Ionic Liquids as Green Solvents - American Chemical Society

Chapter 38

The Use of Ionic Liquids in Polymer Gel Electrolytes 1,2

1,2

2,3,

Downloaded by UNIV OF GUELPH LIBRARY on July 18, 2012 | http://pubs.acs.org Publication Date: August 26, 2003 | doi: 10.1021/bk-2003-0856.ch038

Hugh C. De Long , Paul C. Trulove , and Thomas E. Sutto * 1

Air Force Office of Scientific Research, 801 Randolph Street, Arlington, VA 22203-1977 Naval Research Laboratory, Chemistry Department Building 207, Department 6170, 4555 Overlook Avenue SW, Washington, DC 20375 U.S. Naval Academy, Chemistry Department, Annapolis, MD 21402 2

3

Polymer gel electrolytes composed of 1,2-dimethyl-3-n-alkyl-imidazolium bis-trifluoromethanesulfonylimide (alkyl = propyl or butyl) and polyvinylidenedifluoro-hexafluoropropylene are characterized by ac-impedance and cyclic voltammetry. Two electrode charge-discharge experiments were also performed using graphitic paper or Li metal as the anode, and polymer composites of LiMn O , LiCoO , or V O as cathodes. Results indicated that the polymer composite gel electrolytes were stable for over 50 cycles when used in direct contact with L i metal. High efficiencies and low voltage drop-offs indicate that polymer gel composite electrodes composed of these ionic liquids are a viable alternative to the more common organic solvent electrolytes. 2

478

4

2

2

5

U.S. government work. Published 2003 American Chemical Society In Ionic Liquids as Green Solvents; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

479


Introduction By now, any reader of this collection of work should be extremely familiar with the burgeoning field of both ionic liquids as well as their numerous benefits to green chemistry. Ionic liquids themselves exist as a subset of the science of molten salts, which are already extensively used in industry because of their unique physical properties. It would go well beyond the required length of this paper to describe all of the work that has gone into molten salt (and ionic liquid) techniques and technology, as well as the scientists responsible for all of these developments. Suffice it to say that some of these scientists, and their associated research institutes, serve as the heads and homes of several extended families of researchers devoted to molten salts. In order to gain a better understanding of this, the reader is strongly urged to look through the entire thirteen volume series on Molten Salts published by The Electrochemical Society in which the evolution of novel ideas that have led to the recent surge in interest in molten salts can be traced. Additionally, another excellent text would be David Lovering's, "Molten Salt Technology", published in 1982, which summarizes the state of the ecological, economical and technical advantages of molten salts at that time, and which are still true today. Of all the potential applications for molten salts, one of the least investigated is their use as replacements for some of the more common organic electrolytes in Li-ion batteries. The very physical properties that make them useful as electrolytes for electrochemical deposition techniques, or purification media for plating out even refractory metals, also make them the ideal choice as electrolytes in the rapidly growing area of high energy density power sources. However, in terms of actual attractiveness to the battery industry, it will be important to significantly narrow the scope of molten salts studied. Since most of these systems need to operate in the ambient temperature regime, the molten salt family of choice is clearly the ionic liquids. Furthermore, the question arises as to which ionic liquids to choose. In terms of the anion, this is a relatively simple matter. Although A1C1 \ orother similar types of anions have been studied, they are perhaps not the best choice for commercial Li-ion batteries due to their tendency to evolve heat and HC1 gas upon contact with moisture. Furthermore, some work has even shown that reactions with the A1C1 " anion produce Li-Al alloy coatings on graphitic electrodes rather than allowing for L i intercalation. Thus, for Li-ion battery systems, the best anions to chose from are predominately the air and water stable fluorinated anions, such as BF ", PF ', or the bis-trifluoromethanesulfoylimide (TFSF) anion. Before we can decide which of these to use, it is important to discuss another significant aspect of ionic liquids as they apply to power sources. 4

1

4

2

4

6

In Ionic Liquids as Green Solvents; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

480 Perhaps the crucial difference between these ionic liquid electrolytes and the more commonly used organic carbonates is the presence of the charged imidazole ring and the anion, both of which move, diffuse, and interact with the electrodes when a current is applied. This can, for electrode materials of a layered structure type such as in the case of graphitic or TiS electrodes, lead to a unique type of battery in which the ionic liquid itself serves as both the electrolyte, and the source of the intercalative guest species. This configuration is termed the dual intercalating molten electrolyte, DIME, batteries. " Common sense tells us that these DIME batteries will not exhibit high enough energy densities to compete with Li-ion batteries, due to the much larger size of the current carrying species. However, they do represent the simplest form of a battery system, as well as present a very basic test case to explore the electrochemical stability of a variety of cations and anions during extensive series of charge-discharge experiments. In regards to the anion stability, experiments have shown that the BF " anion is an extremely poor choice due to its electrochemical instability especially when intercalated into graphitic materials. Furthermore, although the PF " anion does behave moderately well, it does create extremely viscous and less ionically conductive ionic liquids. " However, the TFSI" anion has performed admirably well in the basic DIME system, and seems a good anion to chose. In term of the cations to choose from, the substituted imidazoles form ionic liquids with significantly large electrochemical windows, " and four of the more common types of these cations are shown in figure 1. DIME studies have also produced significant evidence that although the l-ethyl-3methylimidazolium, ΕΜΓ, or l-butyl-3-methylimidazolium, ΒΜΓ, cations are the ones most often studied, they are inherently unstable over prolonged use in DIME systems. In fact, both in DIME studies, and in simple Li-ion battery studies, these cations breakdown over time at the high potentials used in highenergy battery sources, most likely due to the reactivity of the lone hydrogen at the 2-position of the imidazole ring. ' However, the l,2-dimethyl-3-npropylimidazolium, DMPI , or l,2-dimethyl-3-n-butylimidazolium, DMBI , cations have proven to be extremely stable during a number of charge-discharge cycles. ' Thus we will concentrate on the two primary ionic liquids, DMPITFSI and DMBITFSI. The properties of ionic liquids also offer profound advantages when used in polymer gel electrolytes. The discovery of ionically conductive polymers such as polyethylene oxide, PEO, and others led to a wide-ranging amount of research in an attempt to prepare solid state batteries using a pure, solid polymer electrolyte. Unfortunately, these solid polymers are very poor ionic 2


3

6

4

4

6

7

9

10

11

13

3,9

3 7,9

+

3 7,9

1415


+

481

H R,= Methyl, R = H, R = Ethyl-» 2

3

l-ethyl-3-methylimodazolium; ΕΜΓ H

R,= Methyl, R = H, R = Butyl-» 2

3


1 -butyl-3-methy limidazolium; BMI

+

1

Rj= R = Methyl, R = Propyl-> l,2Klimethyl-3-n-propylimidazolium; DMPI 2

+

3

Rj= R = Methyl, R = Butyl1,2-dimethyl-3-n-butylimidazolium; DMBI 2

+

3

Figure L The basic structure of the four most commonly used imidazolium based ionic liquids. 16

conductors at room temperature. In order to overcome this rather significant drawback, investigators have tried to correct this deficiency by creating polymer gel composite electrolytes, in which a liquid electrolyte is added to the polymer in order to improve the ionic conductivity while maintaining the ease of working with a solid polymer. These polymer composite gel electrolytes have been extensively studied, using a variety of polymers, as well as a multitude of electrolytes; the most common of which are the organic carbonates such as propylene carbonate. " Unfortunately, these liquid components suffer from the same drawbacks as when used as pure liquids. Their volatility leads to the loss of electrolyte over time and the eventually loss of ionic conductivity, and it increases the polymer's flammability. In fact, some formulations have added fire-retardants to the polymer matrix. In order to overcome these shortfalls, ionic liquids can be used in polymer composite gel electrolytes allowing for the formation of much safer electrolytes for Li-ion solid state batteries. The last decision to be made is which polymer to use as the polymer component in the gels. Initial studies have already pin-pointed one significant problem in choosing PEO. For PEO polymer composite gel electrolytes, the ionic liquid dissolves the PEO somewhat leading to a dimensionally unstable solid like a soft wax, and the resulting gels are fragile and extremely moisture sensitive. However, polyvinylidenedifluoro-hexafluoropropylene, PVdF-HFP, has behaved admirably well in the case of other ionic liquids. ' ' Therefore, polymer composite gel electrolytes composed of DMPITFSI or DMBITFSI and PVdF-HFP will be studied and characterized for potential use in Li-ion batteries. 17

19

20

21

8 9 22,23


482

Experimental Synthesis of the Initial Substituted Imidazolium Salts. The preparation of DMMC1 involved the direct reaction of 1,2-dimethylimidazole (Aldrich, 98%) and 1-chloropropane (Aldrich, 99%), or 1-chlorobutane (Aldrich 99.5%) in a 1:1.15 molar ratio. The 1,2-dimethylimidazole (m.p. 38 °C) was vacuum distilled at approximately 10" torr at 135 °C in order to begin with as pure a product as possible. This was to avoid the formation of the non-volatile yellowish brown tint that often forms in these molten salts formed around 120130 °C. All subsequent procedures were performed in a dry box ( 0 and H 0 < 1 ppm) unless noted otherwise. The 1,2-dimethylimidazole was melted at 50 °C in the dry box. The starting amount of 1,2-dimethylimidazole, 500 g, and a 15% excess of the corresponding 1-chloro alkane were placed, along with 50 ml of acetonitrile, in a thick-walled, single neck, 2-liter round bottom flask. Once filled, the round bottom flask was removed from the dry box and fitted with a reflux condenser. The solution was degassed several times with dry nitrogen, gradually heated to 60 °C for the DMPIC1 reaction, and 80 °C for the DMBIC1 reaction, and allowed to react for 2 days under nitrogen pressure. After this 2day reaction period, the temperature was increased to 85 °C for the DMPIC1 reaction, and to 95 °C for the DMBIC1 reaction. Both reactions refluxed for an additional 5 days. Cooling the solution produced a white precipitate and a very faint, yellow supernatant. 300 ml of ethyl acetate was added to the round bottom flask to precipitate all of the DMJWC1. The material was filtered and washed with five 100 ml washings of ethyl acetate in order to remove all of the unreacted 1,2-dimethylimidazole. The solid material was dissolved in a minimum amount of hot acetonitrile and quickly crashed out of solution by the addition of a large excess of ethyl acetate, since it was found that slow recrystallization resulted in the formation of slightly yellowed crystals. The final product was a white crystalline material. Finally, the DMJMC1 was heated to 100 °C under a dynamic vacuum (10 torr) for 2 days to remove the volatile contaminants giving a final total yield of approximately 80%. The synthesis of BMIC1 was performed in a manner identical to that for the DMBIC1, except that 1-methylimidazole was distilled and used in place of the 1,2-dimethylimidazole. Preparation of the ionic liquids. The preparation of DMJWTFSI was done by an ion exchange of the respective chloride salt with LiTFSI (Aldrich, 99.98%, dried for 12 hours at 80 °C in a vacuum oven at 10" torr) in acetonitrile. For a typical reaction, one mole of the DMJHC1 was placed in a 1-liter reaction flask fitted with a threaded Teflon plug and dissolved in a minimum amount of acetonitrile. To this solution, an equal molar amount of the LiTFSI was added. The flask was sealed and allowed to stir at room temperature for 7 days. After which, the flask was removed from the dry box for die last purification steps.


3

2

3

3


2

483 Preparation of the BMIPF was done in a similar manner, except that NH PF (Aldrich, 99.5%) was used instead of LiTFSI. Upon completion of the anion exchange, the solid material (LiCl or NH C1) was removed by vacuum filtration using a glass frit of medium pore size. Subsequently, all of the acetonitrile was removed by rotavaporization, causing much of the remaining LiCl to precipitate out of solution. This impure form of the DMJHTFSI or BMIPF was washed seven times with 200 mL of distilled water. Since the DMMTFSI and BMIPF are hydrophobic, these solutions separated into layers, with the LiCl or NH C1 being transferred in the water layer, which was discarded. The solution was then dried for 12 hours at 85 °C under a dynamic vacuum (10" torr) for 1 day to remove most of the water. 300 ml of acetonitrile, 15 g of decolorizing carbon and 15 g of neutral alumina were added to the solution, which was allowed to stir at room temperature in a sealed reaction flask for 24 hours. The carbon black and alumina were removed by successive filtration. Prior to the final filtration, the solution volume was reduced by half using a rotovap, and re-diluted with acetone. These ionic liquid/acetonitrile/acetone solutions were then filtered through a medium pore size glass frit, a 1 μ filter disc (Whatman, PTFE Membrane), a 0.45 μfilterdisc (Whatman, PTFE Membrane), and a 0.2 μ filter disc (Whatman, PTFE Membrane). For the final step in the purification process, the molten salt solutions were heated to 85 °C under a dynamic vacuum (10' torr) for 2 days to remove the volatile components and any remaining water. The final molten salts were clear and colorless. Preparation of the lithium containing ionic liquids. LiTFSI (Aldrich, 99.98%, dried for 12 hours at 80 °C in a vacuum oven at 10 torr) was added to both the DMPITFSI and DMBITFSI to form 1.0 M solutions. For these solutions, intermittent sonication and heating to 50 °C was used to facilitate dissolution of the lithium salt. In every case, the 1.0 M L i ion solutions were slightly more viscous, but still clear and colorless. Preparation of the 1.0M LiPF (Aldrich, 99.95%) in BMIPF was done in a similar manner. Preparation of the polymer gel electrolytes. The polymer gels were prepared in an Ar dry box (H 0 and 0 < lppm). The PVdF-HFP (Kynar 2801-00 from Elf Atochem) was dissolved using 4-methyl-2-pentanone, 4M2P, (98%, Aldrich) as the solvent. Previous work has shown that the nominal composition of these gels should be between 70-85% ionic liquid; therefore, gels that were 75% 1.0 M Li/DMjRITFSI or BMIPF and 25% PVdF-HFP were prepared. For each sample, 0.625 grams of the polymer was added to 10 mL of the 4M2P solvent. The solvent/polymer mixture was sonicated for 2 minutes at high power, allowed to cool, and then resonicated for 2 additional minutes in order to completely dissolve the polymer in the solvent. 1.875 grams of the ionic liquid were added, and sonication was continued, with intermittent pauses to allow for sample cooling, until there was a marked drop-off in the sound of 6

4

4

6

6

4


3

3

3

6

6

2

2

6


6

484 cavitation produced. The resulting ionic liquid/polymer solution was removed from the sonicator and allowed to stir at 60 °C for 15 to 30 minutes under Ar until the mixture became extremely viscous. It should be noted here that the 4M2P/PVdF-HFP/Li/BMIPF solution was white and non-transparent, the 4M2P/PVdF-HFP/Li/DMPITFSI solution was slightly opaque but still transparent, while the solution of 4M2P/PVdF-HFP/Li/DMBITFSI was absolutely clear, indicating a better mixing of the components. The liquid composite material was removed from the heat, poured into a 2.5" diameter A l weighing boat, and cured at room temperature for 24 hours in the dry box. The A l dishes were subsequently placed in a dessicator, and dried in a vacuum oven an additional 24 hours under a dynamic vacuum (approximately 10' torr) at 75 °C. For all polymer gels prepared, the thickness of the gels was typically between 1-2 mm. The 1.0 M Li/BMIPF gel was flexible, but solid white in color. The 1.0 M Li/DMPITFSI gel was flexible and slightly transparent, while the 1.0M Li/DMBITFSI gel wasflexibleand completely transparent. Preparation of the polymer composite electrodes. For solid-state test cells, the various electrodes were prepared from polymer solutions of the electrode material. 1.4 grams of the primary electrode materials, L i M n 0 (spinel phase, Aldrich, 99.5%), LiCo0 (Aldrich 99%), or V 0 (Aldrich, 99%), 0.2 grams of graphite (Aldrich, 99.995%, 1-2 micron particle size) and 0.4 grams of PVdF-HFP were mixed with 10 mL 4M2P and sonicated in 2 minute intervals for a total of 14 minutes. This thick solution was carefully brushed on one side of a large rectangle of graphitic paper (Toray Industries, TGP-H-090, 0.01mm thick). The paper was dried in an oven at 110 °C for 12 hours under an active vacuum (10" torr). Electrochemical Techniques. Temperature dependent ac-impedance measurements of the gels utilized a Solartron Si 1260 Gain Phase Analyzer at frequenciesfrom1 Hz up to 3 mHz, and an AC amplitude of 5 mV. The sample holder was a standard T-cell composed of polypropylene and fitted with stainless steel rods in three swage-lock fittings. All measurements started at an initial temperature of 125 °C which was lowered in 10 °C intervals. Between measurements, the cell was thermally equilibrated for one hour. Thickness of the sample was determined by the difference between caliper measurements of the end-to-end length of the stainless steel rods with and without the polymer gel between them. The ionic conductivity was calculated from the measured resistance when the imaginary component at high frequency fell to zero. In order to determine the effects of high voltage charge-discharge cycling, a thick gel composed of 1.0M Li/DMBITFSI, was sandwiched between two pieces of Li metal. Repeated charging and discharging of the cell, in which the Li metal was


6

3

6

2

2

2

5

3


4

485 stripped and deposited over 500 cycles allowed for the monitoring of the change in the ionic conduction of the polymer gel as determined by similar acimpedance measurements. Electrochemical experiments were performed using either an EG&G PAR 273A or 263A Potentiostat/Galvanostat with the M270 ver. 4.30 software. A specially designed flat cell was used for all of the electrochemical measurements. In this cell, the working and counter electrodes were simply V£ diameter cutout discs of either graphitic paper, polymer coated graphitic paper, or L i metal. These were separated using two 14" cutout discs of the polymer gel composite electrolyte. For measurements involving a pseudo-reference electrode, a thin, flattened piece of Pt wire was inserted between the two discs of the polymer gel. This thin, solid-state cell was placed between the two Pt current collectors and the entire cell was clamped with a spring-loaded. For the two electrode measurements, the Pt-wire was simply removed from the assembled cell.


M

Results and Discussion Ionic Conductivity Results. Figures 2 and 3 illustrate the downward shift in ionic conductivity when 1.0 M LiTFSI is added to the ionic liquid, and then after the lithium containing ionic liquid is entrapped in the polymer gel composite. Additionally, Table 1 lists various details of the measured ionic conductivity data for these ionic liquid systems at room temperature. Physically, these polymer gel composites behave as simple, microporous sponges. Although adding LiTFSI results in a drop of approximately 2 mS/cm in the room temperature ionic conductivity, a drop of only 1 mS/cm results from binding them in the polymer composite. It is possible to prepare gels with an even higher percentage of ionic liquid, which results in an even smaller loss in ionic conductivity. However, these gels tear more easily and are also more subject to weeping, or loss of the ionic liquid, as pressure is applied to the cell by the spring loaded clamps. Therefore, the 75% ionic liquid - 25% polymer composite was chosen as the best compromise between ionic conductivity and physical resilience. This drop in ionic conductivity is relatively inconsequential when compared to the ionic conductivity of 0.0005 mS/cm for pure Li-PEO gels, since all of the observed ionic conductivity values of these gels fall within the 25 mS/cm range. Of greater importance will be determination of the L i ion transport properties in these gels, and current NMR experiments are underway to determine to what degree the L i ion, the imidazolium ring, and the TFSF are responsible for the observed conductivities. 16

+



486

2.5

2.7

2.9

3.1

3.3

3.5

3.7

3.9

1000/τ °κ Figure 2. Temperature Dependent Ionic Conductivity of A) Pure DMPITFSI, B) l.OMLi/DMPITFSI, C) Polymer Gel of 75% 1.0 MLi/DMPITFSI and 25% PVdF-HFP

2.5

2.7

2.9

3.1

3.3

3.5

3.7

3.9

1000/T °K Figure 3. Temperature Dependent Ionic Conductivity of A) Pure DMBITFSI, B) 1.0 M Li/DMBITFSI, C) Polymer Gel of 75% 1.0 M Li/DMBITFSI and 25%PVdF-HFP


487 Table 1. Ionic Conductivity Data (± 0.05 mS/cm) at 22 °C of the various DMRITFSI Ionic Liquid Systems As a 75% Polymer Gel Composite WithLt Without LC 3.89 2.88 3.17 2.86 2.64 3.54

Ionic Pure Ionic Liquid Conductivity in mS/cm Without LC WithLt 5.31 3.87 DMPITFSI 4.44 DMBITFSI 3.65 BMIPF 5.28 3.47 Downloaded by UNIV OF GUELPH LIBRARY on July 18, 2012 | http://pubs.acs.org Publication Date: August 26, 2003 | doi: 10.1021/bk-2003-0856.ch038

6

Finally, a detailed analysis of the stability of the polymer gel against metallic Li was studied by sandwiching a gel composed of 1.0M Li/DMBITFSI between two discs of Li metal. The reaction between the polymer gel and Li metal will result in the formation of a solid electrolyte interface at both L i electrodes. This cell was cycled 500 times using a constant current of 250 μΑ/cm to strip-off and deposit Li. Caliper measurements gave the gel thickness as 1.65 mm thick and initial ac-impedance measurements of the cell indicated that the ionic conductivity was 2.71 mS/cm (as determined by the point at which the imaginary component of the conductivity fell to zero at highfrequency).As the cell was cycled, periodic measurements of the ac-impedance of the cell were performed. As shown infigure4, after a rapid drop in the ionic conductivity of approximately 20%, a constant value of 2.2 mS/cm was observed. Thus indicating that although the layer between the gel and the Li metal electrode was somewhat less conducting than the gel itself, it did not passivate the Li surface rendering it inaccessible to removal and deposition processes. 2

2.00 0

100

200

300

400

500

# of Cycles Figure 4. Ionic Conductivity dependency for the 75% 1.0 MLi/DMBITFSI Polymer gel between 2 discs ofLi metal. Charging and discharging current was 50 μΑ/cm 2


488 Electrochemical Characterizations of the Polymer Gel Electrolytes. The cyclic voltammogram of the Li/DMPITFSI gel electrolyte, figure 5 curve C, using a disc of graphitic paper as the counter and a disc of Pt as the working electrode indicated an electrochemical window of approximately 4.85 volts (a similar cyclic voltammogram of the Li/DMBITFSI gel was also observed but not shown since they were nearly identical.) Also shown are the cyclic voltammograms of the polymer gel composites between two discs of graphitic carbon. The observed intercalation and deintercalation of the cation and the anion indicates that a type of DIME cell can be formed in which the L i , the DMI?r cation, and the TFSI* anion intercalate into graphitic paper. Downloaded by UNIV OF GUELPH LIBRARY on July 18, 2012 | http://pubs.acs.org Publication Date: August 26, 2003 | doi: 10.1021/bk-2003-0856.ch038

+

1

1

1

1

1

1

1

1

1

1

1

-0.012 ' • ' ' ' 2.8 1.8 0.8 -0.3

1

1

1

1

1

1

1

1

-1.3

1

1

-2.3

1

1 1

-3.3

-4.3

E(V) vs Pt Wire Figure 5. Cyclic Voltammograms ofA) the gel containing 1.0 M Li/DMPITFSI, B) the gel containing 1.0 M Li/DMBITFSI, using two discs of Graphitic Paper as the Working and Counter Electrodes and C) a gel containing 1.0 M Li/DMPITFSI using a disc of Pt Foil as the working electrode and a disc of graphitic paper as the counter. The pseudo-reference was Pt wire, and the scan rate was 10 m V/sec.

Thus, by removing the Pt wire, a simple two-electrode battery could be prepared, and the charge-discharge behavior of this simple system is shown infigure6. The resulting 1.0 volt battery exhibits only 65% efficiency, and does show significant voltage drop-off as the system discharges. Previous work has shown


489 that the cationic intercalation and deintercalation exhibits over 80% efficiencies when measured in a 3-electrode cell, while only 60% efficiency or less is measured for the TFSI" anion. Thus, the inherent drawback in the DIME system is that the anion charge-discharge efficiencies appear to be the limiting agent.


21

0.0 I

ι

ι

0

ι

ι

200

400

ι

ι

I

ι

600

800

T(s) 2

2

Figure 6. A) Charge (1 mA/cm ) and B) Discharge (0.25 mA/cm ) for two discs ofgraphitic paper and the 75% 1.0M Li/DMPITFSI Polymer Gel Electrolyte L i Metal-Metal Oxide Battery Test Cells. For the next series of charge-discharge experiments, polymer composites of L i M n 0 and LiCo0 were used as the cathode, discs of pure Li Metal (Aldrich 99.9%, 0.75mm thick) were the anode material, and single thin discs of the polymer composite gels served as the solid electrolyte. In these experiments, the system was first discharged for a 2 hour period, and then subjected to 50 charge-discharge cycles. As can be seen in Figure 7 and Figure 8, both behaved extremely well. For figure 7, the Li/DMPITFSI gel exhibited very little voltage drop as well as high efficiency for LiMn 0 . Similarly, for a cell using LiCo0 , Li metal and Li/DMBITFSI, Figure 8, the charging curve was nearly linear, although more of a voltage drop-off over time was observed in the discharge curve. Overall, for both systems, the voltage plateaus and efficiencies did not varyfromthe 10 cycle up to the 50 cycle, and it is the 50 cycle that is shown. Although a more detailed investigation into the behavior of the metal oxide electrodes is required, it is clear that these ionic liquid gel electrolytes can be used for conventional 4.0V Li-ion batteries. Furthermore, after the charge-discharge 2

2

4

4

2

2

th

th

th



490

1000

2000

3000

4000

T (sec) Figure 7. A.) Charging and B)Discharging at 40 μΑ/cm* For a Li metal, LiMn 0 battery. The 5(f Cycle is shown 2

4

h

3.00 1000

2000 Τ (sec)

3000

4000

Figure 8. A) Charging and B)Discharging at 40 For a Li metal, L1C0O2 battery. The 50 Cycle is shown. th


491 experiments were complete and the solid-state cells dismantled, the polymer gels were examined and found to still be flexible, transparent discs. Also, the L i metal discs showed what appeared to be a only a light tarnishing. However, as shown in figure 9, a similar experiment using V 0 and Li metal electrodes was performed using a gel composed of 75% 1.0 M Li/BMIPF and 25% PVdF-HFP or a gel containing 75% 1.0 M Li/DMBITFSI and 25% PVdF-HFP. The gel with BMIPF performed poorly by the 50 cycle, while the cell using the gel electrolyte of Li/DMBITFSI exhibited high efficiency. The degradation of the polymer gel using BMIPF can clearly be seen in figure 10, which shows the change in efficiencies over the entire 50 cycles. After these experiments, when the cell was dismantled, the polymer gel composite disc made of 1.0 M Li/BMIPF had clearly reacted and turned dark brown and the Li metal surface was covered with a yellow film, indicating that the ionic liquid had likely reacted with the L i metal possibly via the hydrogen atom at the 2-position of the imidazole ring. 2

5

6

th

6


6

6

3.80

Τ (sec) Figure 9. A) Charge and B) Discharge behavior ofV 0 vs. Li metal using a gel composed of 1.0 M Li/DMBIPF , and the C) Charge and D) Discharge behavior of V 0 vs. Li Metal using a gel composed of 1.0 MLi/BMIPF . Charging and discharging currents were 50 μΑ/cm . 50^ Cycle is shown. 2

5

6

2

5

6

2


492


100

10 : ο

ι

0

—

.

—

10

—

,

—

,

—

ι

—

20 30 # o f Cycles

i

—

ι

—

40

,

—

I

50

Figure 10. Percent discharge efficiencies vs. # of cycles for cells ofLi metal and V 0 using a A.) Polymer gel composite electrolyte of 75% 1.0 MLi/BMIPF and 25% PVdF-HFP or B.JPolymer gel composite electrolyte of 75% 1.0 M Li/DMBITFSI and 25% PVdF-HFP. 2

5

6

Conclusions The above experiments provide strong evidence that it is entirely possible and feasible to use the DMitlTFSI ionic liquids as electrolytes in L i ion batteries. Although DIME cells did not perform as well as has been reported for ionic liquids using the AlCU" anion, the Li/DMJNTFSI ionic liquids performed as well as any of the more common electrolytes using L i metal and three of the most common metal oxide based cathode materials. Since cycling over 500 times indicated that extensive degradation of the polymer gels did not result in the build up of a non-conductive film on the L i metal, it seems likely a rechargeable Li-metal battery may be possible using these nonflammable polymer gel electrolytes Further work is now needed to answer several important questions. First and foremost, the primary current carrier needs to be determined, and NMR experiments are currently underway to determine the transport properties of these and other ionic liquid composite gel electrolytes. Secondly, the film that


493


did form over the L i metal during the 500 cycle experiment needs to be well characterized, since this could pinpoint specific weakness of the ionic liquid that could be addressed by further chemical manipulation of either die cation or the anion. Additionally, the actual capacity of the metal oxide cathodes need to be determined to see if any type of secondary limiting interactions occur between the ionic liquid and the metal oxides. However, it seems clear that the use of ionic liquids, with their superior environmental and physical properties are viable alternatives to the other, volatile andflammableorganic electrolytes.

Acknowledgments This work was sponsored by the A i r Force Office o f Scientific Research, and space was provided by the Naval Research Laboratory. Opinions, interpretations, conclusions and recommendations are those o f the authors and are not necessarily endorsed by the United States A i r Force.

References th

1. F. A. Cotton and G. Wilkinson, Inorganic Chemistry 4 ed. p.385, 1980. 2. R. T. Carlin, J. Fuller, W. K. Kuhn, M . J. Lysaght, P. C. Trulove, J. Applied Electrochem. 26 (11) 1147, (1996). 3. R. T. Carlin, H. C. DeLong, J. Fuller, and P. C. Trulove, J. Electrochem . Soc. 141, L73, (1994). 4. T. E. Sutto, P. C. Trulove, H. C. DeLong, in the Proceeding of the 12 International Conference on Molten Salts, P. C. Trulove, H. C. De Long, G. R. Stafford, S. Deki, Editors PV 99-41, p. 32, The Electrochemical Society, Inc., Pennington, NJ (2000). 5. T. E. Sutto, P. C. Trulove, H. C. DeLong, in the Proceeding of the 12 International Conference on Molten Salts, P.C. Trulove, H. C. De Long, G. R. Stafford, S. Deki, Editors PV 99-41, p. 43, The Electrochemical Society, Inc., Pennington, NJ (2000). 6. T. E. Sutto, K. D. Sienerth, H. C. DeLong and P. C. Trulove, in the Proceedings of the 12 International Conference on Molten Salts, P.C. Trulove, H. C. De Long, G. R. Stafford, S. Deki, Editors PV 99-41, p. 54, The Electrochemical Society, Inc., Pennington, NJ (2000). 7. T. E. Sutto, P. C. Trulove, H. C. DeLong, in Proceeding of the EUCHEM Conference on Molten Salts, R. W. Berg and H. A. Hjuler, Editors; Elsevier, Paris, p. 511 (2000). th

th

th


494 8. T. E. Sutto, H. C. DeLong, and P. C. Trulove, Zeitschrift für Naturforschung A, in press. 9. T. E. Sutto, H. C. DeLong, and P. C. Trulove, Zeitschrift für Naturforschung A, accepted. 10. T. E. Sutto, H. C. DeLong and P. C. Trulove, in the Proceedings of the 13 International Conference on Molten Salts, The Electrochemical Society, Inc., Pennington, NJ (2002), in press. 11. R. T. Carlin, H. C. De Long, J. Fuller, and P. C. Trulove, J. Electrochem. Soc. 141, L73, (1994). 12. J. Fuller, A. C. Breda, and R. T. Carlin, J. Electrochem . Soc. 144, p. L68, (1997). 13. C. L. Hussey in Chemistry of Nonaqueous Solvents. A. Popov and G. Mamantov, Editors. Chap. 4, V C H Publishers, New York (1994). 14. M . Watanbe, M . Kanda, H. Matsuda, K. Tsumemi, E. Tsuchida and I. Shinohara, Macromol. Chem. Rapid Commun., 2, p. 741 (1981). 15. M . Alamgir and Κ. M . Abraham, "Room Temperature Polymer Electrolytes", Lithium Batteries, Ed. G. Pistoia, Industrial Chemistry Library, Vol. 5, (1994). 16. O. Bohnke, C. Rousselot, P. A. Gillet and C. Truche, J. Electrochem. Soc., 139, p. 1862 (1992). 17. F. Lemaitre-Auger and J. Prud'homme, Electorchem. Acta, 46, Issue 9, pg. 1359 (2000). 18. K. Hayamizu, Y. Aihara, S. Arai and C. Garcia-Martinez, J. Phys. Chem., 103, p. 519 (1999). 19. S. Narang, S. Ventura, B. Dougherty, M . Zhao, S. Smedley and G. Koolpe; SRI International, U. S. Patent 5,830,600, November 3, (1998). 20. S. C. Ventura, S. C. Narang, G. Hum, P. Liu, P. Ranganathan, and L. Sun; SRI International, U. S. Patent 5,731,104, March 24, (1998). 21. T. E. Sutto, H. DeLong and P. C. Trulove, in the 13 International Conference on Molten Salts, The Electrochemical Society, Inc., Pennington, NJ (2002), in press. 22. J. Fuller, A. C. Breda, and R. T. Carlin, J. Electrochem . Soc. 144, L68, (1997). 23. J. Fuller, A. C. Breda, and R. T. Carlin, J. Electroanalytical Chem. 459, 29 (1998).


th

th


Ionic Liquids as Green Solvents - American Chemical Society

Recommend Documents