Synthesis of Hydrophobic Ionic Liquids for

0 downloads 0 Views 226KB Size Report
Room-temperature ionic liquids (RTILs) are an interesting new class of room-temperature ... fully demonstrated that RTILs as N-methyl-N-alkylpyrrolidinium bis.

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲

A1685

0013-4651/2006/153共9兲/A1685/7/$20.00 © The Electrochemical Society

Synthesis of Hydrophobic Ionic Liquids for Electrochemical Applications Giovanni B. Appetecchi, Silvera Scaccia, Cosimo Tizzani, Fabrizio Alessandrini, and S. Passerini*,z Ente Per le Nuove Tecnologie, l’Energia e l’Ambiente (ENEA), IDROCOMB Centro Ricerche Casaccia, Rome 00060, Italy In this work is described an improved synthesis of hydrophobic room-temperature ionic liquids 共RTIL兲 composed of N-methylN-alkylpyrrolidinium 共or piperidinium兲 cations and 共perfluoroalkylsulfonyl兲imide, 关共CnF2n+1SO2兲共CmF2m+1SO2兲N−兴, anions. The procedure described allows the synthesis of hydrophobic ionic liquids with the purity required for electrochemical applications such as high-voltage supercapacitors and lithium batteries. This new synthesis does not require the use of environmentally unfriendly solvents such as acetone, acetonitrile, and alogen-containing solvents that are not suitable for industrial applications. Only water and ethyl acetate are used throughout the entire process. The effect of the reaction temperature, time, and stoichiometry in the various steps of the synthesis has been investigated. With an iterative purification step performed at the end of the synthesis, ultrapure, clear, colorless, inodorous RTILs were obtained. The final vacuum drying at 120°C gave RTILs with a moisture content below 10 ppm. Details for the synthesis of N-butyl-N-methylpyrrolidinium bis共trifluoromethansulfonyl兲imide 共PYR14TFSI兲 are reported. The overall yield for the synthesis of this ionic liquid was above 86 wt %. Electrochemical tests performed on this material are also reported. © 2006 The Electrochemical Society. 关DOI: 10.1149/1.2213420兴 All rights reserved. Manuscript submitted December 22, 2005; revised manuscript received March 3, 2006. Available electronically July 5, 2006.

Room-temperature ionic liquids 共RTILs兲 are an interesting new class of room-temperature fluids. The main advantages of RTILs toward organic solvents are: nonflammability, no detectable vapor pressure up to 200°C, high chemical and thermal stability, and, in some cases, hydrophobicity. Therefore, RTILs have attracted attention for use as “green” solvents for chemical reactions, biphasic catalysis, chemical synthesis, and separations.1-5 They are also under investigation for applications as advanced high-temperature lubricants and transfer fluids in solar thermal energy systems. More recently, RTILs have been largely investigated as electrolytes 共or electrolyte components兲 for electrochemical devices including rechargeable lithium batteries, fuel cells, double-layer capacitors, hybrid supercapacitors, photoelectrochemical cells, and electrodeposition of electropositive metals6-15 due to their high ionic conductivity and electrochemical stability. In previous papers,13-16 we successfully demonstrated that RTILs as N-methyl-N-alkylpyrrolidinium bis 共trifluoromethanesulfonyl兲imides, PYR1NTFSI 共N = propyl or butyl兲, allow to overcome the conductivity drawback of solvent-free polymer electrolytes. Addition of PYR1NTFSI to poly共ethylene oxide兲 共PEO兲-based membranes enhances the ionic conductivity up to about 10−3 S cm−1at room temperature.16 PEO–LiX–RTIL electrolyte systems were tested at low–medium temperatures 共from 20 to 40°C兲 in dry, all-solid-state, Li/V2O5 and Li/LiFePO4 polymer batteries that delivered large capacity with high reversibility and very good cyclability. However, RTILs in the purity level required for electrochemical applications are not widely available commercially. For this reason many researchers in the electrochemistry community are getting involved in the synthesis of RTILs. The formerly developed synthesis16 of the PYR1NTFSI ionic liquids involved the use of acetone. However, acetone is an undesirable solvent for industrial applications because of its high volatility. In this paper, we propose an improved procedure for the synthesis of hydrophobic RTIL, composed of N-methyl-N-alkylpyrrolidinium or piperidinium 共P1N兲 cations and perfluoroalkylsulfonylimide 关共CnF2n+1SO2兲共CmF2m+1SO2兲N−, PFSI兴 anions, that involves ethyl acetate and water as the only solvents. Experimental N-methylpyrrolidine 共97 wt %兲, 1-Iodobutane 共99 wt %兲, ethyl acetate 共ACS grade, ⬎99.5 wt %兲, activated carbon 共Darco-G60兲,

* Electrochemical Society Active Member. z

E-mail: [email protected]

Table I. Weight of all chemicals used in the preparation of a 250 g batch of PYR14TFSI. Chemicals

Weight

Notes

Synthesis of the PYR14I precursor 60 PYR1 1-Iodobutane Ethyl acetate

10 wt % PYR1 excess

114 195

reagent/ethyl acetate volume ratio equal to 1.0/1.5

162 PYR14I Rinsing of the PYR14I precursor Ethyl acetate 294 共⫻3 rinses兲

0.2 L of ethyl acetate per 100 g of PYR14I

Synthesis of the PYR14TFSI ionic liquid 172.5 LiTFSI H 2O

162

PYR14TFSI

252

Rinsing of the PYR14TFSI ionic liquid H 2O 302 共⫻4–5 rinses兲

107 g of LiTFSI per 100 g of PYR14I 0.1 L of H2O per 100 g of PYR14I 157 g of PYR14TFSI per 100 g of PYR14I 0.12 L of H2O per 100 g of PYR14TFSI

Purification of ionic liquid PYR14TFSI Activated carbon 48 Alumina

120

Ethyl acetate

150

Rinsing of carbon and alumina Ethyl acetate 342

20 g of carbon per 100 g of PYR14TFSI 50 g of alumina per 100 g of PYR14TFSI 0.1 L of ethyl acetate per 100 g of PYR14TFSI 0.1 L of ethyl acetate per 13 g of carbon

and alumina 共acidic, Brockmann I兲 were purchased by Aldrich and used as received. LiTFSI salt 共99.9 wt %兲 was purchased from 3M and used as received. H2O was deionized by a Millipore ion resin exchange deionizer. Typically, 250 g batches of ionic liquids were prepared. Table I reports the amount of all chemicals used in the synthesis of a batch of PYR14TFSI. A 1 L, glass, sealed reactor was consecutively used for the synthesis of the precursor and the ionic liquid, thus avoiding the exposure of the reagent materials to the ambient. The operating temperature of the reactor was controlled by using an oil bath connected to

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

A1686

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲

Figure 1. Schematic representation of the synthesis process of the N-methyl-N-alkyl pyrrolidinium bis共trifluoromethansulfonyl兲imide, PYR1NTFSI, ionic liquid. EtAc = ethyl acetate, AC = activated carbon. Thick arrows indicate solids or slurries and thin arrows indicate liquids.

the reactor. The filtration steps were performed using a Millipore vacuum filter system; a water pump 共20 mTorr兲 was used to generate vacuum. The separation of the rinsing fluid 共water兲 from the ionic liquid was performed by aspiration. The ethyl acetate solvent was evaporated using a Resona LaboRota C-311 rotary evaporator connected with a water pump. The solvent obtained by the vacuum distillation was recycled into the process. The final drying step of the ionic liquids was performed in an ISCO NSV 9080 vacuum oven connected with an Edwards E1M40 one-stage high-vacuum pump 共10−2 mTorr兲 at 80°C for at least 12 h and then at 120°C for 24 h. The vacuum oven was located in a dry room 关relative humidity 共RH兲 ⬍ 0.1%兴. This allowed the handling of the materials in a dry environment. The materials were finally stored in sealed glass vials. A screening of all impurities present in the ionic liquid was performed by X-ray fluorescence spectroscopy using a Philips PW 2404 spectrometer. The concentration of cation impurities in the aqueous phases and the ionic liquid was checked by atomic absorption analysis using a SpetcrAA 共model 220兲 atomic absorption spectrometer. The water content in the ionic liquids was measured using the standard Karl Fisher method. The titrations were performed by an automatic Karl Fisher titrator 共Titralab 90, Radiometer, Copenhagen, Denmark兲 in dry room at 20°C. The Karl Fisher titrant was a twocomponent 共Hydranal 34800 and Hydranal 34811兲 reagent provided from Aldrich with a water equivalent of 2.00 ± 0.02 mg mL−1 at 20°C. Evaluation of the electrochemical stability window of the ionic liquids and their mixtures with a lithium salt was carried out by linear sweep voltammetry 共LSV, scan rate 5 mV s−1兲. The samples

were loaded in a three-electrode cell with platinum working electrode, carbon paper counter electrode, and Ag quasireference electrode whose potential was measured by adding the highly reversible redox couple ferrocene/ferrocinium 共Fc/Fc+兲 to the medium. Cyclic voltammetry 共CV兲 experiments were performed with the same cell structure to investigate the presence of irreversible reactions associated to impurities. LSV and CV tests were performed using a Schumberger Solartron Electrochemical Interface 1287 controlled by a software developed at ENEA. Results and Discussion The most stringent requirement considered in designing the new synthesis of the hydrophobic ionic liquids based on N-methylN-alkylpyrrolidinium or piperidinium 共P1N兲 cations and 共perfluoroalkylsulfonyl兲imide 关共CnF2n+1SO2兲共CmF2m+1SO2兲N−, PFSI兴 anions 共Fig. 1兲 was the solvent restriction. Our previous synthesis procedure16 made use of acetone for the purification of the PYR1NI precursor by recrystallization. PYR1NI was dissolved at high temperature 共80–100°C兲 in acetone and then slowly recrystallized at ambient or subambient temperatures.13 However, tests carried out in our laboratory 共data not reported here兲 have indicated that the presence of acetone traces in the ionic liquids might occasionally result in the formation of unwelcome, yellowish compounds during the final purification even at moderate temperatures 共⬎50°C兲. Other procedures reported in literature make use of dichloromethane or acetonitrile.11-17 None of these solvents is welcome in the chemical industry because of volatility 共acetone兲, safety, and environmental issues 共dichloromethane and acetonitrile兲. The synthesis here proposed was developed considering the use

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲

A1687

Figure 2. Structure of the PYR14TFSI ionic liquid.

of ethyl acetate and water as solvents. No other organic compounds were used throughout the synthesis. In addition, the recycling of the organic solvent was done even during the development of the procedure in the lab scale. Also, particular care was taken in determining the minimum excess of reagents and reaction conditions 共temperature and time兲 necessary to optimize the yield, thus producing the minimum amount of waste. The synthesis procedure was also optimized for the use of a single reactor.

Figure 4. Variation of the precursor, PYR14I, synthesis yield as a function of the N-methylpyrrolidine, PYR1, excess vs the stoichiometric amount. The synthesis yield was determined at two different temperatures and various times.

Synthesis of hydrophobic ionic liquids.— The synthesis of the hydrophobic ionic liquids, P1NPFSI, was performed through three steps: 共i兲 synthesis of the precursor P1NI, 共ii兲 synthesis of P1NPFSI, and 共iii兲 purification of P1NPFSI. The overall synthesis process is schematized in Fig. 1. In the following, the results of the investigation performed to optimize the synthesis of PYR14TFSI starting from N-methylpyrrolidine, 1-iodobutane, and lithium bis共trifluoromethansulfonyl兲imide are reported. The structure of the PYR14TFSI ionic liquid is shown in Fig. 2. The effect of the processing temperatures and times, the weight ratio of the reagents, and the amount of the purifying materials and solvents was investigated. The results are illustrated in Fig. 3-7. The weight of reagents, purification materials, and solvents used in the synthesis of a 250 gram batch is reported in Table I. However, the same procedure has been successfully used for the synthesis of several other ionic liquids + composed of 共P1N 兲 cations 共with the alkyl chain length ranging from 1 to 10兲 and 共PFSI−兲 anions such as 共CF3SO2兲2N− 共TFSI兲, 共C2F5SO2兲2N− 共BETI兲, 共C4F9SO2兲共CF3SO2兲N−, and others. Synthesis of the PYR14I precursor.— The precursor, PYR14I, was synthesized from N-methylpyrrolidine, PYR1, and 1-iodobutane by the following reaction

The reaction was performed in the presence of a moderate PYR1 excess with respect to the stoichiometric amount required 共see later兲. Although Reaction 1 could be performed by directly mixing the two reagents in a reactor, ethyl acetate, which is a good solvent for the reagents, was used to dilute the heat delivered during the reaction. This precaution is very important, especially when using iodoalkyl compounds with a long 共iodopentane and longer兲 or a short 共iodomethane or iodoethane兲 alkyl chain because the reaction with PYR1 is highly exothermic due to a large energy release 共long chain兲 or very fast kinetics 共short chain兲. For this reason, the reagents were first dissolved in ethyl acetate 共reagent to ethyl acetate volume ratio of about 2:3兲 and then mixed in a reactor. The reaction temperature was set at a value ranging from 20 to 50°C. The reaction time was also varied from 2 to 24 h. The effect of the reaction temperature and time on the yield of the process was investigated 共see later兲. Upon Reaction 1 a white, crystalline precipitate of PYR14I precursor was obtained. The exposure to sunlight of PYR14I was minimized to avoid oxidation iodide 共I−兲 to iodine 共I2兲. The precursor,

Figure 3. Variation of the precursor, PYR14I, synthesis yield as a function of the temperature and processing time. A 10 wt % excess of N-methylpyrrolidine, PYR1, with respect to the stoichiometric amount of 1-iodobutane was used.

Figure 5. Yield vs processing time dependence of the iodide/ bis共trifluoromethansulfonyl兲 imide, 共I兲− /共TFSI兲−, anion exchange process. A LiTFSI excess equal to 3 wt % of the stoichiometric amount was used. T = 20°C.

PYR1共liquid兲 + 1-Iodobutane共liquid兲 → PYR14I共solid兲

关1兴

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

A1688

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲 45°C. A 90 wt % yield was observed upon a 2 h reaction time at 50°C. Higher processing temperatures were not investigated to avoid getting close to the boiling temperature of the solvent 共ethyl acetate兲. In addition, at temperatures above 50°C a coloration of the solution was observed due to undesired side 共parasitic兲 reactions. In Fig. 4 the yield of the PYR14I precursor is reported as a function of the PYR1 weight excess 共vs the stoichiometric amount of 1-iodobutane兲 used in the synthesis. The effect of the PYR1 excess was evaluated on two sets of synthesis performed in different reaction conditions 共2 h at 50°C and 5 h at 45°C兲. The results evidence a moderate enhancement of the PYR14I synthesis yield up to a 10 wt % PYR1 excess. Increasing the PYR1 excess from 3 to 10 wt % resulted in a yield increase from 85% to 90% in the synthesis at 50°C for 2 h and from 90 to 100% in the synthesis at 45°C for 5 h. The use of higher PYR1 excess did not lead to any practical effect on the process yield. The PYR1 excess can be easily recycled by distillation of the ethyl acetate because it is highly soluble in this solvent. Therefore, the synthesis process of Pyr14I precursor can be practically driven with a 100% yield.

Figure 6. Yield of PYR14TFSI ionic liquid purification step as a function of the ethyl acetate used for the rinsing. The yield is reported as a function of the ethyl acetate/共carbon + Al2O3兲 weight ratio.

Synthesis of the PYR14TFSI ionic liquid.— PYR14TFSI ionic liquid was synthesized from PYR14I precursor and lithium bis共trifluoromethansulfonyl兲imide 共LiTFSI兲 by the following reaction PYR14I共solid兲 + LiTFSI共aqueous兲 → PYR14TFSI共liquid兲 + LiI共aqueous兲

insoluble in ethyl acetate, was easily separated from the solution by vacuum filtration. Impurities and excess reagents soluble in ethyl acetate were removed by repeatedly rinsing the precipitate until the liquid phase was clear and colorless. Figure 3 illustrates the yield of the PYR14I precursor synthesis as a function of the temperature and processing time. For these tests a PYR1 excess of 10 wt % with respect to the stoichiometric amount of 1-iodobutane was used. The data clearly show that the processing temperature largely affects the kinetic of the precursor synthesis 共see Reaction 1兲. At lower temperatures 共from 20 to 30°C兲 less than 70 wt % of the stoichiometric amount of PYR14I was obtained upon 5 h reaction time. A prolonged reaction time 共24 h兲 was required to increase the yield up to reasonable values 共⬎95 wt %兲. At higher temperatures 共ⱖ40°C兲 more than 90 wt % of the stoichiometric amount of PYR14I was obtained in a relatively short time. After 5 h of reaction, the yield approached 93 wt % at 40°C and 100 wt % at

关2兴

The lithium salt was first dissolved in deionized water to produce a 3.5 M solution 共corresponding to 1 mL of H2O per gram of LiTFSI兲. The solution was deareated by nitrogen bubbling to minimize the amount of dissolved oxygen that would cause the oxidation of iodide to iodine. The LiTFSI solution was then added to the PYR14I solid precursor 共kept in a small amount of ethyl acetate兲. The reactor was also previously deareated and nitrogen was blown over the liquid phase throughout the addition. The addition of the aqueous solution 共containing LiTFSI兲 to PYR14I resulted in the dissolution of the latter immediately followed by its reaction with LiTFSI to form hydrophobic PYR14TFSI and hydrophilic LiI. The rapid formation of two liquid phases clearly indicated that the reaction proceeded quickly. However, to facilitate the anion exchange reaction the two phases were vigorously stirred at room temperature. After a selected time the stirring was interrupted and the phase separation took place in a few minutes. The

Figure 7. Summary of the yields and losses detected during the PYR14TFSI synthesis. The data given for each synthesis step refer to the expected yield for that step. The overall process yield is reported in the last panel.

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲 Table II. Yield of the anion exchange process as a function of the LiTFSI excess used. LiTFSI excess 共wt %兲

Yield 共%兲

0 3 10 30 100

93.5 ± 0.7 94.6 ± 0.3 92.0 ± 0.7 90.4 ± 0.7 ⬍50

upper phase was mostly composed of water, lithium iodide 共LiI兲, and excess LiTFSI. The lower phase was composed mostly of PYR14TFSI ionic liquid with a small amount of ethyl acetate and traces of water and lithium salts 共i.e., LiI and LiTFSI兲. Typically, the upper 共aqueous兲 phase was clear and colorless while the lower was clear with a slightly yellowish coloration. The anion exchange reaction, in which the iodide anions are replaced by the TFSI anions 共see Reaction 2兲 to form PYR14TFSI, is driven by the intrinsic hydrophobicity of both the anion 共TFSI−兲 and the cation 共PYR+14兲. These ions, in which the charge is well-shielded by hydrophobic groups 共PYR+14兲 or extensively delocalized 共TFSI−兲, do not easily form hydrogen bonds with water molecules and tend to separate from the aqueous phase, forming a second 共denser兲 liquid phase. The disappearance of the product from the aqueous solution drove the exchange reaction to completion. After removal of the upper aqueous phase, the PYR14TFSI ionic liquid was washed several times with deionized water to remove water-soluble salts and impurities such as LiI and excess LiTFSI. After rinsing, a clear and colorless 共or slightly yellowish兲 liquid containing PYR14TFSI ionic liquid and a small amount of ethyl acetate was obtained. The rinsing was carried out at 20°C to limit the loss of ionic liquid, whose solubility in water is very low at room temperature but increases with temperature. The dependence of the process yield vs exchange time is reported in Fig. 5. After 10 min of exchange time more than 85 wt % of the expected stoichiometric amount of PYR14TFSI was obtained. After 1 h the yield of the anion exchange process leveled to 95 wt % of the expected stoichiometric amount from Reaction 2. No practical increase of the yield was observed at longer process times 共240 h兲. Table II reports the yield of the PYR14TFSI synthesis as a function of the LiTFSI excess. A slight increase from 93.5 to 94.6 wt % yield was observed in passing from 0 to 3 wt % LiTFSI excess. A further increase of the LiTFSI excess 共up to 30 wt %兲 led to a decrease of the anion exchange yield. These results suggested the use of a very limited excess of LiTFSI, i.e., ranging from 0 to 3 wt % of the stoichiometric amount, because LiTFSI is the most expensive reagent used in the synthesis and, more importantly, a reduction of yield was found at larger LiTFSI excess. In the extreme case where 100 wt % excess of LiTFSI was used, more than 50 wt % of ionic liquid dissolved in the aqueous phase. This behavior is due to the fact that even if PYR14TFSI is insoluble in water 共at room temperature兲 its anion 共TFSI−兲 bonds very strongly to lithium cations. Raman18 and nuclear magnetic resonance 共NMR兲19 measurements have shown that 1 lithium ion can coordinate up to 3 TFSI− anions in LiTFSI-PYR13TFSI mixtures. The strong coordination of lithium ions by water molecules and TFSI− anions is the driving force that causes the partial dissolution of PYR14TFSI in the aqueous phase. The higher the amount of lithium in the aqueous phase, the lower the yield of the anion exchange process. The coordination of Li cations and TFSI− anions is also the reason for the presence of lithium salts 共i.e., unreacted LiTFSI and LiI兲 in the ionic liquid phase. These salts were removed from PYR14TFSI through rinsing with deionized water at room temperature. The water/ionic liquid weight ratio was fixed to 1.2 for each rinsing step. Table III reports the lithium concentration in the aqueous phase collected after the anion exchange and the rinsing steps. A

A1689

Table III. Lithium concentration in the aqueous phase collected after the anion exchange process and the rinsing steps. The water/ionic liquid weight ratio was fixed to 1.2 in each rinsing step. The lithium concentration in the ionic liquid was found to be below the detection limit of the instrument. The value reported indicates the lowest limit detectable. Concentration 共mg/l兲

Lithium content in: Aqueous phase from the anion exchange process Aqueous phase after rinsing step 共#兲

Ionic liquid after rinsing

14,000 ± 300 1st 2nd 3rd 4th 5th

1,500 ± 30 150 ± 3 21.0 ± 0.4 1.00 ± 0.02 0.100 ± 0.002 ⬍0.003

lithium concentration of 14,000 mg L−1 was detected in the aqueous phase after the anion exchange process, corresponding to 89.3 wt % of overall lithium extracted. A fast and progressive decay of lithium concentration from 1,500 to 0.1 mg L−1 was observed in the aqueous phase after the first and the fifth rinsing step, respectively. Practically, more than 99.99 wt % of the overall lithium is extracted after three rinsing steps. However, due to the fact that LITFSI is highly soluble in PYR14TFSI, two further rinsing steps were performed to achieve an extremely low lithium content in the ionic liquid. The lithium concentration in the ionic liquid after five water rinsing steps is found to be lower than 0.003 mg/L 共2 ppm兲. Purification of the PYR14TFSI ionic liquid.— The purification was performed by intimately mixing the PYR14TFSI with activated carbon 共AC兲 and alumina 共Al2O3兲 materials which are able to trap impurities. First of all, further ethyl acetate was added to the ionic liquid to decrease viscosity. Activated carbon was added to the solution of PYR14TFSI in ethyl acetate and the resulting slurry was stirred for 10 h at 70–75°C, i.e., just below the boiling point of ethyl acetate. Tests performed at lower temperature and shorter stirring time gave slightly yellow-colored ionic liquids. Upon cooling to room temperature, the suspension was vacuum-filtered and the liquid fraction 共containing PYR14TFSI and ethyl acetate兲 and the solid fraction 共containing carbon with some amounts of adsorbed PYR14TFSI and ethyl acetate兲 were collected separately. Successively, alumina was added to the liquid fraction that was then stirred for at least 5 h at room temperature. The activated alumina was then separated from the liquid phase by vacuum-filtering. The liquid fraction collected was placed in a rotary evaporator at 80°C under vacuum 共20 mTorr兲 to separate the ethyl acetate 共that was recycled into the process兲 from the pure PYR14TFSI ionic liquid. This represents the purest fraction 共labeled A in Fig. 1兲 of the ionic liquid produced in the process. The solid fractions separated from the two purification steps were rinsed in sequence with ethyl acetate to recover the ionic liquid trapped in the solid phases. The PYR14TFSI 共labeled C in Fig. 1兲 was separated from the rinsing liquid as described above 共evaporation of ethyl acetate under vacuum at 80°C兲. For analysis purposes a small fraction of the rinsing liquid 共labeled B in Fig. 1兲 was collected and analyzed after evaporation of the ethyl acetate and drying. Finally, the activated carbon and alumina solid fractions were rinsed with a larger amount of ethyl acetate to extract all the ionic liquid trapped. The resulting solution 共labeled D in Fig. 1兲 showed a slightly yellow coloration, thus indicating that part of the impurities were also extracted from activated carbon and alumina. To increase the overall yield of the synthesis, this fraction was further purified following the same procedure described above. For analysis purposes a small fraction of solution D was collected and analyzed after evaporation of the ethyl acetate and drying.

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲

A1690

Table IV. Chemical analysis of the PYR14TFSI fractions collected throughout the purification process.

Sample Theoretical A B C D

C 共wt %兲 31.3 30.9 31.1 30.9 30.7

± ± ± ±

0.05 0.05 0.05 0.05

H 共wt %兲 4.78 4.67 4.71 4.65 4.63

± ± ± ±

0.01 0.01 0.01 0.01

The different fractions of PYR14TFSI ionic liquid were dried under high vacuum 共10−2 mTorr兲 at 80°C for at least 12 h and then at 120°C for 24 h, thus obtaining RTILs with a moisture content below 10 ppm. The materials were stored in sealed glass tubes in a controlled environment 共dry-room, RH ⬍0.1%兲 at 20°C. Overall, the synthesis process described above gives clear and colorless PYR14TFSI ionic liquid. The amounts of activated carbon and alumina used to purify PYR14TFSI were selected on the basis of an iterative process in which the amount of the purifying materials was increased until a clear and colorless solution of PYR14TFSI in ethyl acetate was obtained. The optimum C/PYR14TFSI weight ratio was found to be 0.2, while the optimum Al2O3 /PYR14TFSI weight ratio was found to be 0.5. These two values were found to give good results for all the hydrophobic ionic liquids synthesized. Higher weight ratios are not recommended because the fraction of ionic liquid obtained during the first filtration is strongly reduced and larger amounts of ethyl acetate are required for the further extraction of the ionic liquid. With the above-mentioned weight ratios about 47 wt % of PYR14TFSI was recovered in the first filtration 共see Fig. 1兲. The chemical analysis of this fraction 共see Table IV兲 showed the absence of impurities in any relevant quantity. Only chlorine and silicon were found by RFA analysis at levels below 100 ppm. The rinsing of carbon and alumina to recover the second fraction of ionic liquid is also critical because the amount of ethyl acetate needs to be dosed accurately to recover most of the ionic liquid without extracting the impurities. In Fig. 6 is plotted the yield, referred to the expected PYR14TFSI ionic liquid stoichiometric amount, as a function of the ethyl acetate content used for the rinsing of carbon and alumina. The data are reported as a function of the ethyl acetate/共C + Al2O3兲 weight ratio. The 共C ⫹ Al2O3兲/ PYR14TFSI and the carbon/alumina weight ratios were fixed to 0.7 and 0.4, respectively. The first data point, 46.8 wt %, refers to the amount of ionic liquid that was obtained directly by the filtration under vacuum of the carbon and alumina slurries 共without any rinsing兲. The rinsing of carbon and alumina allowed the enhancement of the yield of the purification step up to about 89 wt %. The yield leveled off at 89 wt % for an ethyl acetate/共C + Al2O3兲 weight ratio of 1.5, corresponding to about 1.7 mL of ethyl acetate per gram of C ⫹ Al2O3. No relevant amounts of PYR14TFSI were extracted at higher ethyl acetate/共C + Al2O3兲 weight ratios. The chemical analysis showed that the PYR14TFSI samples taken after rinsing the carbon and alumina materials 共B and C in Fig. 1兲 also had impurity levels below 100 ppm. A final fraction of ionic liquid was recovered by an extensive rinsing of carbon and alumina with ethyl acetate 共0.75 L of ethyl acetate per 100 g of carbon兲. However, the ionic liquid collected 共sample D in Fig. 1兲 had a yellowish coloration, thus indicating the partial release of impurities by carbon and alumina. This fraction is usually repurified, although the chemical analysis did not indicate a substantial contamination. Figure 7 reports the weight percent of the ionic liquid fractions obtained or lost throughout the entire synthesis process. The precursor synthesis yield of 100% was reached using an excess 共10 wt %兲 of PYR1. The anion exchange process showed a yield of 94.3 wt % using a moderate excess 共3 wt %兲 of LiTFSI. The purification step showed a yield of 91.3 wt %, of which about 47 wt % was obtained

N 共wt %兲 6.63 6.66 6.67 6.66 6.70

± ± ± ±

0.01 0.01 0.01 0.01

Impurities 共⬍100 ppm兲 共* = 2 ppm; ** = 40 ppm兲 Li共*兲, Cl, Si Li共*兲, Cl, Si, Mo Li共*兲, Al, Mo Li共*兲, Na, Fe 共**兲, Si, Mo

directly through the first filtration 共sample A兲, while more than 42 wt % of PYR14TFSI was recovered after carbon and alumina rinsing 共sample C兲. A small fraction 共2.2 wt %兲 was recovered through the extensive rinsing of carbon and alumina with ethyl acetate 共sample D兲. The results of the elemental analysis and the impurities performed on the four fractions collected during the synthesis are reported in Table IV. All samples showed a slight defect of C and H and a slight excess of N. Because the defecting elements are present only 共H兲 or mostly 共C兲 in the cation 共PYR+14兲 and the impurities detected are cations, it is reasonable to assume that the deviations evidenced by the elemental analysis are due to the presence of M+TFSI− salts 共where M+ is Li+, Na+, etc.兲 in the ionic liquid. In fact, the presence of only 100 ppm 共0.01 wt %兲 of Li+ would correspond to an excess 0.36 wt % of TFSI−. Overall, the PYR14TFSI synthesis procedure gave 86.4 wt % of the expected amount. The ionic liquid losses were also localized; 5.4 wt % of PYR14TFSI was lost by dissolution in water during the anion exchange, while another 8.2 wt % of PYR14TFSI was lost in the purification step 共trapped into carbon and alumina兲. However, these fractions of PYR14TFSI could be 共should be兲 recovered by the appropriate processing of the rinsing water waste and the carbon and alumina waste produced in the synthesis. The purity of PYR14TFSI has been investigated also by means of electrochemical measurements. Figure 8 reports the LSVs performed with platinum electrode to investigate the anodic and the cathodic PYR14TFSI 共solid line兲 electrochemical stability at 20°C. Separate experiments were conducted for the cathodic and anodic electrochemical stability measurements by scanning from the opencircuit potential to negative 共cathodic兲 or positive 共anodic兲 voltages 共using clean electrodes and a fresh sample in each case兲.

Figure 8. LSVs of pure PYR14TFSI 共solid line兲 and PYR14TFSI/LiTFSI 共PYR+14 /Li+ = 9兲 mixture 共dotted line兲 at 20°C on a platinum working electrode. A carbon paper disk and a silver strip were used as the counter and the reference electrode, respectively. Scan rate: 5 mV s−1. The potential is given with reference to the Fc+ /Fc couple.

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 153 共9兲 A1685-A1691 共2006兲

Figure 9. Lithium/plating process of PYR14TFSI-LiTFSI 共PYR+14 /Li+ = 9兲 mixture on a platinum working electrode at 20°C. Carbon paper and silver foil were used as the counter and the reference electrodes, respectively. Scan rate: 5 mV s−1. The potential is given with reference to the Fc+ /Fc couple.

The anodic current flow observed at voltages above 2.25 V 共vs Fc+ /Fc0兲 corresponds to the oxidation of the TFSI− anion. No other features are observed at lower voltages during the anodic scan, thus excluding the presence of impurities that are oxidized within the anodic electrochemical stability limit of the ionic liquid. Some current flow is observed during the cathodic scan at about −1.25 V 共vs Fc+ /Fc0兲, i.e., well above the extensive decomposition 共reduction兲 process that takes place at −3.1 V 共vs Fc+ /Fc0兲. The nature of the processes taking place is not well known but is related with the PYR+14 cation and not with the reduction of impurities.20 In fact, the addition of LiTFSI 共PYR+14 /Li = 9兲 resulted in the substantial disappearance of the current features in the voltage region between −1.25 and −3.0 V 共vs Fc+ /Fc0兲 共see dotted curve in Fig. 8兲. In the presence of LiTFSI the cathodic stability was largely enhanced to the point that lithium metal could be plated on the platinum working electrode, as indicated by the cathodic peak shown at −3.2 V 共vs Fc+ /Fc0兲. At more cathodic potentials the current associated to the lithium plating process assumed the limiting current value until the extensive decomposition of the PYR+14 cations took place at −4.2 V 共vs Fc+ /Fc0兲. However, if the cathodic limit was set above such a value, the lithium plating process was found to be reversible, as shown by the CV presented in Fig. 9. Such a good lithium plating– stripping reversibility allows to exclude the presence of impurities in PYR14TFSI because lithium plating is the most cathodic electrochemical process.

A1691

Conclusions An improved process has been developed to synthesize hydrophobic RTIL composed of N-methyl-N-alkylpyrrolidinium 共or piperidinium兲 cations and 共perfluoroalkylsulfonyl兲imide 关共CnF2n+1SO2兲 − 共CmF2m+1SO2兲N−兴 anions. The preparation procedure is suitable for lab-scale and industrial applications and requires only ethyl acetate and water as solvents. The process allowed to obtain clear, colorless, inodorous, high-purity ionic liquid 共⬎99.5 wt %兲 with a water content below 10 ppm and an overall yield above 86 wt %. Electrochemical measurements performed on three-electrode cells containing PYR14TFSI and its mixture with LiTFSI as the electrolyte have confirmed the high purity of the ionic liquid by showing the possibility of reversibly reducing lithium ions to lithium metal. Acknowledgment The authors thank Dr. A. Prodi-Schwab of Degussa for the X-ray fluorescence measurements and MIUR 共Italian Ministry of Education, University and Research–FISR program兲 for financial support. Italian National Agency for New Technologies, Energy and the Environment assisted in meeting the publication costs of this article.

References 1. P. Wasserscheid and W. Keim, Angew. Chem., Int. Ed., 39, 3772 共2000兲. 2. Ionic Liquids-Industrial Applications to Green Chemistry, R. D. Rogers and K. R. Seddon, Editors, ACS Symposium Series 818; Oxford University Press, New York 共2002兲. 3. M. J. Earle and K. R. Seddon, Pure Appl. Chem., 72, 1391 共2000兲. 4. J. L. Anderson, J. Ding, T. Welton, and D. W. Armstrong, J. Am. Chem. Soc., 124, 14247 共2002兲. 5. J. Dupont, R. F. de Souza, and P. A. Z. Suarez, Chem. Rev. (Washington, D.C.), 102, 3667 共2002兲. 6. A. I. Bhatt, I. May, V. A. Volkovich, M. E. Hetherington, B. Lewin, R. C. Thied, and N. Ertok, J. Chem. Soc. Dalton Trans., 2002, 4532. 7. S. Panozzo, M. Armand, and O. Stephan, Appl. Phys. Lett., 80, 679 共2002兲. 8. P. Wang, S. M. Zakeeruddin, I. Exnar, and M. Gratzel, Chem. Commun. (Cambridge), 2002, 2972. 9. J. Fuller, A. C. Breda, and R. T. Carlin, J. Electroanal. Chem., 459, 29 共1998兲. 10. H. Nakagawa, S. Izuchi, K. Kunawa, T. Nukuda, and Y. Aihara, J. Electrochem. Soc., 150, A695 共2003兲. 11. H. Sakaebe and H. Matsumoto, Electrochem. Commun., 5, 594 共2003兲. 12. A. Noda, M. A. B. H. Susan, K. Kudo, S. Mitsushima, K. Hayamizu, and M. Watanabe, J. Phys. Chem. B, 107, 4024 共2003兲. 13. J.-H. Shin, W. A. Henderson, and S. Passerini, J. Electrochem. Soc., 152, A978 共2005兲. 14. J.-H. Shin, W. A. Henderson, and S. Passerini, Electrochem. Solid-State Lett., 8, A125 共2005兲. 15. J.-H. Shin, W. A. Henderson, G. B. Appetecchi, F. Alessandrini, and S. Passerini, Electrochim. Acta, 50, 3859 共2005兲. 16. W. A. Henderson and S. Passerini, Chem. Mater., 16, 2881 共2004兲. 17. D. R. MacFarlane, P. Meakin, J. Sun, N. Amini, and M. Forsyth, J. Phys. Chem., 103, 4164 共1999兲. 18. M. Castriota, T. Caruso, R. G. Agostino, E. Cazzanelli, W. A. Henderson, and S. Passerini, J. Phys. Chem. A, 109, 92 共2005兲. 19. I. Nicotera, C. Oliviero, W. A. Henderson, G. B. Appetecchi, and S. Passerini, J. Phys. Chem. B, 109, 22814 共2005兲. 20. N. Byrne, P. C. Howlett, D. R. MacFarlane, and M. Forsyth, Adv. Mater. (Weinheim, Ger.), 17, 2497 共2005兲.

Downloaded 10 Jan 2011 to 192.107.76.2. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp