a crucial role on the MnO2 electrochemical performances [19]. Indeed, depending on the structural characteristics of the polymorph, too large electrolytic cations may degrade the electrode material behavior by the limitation of the associated intercalation/insertion in too small crystallographic sites and lead to a strong discrepancy between surface and bulk contributions. In contrast, small cations should be able to “probe” either surface and bulk electroactive sites. Moreover, cation access to the bulk is controlled by the solid-state diffusion while much faster surface electrosorption kinetics are governed by the ion diffusion at the solid-liquid interface. As such, it is anticipated that the surface and bulk capacitive contributions should by highly dependent on the charge/discharge rate [20]. The goal of this study is to highlight the contributions of surface and bulk to the overall capacitance by using different electrolytes and different MnO2 polymorphs. These MnO2 polymorphs present different crystallographic structures and different ion exchange properties. The behavior in cyclic voltammetry of amorphous MnO2, birnessite, cryptmelane and spinel phases have been compared at various scan rates, in two aqueous electrolytes of distinct cation sizes, Li2SO4 and (NMe4)2SO4, used at the same 1 M concentration.
2. Experimental All chemical reagents were analytical grade and were used as purchased without further purification. Several techniques were used to prepare a set of MnO2 based materials with various microstructures. All aqueous solutions were prepared using deionized water (18 MΩ, obtained from an ELGA Lab Water deionized system). 2.1 Synthesis of amorphous MnO2 Amorphous phase of manganese dioxide was prepared by the co-precipitation route via the reaction between KMnO4 and MnSO4.H2O as detailed in the following: 0.013 moles KMnO4
(99%, Alfa Aesar) were dissolved in 100 mL of deionized water (resistivity > 18 MΩcm, from an ELGALab Water deionized system) kept at 20 °C (water bath) to avoid any solution overheating. Separately, 0.019 moles of MnSO4.H2O (99%, VWR) was dissolved in 100 mL of deionized water. This latter solution was added drop-wise to the former MnO4- solution under stirring. Stirring was kept for 1 h after complete mixing of the solutions. Obtained suspension was filtered on Buchner, and washed many times with distilled water. Final product was dried at 90 ° C in air overnight. 2.2 Synthesis of birnessite phase MnO2 Firstly, a solution (A) was prepared by the dissolution of 0.04 moles of KMnO4 (99% Alfa Aesar) ad 1.2 moles of NaOH (ACS grade, VWR) in 400 mL of distilled water. A solution (B), prepared with 0.112 moles de MnCl2.4H2O (99%, Aldrich) in 400 mL of water, was added to the solution (A), drop by drop, and under stirring in an ice bath. Obtained suspension has kept under stirring, at room temperature for 24 h and then was washed with water and dried at 90 °C overnight. 2.3 Synthesis of cryptomelane phase MnO2 The cryptomelane were obtained from the ignition in air of MnO2 birnessite at 400 °C for 60 h. 2.4 Synthesis of spinel phase MnO2 In a first time, LiMn2O4 was prepared by dissolution of 0.01 mole of LiNO3 (99%, Fisher Scientific) and 0.02 moles of Mn(NO3)2·2H2O (ACS grade, Fisher Scientific) in a 100 mL mixture of citric acid (ACS grade, Aldrich) and ethylene glycol (ACS grade, Aldrich) (molar ratio 1:4). Then, the solution was heated at 90 ◦C for 30 min and the temperature was subsequently increased to 140 °C in order to activate esterification and remove the ethylene glycol excess until a coloration of the solution. This solution was heated at 180 °C under vacuum and brown slurry was obtained. The resulting slurry was annealed at 250 °C in air leading to LiMn2O4. Finally, the MnO2 spinel was synthesized by hydrolysis of LiMn2O4 in a 0.5 M HCl (ACS grade, Aldrich) aqueous solution for 24 h. 2.5 Material characterization
The crystalline structure of the materials was investigated with X-ray diffraction using a PANanalytical X’Pert Pro powder X-ray diffractometer (Cu Kα1 radiation) in a BraggBrentano configuration. The Brunauer – Emmett – Teller (BET) surface area and pore volume were measured with N2 sorption at liquid nitrogen temperature (77 K) using a Micrometrics ASAP 2020 instrument. Samples were degassed at 200 °C for 5 h in flowing nitrogen prior to measurements to remove guest molecules. The total surface area was determined by the multipoint Brunauer-Emmett-Teller (BET) method.
2.6 Electrode preparation and electrochemical characterization Prepared MnO2 powders and PTFE (from 60% wt PTFE dispersion in water) in a 90/10 ratio, were manually mixed in acetone until a homogeneous slurry was obtained. This slurry was spread on a glass plate and rolled several times to obtain thick films. Square pieces of films, about 1 cm2 in surface area and 150 µm in thickness, were cut out and pressed at 10 tons for a few minutes in a stainless steel grid used as current collector. In the whole series of prepared electrodes, MnO2 loadings ranged from 9.1 to 12.3 mg.cm-2. Before electrochemical characterization, the electrodes were dried overnight under vacuum at room temperature and immersed, under vacuum, in the electrolyte solutions for 4 h. Electrochemical measurements were carried out using a Biologic VMP3 potentiostat controlled by EC-Lab V10.38 software. The electrochemical cell used in a conventional three-electrode setup was composed of a Pt counter electrode, an Ag/AgCl (3 M KCl) reference electrode and the composite electrode as the working electrode.
3. Results and discussion
Crystallographic structures of birnessite, cryptomelane and spinel forms of MnO2 are schematized in Figure 1 (a), (b) and (c), respectively. These structural arrangements are built on MnO6 octahedra building blocks sharing faces, edges or corners. Birnessite is obtained by slow reduction of MnO4+ by MnCl2 leading to a layered form of MnO2 made by stacking edge sharing MnO6. The presence of Na+ cations in the interlayer gap of about 7 Å accounts for the birnessite structure stability. Heating birnessite at 400 °C under air leads to the cryptomelane form of MnO2. During this heat treatment, part of Na+ is extracted from the layered pristine structure. Destabilized, the layered structure is converted as a channeled one. Resulting
channels, made of 2 x 2 arrangements of MnO6 octahedra sharing their edges, have side dimensions of 5 Å and 6.3 Å in diagonal. The spinel form of MnO2 is a tridimensional lithiated phase, formed by 1 x 1 channels built on octahedra sharing edges. The resulting tetrahedric site is about 2.9 Å in size.
Thanks to its layered structure, 2D-MnO2 birnessite allows a good mobility of the cations into the bulk. Because of more constrained structural arrangements, the progressive ionic conductivity loss from birnessite to cryptomelane and then spinel is expected to impact on the Faradaic bulk contribution to the capacitance and favors surface reactions [10]. XRD pattern of the various MnO2 phases synthesized during this study, amorphous MnO2 (a), birnessite (b), cryptomelane (c) and spinel (d) are shown in Figure 2. Amorphous behavior of MnO2 is confirmed by the presence of broad and low intensity diffraction peaks (a). Diffractogramm (b) shows similar features as (a) but with more defined diffraction peaks at 2θ angles of 12.2°, 24.7°, 36.9°. These peaks are unambiguously assigned to the monoclinic (C2/m) layered structure of the MnO2 birnessite polymorph (JCPDS 43-1456) [12,21]. In such a structure, layers are containing a mixture of Mn(IV) and Mn(III) [22–24]. The peak at 12.2° (2θ) corresponds to the (001) planes and to the interlayer distance at 7.0 Å. Its shape and intensity, as well as angular position, are highly sensitive to the phase crystallinity and, nature and content of the alkali cations intercalated in between the layers. The (002) peak at 24.5° (2θ) is also related to the interlayer spacing. Cryptomelane XRD pattern in Figure 2(c) is very different from those of the other MnO2 polymorphs in the series. Indeed, this diffractogramm present narrow and sharp peaks, indicating that the birnessite heating increases the crystallinity of the resulting cryptomelane. Main peaks are at 2θ angles of 12.7°, 18.1°, 28.7°, 37.2°, 41.9° and 49.5° (JCPDS 29-1020) corresponding to a 2 x 2 channeled structure in the I4/m group [25]. Spinel XRD pattern in Figure 2(d) is characteristic of a lower crystallinity. Nevertheless, main diffraction peaks at the 2θ angles of 19.3°, 37.3°, 38.9°, 45.2° and 49.5° (JCPDS 44-0992) confirm the tridimensional cubic structure, assigned to the Fd3m group [14]. The peaks at about 37° (2θ), shared by the diffractogramms in the series, is commonly observed in manganese oxide diffraction patterns since it is related to interatomic distances in materials based on edge-sharing MnO6 octahedra.
In this study, two different electrolytic salts were used, Li2SO4 and (NMe4)2SO4, at a 1 M concentration. Each salt obviously presents distinct characteristics. Some of them are related to the cation size which is given in Table 1. In aqueous solution, such monovalent cations are usually surrounded by one or more hydration shell. Li+ cation, the smallest of both, is the most polarizing and builds a hydration shell made of four water molecules in a tetrahedral symmetry. However, interactions are so strong that additional hydration shells surround the first one [26].
Actually, Li+ locally modifies the structure of the water molecules by breaking hydrogen bonds, and promotes the local water environment as structured hydration shells. This explains why its hydrated radius (3.82 Å) is very large before its radius as bare cation (0.68 Å). In contrast, NMe4+ is a destructuring cation, as it hardly forms any hydration shells, leading to a lower solvation than Li+ [26,27]. Table 2 shows the cavity sizes of prepared manganese oxides and a rapid comparison with cation size may stress some steric issues unfavorable to intercalation/insertion into the material bulk. It seems however important to keep in mind that the radius of hydrated cation is not the effective radius when intercalated/inserted/confined in the material structure [10,28,29]. Indeed, many authors have shown that ions could be desolvated in confined media [28,29]. As such, in the material bulk, Li+ is smaller (1.95 Å radius) than NMe4+ (3.22 Å radius). It is also obvious that, because of its large size, tetramethylammonium ion, even as desolvated cation, cannot access the bulk of any of the manganese dioxide crystallized polymorphs in the prepared series but birnessite for which a 7 Å interlayer space can a priori afford the intercalation of a 6.45 Å diameter NMe4+ bare cation. Cyclic voltammograms (CVs) for the various prepared MnO2 polymorph electrodes, amorphous, birnessite, cryptomelane and spinel, are shown in Figure 3A, B, C and D respectively. These were measured at various scan rates from 5 to 200 mV.s-1 in a three electrode configuration from 0 to 0.85 V (vs AgAgCl) in 1 M Li2SO4 (a) and (NMe4)2SO4 (b) aqueous electrolytes. The roughly rectangular shape of the CVs is characteristic of the pseudocapacitive behaviour of the prepared electrode materials. As electrode materials, amorphous and birnessite phases of MnO2 are presenting the most capacitive electrochemical signatures, either in Li2SO4 or (NMe4)2SO4 -based electrolytes. Some Faradaic contributions are visible, especially for birnessite (Figure 3B) and cryptomelane (Figure 3C) -based
electrodes in Li2SO4 electrolyte. These CVs highlight the influence of the electrolyte nature on the electrochemical behaviour of each electrode. Generally speaking, prepared materials show greater capacitances when operated in Li2SO4 electrolyte. However, even if the electrode capacitance is lower in (NMe4)2SO4 -based electrolyte, pseudocapacitive behaviour is confirmed in this electrolyte too. To highlight the effect of the electrolyte nature on the electrochemical behaviour of the prepared MnO2 phases, Figure 4 shows the 50th voltammetric cycle of the prepared electrode materials carried out at a scan rate of 5 mV.s-1 in 1M aqueous Li2SO4 (solid line) and (NMe4)2SO4 (dashed line). As already mentioned, depicted voltammetric cycles are more or less large but remain characteristic of a pseudocapacitive behaviour. Switching from Li2SO4 and (NMe4)2SO4 electrolyte, induces several changes in terms of shape and amplitude (current density) of the measured CVs, depending on the MnO2 phase considered. For example, the capacitive behaviour of the birnessite based electrode is greater in Li2SO4 than in (NMe4)2SO4 electrolyte (Figure 4b). In the former electrolyte, the cycle is more rectangular but especially larger, corresponding to higher current densities than in the latter. Similar current density fading can also be observed for the other electrode materials in the series. The electrolyte nature also impact on the eventual Faradaic contributions such as those observed for the cryptomelane electrode material. When operated in Li2SO4, the corresponding voltammogram in Figure 4c shows redox peaks, highlighting the contribution of Faradaic phenomena involved in the charge storage. These peaks, at about 0.5 and 0.2 V (vs AgAgCl), are hardly visible when the electrode is operated in 1 M aqueous (NMe4)2SO4 electrolyte. Table 3 lists the specific capacitances measured by CV at 5 mV.s-1 for the various prepared MnO2 based electrodes in both electrolytes. As suggested by the CV shapes in Figure 4, the capacitance losses are not the same for all the samples when switching from Li2SO4 to (NMe4)2SO4 -based electrolytes. For the spinel and amorphous MnO2, the capacitance losses are about 26 and 30 %, respectively. For birnessite and cryptomelane based electrodes, the capacitance losses are more marked. The capacitance of the birnessite based electrode at 200 F.g-1 in Li2SO4, drastically decreases when operated in (NMe4)2SO4 electrolyte (112 F.g-1 ). The corresponding capacitance loss is about 44 % and is accompanied by a strong deformation of the voltammogram shape. Cryptomelane electrode also presents a lower capacitance when used in (NMe4)2SO4 electrolyte. When operated in Li2SO4, the
corresponding capacitance is about 102 F.g-1 and falls down to 55 F.g-1 when in (NMe4)2SO4 electrolyte, corresponding to a 45 % loss.
In regards to the scan rate, a similar behavior is observed for the prepared electrodes in both electrolytes. Measured CVs remain characteristic of pseudocapacitive materials as scan rate is increased but the corresponding current densities are strongly affected. Figure 5 shows the changes in the relative capacitance with the scan rate. Whatever the considered MnO2 phases, capacitance is highly dependent on the scan rate and drastically decreases until 50 mV.s-1. For example, from 5 to 50 mV.s-1, up to 95 % capacitance is lost for amorphous and spinel MnO2 phases in Li2SO4 -based electrolyte. Except for the birnessite based electrode, the relative capacitances at high scan rate are higher in (NMe4)2SO4 electrolyte than in Li2SO4 electrolyte. This lower capacitance loss can be explained by (NMe4)+ being too large to be intercalated/inserted in the cryptomelane and spinel structures. The corresponding capacitance, in these cases, is originating from surface contributions, eventually limited by the ion diffusion at the solid/liquid interface.
Figure 6 is an alternative way to compare the capacitance fading of the prepared electrodes in Li2SO4 and (NMe4)2SO4 electrolytes. Using such a logarithmic scale, resulting curves are roughly linear and their slopes dependent on the material nature and electrolyte. Calculated slopes are given in the table 4. Electrodes made of amorphous (a) and birnessite (b) MnO2 are characterized by very similar slopes when cycled in Li2SO4 or (NMe4)2SO4 electrolytes, respectively at - 0.58 and - 0.53 for amorphous MnO2, and - 0.42 and - 0.46 for birnessite. These slight slope variations suggest that the behavior of Li+ and NMe4+ cations towards both amorphous and birnessite MnO2 is about the same when scan rate is increased. Slopes extracted from measurements done with cryptomelane and spinel -based electrodes (Figures 6c and d, respectively) present larger discrepancies when cycled in Li2SO4 and (NMe4)2SO4 electrolytes as a proof of charge storage mechanisms being different in these materials depending on the electrolyte. For these two electrode materials, the line slopes obtained from cycling in (NMe4)2SO4 electrolyte are lower than those in Li2SO4 electrolyte. This difference in scan rate capability comes from the charge storage in Li2SO4 involving the intercalation/insertion of Li+ in the structure of the material, which is drastically altered at high scan rate because of a limited solid state diffusion in the materials. In contrast, NMe4+ is
too large to diffuse into the material structure and its interaction, limited to the surface of the electrode material, is responsible for the lower capacitance than in Li2SO4 electrolyte at low scan rate but also less or un-affected by eventual diffusion limitations at high scan rate. In contrast, the slope from measurements with birnessite MnO2 based electrode in (NMe4)2SO4 is slightly greater than in Li2SO4 electrolyte. As already mentioned, the interlayer distance of about 7 Å is large enough for this material to host either Li+ or NMe4+ cations. Obviously, the larger size of the latter is less favourable to diffusion in the solid and the impact of the increased scan rate on the measured capacitance is consequently greater. Finally, despite a “structure” that can be described as a disordered birnessite, the electrochemical behavior of amorphous MnO2 is mostly led by surface interactions and the measured slope is slightly greater in Li2SO4 than in (NMe4)2SO4 electrolyte. In summary, in the case of (NMe4)2SO4 electrolyte, the charge storage mechanism is less dependent on diffusion properties of the cations into the bulk of the prepared electrode materials but in MnO2 birnessite, and the effect of the scan rate on the capacitance is lesser. In contrast,
because
of
storage
mechanisms
involving
solid
state
diffusion
and
intercalation/insertion phenomena of Li+, capacitance is more scan rate sensitive when using Li2SO4 electrolyte. Using such logarithmic scale, it is possible to express the specific capacitance C of the electrode materials as a linear function of the scan rate v by the following equation [30]: log C= a log v + logC 1
Where a is the line slope discussed above, and C1 the capacitance value at 1 mV.s-1 scan rate. As such, a linear relationship between the specific capacitance, C and the opposite of the scan rate square root, v-1/2, can be expressed by the following equation: C = constante × v
− 1 /2
+C ∞
In this equation, the intercept with the C axis, C ∞ , is the capacitance at infinite scan rate. At such infinite scan rate, diffusion of electrolytic species into the electrode material bulk cannot be considered and C ∞ corresponds to the contribution of the surface material only to the overall specific capacitance. The corresponding plots are depicted in figure 7. In a similar way, the capacitance inverse, C-1, can be expressed as a function of the scan rate square root to obtain, at the intercept, the material capacitance at infinitely low scan rate thanks to the following equation: −1
C = constante × v
1/ 2
−1
+ C0
−1 Here, C0 corresponds to the specific capacitance inverse at infinitely low scan rate. This
capacitance can be assigned to the total capacitance that can be developed by the electrode material, involving both surface and bulk contributions, without any diffusion limitations neither at liquid nor solid states. These two equations have already been used, in particular by Ardizzione et al, for the study of RuO2 capacitance contributions in the more or less accessible regions of the electrode material [31,32]. Figures 7 and 8 show the capacitance changes as a function of the inverse of the scan rate square root (C = f (v-1/2)) and the capacitance inverse as a function of the scan rate square root (C-1 = f (v1/2)), respectively. The surface contribution to the capacitance of the electrode material C∞ is extracted from figure 7 whereas figure 8 highlights the total capacitance of the electrode material, C0. Thanks to these values, the capacitance material core contribution Cbulk can be easily calculated by considering the total capacitance as the sum of the surface and bulk capacitances and the following: Cbulk = C0 - C∞. Table 5 gives both C∞ and C0 capacitance values with their uncertainty values (automatically generated by OriginLab software). From the graphs in Figure 7, the plot intercepts correspond to very small surface capacitances, ranging from 2 to 8 F.g-1, in the whole series of electrode materials and electrolytes. Spinel C∞, at about 2 F.g-1, is the lowest in the series. In a first approximation, C∞ seems to hardly depend on the electrolyte and polymorph natures.
The total capacitance developed by the electrodes is extracted from the graphs in Figure 8. At such an infinitely low scan rate, the kinetic transport and the ion diffusion into the electrolyte and the material do not alter the charge storage mechanism and both surface and bulk contribution maximize the overall capacitance of the considered electrode material. Total capacitance values are way greater than measured surface capacitances, in some cases, by more than two orders of magnitude. At such a very low scan rate, bulk contribution to the overall capacitance is the most important but seems to strongly depend on the considered electrode materials in the series. Specific capacitance for birnessite phase is calculated at about 1210 and 1159 F.g-1 in Li2SO4 and (NMe4)2SO4 respectively. These remarkable values close to the theoretical capacitance of MnO2 at about 1300 F.g-1, demonstrate that birnessite structure allows enhanced cationic exchanges regardless of the considered electrolyte. In this case, Faradaic phenomena are dominating and intercalation processes are favored. For other
prepared MnO2 phases, the total capacitances are lower (from 200 to 700 F.g-1) and drastically decrease when used in (NMe4)2SO4 instead of Li2SO4 electrolyte. This later point expresses the difficulty or impossibility for the large NMe4+ cations to be intercalated/inserted into the structure of either amorphous, cryptomelane or spinel MnO2 prepared materials.
On the other hand, considering this size effect and the surface to solely contribute to the capacitance, the total capacitances (C0) developed by cryptomelane and spinel in (NMe4)2SO4 electrolyte should be equal to the surface capacitance determined at infinitely high scan rate (C∞). This is obviously not true since for cryptomelane, C0= 196 F.g-1 and C∞= 6 F.g-1 while for spinel, C0= 204 F.g-1 and C∞= 2 F.g-1. A first hypothesis to account for these discrepancies could be that, contrarily to what was supposed earlier, C∞ capacitance, as calculated here, does not represent the total contribution of the material surface. The very low capacitance values C∞, listed in the Table 5, could actually be assigned to the unique double layer contribution to the capacitance, originating from electrostatic interactions, to differentiate from a Faradaic surface contribution. If these C∞ values correspond to the double layer capacitance developed by MnO2 electrode materials studied here, then it should be possible to compare them to the corresponding specific surface area values listed in the Table 6. Amorphous MnO2, which develops the higher specific surface area about 78 m².g-1, shows a specific capacitance at high scan rate similar to the other materials. Cryptomelane, with a specific surface area about 40 m².g-1, similar to that of spinel, but the half of that of amorphous phase, still presents a specific capacitance about 6 F.g-1 at infinite scan rate. At this stage, it does not seem possible to establish any correlation between these reported capacitance values and the specific surface area developed by these materials. It can also be interesting to compare these surface capacitances to the electronic conductivities of the various polymorphs given in the Table 6. Electronic conductivity values strongly depend on the material nature. While they present a similar C∞ capacitance, birnessite MnO2 appears as 103 less conductive than cryptomelane or spinel forms, while the latter present the lowest C∞ capacitance in the series. Again, the electronic conductivity characteristics of the prepared materials do not fit the values and their assignment to double layer contributions cannot be confirmed. A second hypothesis can be raised for the very low calculated C∞ capacitances. Whatever is the considered material and electrolyte, they all range from 2 to 8 F.g-1. These values do not
depend on the developed surface area of the electrode material nor on its electronic conductivity and are not correlated to the size of the electrolytic cation. They however express the impact of the scan rate on the surface contribution and at this infinitely strong regime, diffusion limitations of the electrolytic species at the liquid state within the electrolyte bulk and/or at the electrode/electrode interface may have to be considered showing that the scan rate increase does not only impact on the material bulk contribution but also on the surface contribution. It remains thus difficult, in regard to the different observations, to establish any correlation between C∞ and a double layer capacitance, and to conclude on the origin of this contribution, although limited, to the MnO2 capacitance. In contrast, since only the surface of cryptomelane or spinel MnO2 is “probed” (and not the bulk) when operated in (NMe4)2SO4 electrolyte, C0 values at about 200 F.g-1 correspond to the Faradaic contribution that can be provided at most by the material surface. Thanks to its “structure” deriving from birnessite, a mixed surface/bulk contribution to the C0 capacitance, calculated at 500 F.g-1, should probably be considered for the amorphous MnO2 based electrode.
4. Conclusion This study is based on the comparison of the electrochemical behavior of four MnO2 based electrode materials, amorphous, birnessite, cryptomelane and spinel, showing various structural arrangements and cation hosting capabilities. To promote cation size effects, two different electrolytic media, 1M aqueous Li2SO4 and (NMe4)2SO4 were used. Material specific capacitances were calculated and expressed as a function of the scan rate used in cyclic voltammetry experiments. Materials showed distinct capacitive behaviors depending on the used electrolyte and specific capacitances were shown to drastically decrease with the scan rate. To differentiate and assign the surface and bulk contributions to the overall capacitance, a diffusion -based modeling was used. The expression of the capacitance as a function of the inverse of the scan rate square root, and the capacitance inverse as a function of the scan rate square root, demonstrated that the total capacitance is balanced by both bulk and surface contributions. This balance strongly depends on the electrode material structure (and cationic exchange capabilities), on the electrolyte nature, and, consequently, on the charge/discharge rate. Although a contribution from the electrochemical double layer cannot be excluded, it is fairly limited to at most 1% of the overall capacitance. The limited
capacitance retention observed upon scan rate originates from the kinetics of Faradaic phenomena involved in their charge storage mechanisms. These phenomena are mostly influenced by the ion mobility into the material structure.
Acknowledgements The Centre National de la Recherche Scientifique is gratefully acknowledged for funding. This work has been performed with the support of the PROGELEC program of the Agence Nationale de la Recherche (ANR) through the FlexCap project (grant #073616). More information available at www.flexcap.fr.
References 1.
Ghodbane, O., Ataherian, F., Wu, N.-L. & Favier, F. In situ crystallographic investigations of charge storage mechanisms in MnO2-based electrochemical capacitors. J. Power Sources 206, 454–462 (2012).
2.
Augustyn, V., Simon, P. & Dunn, B. Pseudocapacitive oxide materials for high-rate electrochemical energy storage. Energy Environ. Sci. 7, 1597 (2014).
3.
Conway, B. E., Birss, V. & Wojtowicz, J. The role and utilization of pseudocapacitance for energy storage by supercapacitors. J. Power Sources 66, 1–14 (1997).
4.
Jo, C., Hwang, I., Lee, J., Lee, C. W. & Yoon, S. Investigation of pseudocapacitive charge-storage behavior in highly conductive ordered mesoporous tungsten oxide electrodes. J. Phys. Chem. C 115, 11880–11886 (2011).
5.
Coustan, L. & Favier, F. Microwave-Assisted Decoration of Carbon Substrates for Manganese Dioxide-Based Supercapacitors. J. Electrochem. Soc. 162, A5133–A5139 (2015).
6.
Cormie, A., Cross, A., Hollenkamp, A. F. & Donne, S. W. Cycle stability of birnessite manganese dioxide for electrochemical capacitors. Electrochim. Acta 55, 7470–7478 (2010).
7.
Toupin, M., Brousse, T., Belanger, D. & Be, D. Charge Storage Mechanism of MnO Electrode Used in Aqueous Electrochemical Capacitor Charge Storage Mechanism of MnO 2 Electrode Used in Aqueous Electrochemical Capacitor. Chem. Mater. 3184– 3190 (2004). doi:10.1021/cm049649j
8.
Brousse, T., Toupin, M., Dugas, R., Athouel, L., Crosnier, O. & Belanger, D. Crystalline MnO[sub 2] as Possible Alternatives to Amorphous Compounds in Electrochemical Supercapacitors. J. Electrochem. Soc. 153, A2171 (2006).
9.
Ghodbane, O., Louro, M., Coustan, L., Patru, A. & Favier, F. Microstructural and Morphological Effects on Charge Storage Properties in MnO2-Carbon Nanofibers Based Supercapacitors. J. Electrochem. Soc. 160, A2315–A2321 (2013).
10.
Ghodbane, O., Pascal, J.-L., Fraisse, B. & Favier, F. Structural in situ study of the thermal behavior of manganese dioxide materials: toward selected electrode materials for supercapacitors. ACS Appl. Mater. Interfaces 2, 3493–505 (2010).
11.
Julien, C. ., Massot, M. & Poinsignon, C. Lattice vibrations of manganese oxides. Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 60, 689–700 (2004).
12.
Athouël, L. Moser, F., Dugas, R.,Crosnier, O., Bélanger, D &Brousse, T. Birnessite as Possible Candidate for Hybrid Carbon/MnO2 Electrochemical Capacitor. ECS Trans 112, 7270–7277 (2008).
13.
Ghodbane, O., Pascal, J.-L. & Favier, F. Microstructural effects on charge-storage properties in MnO2-based electrochemical supercapacitors. ACS Appl. Mater. Interfaces 1, 1130–9 (2009).
14.
Lin, Y.-P., Tsai, C.-B., Ho, W.-H. & Wu, N.-L. Comparative study on nanostructured MnO2/carbon composites synthesized by spontaneous reduction for supercapacitor application. Mater. Chem. Phys. 130, 367–372 (2011).
15.
Simon, P., Gogotsi, Y. & Dunn, B. Where do batteries end and supercapacitors begin? Science 343, 1210–1211 (2014).
16.
Winter, M. & Brodd, R. J. What are batteries, fuel cells, and supercapacitors? Chem. Rev. 104, 4245–69 (2004).
17.
Brousse, T., Belanger, D. & Long, J. W. To Be or Not To Be Pseudocapacitive? J. Electrochem. Soc. 162, A5185–A5189 (2015).
18.
Simon, P. & Gogotsi, Y. Materials for electrochemical capacitors. Nat. Mater. 7, 845– 854 (2008).
19.
Chmiola, J. Yushin, G., Gogotsi, Y., Portet, C., Simon, P &Taberna, P L Anormalous increase in carbon capacitance at pore sizes less than 1 nanometer. Science (80-. ). 313, 1760–1763 (2006).
20.
Boisset, A. & Athouël, L. Comparative Performances of Birnessite and Cryptomelane MnO2 as Electrode Material in Neutral Aqueous Lithium Salt for Supercapacitor Application. J. 2, (2013).
21.
Moser, F., Athouël, L., Crosnier, O., Favier, F., Bélanger, D. & Brousse, T. Transparent electrochemical capacitor based on electrodeposited MnO2 thin film electrodes and gel-type electrolyte. Electrochem. commun. 11, 1259–1261 (2009).
22.
Drits, V. A., Silvester, E., Gorshkov, A. I. & Manceau, A. Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite .1. Results from X-ray diffraction and selected-area electron diffraction. Am. Miner. 82, (1997).
23.
Silvester, E., Manceau, A. & Drits, V. A. Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite .2. Results from chemical studies and EXAFS spectroscopy. Am. Miner. 82, (1997).
24.
Ramstedt, M., Shchukarev, A. V. & Sjoberg, S. Characterization of hydrous manganite (γ-MnOOH) surfaces- An XPS study. Surf. Interface Anal. 34, 632–636 (2002).
25.
Toupin, M., Brousse, T. & Be, D. Influence of Microstucture on the Charge Storage Properties of Chemically Synthesized Manganese Dioxide. Chem. Mater. 3946–3952 (2002).
26.
Jolivet, J. P. De la solution à l’oxyde. Inter Ed. Editions C, (1994).
27.
Ohtaki, H. & Radnai, T. Structure and dynamics of hydrated ions. Chem. Rev. 93, 1157–1204 (1993).
28.
Lynden-Bell, R. & Rasaiah, J. Using simulation to study solvation in water. J. Chem. Phys. 105, 9266 (1996).
29.
Tanimura, A., Kovalenko, A. & Hirata, F. Molecular theory of an electrochemical double layer in a nanoporous carbon supercapacitor. Chem. Phys. Lett. 378, 638 (2003).
30.
Fregonara, G., Fregonara, G., Trasatti, S. & Trasatti, S. ‘Inner’ and ‘Outer’ Active Surface Electrodes. 5–9 (1989).
31.
Athouel, L. Arcidiacono, P., Ramirez-Castro, C., Crosnier, O., Hamel, C. Investigation of cavity microelectrode technique for electrochemical study with manganese dioxides Electrochim. Acta 86, 268–276 (2012).
32.
Ardizzione, S., Fregonara, G. & Trasatti, S. “Inner” and “outer” active surface of RuO2 electrodes. Electrochim. Acta 35, 263 (1990).
Figure 1: Crystallographic structures of birnessite (a), cryptomelane (b) and spinel (c) MnO2 phases
Figure 2: XRD patterns of MnO2 polymorphs: amorphous (a), birnessite (b), cryptomelane (c) and spinel (d)
Figure 3: Cyclic voltammograms for the various prepared MnO2 electrodes: amorphous (A), birnessite (B), cryptomelane (C) and spinel (D) for Li2SO4 (a) and (NMe4)2SO4 (b) at scan rates from 5 to 200 mV.s-1
Figure 4: Cyclic voltammograms at the 50th cycle in both electrolytes at 5 mV.s-1 for amorphous (a), birnessite (b), cryptomelane (c) and spinel (d)
Figure 5: Relative capacitance upon scan rate for amorphous (a), birnessite (b), cryptomelane (c) and spinel (d) MnO2 based electrodes
Figure 6: Specific capacitance upon scan rate in log/log scale for amorphous (a), birnessite (b), cryptomelane (c) and spinel (d) in logarithmic scale
Figure 7: Specific capacitance as a function of the inverse of the square root scan rate for the various prepared electrode materials operated in Li2SO4 and (NMe4)2SO4 electrolytes
Figure 8: Inverse of the specific capacitance as a function of the scan rate square root for the various prepared electrode materials operated in Li2SO4 and (NMe4)2SO4 electrolytes
Table 1: Radii of the cations in used electrolytes Cation
Li+
Bare cation radius (Å) 0.68
Hydrated cation radius (Å) 3.82
1.95
3.22
3.22
3.22
(NMe4)+
Radius with first hydration shell (Å)
Table 2: Cavity sizes in MnO2 polymorphs MnO2 polymorph Cavity dimension (Å) Birnessite
7
Cryptomelane
5 x 6.2
Spinel
2.9
Table 3: MnO2 polymorph specific capacitances from measurements at 5 mV.s-1 in Li2SO4 and (NMe4)2SO4 1 M aqueous electrolytes Li2SO4 (NMe4)2SO4
Amorphous 139 F.g‐1 97 F.g‐1
Birnessite 200 F.g‐1 112 F.g‐1
Cryptomelane 102 F.g‐1 55 F.g‐1
Table 4: log C = f(log v) line slopes for MnO2 based electrodes Amorphous Birnessite Cryptomelane Spinel
Li2SO4 ‐0.58 ‐ 0.42 ‐ 0.52 ‐ 0.59
(NMe4)2SO4 ‐0.53 ‐ 0.46 ‐ 0.36 ‐ 0.39
Spinel 78 F.g‐1 56 F.g‐1
Table 5: C∞ and C0 capacitances extracted from graphs 7 and 8 for amorphous, birnessite, cryptomelane and spinel MnO2 phases MnO2 polymorph Amorphous Birnessite Cryptomelane Spinel
Li2SO4 C∞ (F.g‐1) 6.1 ± 1.4 7.7 ± 1.4 6.3 ± 1.1 2.1 ± 0.3
(NMe4)2SO4 C∞ (F.g‐1) C0 (F.g‐1) 5.8 ± 0.9 490 ± 87 6.4 ± 1.0 1159 ± 166 6.3 ± 1.1 196 ± 25 1.6 ± 0.3 204 ± 29
C0 (F.g‐1) 709 ± 95 1210 ± 181 476 ± 68 416 ± 64
Table 6 Specific surface areas and electronic conductivities of the studied MnO2 polymorphs MnO2 polymorph Specific surface area (m².g‐ 1 ) Electronic conductivity S.cm‐1
Amorphous
Birnessite
Spinel
40
Cryptomelan e 50
78 2 .10‐5
6 .10‐6
9 .10‐3
3 .10‐3
50