Article pubs.acs.org/JPCB
Influence of Zn2+ and Water on the Transport Properties of a Pyrrolidinium Dicyanamide Ionic Liquid T. J. Simons,† P. M. Bayley,† Z. Zhang,§ P. C. Howlett,† D. R. MacFarlane,‡ L. A. Madsen,§ and M. Forsyth*,† †
Australian Centre for Electromaterials Science (ACES), Institute for Frontier Materials (IFM), Deakin University Burwood Campus, Burwood 3125, Australia ‡ Australian Centre for Electromaterials Science (ACES), School of Chemistry, Monash University, Clayton 3800, Victoria, Australia § Department of Chemistry and Macromolecules and Interfaces Institute, Virginia Tech, Blacksburg, Virginia 24061, United States S Supporting Information *
ABSTRACT: In order to expand our understanding of a potential zinc-based battery electrolyte, we have characterized the physical and transport properties of the ionic liquid (IL) 1-butyl-1-methylpyrrolidinium dicyanamide ([C4mpyr][dca]) containing various levels of both Zn2+ and H2O. Detailed measurements of density, viscosity, conductivity, and individual anion and cation diffusion coefficients using pulsed-field-gradient (PFG) NMR combined with NMR chemical shifts and spin−lattice relaxation (T1) NMR experiments provide insights into the motion and chemical environment of all molecular species. We find that the various techniques for probing ion transport and dynamics form a coherent picture as a function of electrolyte composition. Zn2+ addition causes a moderate reduction in the self-diffusion of the IL anion and cation, whereas the addition of H2O increases ion mobility by increasing the liquid’s overall fluidity. Temperaturedependent 13C T1 experiments of the dca carbon analyzed using Bloembergen−Purcell− Pound fits show monotonic slowing of anion dynamics with Zn2+ addition, suggesting increased Zn2+/dca− association. T1 experiments show minimal change in the spin−lattice relaxation of cation or anion upon H2O addition, suggesting that H2O is playing no significant role in Zn2+ speciation. Finally, we employ a novel electrophoretic NMR technique to directly determine the electrophoretic mobility of the C4mpyr cation, which we discuss in the context of impedance-based conductivity measurements.
■
INTRODUCTION Currently, there is increasing demand for high-capacity energy storage devices that are sustainable, affordable, and environmentally responsible.1 Zn-based batteries, including Zn−air batteries, show promise for commercialization due to the low reactivity, relatively low cost, existing recycling infrastructure, and nontoxicity of zinc metal.2−6 However, in the case of Zn− air batteries there are many obstacles to overcome before a commercially available rechargeable (secondary) device can be produced. These issues include electrolyte evaporation, nonuniform deposition during recharge, poor cycle life, and poor efficiency. 7,8 While many researchers have been researching novel oxygen-reduction catalysts2,9,10 and zinc anode compositions,6,11 many of these problems can also be addressed through novel electrolyte design. Ionic liquids (ILs) are an established class of compounds whose properties can include low volatility, ionic conductivity, thermal stability, electrochemical stability, and high metal−salt solubility.12 These properties have made ILs the subject of much scrutiny for applications in electrochemical devices such as lithium−metal batteries,13,14 supercapacitors,15−17 and fuel cells.12 Their common low-volatility, conductivity and ability to electrodeposit metals have recently made various ILs into © 2014 American Chemical Society
appealing candidates for novel electrolytes in zinc-based batteries. Of the many anions that have been investigated, one of the most popular for electrochemical processes has been the bis(trifluoromethanesulfonyl)imide (NTf2) anion. Several groups, including those of Tulodziecki et al.,18 Chen et al.,19 Doan et al.,20 and Azaceta et al.,21 have investigated the electrochemistry and deposition morphology of Zn,19 ZnO,18,21 and alloys of Zn19 in various NTf2 ILs. Despite the low viscosity and high stability of ILs based on the NTf2 anion, deposition and stripping currents of Zn electrochemistry do not suggest outstanding performance in an electrochemical device such as a battery or supercapacitor. The economic and environmental considerations of using NTf2 in a mass produced device have also driven the authors to consider alternate ILs with nonhalogenated anions in this work. One class of ILs that has been extensively studied in relation to zinc are those containing the dicyanamide (dca) anion.22,23 Deng et al. first investigated the electrochemistry of the Zn0/ Received: February 16, 2014 Revised: April 8, 2014 Published: April 8, 2014 4895
dx.doi.org/10.1021/jp501665g | J. Phys. Chem. B 2014, 118, 4895−4905
The Journal of Physical Chemistry B
Article
Zn2+ couple in both imidazolium24 and pyrrolidinium25 dca, after which Xu et al.26 fully characterized the mechanisms of deposition on Pt electrodes. Subsequently, Simons et al.27 investigated the importance of the added Zn2+ salt, concluding that Zn(dca)2 yielded a deposition/stripping Coulombic efficiency of 85% and current densities 10 times greater than those found in the studies of Deng and Xu, who used ZnCl2 and Zn(NTf2)2, respectively. Since then, our group has investigated Zn(dca)2 in dca-based ILs both electrochemically and spectroscopically, concluding that formation of a Zn(dca)x2‑x complex anion is responsible for the excellent electrochemical properties observed thus far.28,29 Nuclear magnetic resonance (NMR) has been used extensively for IL-based electrolyte characterization.30−36 Noda et al.34 have used pulsed-field-gradient (PFG) diffusion NMR to investigate the proton conduction and diffusion mechanism of a protic ionic liquid, imidazolium NTf2. A chemical shift analysis of the 1H spectra provided evidence that the proton was transported via the imidazole in a “vehicularlike” fashion through the electrolyte. Similar self-diffusion measurements on the nonprotic 1-ethyl-3-methylimidazolium NTf2 IL have been performed by probing the 19F nuclei for the anion and 1H for the cation to infer diffusion-based transference numbers for the species involved. These measurements have been paired with conductivity measurements to provide evidence of ion association in the electrolyte solutions, which may decrease an electrolyte’s efficiency in a battery. Hou et al.33 conducted a detailed study of ion associations in pure ILs and in ILs absorbed into polymer electrolytes in order to advance more quantitative ideas on the clustering of IL ions due to variations of specific intermolecular interactions. A recent study by Zhang and Madsen37 has used electrophoretic NMR to observe separate cation and anion mobilities in two pure ILs, which verify pure vehicular ion transport in these systems and provide a new method to investigate electrolyte transport. Bayley et al.30 effectively used T1 relaxation experiments to show that tetrahydrofuran can be used to control lithium speciation in an IL electrolyte for lithium−metal batteries. In particular, the combination of variable temperature T1 and diffusion measurements aided in understanding the role of these additives in lithium complexation. Hayamizu et al.31,32 also combined both physical and spectroscopic techniques to investigate the librational motions of imidazolium cations and information on the correlation time of a lithium jump in a series of NTf2 and bis(fluorosulfonylimide) IL−lithium salt mixtures. In order to further our understanding of the physical properties and molecular dynamics of the species present in the [C4mpyr][dca]/Zn(dca)2 electrolyte systems, we have investigated these aspects through the use of PFG diffusion (1H and 13 C nuclei), 13C chemical shifts, and 13C T1 NMR experiments on the [C4mpyr][dca] IL. We have analyzed both the physical properties and the relaxation times of the nuclei to understand the effects of both H2O and Zn2+ concentration, with a view to understanding the potential changes in speciation ion aggregation that would occur in a real IL-based zinc battery system.
Figure 1. Chemical structure of [C4mpyr][dca].
Zn(dca)2 was synthesized as reported previously38 by combining aqueous solutions of Zn(NO3)2·6H2O (98%, Sigma-Aldrich), and Na(dca) (96%, Alfa Aesar), filtering the resulting white precipitate, washing with distilled water, and drying under high vacuum for 2 days at 40 °C. Solution Preparation. Samples were prepared by adding dry Zn(dca)2 at the correct ratio to the neat IL inside an Arfilled glovebox. For samples containing 3 wt % of H2O, distilled water was added to the sample by a 10 or 100 μL auto pipet (Lab Co.) and then the water content verified by a Model 756 Karl Fischer Coulometer (Metrohm) using Hydranal Coulomat AG titrant. Table 1 displays the composition of all [C4mpyr][dca]/Zn(dca)2 samples. Table 1. Composition of Ionic Liquid Samples under Studya sample a
1 2 3a 4 5 a
χZn2+
IL: Zn2+ mole ratio
dca:Zn2+ mole ratio
0 5 9 20 29
20 10 4 2.5
22 12 6 4.5
Samples that were also made containing 3 wt % H2O.
Density. Density measurements were performed on an Anton Parr DMA 5000 density meter over a temperature range of 20−70 °C at 10 °C intervals. Viscosity. Viscosity measurements were performed on an Anton Parr Lovis 2000 M falling ball viscometer using either 2.5 or 1.8 mm internal diameter capilliary tube at an angle of 60° containing the sample and a 1.5 mm stainless steel ball with a density of 7.69 g.cm−3, loaded inside a nitrogen glovebox. Conductivity. Ionic conductivity was measured by AC impedance techniques using a SP-200 impedance/frequency response analyzer (Biologic, USA). A sample of approximately 2 mL was measured over a temperature range of 25−70 °C at 5 °C intervals over the frequency range 0.1 Hz to 1 MHz using an alternating voltage of 0.1 V amplitude. Samples were contained in a custom-built dip-cell probe containing two platinum wires sheathed in glass (Monash Scientific, Australia), sealed with a rubber O-ring, and placed into a cavity in a brass block. Desired isothermal temperature points were reached by ramping at a steady rate of 0.2 °C/min using a Eurotherm 2204E temperature controller, with temperature measured using a Type T thermocouple inserted in a brass block adjacent to the cell. The cell constant was determined using a solution of 0.01 M KCl at 25 °C. Data acquisition commenced following isothermal equilibration for 20 min at each temperature point, and cell resistance was recorded as a function of frequency. The resistance (Ω), and the derived conductivity (S·cm−1), were
■
EXPERIMENTAL SECTION Materials. 1-Butyl-1-methylpyrrolidinium dicyanamide (99.9%, Iolitech, [C4mpyr][dca] (Figure 1)) was stored and handled in an argon-filled glovebox and was used as received. DMSO-d6 (98%, Cambridge Isotope Laboratories) was also used as received. 4896
dx.doi.org/10.1021/jp501665g | J. Phys. Chem. B 2014, 118, 4895−4905
The Journal of Physical Chemistry B
Article 13
C and 1H nuclei. The relaxation delay was set to a minimum of 5 × T1, which was in the range of 300−3000 ms for 13C. Previous studies have shown that values for D can be dependent on the value for Δ used in the experiment when studying ILs;32,39−41 however, this effect has apparently only been observed when using the PGSE-based (rather than PGSTE) pulse sequences. This phenomenon has been attributed to heterogeneity in the viscous IL system under study.41 In order to demonstrate that D does not depend on Δ, the C4mpyr cation of the most viscous sample ([C4mpyr][dca] + 29 mol % Zn(dca)2) was measured using 1H NMR and varying values for Δ (from 20 to 160 ms). The values for D obtained were found to show to no trend with changing Δ, and were within the experimental error (
Influence of Zn2+ and Water on the Transport Properties of a Pyrrolidinium Dicyanamide Ionic Liquid T. J. Simons,† P. M. Bayley,† Z. Zhang,§ P. C. Howlett,† D. R. MacFarlane,‡ L. A. Madsen,§ and M. Forsyth*,† †
Australian Centre for Electromaterials Science (ACES), Institute for Frontier Materials (IFM), Deakin University Burwood Campus, Burwood 3125, Australia ‡ Australian Centre for Electromaterials Science (ACES), School of Chemistry, Monash University, Clayton 3800, Victoria, Australia § Department of Chemistry and Macromolecules and Interfaces Institute, Virginia Tech, Blacksburg, Virginia 24061, United States S Supporting Information *
ABSTRACT: In order to expand our understanding of a potential zinc-based battery electrolyte, we have characterized the physical and transport properties of the ionic liquid (IL) 1-butyl-1-methylpyrrolidinium dicyanamide ([C4mpyr][dca]) containing various levels of both Zn2+ and H2O. Detailed measurements of density, viscosity, conductivity, and individual anion and cation diffusion coefficients using pulsed-field-gradient (PFG) NMR combined with NMR chemical shifts and spin−lattice relaxation (T1) NMR experiments provide insights into the motion and chemical environment of all molecular species. We find that the various techniques for probing ion transport and dynamics form a coherent picture as a function of electrolyte composition. Zn2+ addition causes a moderate reduction in the self-diffusion of the IL anion and cation, whereas the addition of H2O increases ion mobility by increasing the liquid’s overall fluidity. Temperaturedependent 13C T1 experiments of the dca carbon analyzed using Bloembergen−Purcell− Pound fits show monotonic slowing of anion dynamics with Zn2+ addition, suggesting increased Zn2+/dca− association. T1 experiments show minimal change in the spin−lattice relaxation of cation or anion upon H2O addition, suggesting that H2O is playing no significant role in Zn2+ speciation. Finally, we employ a novel electrophoretic NMR technique to directly determine the electrophoretic mobility of the C4mpyr cation, which we discuss in the context of impedance-based conductivity measurements.
■
INTRODUCTION Currently, there is increasing demand for high-capacity energy storage devices that are sustainable, affordable, and environmentally responsible.1 Zn-based batteries, including Zn−air batteries, show promise for commercialization due to the low reactivity, relatively low cost, existing recycling infrastructure, and nontoxicity of zinc metal.2−6 However, in the case of Zn− air batteries there are many obstacles to overcome before a commercially available rechargeable (secondary) device can be produced. These issues include electrolyte evaporation, nonuniform deposition during recharge, poor cycle life, and poor efficiency. 7,8 While many researchers have been researching novel oxygen-reduction catalysts2,9,10 and zinc anode compositions,6,11 many of these problems can also be addressed through novel electrolyte design. Ionic liquids (ILs) are an established class of compounds whose properties can include low volatility, ionic conductivity, thermal stability, electrochemical stability, and high metal−salt solubility.12 These properties have made ILs the subject of much scrutiny for applications in electrochemical devices such as lithium−metal batteries,13,14 supercapacitors,15−17 and fuel cells.12 Their common low-volatility, conductivity and ability to electrodeposit metals have recently made various ILs into © 2014 American Chemical Society
appealing candidates for novel electrolytes in zinc-based batteries. Of the many anions that have been investigated, one of the most popular for electrochemical processes has been the bis(trifluoromethanesulfonyl)imide (NTf2) anion. Several groups, including those of Tulodziecki et al.,18 Chen et al.,19 Doan et al.,20 and Azaceta et al.,21 have investigated the electrochemistry and deposition morphology of Zn,19 ZnO,18,21 and alloys of Zn19 in various NTf2 ILs. Despite the low viscosity and high stability of ILs based on the NTf2 anion, deposition and stripping currents of Zn electrochemistry do not suggest outstanding performance in an electrochemical device such as a battery or supercapacitor. The economic and environmental considerations of using NTf2 in a mass produced device have also driven the authors to consider alternate ILs with nonhalogenated anions in this work. One class of ILs that has been extensively studied in relation to zinc are those containing the dicyanamide (dca) anion.22,23 Deng et al. first investigated the electrochemistry of the Zn0/ Received: February 16, 2014 Revised: April 8, 2014 Published: April 8, 2014 4895
dx.doi.org/10.1021/jp501665g | J. Phys. Chem. B 2014, 118, 4895−4905
The Journal of Physical Chemistry B
Article
Zn2+ couple in both imidazolium24 and pyrrolidinium25 dca, after which Xu et al.26 fully characterized the mechanisms of deposition on Pt electrodes. Subsequently, Simons et al.27 investigated the importance of the added Zn2+ salt, concluding that Zn(dca)2 yielded a deposition/stripping Coulombic efficiency of 85% and current densities 10 times greater than those found in the studies of Deng and Xu, who used ZnCl2 and Zn(NTf2)2, respectively. Since then, our group has investigated Zn(dca)2 in dca-based ILs both electrochemically and spectroscopically, concluding that formation of a Zn(dca)x2‑x complex anion is responsible for the excellent electrochemical properties observed thus far.28,29 Nuclear magnetic resonance (NMR) has been used extensively for IL-based electrolyte characterization.30−36 Noda et al.34 have used pulsed-field-gradient (PFG) diffusion NMR to investigate the proton conduction and diffusion mechanism of a protic ionic liquid, imidazolium NTf2. A chemical shift analysis of the 1H spectra provided evidence that the proton was transported via the imidazole in a “vehicularlike” fashion through the electrolyte. Similar self-diffusion measurements on the nonprotic 1-ethyl-3-methylimidazolium NTf2 IL have been performed by probing the 19F nuclei for the anion and 1H for the cation to infer diffusion-based transference numbers for the species involved. These measurements have been paired with conductivity measurements to provide evidence of ion association in the electrolyte solutions, which may decrease an electrolyte’s efficiency in a battery. Hou et al.33 conducted a detailed study of ion associations in pure ILs and in ILs absorbed into polymer electrolytes in order to advance more quantitative ideas on the clustering of IL ions due to variations of specific intermolecular interactions. A recent study by Zhang and Madsen37 has used electrophoretic NMR to observe separate cation and anion mobilities in two pure ILs, which verify pure vehicular ion transport in these systems and provide a new method to investigate electrolyte transport. Bayley et al.30 effectively used T1 relaxation experiments to show that tetrahydrofuran can be used to control lithium speciation in an IL electrolyte for lithium−metal batteries. In particular, the combination of variable temperature T1 and diffusion measurements aided in understanding the role of these additives in lithium complexation. Hayamizu et al.31,32 also combined both physical and spectroscopic techniques to investigate the librational motions of imidazolium cations and information on the correlation time of a lithium jump in a series of NTf2 and bis(fluorosulfonylimide) IL−lithium salt mixtures. In order to further our understanding of the physical properties and molecular dynamics of the species present in the [C4mpyr][dca]/Zn(dca)2 electrolyte systems, we have investigated these aspects through the use of PFG diffusion (1H and 13 C nuclei), 13C chemical shifts, and 13C T1 NMR experiments on the [C4mpyr][dca] IL. We have analyzed both the physical properties and the relaxation times of the nuclei to understand the effects of both H2O and Zn2+ concentration, with a view to understanding the potential changes in speciation ion aggregation that would occur in a real IL-based zinc battery system.
Figure 1. Chemical structure of [C4mpyr][dca].
Zn(dca)2 was synthesized as reported previously38 by combining aqueous solutions of Zn(NO3)2·6H2O (98%, Sigma-Aldrich), and Na(dca) (96%, Alfa Aesar), filtering the resulting white precipitate, washing with distilled water, and drying under high vacuum for 2 days at 40 °C. Solution Preparation. Samples were prepared by adding dry Zn(dca)2 at the correct ratio to the neat IL inside an Arfilled glovebox. For samples containing 3 wt % of H2O, distilled water was added to the sample by a 10 or 100 μL auto pipet (Lab Co.) and then the water content verified by a Model 756 Karl Fischer Coulometer (Metrohm) using Hydranal Coulomat AG titrant. Table 1 displays the composition of all [C4mpyr][dca]/Zn(dca)2 samples. Table 1. Composition of Ionic Liquid Samples under Studya sample a
1 2 3a 4 5 a
χZn2+
IL: Zn2+ mole ratio
dca:Zn2+ mole ratio
0 5 9 20 29
20 10 4 2.5
22 12 6 4.5
Samples that were also made containing 3 wt % H2O.
Density. Density measurements were performed on an Anton Parr DMA 5000 density meter over a temperature range of 20−70 °C at 10 °C intervals. Viscosity. Viscosity measurements were performed on an Anton Parr Lovis 2000 M falling ball viscometer using either 2.5 or 1.8 mm internal diameter capilliary tube at an angle of 60° containing the sample and a 1.5 mm stainless steel ball with a density of 7.69 g.cm−3, loaded inside a nitrogen glovebox. Conductivity. Ionic conductivity was measured by AC impedance techniques using a SP-200 impedance/frequency response analyzer (Biologic, USA). A sample of approximately 2 mL was measured over a temperature range of 25−70 °C at 5 °C intervals over the frequency range 0.1 Hz to 1 MHz using an alternating voltage of 0.1 V amplitude. Samples were contained in a custom-built dip-cell probe containing two platinum wires sheathed in glass (Monash Scientific, Australia), sealed with a rubber O-ring, and placed into a cavity in a brass block. Desired isothermal temperature points were reached by ramping at a steady rate of 0.2 °C/min using a Eurotherm 2204E temperature controller, with temperature measured using a Type T thermocouple inserted in a brass block adjacent to the cell. The cell constant was determined using a solution of 0.01 M KCl at 25 °C. Data acquisition commenced following isothermal equilibration for 20 min at each temperature point, and cell resistance was recorded as a function of frequency. The resistance (Ω), and the derived conductivity (S·cm−1), were
■
EXPERIMENTAL SECTION Materials. 1-Butyl-1-methylpyrrolidinium dicyanamide (99.9%, Iolitech, [C4mpyr][dca] (Figure 1)) was stored and handled in an argon-filled glovebox and was used as received. DMSO-d6 (98%, Cambridge Isotope Laboratories) was also used as received. 4896
dx.doi.org/10.1021/jp501665g | J. Phys. Chem. B 2014, 118, 4895−4905
The Journal of Physical Chemistry B
Article 13
C and 1H nuclei. The relaxation delay was set to a minimum of 5 × T1, which was in the range of 300−3000 ms for 13C. Previous studies have shown that values for D can be dependent on the value for Δ used in the experiment when studying ILs;32,39−41 however, this effect has apparently only been observed when using the PGSE-based (rather than PGSTE) pulse sequences. This phenomenon has been attributed to heterogeneity in the viscous IL system under study.41 In order to demonstrate that D does not depend on Δ, the C4mpyr cation of the most viscous sample ([C4mpyr][dca] + 29 mol % Zn(dca)2) was measured using 1H NMR and varying values for Δ (from 20 to 160 ms). The values for D obtained were found to show to no trend with changing Δ, and were within the experimental error (
