and the Formation of a New Ionic Liquid
David M. Ryan,a,* Thomas L. Riechel, a,* and Thomas Weltonb,* aDepartment of Chemistry and Biochemistry, Miami University, Oxford, Ohio, USA
bDepartment of Chemistry, Imperial College, London, U.K.
The electrochemistry of several vanadium oxides and oxyhalides has been examined in room temperature ionic liquids made by mixing 1-ethyl-3-methylimidazolium chloride (EMIC) and aluminum chloride. By varying the mole ratio of the two components, Lewis basic, neutral and acidic ionic liquids were made, while the addition of NaCl to an acidic ionic liquid resulted in a Lewis buffered, neutral solvent. Because of the ionic nature of these liquids, most oxides have very low solubility: V2O3, V2O4 and VOSO4 are insoluble, while V2O5 and Na3VO4 are slightly soluble with solubility limits of less than 5 mM. The solubilities of the salts NaVO3 and NH4VO3 and the oxyhalides VOCl3 and VOF3 are significantly higher. In acidic ionic liquids V2O5, NaVO3, Na3VO4 and NH4VO3 all exhibit three irreversible reduction waves between 1.7 and 0.5V vs. a reference electrode consisting of an aluminum wire in a 0.6 mole fraction AlCl3 ionic liquid. VOCl3 and VOF3 gave similar reduction peaks suggesting that the oxides may have reacted with the ionic liquid to form oxychlorides. A new ionic liquid was formed by using VOCl3 as the Lewis acid instead of AlCl3:
EMIC + VOCl3 [EMI]+ + [VOCl4]-
In acetonitrile [VOCl4]- was found to undergo a one electron reversible reduction to [VOCl4]2-:
[VOCl4]- + e- [VOCl4]2- Introduction Room temperature ionic liquids made by mixing 1-ethyl-3-methylimidazolium chloride and AlCl3 (EMIC/AlCl3) are being investigated by several groups as possible battery electrolytes for use in cells with alkali metal anodes. 1,2,3,4 By varying the ratio of the organic salt (EMIC) to the inorganic salt (AlCl3), Lewis basic, neutral, and acidic liquids can be formed. The composition of such ionic liquids is represented by the value of N, the mole fraction of AlCl3. Because these ionic liquids are formed when two solids are mixed, no solvent is used to dissolve the salts. When N=0.5, a neutral liquid is formed with [EMI]+ and [AlCl4]- in a one to one mole ratio. If N < 0.5, excess chloride from the EMIC will be present and a Lewis basic liquid is formed. If N > 0.5, polymeric forms of [AlCl4]- occur and a Lewis acidic liquid results. Such liquids have several advantages including wide voltage windows, wide liquid temperature ranges, and high conductivities when compared to other solvents and electrolytes or conducting polymers. The values of these parameters depend on the composition. For example, compositions between about N=0.35 and 0.65 all yield liquids at, or somewhat below, room temperature (25oC). The properties of these liquids and their use as solvents for the study of transition metal compounds have been summarized in recent reviews. 5,6,7
Since the nonaqueous nature of these liquids allows the use of alkali metals, lithium and sodium have been investigated as possible anodes for cells using them as electrolytes. Most work has focused on the neutral ionic liquids because these provide the widest voltage window, slightly greater than 4V. This means that if reversible couples for the anode and cathode can be found to operate near the voltage limits, a cell of about 4V would be possible. Lipsztajn and Osteryoung 8were the first workers to report the plating and stripping of Li on tungsten in a neutral ionic liquid. Yu, Winnick and Kohl 9 reported the reversible plating and stripping of sodium at a mercury drop attached to a gold disk. Scordilis-Kelley, Fuller, Carlin and Wilkes 10 studied the alkali metals in ionic liquids buffered with NaCl in order to maintain the wide voltage window during charge and discharge cycles. They found that only lithium could be deposited on a bare tungsten electrode, as the reduction potentials for sodium and potassium fell outside the voltage window. By adding protons to the buffered ionic liquid, Riechel and Wilkes11found that the voltage window could be widened by about 300mV, allowing the reversible plating and stripping of sodium at Pt and W electrodes.
Because of instability of the proton concentration in the ionic liquid studied by Riechel and Wilkes, other reagents for widening the voltage window were sought. Piersma, Ryan, Schumacher and Riechel Error: Reference source not found found that addition of triethanolamineHCl produced liquids that were stable for months. Gray, Kohl and Winnick Error: Reference source not foundproduced stable liquids by pressurizing cells with HCl(g) . Fuller, Carlin and Osteryoung Error: Reference source not found,Error: Reference source not foundhave successfully modified these liquids by the addition of SOCl2. Whereas lithium has been shown to form dendrites when plated in these electrolytes, sodium forms smooth layersError: Reference source not found. Since sodium is more stable than lithium, provides a higher potential, and does not exhibit the problem of dendrite formation, it may be a better candidate for an anode in this electrolyte.
In contrast to the issue of a suitable anode, little work has been done to identify suitable cathode materials. There are only a few reports in the literature of vanadium compounds studied in EMIC/AlCl3 ionic liquids12,13,14. We present here a study of the electrochemistry of several vanadium oxides and oxyhalides in EMIC/AlCl3 electrolytes in order to evaluate their feasibility as cathode materials. We also report the formation of new room temperature ionic liquids made by mixing EMIC with VOCl3 or VOF3, and describe the reversible electrochemistry of a vanadium (V)/(IV) couple.
Instrumentation.- All electrochemical experiments were carried out using a three electrode cell set up in a Vacuum Atmospheres glove box under a dry nitrogen atmosphere. The electrochemical measurements were made with either an EG&G Princeton Applied Research (PAR) Model 263A potentiostat/galvanostat interfaced to a personal computer running PAR Model 270 software, or a Bioanalytical Systems (BAS) Model 100A electrochemical analyzer. A BAS platinum disk working electrode (Model MF 2013, 2.01 mm2) was used for voltammetry, while a Pt mesh working electrode was used for coulometry. A Pt foil auxiliary electrode was used for voltammetry and a Pt foil, isolated from the test solution by means of a fine porosity fritted tube, was used as an auxiliary electrode for coulometry. The reference electrode for both voltammetry and coulometry in the ionic liquids consisted of an Al wire in an N=0.60 EMIC/AlCl3 ionic liquid, contained in a Pyrex tube with a porous tip to provide solution contact. The reference electrode for work done in acetonitrile was a silver wire in 0.1M tetrabutlyammonium percholrate (TBAP). The scan rate for all voltammograms was 100mV/s, unless otherwise indicated. UV-Visible spectra were recorded using a Hewlett Packard Model 8452A diode array spectrometer and quartz cuvettes. Slightly soluble samples were decanted or filtered before recording spectra. Conductivity measurements were made with an Extech Instruments Oyster conductivity meter kit. The instrument was calibrated with a KCl solution outside of the glove box in order to determine the cell constant. Then the clean, dry cell was taken inside the glove box for the ionic liquid measurements.
Reagents and Synthesis.- The components for the ionic liquids were synthesized and purified using published methods. 1-Ethyl-3-methylimidazolium chloride (EMIC) was synthesized from 1-methylimidazole and chloroethane (Aldrich)15. Purification of AlCl3 (Fluka) was carried out by recrystallization in a tube furnace. Neutral ionic liquids were made in a glove box by mixing carefully weighed amounts of EMIC and AlCl3. Traces of moisture arising from the synthesis of EMIC were reduced by the addition of ethylaluminum dichloride (Aldrich)16. Sometimes the ionic liquids were further purified by electrolyzing at 1.0V overnight using a Pt mesh working electrode and a large Pt flag auxiliary electrode. It was concluded that the residual moisture had been removed when reduction peaks normally observed at – 0.5V and –1.2V in the voltammograms were decreased to a few microamps. Acidic ionic liquids were made by adding excess AlCl3 to a neutral liquid, while basic liquids were made by adding excess EMIC. The vanadium reagents V2O3, V2O4, V2O5, VOSO4, Na3VO4, VOCl3 and VOF3 (Aldrich), NH4VO3 (Alfa Aesar), and NaVO3 (GFS Chemicals) were used as received.
EMIC/VOCl3 ionic liquids were prepared by mixing 1:1 stoichiometric amounts of VOCl3 and EMIC in a small glass flask and stirring with a Teflon covered magnetic stir bar in the glove box.
The preparation of [EMI]2[VOCl4] has been previously published Error: Reference source not found,17.
Results and Discussion
Oxides.- The solubilities of seven vanadium oxides (V2O5, V2O4, V2O3, NaVO3, Na3VO4, NH4VO3 , and VOSO4) were evaluated in various EMIC/AlCl3 ionic liquids. The compositions were basic (N=0.45), neutral (N=0.50), neutral buffered (N=0.50, resulting from the addition of excess NaCl to an N=0.55 ionic liquid), and acidic (N=0.55).
Solubility at room temperature was determined by trying to dissolve enough of each sample in 11 g of an ionic liquid to create a 5 mM solution. Table 1 shows the solubilities of the materials surveyed. In this paper, a material is defined as soluble when a solution greater than or equal to 5 mM can be made. It is considered slightly soluble if it will form a solution of less than 5 mM that displays significant color or electrochemical activity. Solution formation was aided by vigorous stirring with a magnetic stirrer. Some solids were difficult to dissolve, requiring several days of continuous stirring. In some cases stirring only reduced the particle size of the material resulting in a suspension. Since the particles settled out when stirring was stopped, these materials were judged insoluble. Sometimes the sample/solvent suspension was also subjected to mild heating (81oC) or prolonged sonication (about three hours) in an effort to make the sample dissolve. The most soluble compounds were NaVO3 and NH4VO3. V2O5 and Na3VO4 were slightly soluble, while the other compounds were insoluble.
Because of its reported use in xerogel cathodes, 18,19,20,21,22,23 V2O5 was of greatest interest. In spite of its low solubility, characteristic redox peaks were observed. Figure 1A shows the cyclic voltammogram of V2O5 in a basic ionic liquid (N=0.45). The only significant feature is an irreversible reduction peak at about 0.3V. In a neutral (N=0.50) ionic liquid, (Figure 1B) a similar reduction peak appears near 0.5V and is accompanied by an oxidation peak at 1.3V. The voltammogram in a neutral, NaCl buffered liquid (not shown) exhibits the same reduction and oxidation peaks as in the neutral solvent. The voltammogram of V2O5 in an acidic liquid (N=0.55) (Figure 1C) is different in that two distinct reduction peaks are observed, one at 0.5V as in the other solvents, and a second peak at a much more positive potential of 1.7V. A third, less distinct reduction peak was observed at about 1.4V. Note also that the open circuit potential for this solvent is almost 2V.
The anodic and cathodic scan limits were chosen for several reasons. The negative limit of the electrochemical window for acidic liquids is at 0V and is determined by the plating of aluminum. Therefore, acidic liquids were not scanned more negative than 0V. For neutral or basic liquids the cathodic limit is about –2V. However, there is no electrochemistry for any of the oxides observed more negative than 0V. In the range of 0V to –2V only small moisture reduction peaks are observed. Basic liquids have an electrochemical window with an anodic limit of about 1.2V, which is determined by the oxidation of chloride to chlorine. Therefore, basic ionic liquids were not scanned more positive than 1.2V. Neutral and acidic liquids have a limit of about 2V and the positive potential region is where most of the electrochemistry of the vanadium oxides was observed.
Controlled potential coulometry was carried out on V2O5 in the acidic ionic liquid, as the solubility of V2O5 appeared to be highest in this solvent (compare the current scales in Figure 1). Because the concentration of the oxide was uncertain, no quantitative data can be reported, but upon electrolysis at 1.55V the solution color changed from an intense yellow to green, which is often associated with vanadium(IV) species. Subsequent voltammograms no longer showed the reduction peak at 1.7V. This suggests a 1e- reduction to vanadium(IV).
A subsequent controlled potential coulometry experiment performed at 0.3V resulted in no further color change and the reduction peak at 0.5V was absent from the resulting voltammogram. Thus, the two reduction peaks observed in the acidic liquid appear to correspond to two different vanadium(V) species each undergoing a 1e- reduction, rather than the sequential reduction of vanadium(V) to vanadium(IV) and vanadium(III). This is consistent with the fact that only the second peak (reduction near 0.5V) is observed in basic and neutral ionic liquids.
Sodium metavanadate, NaVO3, dissolved at 5mM concentration in neutral ionic liquid after stirring overnight and resulted in an orange solution. NaVO3 is soluble in neutral, neutral buffered and acidic liquids, but is only slightly soluble in basic liquids. The voltammogram of NaVO3 in neutral liquids is similar to that for V2O5 as shown in Figure 1B. There is a reduction peak at 0.5V and an oxidation peak at about 1.3V. The voltammogram of NaVO3 in an acidic ionic liquid is shown in Figure 2A. Like V2O5 (Figure 1C), three reduction peaks are observed, shifted slightly to 1.2, 1.0, and 0.6V. Comparison of the two voltammograms suggests that these peaks correspond to the same species. Sodium orthovanadate, Na3VO4, is less soluble than NaVO3, but most of the same features can be seen in the corresponding voltammograms. In the neutral ionic liquid the reduction peak is at 0.5V and the oxidation peak is at 1.2V. The voltammogram of Na3VO4 in the acidic solvent is shown in Figure 2B, with only small shifts in the peaks observed. (Because of the small currents in these figures, a voltammogram of a blank acidic ionic liquid recorded at the same current sensitivity is displayed in Figure 2C for comparison.)
Ammonium vanadate, NH4VO3, was soluble in all of the liquids surveyed. The voltammograms of the bright yellow solution that resulted from addition to an acidic solvent (N=0.55) yielded three reduction peaks as shown in Figure 3. The data summarized in Table 2 shows the similarity of the reduction potentials for V2O5, NaVO3, Na3VO4 and NH4VO3 in acidic ionic liquids (Figures 1C, 2A, 2B and 3, respectively). All exhibit three irreversible reductions in the range 1.7 to 0.5V.
VOCl3 and VOF3in EMIC/AlCl3 Ionic Liquids.- Figure 4 shows a voltammogram of VOCl3 in a neutral ionic liquid. The voltammogram of VOCl3 in basic ionic liquid is not shown but is similar to the voltammogram of VOCl3 in the neutral liquid. The main feature of these voltammograms is the reduction peak at about 0.4V. The reduction peaks at about 0.0 and –0.5V in Figure 4 are due to residual moisture in the solvent.
Controlled potential coulometry (CPC) was done to determine what species were responsible for the reduction peak at 0.4V. A 7mM VOCl3 neutral ionic liquid was electrolyzed for 22 hours at 0.25V. A voltammogram recorded before electrolysis was like that shown in Figure 4. After electrolysis the 0.4V peak was gone (Figure 5) and the rest potential, as well as the remaining proton peaks, had moved more negative. The color of the initial solution was dark brown but after electrolysis for nearly three hours it turned to a light green and after overnight electrolysis it was a light blue. These color changes corresponded to a total of 1.2 equivalents of electrons per vanadium. Spectra recorded before and after electrolysis are shown in Figure 6. These demonstrate the loss of the 470nm peak. The appearance of the low absorbance spectrum after electrolysis is consistent with the reduction of a V(V) (d0) species to a lower oxidation state V(IV) (d1) species. Thus, it is believed that the reduction observed in the various ionic liquids at 0.4V is the reduction of V(V) to V(IV).
Figure 7A shows a voltammogram of 7mM VOCl3 in acidic ionic liquid (N=0.55) which exhibits three characteristic reduction peaks at 1.3, 0.8, and 0.4V. After electrolysis for 20 hours at 1.0V no reduction peaks remain (Figure 7B). Note in particular the absence of the reduction peak expected at 0.4V. This suggests that all of the V(V) was reduced as a result of the electrolysis at 1.0V and thus, there was no V(V) left to generate the peak expected at 0.4V. The solution changed from dark brown to dark purple after electrolysis. These results demonstrate that there are two or three V(V) species that can be reduced, as was shown for V2O5 in acidic ionic liquids.
Similar experiments were performed with VOF3. VOF3 was much easier to work with because it was a stable solid. The solid was yellow-orange, very similar in appearance to V2O5, and was soluble in neutral and acidic EMIC/AlCl3 ionic liquids. A voltammogram of 7mM VOF3 in neutral ionic liquid looks much like that of VOCl3 in the same solvent (Figure 4). The most important feature is the characteristic reduction peak at 0.4V. A spectrum of this liquid exhibited a peak at 480nm that is similar to that in the spectrum for VOCl3 (Figure 6) at 470nm.
CPC was performed on the neutral VOF3 solution and the results were similar to those observed for VOCl3, except that two new weak peaks in the visible spectrum appeared. A spectrum of VOF3 after overnight electrolysis at 0.0V exhibited a new peak at 560nm, while the peak at 480nm was gone. There was also a new small peak at 720nm, similar to one previously reported as due to the presence of V(IV)24. The electrolyzed solution was deep purple. The purple color could also be due to VCl3 or [VCl6]3-, if the reduction of VOCl3 or VOF3 led to vanadium(III) products.13, 24
A VOF3 solution in acidic ionic liquid was also examined. The voltammogram for this solvent was like that for VOCl3 (Figure 7A) with three reduction peaks at 1.6, 1.2, and 0.6V.
EMIC/VOX3 Ionic Liquids.-When an equimolar mixture of EMIC and VOCl3 was stirred in a small Erlenmeyer flask under a dry nitrogen atmosphere, a homogeneous, dark red, almost black, liquid resulted. It is proposed that VOCl3 and EMIC react in a manner analogous to the reaction of EMIC and AlCl3:
EMIC + VOCl3 [EMI]+ + [VOCl4]-
It is also proposed that [VOCl4]- undergoes dissociation:
[VOCl4]- VOCl3 + Cl-
This reaction is analogous to:
[AlCl4]- AlCl3 + Cl-
in EMIC/AlCl3 ionic liquids, and
H2O H+ + OH-
in aqueous solutions. In each of these cases, the equilibrium lies far to the left. This suggests that a new room temperature ionic liquid, EMIC/VOCl3, is formed. The conductivity of the new liquid was 2mS (n=3) at 20oC. This value can be compared to 4 mS for a sample of neutral EMIC/AlCl3 liquid that was measured as a control. Figure 8 shows a cyclic voltammogram of a slightly basic (Nvthemole fraction VOCl3) EMIC/VOCl3 ionic liquid. The main features observed during the negative scan were large reductions at 0.5V and –0.6V. The current then forms a plateau. During the positive scan, an oxidation peak was observed at 1.6V, which indicates the presence of chloride. The anodic limit is about 2.0V. After the experiment, when the Pt working and the Pt auxiliary electrodes were being washed, it was observed that both Pt surfaces were coated with a thick, dark film that turned brick red briefly when being washed off with tap water. This film could account for the passivation of the electrodes during cyclic voltammetry.
As it was noticed that there was some undissolved EMIC still present, additional VOCl3 was added to make the liquid acidic (Nv > 0.5). This led to a completely liquid sample. A voltammogram for this liquid is shown in Figure 9. In contrast to the basic case, the acidic liquid showed more peaks, but their resolution was poor. The positive limit of this liquid was about 0.5V. The currents were very high (milliamps instead of microamps) indicating that the solvent itself was being reduced. The Pt working electrode was passivated at potentials beyond –2.0V.
The neat EMIC/VOCl3 ionic liquid was so intensely colored that a visible spectrum could not be recorded directly. Instead, a solution was prepared by dissolving 0.4 grams of the liquid in 100 grams of CH3CN. The spectrum (Figure 10) showed a strong absorption peak at 475nm and a very weak peak at about 660nm, and is very similar to the spectra of VOCl3 (Figure 6) and VOF3 recorded in EMIC/AlCl3 ionic liquids. These spectra match closely the spectrum for [VOCl4]- in acetonitrile reported by Zhang and Holm25. These workers reported a high absorbance peak at 487nm and a weak absorbance at 650nm. Thus, we believe that [VOCl4]- is the dominant vanadium species in the EMIC/VOCl3 ionic liquid.
For comparison to VOCl3, VOF3 was mixed with EMIC under dry nitrogen in the glove box. Since the mixtures were made from two solids, they were easy to weigh accurately and were made in three ratios: 1:1 VOF3:EMIC (NF=0.5, NF=mole fraction of VOF3, (neutral)), 2:3 VOF3:EMIC (NF=0.40, (basic)), and 3:2 VOF3:EMIC (NF=0.60, (acidic)). The reagents were purposely not mixed to homogeneity so the contrasting colored materials could be observed. The mixtures all turned to a uniform, dark brown solid when allowed to sit for a month.
After one month, the bottles containing the three mixtures were placed on a hotplate in the glove box and heated slowly. In each case the solid mixtures turned to a dark red liquid, similar in appearance to the EMIC/VOCl3 ionic liquids. The mixtures melted between 29 to 32oC. After the mixtures became liquid they remained liquid, even after cooling to room temperature. These results suggest the following reaction takes place:
EMIC + VOF3 EMI]+ + [VOF3Cl]-
Thus, VOF3 appears to form a mixed halide ionic liquid with EMIC.
Since the working electrode is passivated in neat EMIC/VOCl3 ionic liquids (Figures 8 and 9), the electrochemistry of the new liquids was also examined in acetonitrile. Figure 11A shows a voltammogram made with 11mM EMIC/VOCl3 dissolved in a solution of 0.1M tetrabutylammonium perchlorate (TBAP) in dry CH3CN. The working electrode was a Pt disk and the reference electrode was a silver wire immersed in 0.1M TBAP in CH3CN. A series of voltammograms made to different anodic and cathodic limits demonstrated that the reduction at –0.5V and the oxidation at –0.2V were related. Further, the reduction observed at 0.7V appeared to be coupled with the oxidation observed at 0.9V. It is believed that this couple is the reduction of V(V) to V(IV) and oxidation back to V(V). Note that this was not observed as a reversible reaction in any of the EMIC/AlCl3 ionic liquids.
Figure 11B shows the results of a controlled potential coulometry experiment that was performed on this solution at 0.2V for one hour. The large, well defined reduction peaks at 0.7V and –0.5V observed in Figure 11A are gone. The oxidation peak at approximately 1V is now larger and there are several additional oxidation peaks more positive than 1V. The original solution was dark red, exhibiting an intense absorbance peak at 480nm (Figure 12A). After electrolysis the solution was light blue-green and the absorbance at 480nm was replaced by a low absorbance at 720nm (Figure 12B). Dent and coworkersError: Reference source not found found that [VOCl4]- was easily reduced to [VOCl4]2-. Also, Zhang and HolmError: Reference source not found reported the spectrum of [VOCl4]2- in acetonitrile, which exhibits an absorbance peak at 734nm. Hitchcock and coworkers12 reported that the visible spectrum of [EMI]2[VOCl4] in a basic EMIC/AlCl3 ionic liquid showed a major absorbance peak at 739nm. These data suggest that the peak at 0.7V is the reduction of [VOCl4]- to [VOCl4]2-.
Figure 13 is a cyclic voltammogram of 25mM [EMI]2[VOCl4] dissolved in a solution of 0.1M TBAP in CH3CN. Note that the voltammogram of this V(IV) compound is very similar to that for the reduction product of the new ionic liquid, EMIC/VOCl3, shown in Figure 11B. Controlled potential coulometry was performed at 1.0V to oxidize the [EMI]2[VOCl4] solution. The bright green solution turned to red or orange-red after only about 0.5 equivalents of electrons had passed. This color change was expected for a V(IV) compound oxidized to V(V), but the electrolysis current remained high. When one equivalent of electrons had passed, the red-orange color began to fade. After a total of three equivalents of electrons had passed, the solution was a faint green. After exhaustive oxidation, a dark precipitate was observed in the cell. The resulting voltammogram showed no distinct peaks. This result and the observed colored precipitate suggest that much of the vanadium precipitated. The second and third equivalents of electrons passed are believed to be due to the oxidation of chloride ion to chlorine gas, which was lost from the solution.
Spectra were recorded for the [EMI]2[VOCl4] solution before and after oxidation. The spectrum of the solution of the green V(IV) compound before electrolysis (Figure 14A) exhibits an intense absorbance peak at 720nm and is similar to the spectrum resulting from the one electron reduction of the EMIC/VOCl3 ionic liquid (Figure 12B). Figure 14B is the spectrum of the [EMI]2[VOCl4] solution after oxidation. The spectrum shows a dominant absorbance at 480nm, and is nearly identical to the spectrum for the initial solution of the new ionic liquid, EMIC/VOCl3 (Figure 12A).
The similarity of the spectra shown in Figures 12A and 14B, and those in Figures 14A and 12B clearly indicate that the vanadium(V) species present in the new liquid is related to the previously characterized vanadium(IV) compound, [EMI]2[VOCl4]. The reduction of the V(V) species is a one electron process. The oxidation of the V(IV) species involves the one electron oxidation of vanadium and the oxidation of chloride. If the mixture of EMIC and VOCl3 produces [VOCl4]-, as suggested earlier, then the electrochemical and spectroscopic data suggest the following reversible couple:
[VOCl4]- + e- [VOCl4]2-.
When [VOCl4]2- was oxidized, the first equivalent of electrons was consumed by the oxidation of V(IV) to V(V) and explains the first color change. The continuing high oxidation current suggests that a second process occurred. We believe that the V(V) product, [VOCl4]-, dissociated producing Cl-, which was oxidized to Cl2 gas:
[VOCl4]- VOCl3 + Cl-
2Cl- Cl2(g) + 2e-
VOCl3 may have been further degraded, as the solution ultimately turned nearly colorless, and yielded a precipitate that may have been a vanadium oxide.
Of the seven oxides of vanadium studied only four showed any solubility in EMIC/AlCl3 ionic liquids: V2O5, NaVO3, Na3VO4, and NH4VO3. The remaining three, V2O4, V2O3, and VOSO4 were insoluble. Two solubility trends were found. First, the oxides were more soluble in Lewis acidic liquids than in neutral or basic liquids. This was particularly evident for the sodium salts examined. Secondly, it was found that the ionic compounds were more soluble than the covalent compounds. In neutral and acidic solvents NaVO3, Na3VO4, and NH4VO3 are believed to exist as ion pairs, while in basic solvents NaCl precipitates26:
NaVO3 + Cl- NaCl(s) + VO3-.
An irreversible reduction peak was observed for each vanadium(V) oxide in all ionic liquids in the range from 0.9 to 0.2V. In acidic liquids two additional irreversible reduction peaks were observed, one in the range 1.7V to 1.2V, and a smaller peak in the range 1.4 to 0.8V. Thus, it appears that three vanadium(V) species are present and each undergoes reduction to vanadium(IV). In general, this agrees with the work of Bell and coworkersError: Reference source not found. They found that at temperatures greater than 70oC V2O5 exhibited significant solubility, and various vanadium(V) species were formed depending on the Lewis acidity of the liquid. In acidic ionic liquids they reported that V2O5 reacted to form VOCl3. The voltammograms of V2O5 and the other oxides that we have reported here also suggest that reaction with the solvent leads to various oxychloride species.
This project was initiated to assess the use of vanadium oxides as cathode materials in room temperature ionic liquid electrolytes. V2O5 has already been used successfully as a cathode material for high temperature thermal batteries where it is reduced to V2O4 27,28,29,30,31. Recently, V2O5 has been used as a cathode material in room temperature ionic liquid electrolyte cells 32. However, with these cells the V2O5 was “lithiated” (LixV2O5) and the cathode functioned as a “rocking chair” electrode, alternately intercalating Li+ ions into and back out of the V2O5 structure. Our work has demonstrated that V2O5 and the ionic materials NaVO3, Na3VO4, and NH4VO3 are too soluble to be considered as cathode materials for practical batteries without separators. This is particularly true when acidic liquids are used because solubility and reactivity was found to be highest in these solvents. The insoluble materials V2O4, V2O3, or VOSO4 could be considered if solubility is the only criterion for selection. The sodium salts were not very soluble in the NaCl buffered solvents so they may be considered as possible candidates for cathodes in those media. When voltammetry was performed on the soluble materials it was found that the reductions were irreversible, so these materials would not function well as cathodes for rechargeable batteries even if the solubility problems were overcome. However, it may be possible to utilize some of these materials if the solubility and migration of these active materials is controlled by inclusion in an insoluble, conductive matrix, such as a sol-gel.
The electrochemistry and visible spectroscopy of VOCl3 and VOF3 in EMIC/AlCl3 ionic liquids have also been investigated as these compounds are very soluble in EMIC/AlCl3 solvents. Cyclic voltammograms of VOCl3 in acidic media exhibit three irreversible reduction peaks that fall in the same potential range as the reductions observed for the oxides. The reduction peaks for VOCl3 appear to represent the same vanadium(V) species and the coulometric and spectroscopic data show that a one electron reduction leads to vanadium(IV) products. Thus, we conclude that the oxides react with the acidic ionic liquid (probably [Al2Cl7]-) to form oxyhalides. The experiments with VOF3 suggest that mixed halide products are formed in an analogous fashion.
A new ionic liquid has been prepared by mixing Lewis acidic VOCl3 with Lewis basic EMIC. The dark red liquid that results is easily reduced and has a conductivity close to that of a neutral EMIC/AlCl3 ionic liquid. Mixing VOF3 and EMIC resulted in a dark brown solid with a melting point slightly above room temperature. This suggests the formation of a second new ionic liquid, perhaps containing mixed halides.
Because experiments carried out in the neat EMIC/VOCl3 ionic liquid appeared to passivate the electrodes, this liquid was examined in acetonitrile. Comparison of the electrochemical and spectroscopic data to that of a vanadium(IV) species, [EMI]2[VOCl4], showed that [VOCl4]- and [VOCl4]2- form a reversible couple. Although the neat ionic liquid may be of limited use as a battery electrolyte because of its narrow voltage window, this solvent does stabilize some vanadium oxyhalides not easily studied in other media.
This work was supported by an Academic Challenge grant from the State of Ohio to Miami University. Thanks are given to Dr. Vladimir Katovic, Chemistry Department, Wright State University, for help with the conductivity measurements and to the Air Force Research Laboratory, Wright-Patterson Air Force Base, OH for support to D.M.R. while conducting this study.
S = Soluble ( 5 mM)
SS = Slightly Soluble (< 5 mM)
I = Insoluble
Table 1. Solubilities at room temperature for vanadium oxides and oxide salts in EMIC/AlCl3 ionic liquids.
Table 2. Reduction peaks for V(V) oxides and oxide salts in acidic EMIC/AlCl3
Reduction peaks in acidic EMIC/AlCl3 ionic liquids
Figure 1. Cyclic voltammograms of V2O5 in (A) basic (N=0.45), (B) neutral (N=0.50) and (C) acidic (N=0.55) EMIC/AlCl3 ionic liquids.
Figure 2. Cyclic voltammograms in an acidic (N=0.55) EMIC/AlCl3 ionic liquid of (A) NaVO3 and (B) Na3VO4; and (C) a similar N=0.55 blank ionic liquid.
Figure 3. A cyclic voltammogram of NH4VO3 in acidic (N=0.55) EMIC/AlCl3 ionic liquid.
Figure 4. A cyclic voltammogram of VOCl3 in neutral EMIC/AlCl3 ionic liquid.
Figure 5. A cyclic voltammogram of 7mM VOCl3 in neutral EMIC/AlCl3 ionic liquid recorded after a 22 hour electrolysis at 0.25V.
Figure 6. Visible spectra of VOCl3 in neutral EMIC/AlCl3 ionic liquid; before electrolysis (V(V) spectrum), and after electrolysis (V(IV) spectrum).
Figure 7. Cyclic voltammograms of VOCl3 in acidic EMIC/AlCl3 ionic liquid; (A) prior to electrolysis at 1.0V and (B) after electrolysis for 20 hours.
Figure 8. A cyclic voltammogram of a slightly basic EMIC/VOCl3 ionic liquid.
Figure 9. A cyclic voltammogram of an acidic EMIC/VOCl3 ionic liquid.
Figure 10. Visible spectrum of 0.4 g of neutral EMIC/VOCl3 ionic liquid in 100 g CH3CN.
Figure 11. Cyclic voltammograms of 11mM neutral EMIC/VOCl3 ionic liquid in 0.1M TBAP in CH3CN; (A) before electrolysis, and (B) after electrolysis at 0.2V.
Figure 12. Visible spectra of 11mM neutral EMIC/VOCl3 ionic liquid in 0.1M TBAP in CH3CN; (A) before electrolysis, and (B) after electrolysis at 0.2V.
Figure 13. A cyclic voltammogram of 25mM [EMI]2[VOCl4] in 0.1M TBAP in CH3CN.
Figure 14. Visible spectra of 25mM [EMI]2[VOCl4] in 0.1M TBAP in CH3CN; (A) before electrolysis and (B) after electrolysis at 1.0V.
1. B. J. Piersma, D. M. Ryan, E. R. Schumacher and T. L. Riechel, J. Electrochem. Soc., 143 (1996) 908.
2. G. E. Gray, P. A. Kohl and J. Winnick, J. Electrochem. Soc., 142 (1995) 3636.
3. J. Fuller and R. T. Carlin, J. Electrochem. Soc., 143 (1996) L145.
4. J. Fuller, R. A. Osteryoung and R. T. Carlin, J. Electrochem. Soc., 142 (1995) 3632.
5. C. L. Hussey in Chemistry of Nonaqueous Solutions, Current Progress, G. Mamantov and A. I. Popov, Eds., VCH Publisher, New York (1994) Chap. 4.
6. R. T. Carlin and J. S. Wilkes in Chemistry of Nonaqueous Solutions, Current Progress, G. Mamantov and A. I. Popov, Eds., VCH Publisher, New York (1994) Chap. 5.