Post on 09-Jan-2020
1
Investigations of Ionic Liquid Electrolytes for Li-Air Batteries
A thesis presented
by
Jaehee Hwang
to
the Department of Chemistry and Chemical Biology
In partial fulfillment of the requirements for the degree of
Master of Science
in the field of
Chemistry
Northeastern University
Boston, Massachusetts
January, 2012
2
Investigations of Ionic Liquid Electrolytes for Li-Air Batteries
A thesis presented
by
Jaehee Hwang
ABSTRACT OF THESIS
Submitted in partial fulfillment of the requirements
for the degree of Master of Science
in the Graduate School of Northeastern University
Northeastern University
Boston, Massachusetts
January, 2012
3
Abstract
Ionic liquids are molten salts at room temperature that possess many unique properties1. There has
been continued increase of interest in ionic liquids because their properties make them attractive
for many applications including electrocatalysts, electrodeposition, biosensors, and electrolytes
for ionic and electronic devices, including batteries and capacitors2. Their low volatility, high
ionic conductivity, and large electrochemical window make them ideal candidates for next
generation electrolytes for primary and secondary lithium batteries. The lithium-O2 battery, with a
theoretical specific energy of 5200 Wh/Kg3, ten times as high as Li-ion technology, faces many
challenges in development before it can become a practically viable power source. The role of
electrolyte is crucial in this system and in particular it is necessary to understand reversible O2
electrochemistry in the electrolyte. Therefore, we have been investigating a series of advanced
non-aqueous electrolytes for the Li-O2 battery in our Laboratory at Northeastern University
Center for Renewable Energy Technology. Our recent studies on the influence of non-aqueous
solvents on the electrochemistry of oxygen in the rechargeable lithium-air battery showed that the
solvent and the supporting electrolyte cations in the electrolyte act in concert to influence the
nature of reduction products and the rechargeability4.
In the first part of this thesis, Nuclear Magnetic Resonance (NMR) spectroscopy was employed to
determine the relative Lewis acidity of ionic salt cations including room temperature ionic liquids,
and the Lewis basicity of organic solvents. The Varian 400MHz and 500MHz NMR instrument at
Northeastern was used for conducting 13C NMR and 7Li NMR chemical shift studies respectively.
The 13C NMR provided a linear plot of chemical shift (ppm) versus the Lewis acidity of cations of
the selected salts including ionic liquids dissolved in the non-aqueous organic electrolyte,
propylene carbonate. From this a Lewis acidity scale for Li salt cations and two ionic liquids has
been developed. The 7Li NMR study resulted in no correlation between the chemical shifts and
Gutmann’s donor number, a measure of the Lewis basicity, of various solvents selected. The
4 challenges of 7Li NMR technique including solvent effect and choosing a proper reference
material are presented and several experimental approaches taken in attempt to overcome those
challenges are discussed. 13C NMR longitudinal relaxation time measurement by inversion
recovery method was further conducted with Varian 500 MHz NMR and its preliminary results
are reported.
In the second investigation we synthesized and fully characterized 1-ethyl-3-methyl-imidazolium
(EMITFSI) Ionic Liquid and studied the electrochemistry of O2 on this electrolyte5. In this thesis,
the synthesis of 1-butyl-1-methyl-pyrrolidinium bis-(triflouromethanesulfonyl) imide
(PYR14TFSI) ionic liquid6 and the discharge profile of the Lithium- air pouch cell utilizing these
ionic liquids as electrolytes are introduced. The battery discharge capacities obtained in these
lithium air pouch cells will also be discussed.
5 Acknowledgements
I would like to first thank Prof. K.M Abraham, for providing me great support and advice
throughout my Master of Science degree research. His passion, enthusiasm and curiosity on
imperative scientific topics on batteries has inspired and motivated me to conduct the research
that I have. It is my honor to have him, a very well acknowledged and influential individual in the
battery community, as my research advisor and closely work with. I am always grateful for this
opportunity and it has been a wonderful learning experience.
I thank Prof. Sanjeev Mukerjee for his patience and understanding his students’ need. I am
thankful for his willingness and time to listen to my concern with any issue related to my
University program as well as the post-graduate plans and providing me his thoughts on what
would be the best for me. He has let me join his laboratory for my senior’s research, and his
laboratory has become my favorite place and time spent on Campus during my 5years of BS-MS
program for not only as a learning place, but also the place I have developed a professional and
personal relationship with many great scientists; Mathew Trahan, Chris Allen, Kara Strickland,
Mehmet Nurullah Ates, Aditi Halder, Braja Ghosh, Christopher Allen, Daniell Abbott, Iromie
Arunika Gunasekara , Michael Bates, Mohsin Ehteshami, Myoung Seok Lee, Nagappan
Ramaswamy, Qingying Jia, Tetiana Bairachnaya, Urszula Latosiewicz. Thank you to everyone
for their generosity since I joined Prof. Mukerjee’s lab for the encouragement and support they
have provided for me to complete my Master of Science degree successfully.
I would especially like to thank Chris Allen, a 5th year Ph.D student, whom I have worked
with closely and whom I have learned a great deal from during my Masters research. I could have
not asked for a better mentor, and I appreciate his excellent work ethic which motivates me to
continue in this field.
I would also like to thank Dr. Roger Khauz for his advice on experimental design and
discussion on NMR studies that I have conducted, making this 7Li NMR study possible with
6 Varian 500 at Northeastern University. Despite his busy demanding schedule, he has always taken
the time to teach me the fundamentals behind these experiments and accommodated my need to
use the instrument. I am also thankful for Chemistry Department at Northeastern University for
the lab-based experimental learning experience that I have obtained that would help me to further
my career.
Lastly, I want to express my thanks to my family in Seoul, South Korea, for their entire
support and love they have provided me. Although I am not physically near them, I always feel
their care and love for me, and that gives me the confidence and strength to become more
independent in achieving my educational and professional goals. I am always grateful for their
guidance and belief in me and words cannot describe how much I appreciate the role that they
have played.
Partial financial support for this work was provided by US Army CERDEC through
contract ARMY-PT GTS-S-09-1-057.
7
Table of Contents
Abstract 2
Acknowledgements 5
Table of Contents 7
List of Tables and Figures 9
List of Abbreviations and Symbols 10
CHAPTER1: INTRODUCTION AND BACKGROUND 11
1.1 Introduction to non-aqueous electrolytes 11
1.2 Ionic Liquids as solvents for electrochemistry 11
1.3 Ionic liquids as Li battery electrolytes 13
1.4 Nuclear Magnetic Resonance (NMR) spectroscopy to study non-aqueous organic
electrolytes and ionic liquids 14
1.5 Research Objective 15
CHAPTER 2: EXPERIMENTAL METHODS 16
2.1 General Methods 16
2.2 13C NMR measurements including spectrometer and sample preparation 17
2.3 13C NMR Longitudinal Relaxation Time measurement by Inversion Recovery
Method 18
2.4 7Li NMR measurements including spectrometer and sample preparation 18
CHAPTER 3: RESULTS AND DISCUSSION 20
3.1 Preparation and characterization of ionic liquids 20
3.1.1 Synthesis of 1-butyl-1-methyl-pyrrolidinium
bis-(trifluoromethanesulfonyl)imide 20
3.1.2 Characterization of PYR14TFSI 22
3.2 13C NMR analysis on organic electrolytes and Ionic Liquids 24
3.2.1 Correlation between the 13C NMR chemical shifts and acidity of the cation
8 in the electrolytes 24
3.3 Preliminary results of 13C NMR Relaxation T1 measurement 27
3.4 7Li NMR analysis on organic) electrolytes and Ionic Liquids 30
3.4.1 Correlation between 7Li NMR chemical shifts and electrolyte solvent basicity
(as measured by the Gutmann’s Donor Number of solvents) 30
3.5 Lithium air pouch cell fabrication and discharge profile 34
CHAPTER 4: CONCLUSIONS AND FUTURE DIRECTION 37
REFERENCE 38
9
List of Tables and Figures
Chapter 2
Table 2.1. List of salts employed for 13C NMR study and the ionic radius of cations of those salts
Table 2.2. List of selected solvents for 7Li NMR study and the donor number of those solvents
Figure 2.1. NMR Sample Setup (External Referencing)
Chapter 3
Figure 3.1. Picture of purified neat PYR14TFSI Ionic Liquid
Figure 3.2. Proton NMR of neat PYR14TFSI
Figure 3.3. Cyclic voltammogram of PYR14TFSI Oxygen Reduction Reaction (ORR) on glassy
carbon electrode
Figure 3.4 (a)-(b). Composite 13C NMR spectra of selected salts consisting single charged cations
of different sizes dissolved in propylene carbonate
Figure 3.5. Plot of 13C chemical shift on C=O verses ionic radii of cation of each salt dissolved in
Propylene Carbonate organic electrolyte
Figure 3.6. Spectra from an inversion recovery experiment to determine the T1’s for (a) neat
propylene carbonate, (b) propylene carbonate/ 1M LiPF6, (c) propylene carbonate/ 1M EMITFSI,
(d) propylene carbonate/ 1M PYR14TFSI
Figure 3.7. Plots of T1’s (sec) of (a) C=O, (b) C-H, (c) C-H2, (d) C-H3 versus the ionic radius
(pm)
Figure 3.8. Plot of chemical shift verses donor number of selected solvents used in the 7Li NMR
study
Figure 3.9. Proton spectra two TMS peaks
Figure 3.10. Caged lithium ion served as an internal standard, Solvent: Acetone d-6
Figure 3.11. Picture of assembled Li-Air pouch cell
Figure 3.12. Li-Air pouch cell full discharge profile (electrolyte: EMITFSI/0.5M LiTFSI)
10
List of Abbreviation and symbols
RTIL: Room Temperature Ionic Liquid
NMR: Nuclear Magnetic Resonance
13C NMR: 13Carbon Nuclear Magnetic Resonance
7Li NMR: 7Lithium Nuclear Magnetic Resonance
T1: longitudinal relaxation time
CV: Cyclic Voltammetry
PC: Propylene Carbonate
PYR14TFSI: 1-butyl-1-methyl-pyrrolidium bis(trifluoromethanesulfonyl)imide
EMITFSI: 1-ethyl-3-methyl-imidazolium bis(trifluoromethanesulfonyl)imide
LiTFSI: Lithium bis(trifluoromethanesulfonyl)imide
δ: Chemical Shift
11 CHAPTER 1: INTRODUCTION AND BACKGROUND
1.1 Introduction to non-aqueous electrolytes
Non-aqueous electrolytes serve as the medium for the transfer of charges in the form of
ions between negative and positive electrodes in battery cells. They include solutions of alkali or
alkaline earth metal salts dissolved in a variety of non-aqueous solvents including organic and
inorganic solvents, and room temperature ionic liquids. Well known examples are the electrolytes
used in Li-ion batteries consisting of solutions of lithium hexafluophosphate ( LiPF6) in organic
carbonates, for example, ethylene carbonate (EC), propylene carbonate (PC), dimethyl carbonate
(DMC), diethyl carbonate (DEC) or mixtures of them. Ionic liquids, which themselves are liquid
salts, containing organic cations and inorganic or organic anions, are used in Li batteries as
solutions of Li salts. Salt cations in non-aqueous organic electrolytes are generally solvated by
the solvent through acid-base complex formation (or solvation), although the relative acidity of
the cations is not clearly understood especially when the salts contain large tetra-alkyl ammonium
and ionic liquid cations.
1.2 Room temperature ionic liquids (RTILs) as solvents for electrochemistry
RTILs are molten salts consisting of large asymmetrical organic cations and large charge
delocalized anions. There has been a tremendous increase of interest in ionic liquids because of
their unique properties that make them suitable for many applications including electrocatalysis,
electrodeposition, biosensors, and electrolytes for ionic and electronic devices.1. Non-volatility,
high ion concentration and conductivity, and wide electrochemical window make them the
desired advanced future electrolytes. With growing research on the development of lithium
batteries, ionic liquids have become the focus for developing safe electrolytes for both primary
and rechargeable Li batteries due to their high thermal stability, low volatitlity and good chemical
inertness in battery environments. .
12 In mid-1970s, the haloaluminate room temperature ionic liquids (RTILs) were invented.
They were prepared by mixing aluminium halides (most often the chloride) with the
corresponding halide salt of an organic cation, commonly of the alkylpyridinium or 1,3-
dialkylimidazolium and were used as solvents in electrochemistry, electrodeposition,
electropolymerisation, and as electrolytes in electrochemical devices. However, they were
extremely sensitive to atmospheric moisture, which reacts with the aluminium halide species to
produce aluminium oxide halides and protonic impurities2 requiring them to be strictly handled in
the anhydrous environment.
The non-haloaluminate RTILs were first made by incorporating imidazolium salt with
hexaflurophosphate or tetrafluoroborate. Although they are generally unreactive with water and
less sensitive to atmospheric moisture compared to haloaluminate RTILs, they still absorb some
moisture, which alters their physical and chemical properties. For example, it is reported that in a
biphasic 1-butyl-3-methylimidazolium hexafluorophosphate ([BMIM][PF 6])/ acidic aqueous
solution, (PF6-) was decomposed to phosphate (PO43-) 3. Therefore, complete drying of all
moisture is necessary prior to conducting electrochemical experiment, and having hydrophobic
anions, such as bis(trifluoromethylsunfolnyl)imide, in ILs can improve this phenomena4. Due to
their higher stability towards air and water, non-haloaluminate RTILs have been used in recent
electrochemistry studies and application.
Non-haloaluminate ionic liquids are highly polar and have good solvating ability. They
also have high thermal, chemical, and electrochemical stability as well as tunable miscibility with
water and organic solvents. There are a number of possible ways to come up with different
combination of cation and anion. The properties such as viscosity, conductivity, electrochemical
window, solubility, and solvating properties can be predicted from the structure of the RTIL, and
therefore, RTILs can be tuned for specific applications5. Some of the most popular RTILs in
scientific studies have incorporated the cations such as 1-alkyl-3-methylimidazolium, N-methyl-
13 N-alkylpyrrolidinium, and tetraalkylammonium with the anions of the tetrafluoroborate (BF4
-),
hexafluorophosphate (PF6-), and bis(trifluoromethylsulfonyl)imide
[bistriflimide, N(Tf)2-]1.
1.3: Ionic liquids as Li battery electrolytes
RTILs have many desired properties for electrolyte use in Li Battery applications
including cost reduction, increase of the energy density, and safety enhancement of large-format
lithium battery packs for the use of high-voltage, high-energy electrodes that are stable above
4.3V (vs. Li/Li+). The currently optimized electrolytes for Li-ion battery, a mixture of ethylene
carbonate (EC) and a linear carbonate such as diethyl carbonate (DEC) with LiPF6 being used in
small Li-ion batteries, are not sufficient for high-voltage electrode materials due to their thermal
instability6. During the first stage of thermal runaway of a Lithium battery with an organic
electrolyte, the exothermic reaction decomposes the electrolyte and accelerates the rising heat
through the electrolyte-electrode reaction. With volatile solvents, such as propylene carbonate,
dimethyl carbonate, and dimethoxyethane, the inner pressure of the battery can cause case
expansion leading to explosion. Therefore, the high decomposition temperature of RTILs is
attractive to build safer batteries. Some other noble properties of RTILs as electrolytes include
low toxicity, material compatibility, non-flammability, liquidity in a wide temperature range, low
vapor pressure, good thermal and chemical stability, good ionic conductivity, and relatively large
electrochemical window1.
There has been a focus on studying electrolytes comprised of solutions of lithium salts
dissolved in ionic liquids for Li battery applications. Among the many studies, a few are
mentioned below as they are related to the study of PYR14TFSI as a part of this MS research,
described in a later chapter.
Matsumoto created and investigated ILs with aliphatic and asymmetric quaternary
ammonium cations, which were found to have the most cathodic stability. They also showed
14 reversibility by plating/ stripping of the Li metal. Combining with TFSI anion, three different
RTILs, EMI-TFSI, PP13-TFSI, and TMPA-TFSI, were tested for their performance as
electrolytes in Li/LiCoO2 cell. The charge-discharge properties showed that that PP13-TFSI was
cathodically the most stable and delivered the best performance in the cell. This is testimony to
the fact that cathodic stability is a minimum requirement for the acceptable battery’s
performance1,7.
Miura et al. have studied the oxygen reduction in some hydrophobic RTILs including
trimethyl-n-hexylammonium(TMHA)-bis(trifluoromethanesulfonyl)imide(TFSI), 1-butyl-1-
methylpyrrolidinium bromide(BMP)-TFSI, 1-ethyl-3-methyl imidazolium(EMI)-TFSI, and 1,2-
dimethyl-3-propylimidazolium(DMPA)-TFSI. They combined their electrochemistry with
computational work using Gaussian98 to explain why the superoxide is not stable in EMI+ and
DMPI+ containing ILs; mulliken atomic charges atoms in TMHA+ and BMP+ are negative verses
the carbon atoms at 2-, 4-, and 5- positions in the heterocycle of EMI+ and DMPI+ indicate
positive charges. They proposed that because O2�- is highly nucleophilic, it would react with
imdidazolium cations, such as EMI+ and DMPI+ leading to the degradation of the melts8.
1.4 Nuclear Magnetic Resonance (NMR) spectroscopy to study non-aqueous organic
electrolytes and ionic liquids
The understanding of solute-solvent (solvation) interactions is critical in understanding the
properties of electrolytes in solvent media and therefore there has been many theoretical
approaches and spectroscopic methods used to study the solvation. As a spectroscopic method,
Nuclear Magnetic Resonance (NMR) has been a powerful technique that makes it possible to
monitor the nuclei of dissolved electrolyte or the nuclei residing in solvent molecules separately9
without destructing the sample. NMR exploits the behavior of certain magnetically active atoms
when they are placed in a very strong magnetic field.
15 13C NMR using Varian 400MHz was employed in this research to develop the relative
acidity scale of cations in nonaqueous electrolytes by plotting the chemical shift of carbon nuclear
magnetic resonance of carbonyl carbon as a function of ionic radius of the cations. Various
radiofrequency pulse techniques were further implemented to manipulate the magnetization for
longitudinal relaxation time (T1) measurement to better understand the local environments of
nuclei and characterize the motion of different molecular groups in solutions of propylene
carbonate/ 1M salts (Chapter 2.2).
7Li using Varian 500MHz was employed to develop the relative basicity scale of solvent
by plotting the 7Li chemical shift as a function of Gutmann’s donor number.
1.5 Research Objective
The first objective was to successfully synthesize 1-butyl-1-methyl-pyrrolidium
bis(trifluoromethanesulfonyl)imide for electrochemical and NMR studies to be conducted.
The recent study of non-aqueous electrolytes on electrochemistry of oxygen reduction
reaction in the rechargeable lithium air battery has shown that the acidity of cation in supporting
electrolytes and the basicity of solvent influence the discharge products and rechargeability of the
Li-air cell. Related to this was the second objective of this study to employ nuclear magnetic
resonance spectroscopic technique to develop a relative acidity scale of alkali metal, tetraalkyl
ammonium, and ionic liquid cations by observing the 13C chemical shifts of the solvents in which
they are dissolved. The preliminary results on 13C T1 measurement which complement the
chemical shift data are also presented.
16 CHAPTER 2: EXPERIMENTAL METHODS
2.1 General Methods
General NMR sample preparation for all NMR samples employing external referencing in 13C
and 7Li NMR studies were prepared using three components: 1)Wilmad NMR tubes (100mHz,
Economy) with the dimension of OD: 5.0mm, ID: 4.1mm, Length: 20.32cm, 2) PYREX®
capillary melting point tubes with the dimension of ID: 0.8-1.1 mm, Length: 90mm, and 3)
Masterflex® Peroxide-Cured Silicone to hold the capillary melting point tube in place. The
Masterflex® silicone tubing used has the fits snugly inside the Wilmad NMR tube and has a hole
of the size of the melting point capillary outer diameter in the center. It was used to hold the inner
capillary tube in the center of the outer tube to prevent displacement of the capillary during the
spinning and the reduction of homogeneity which would cause line broadening and side bands in
the NMR spectra.
The prepared NMR samples were brought out after they were capped and sealed with Teflon®
tape. They were put in a sealed plastic bag until the NMR spectra were collected.
Figure 2.1: NMR sample setup (External Referencing)
17 Cyclic voltammetry measurement for the synthesized ionic liquid was performed on an Autolab
potentiostat (Eco Chemie B.V.) equipped with a frequency response analyzer for iR correction. A
conventional three-electrode setup was used with a 6mm diameter glassy carbon working
electrode along with a platinum counter electrode. The reference electrode solution contained in a
glass tube was placed close to the working electrode, separated with a Vycor frit. Potentials were
converted to the Li/Li+ reference electrode by measuring the potential of Ag/Ag+ electrode
against a Li foil. Gas purging of the IL solution include one hour of argon gas followed by one
hour of oxygen. The electrochemistry was carried out in a controlled atmosphere glovebox.
Arbin multi-channel battery testing cycler was employed for testing the Lithium air pouch cell.
The current rate of 0.1mA/cm2 was applied for each battery discharge.
2.2 13C NMR sample preparation
The selected organic electrolyte, solvent to dissolve various salts, is propylene carbonate
(anhydrous, 99.7%, Sigma-Aldrich). Solutes with different cations include lithium
hexafluorophosphate (Purolyte), lithium bis(trifluoromethanesulfonyl)imide (Purolyte), sodium
hexafluorophosphate (98%, Sigma-Aldrich), potassium hexafluorophosphate (98%, Sigma-
Aldrich), tetrabutylammonium hexafluorophosphate (Fluka), 1-Ethyl-3-methylimidazolium bis(
trifluoromethylsulfonyl)imide (synthesized in Northeastern University laboratory for
electrochemical advanced power (NULEAP)), and 1-Butyl-1-methylpyrrolidinium bis(
trifluoromethylsulfonyl)imide (synthesized in NULEAP). The concentration of solutions is 1M.
General NMR sample preparation has been employed with external reference, deuterated acetone
with TMS (1% v/v).
Varian 400MHz instrument has been employed and the magnetic field was locked on acetone
with lock power of 20. Each time the NMR sample was inserted, the instrument was locked and
shimmed with spin on. Table 2.1 shows the solutes employed with the ionic radius of each
cations. Each sample was scanned until the carbonyl carbon in acetone solvent peak clearly
appeared to set as a reference (approximately 2500 scans).
18
Table 2.1: Selected solutes with various cations for 13C NMR study. i. Laoire et al.10, ii.
Fawcett, W.R.; Ryan, P.J.11, iii. G.B. Appetecchi et al.12
2.3 13C Longitudinal Relaxation Time (T1) Measurements by Inversion Recovery Method
Identical sample preparation was employed for all 13C NMR samples as described using the exact
same set of the chemicals and components (Table 2.1). The inversion-recovery T1 measurement
was employed using Varian 500. Each time the NMR sample was inserted, the instrument was
locked on acetone d-6 placed in the capillary tube, and shimmed with spinner on. The 90º pulse-
width was calibrated on each sample before measuring T1 and eight acquisition arrays were used.
Two scans were taken for each sample.
2.4 7Li NMR Measurements
The selected solvents used for 7Li NMR study include dimethyl sulfoxide (Sigma-Aldrich),
propylene carbonate (anhydrous, 99.7%, Sigma-Aldrich), 1, 2-dimethoxyethane (anhydrous,
99.5%, Sigma-Aldrich), tetraethyleneglycol dimethyl ether (Fluka), acetonitrile (anhydrous,
99.8%, Alfa Aesar), 1-ethyl-3-methylimidazolium bis( trifluoromethylsulfonyl)imide (synthesized
in NULEAP), and 1-butyl-1-methylpyrrolidinium bis( trifluoromethylsulfonyl)imide (synthesized
in NULEAP).
Salts Cation Ionic Radius
(pm) Lithium Bis(trifluoromethane)sulfonimide (LiTFSI) Li 68 i Lithium hexafluorophosphate (LiPF6) Li 68 i Sodium hexafluorophosphate (NaPF6) Na 98 i Potassium hexafluorophosphate (KPF6) K 133 i Tetrabutylammonium hexafluorophosphate (TBAPF6) TBA 494 i 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide (EMITFSI) EMI 239 ii 1-butyl-1-methyl-pyrrolidinium bis(trifluoromethanesulfonyl)imide ) (PYR14TFSI) PYR 330 iii
19 The internal reference for the 7Li measurements was lithium ion caged in Cryptan 211. First,
0.5M Lithium bis(trifluoromethanesulfonyl)imide salt solutions were prepared with various
solvents and 0.5 molar ratio of Cryptan 211 to Lithium ion was added to the solutions caging half
of lithium ions in the cavity of the Cryptan 211 molecules, which served as the internal reference.
About 700ul of each prepared solutions were transferred into a Wilmad NMR tube with outer
diameter of 5mm. All sample preparation was conducted in the Argon glove box with moisture
level (<1-2ppm).
Varian 500MHz instrument has been employed. Table 2.2 shows the selected solvents with their
donor number. Single scan was taken.
Solvent Donor Number
(kcal/mol) Dimethyl Sulfoxide (DMSO) 29.8 i 1,2-Dimethoxy Ethane/ Proplyene Carbonate (1:1 v/v) (PC:DME) 17.55 * Acetone-d6, TMS (1% v/v) - external standard (AC) 17 ii Tetraethylene glycol dimethyl ether (TEGDME) 16.6 i Propylene Carbonate (PC) 15.1 ii Acetonitrile (ACN) 14.1 i 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide (EMITFSI) 1,4-butyl methylpyrrolidinium bis(trifluoromethanesulfonyl)imide ) (PYR14TFSI)
Table 2.2: Selected solvents possessing different donor numbers. i. Chemistry of Nonaqueous
Solutions13, ii. R.H. Erlich et al.14, *estimated
20 Chapter 3: Results and Discussion
3.1 Preparation and characterization of ionic liquids
3.1.1 Synthesis of 1-Butyl-1-Methyl-Pyrrolidinium bis(trifluoromethanesulfonyl)imide
(PYR14TFSI)
1-methylpyrrolidine, 98% (Alfa Aesar), 1-Iodobutane, 99%, stab. with copper (Alfa Aesar),
lithium bis-(trifluoromethanesulfonyl)imide (LiTFSI), 98+% (Alfa Aesar), ethyl acetate, HPLC
Grade, 99.5+% (Alfa Aesar), decolorizing activated carbon (Sigma-Aldrich), aluminum oxide
activated, neutral, Brockmann I (Sigma-Aldrich) were used as received. H2O was deionized by a
Millipore ion resin exchange deionizer. Typically, 100g batches of PYR14TFSI were produced. A
500mL round bottom flask with rubber stopper was used throughout to produce the ionic liquid
and the temperature was controlled using a temperature controlled oil bath. The ethyl acetate was
evaporated using a Heidolph Laborota-4003 rotary evaporator connected with a vacuum pump
(30mTorr). The final drying step of the purified ionic liquid employed Sargent-Welch Scientific
Company 1402-high vacuum pump (0.1mTorr).
A hydrophobic ionic liquid synthesis proposed by Appetecchi, G. B. et. Al 15 was adopted to
synthesize PYR14TFSI. It only involves two solvents, ethyl acetate and water, throughout the
entire synthesis and produces high over yield above 86%. The synthesis was performed in three
steps: (i) synthesis of the precursor, 1-butyl-1-methyl-pyrrolidinium iodide (PYR14I) (ii) synthesis
of PYR14TFSI, and (iii) purification of PYR14TFSI.
The first step in synthesis of the precursor PYR14I was performed by mixing, 1-methylpyrrolidine,
PYR1 (diluted with ethyl acetate), and 1-iodobutane (diluted with ethyl acetate) and stirring the
mixture at 45˚C for at least 7 hours under Argon.
PYR1(liquid) + 1-Iodobutane(liquid) → PYR14I (white solid)
In order to prevent the oxidation of iodide to iodine, filtration of this white solid precursor was
conducted as quickly as possible, avoiding exposure to light. Residual solvent was removed by
21 transferring the filtered solid precursor into glassware covered with aluminum foil and drying in a
vacuum oven at 110˚C overnight. This provided a higher yield of the precursor, which helped to
accurately determine how much LiTFSI salt to be used for the second step. Using an excess of
LiTFSI is not only costly, but also decreases the overall yield in the second reaction.
The second reaction, synthesis of the PYR14TFSI ionic liquid involved mixing the lithium salt
dissolved in deionized water (3.5M) and the precursor obtained from the first reaction. The
LiTFSI solution was first deaerated by argon bubbling and was transferred to a reactor that had
been flushed with argon. The reagents were mixed at room temperature and were stirred
vigorously for more than 3 hours before the phase separation was performed. The anion exchange
is driven by hydrophobic nature of products and lithium iodide ends up being in the separate
aqueous phase upon completion of the reaction. The liquid-liquid extraction was conducted to
obtain crude PYR14TFSI. The ionic liquid phase was washed at least 5 times with water to reduce
the lithium content.
PYR14I (solid) + LiTFSI (aqueous) → PYR14TFSI(liquid) + LiI (aqueous)
Finally, purification of crude ionic liquid, PYR14TFSI, was conducted by diluting the ionic liquid
with ethyl acetate and mixing with first, decolorizing activated carbon followed by activated
alumina. The mixture of activated carbon and ionic liquid was stirred at 70-75˚C for at least 10
hours. The carbon powder was then vacuum filtered and the activated alumina was then added
into the filtered ionic liquid. This mixture was stirred at room temperature for at least 7 hours to
remove the protic impurities. After 7 hours, the alumina was separated from the liquid phase by
vacuum filtration. The filtrate containing the product was connected to a rotary evaporator at
45°C under vacuum (30mTorr) and the temperature was raised to 100°C to remove the ethyl
acetate. The neat PYR14TFSI ionic liquid was dried under high vacuum (0.1mTorr) at 80°C for t
least 12 hours and then at 120°C for at least 24 hours.
The dried neat ionic liquid produced was transported and kept in the Argon glovebox (<3 ppm
water). The Figure 3.1 shows the purified neat ionic liquid synthesized as a clear liquid.
22
Figure 3.1: Purified neat PYR14TFSI Ionic Liquid
3.1.2 Characterization of PYR14TFSI
The water content in the synthesized PYR14TFSI ionic liquid was measured employing the
standard Karl Fisher method. The titrations were performed with Mettler Toledo Karl Fischer
titrator and it was calibrated and validated with Hydranal 34847 water standard before each water
content measurement. The moisture content was confirmed to be below 25ppm.
The 1H NMR spectrum was taken with Varian 400MHz spectrometer at Northeastern University
to confirm the structure of the ionic liquid (Figure 3.2). PYR14TFSI was dissolved in deuterated
acetone, 99.9 atom % D, contains 0.03% TMS (Sigma-Aldrich), and 700ul of the solution was
transferred into a Wilmad NMR tube (100 MHz, Economy) with an outer diameter of 5mm. The
chemical shifts were consistent with published literature values16. The cyclic voltammogram
(Figure 3.3) was taken with and without the oxygen saturation to confirm the purity of the
synthesized neat ionic liquid. With only argon, no peak is observed as expected, and when the
RTIL was saturated with oxygen, the peaks corresponding to the PYR/ O2- couple appears.
23
Figure 3.2: Proton NMR spectrum of neat PYR14TFSI. The peaks are number from 1 to 7, and assigned to each proton
Volts vs Li/Li+1 2 3 4
mA
/cm
2
-1.5
-1.0
-0.5
0.0
0.5
1.0
1 2 3 4 5 10
Cycle #
No current loss over 10cycles
@ 100 mV/s
Argon
Figure 3.3: Cyclic Voltammogram of PYR14TFSI Oxygen Reduction Reaction (ORR) on glassy
carbon electrode
24 3.2 13C NMR analysis on organic electrolytes and Ionic Liquids
3.2.1 Correlation between the 13C NMR chemical shifts and acidity of the cation in the
electrolytes
In nonaqueous electrolytes, salt cations are generally solvated by the solvent through acid base
complex formation. The 13C NMR chemical shift has been used to probe the interaction of
cations of alkali metal salts and ionic liquids with solvent molecules. Typical 13 C NMR spectra is
presented in figures; Figure 3.4 (a) shows the composite 13C NMR spectra of selected salts
consisting single charged cations of different sizes dissolved in propylene carbonate and Figure
3.4 (b) shows the composite 13C NMR spectra of salts consisting single charged cations of larger
size including the cations of ionic liquids.
Figure 3.4 (a): Composite 13C NMR spectra of 13C NMR solutions; selected salts consist single
charged cations of different sizes (a- None, (neat PC), b- 1M LiPF6, c- 1M LiTFSI, d- 1M NaPF6,
e- 1M KPF6, f- 1M TBAPF6) dissolved in propylene carbonate
25
Figure 3.4 (b): Composite 13C NMR spectra of 13C NMR solutions; selected salts consist single
charged cations of larger sizes (a- None, (neat PC), g- 1M TBAPF6, h- 1M EMITFSI, i- 1M
PYR14TFSI)
The plot of 13C NMR chemical shifts of carbonyl carbon versus ionic radius of the cation of the
salts dissolved in propylene carbonate (PC) are shown in the Figure 3.5. This plot shows that in
LiPF6 and NaPF6 salt solutions (1M), Li+ and Na+ ions are strongly solvated to the carbonyl
oxygen atoms of PC while potassium is weakly coordinated. Cations of ionic liquid behave like
26 TBA cation and they are not coordinated to the carbonyl oxygen atoms.
Figure 3.5: Plot of 13C nuclear magnetic resonance chemical shift versus ionic radius of cation of
salts used in the study
Since all of the ions studied have the same single positive charge, the radius (and hence ionic
volume) is roughly inversely proportional to the charge density, which is a measure of the Lewis
acidity. Therefore, 13C NMR chemical shifts on C(=O) provide a measure of cations’ relative
acidity and the cations of ionic liquid are shown to have the similar acidity as TBA+.
It should be also mentioned that the different nuclear magnetic resonance on carbonyl carbon
between LiPF6 and LiTFSI in PC were observed; the carbonyl carbon chemical shift in case of
LiPF6 in PC shows at higher frequency than the carbonyl carbon chemical shift in case of LiTFSI
in PC due to the nature of different anions. The 13C nuclear magnetic resonance on carbonyl
carbon in LiTFSI in PC is more deshielded compared to LiPF6 in PC because of the lone pair of
electrons present on nitrogen atom in TFSI anion. However, the difference was very small and is
negligible.
3.3 Preliminary results of 13C NMR Relaxation T1 measurement
27 The exponential recovery curve was created using eight data points for each sample in order to
ensure a good fit. Figure 3.6 shows the spectras from the inversion recovery experiment to
determine the T1’s of carbons in propylene carbonate solutions.
Figure 3.6 (a): Spectra from an inversion recovery experiment to determine the T1’s for neat
propylene carbonate (no salt)
Figure 3.6 (b): Spectra from an inversion recovery experiment to determine the T1’s for
propylene carbonate/ 1M LiPF6
28
Figure 3.6 (c): Spectra from an inversion recovery experiment to determine the T1’s for
propylene carbonate/ 1M EMITFSI (12 arrays employed, 2 scans)
Figure 3.6 (d): Spectra from an inversion recovery experiment to determine the T1’s for
propylene carbonate/ 1M PYR14TFSI (9 arrays employed, 2 scans)
29 Figure 3.7 (a)-(d) display the plots of 13C NMR T1 (sec) of each carbon in propylene carbonate
versus the ionic radius (pm) of cations of salts dissolved in propylene carbonate. The results are to
be further analyzed.
Figure 3.7 (a) T1 (C=O) vs. ionic radius of cations
Figure 3.7 (b) T1 (C-H) vs. ionic radius of cations
30
Figure 3.7 (c) T1 (C-H2) vs. ionic radius of cations
Figure 3.7 (d) T1 (C-H3) vs. ionic radius of cations
3.4 7Li NMR analysis on organic electrolytes and Ionic Liquids
3.4.1 Correlation between 7Li NMR chemical shifts and electrolyte solvent basicity (as
measured by the Gutmann’s Donor number of solvents)
The Gutmann’s solvent donor numbers (or donicity), is defined as the negative enthalpy of the 1:1
complex formation between a dilute solution between a given solvent and the standard Lewis
acid, antimony pentachloride, in 1, 2-dichloroethane, is a qualitative measure of the basicity of
31 solvents. It was my interest to see if there is a correlation between the donor numbers and the 7Li
chemical shift as glancing at the lithium chemical shifts in different solvents may immediately
indicate a possible correlation with the salvation ability of these solvents. Thus, acetonitrile, the
least solvating solvent, was expected to have the largest upfield shift, while strongly solvating
solvents such as dimethyl sulfoxide was expected to have the most pronounced downfield shift.
The selected solvents include organic solvents, binary solve mixture, and ionic liquids. The plot,
however, illustrated in Figure 3.8, shows no correlation.
Figure 3.8: Plot of chemical shift verses donor number of selected solvents used in the 7Li NMR
study (1) Acetonitrile, (2) Propylene Carbonate, (3) Tetraethylene glycol dimethyl ether, (4)
Acetone-d6, (5) 1,2- Dimethoxy Ethane/ Propylene Carbonate (1:1 v/v), (6) Dimethyl sulfoxide
Contrary to the case of 7Li NMR, a respectable straight line was observed in a plot of 23Na
chemical shifts vs. donicity of solvents in the study conducted by Popov, Alexander17. The major
reasoning behind is thought to be because sodium nucleus has the paramagnetic screening
constant that is much larger than the diamagnetic screening constant while lithium nucleus has the
similar magnitude for both of paramagnetic and diamagnetic shielding terms that make them to
cancel each other resulting the specific properties of solvent molecules, such as ring currents and
anisotropic terms, to become more important for lithium chemical shifts14.
32 There were several challenges associated with 7Li NMR study which include the solvent effect,
bulk susceptibility correction, and limited sources of information that determines the parameter of
7Li chemical shift.
The greatest challenge in this research was the solvent effect, raised from using various solvents.
Because some solvents are more magnetically active than others, it was difficult to choose a
suitable standard reference. The first attempt to correct for the solvent effect was done by adding
tetramethylsilane (TMS) into 0.5M lithium salt dissolved in each solvent while the external
reference, deuterated acetone already contains 1% TMS (v/v). The proton NMR spectrum was
taken right before each 7Li NMR scans. The difference in chemical shift between two TMS peaks
was considered to be from the result of the solvent effect and the difference in chemical shift
value was applied to correct for the 7Li chemical shift of each sample. In this approach, TMS was
assumed not to interact with samples and not to contribute to the chemical shift change. Figure 3.9
shows an example of proton spectra two TMS peaks. Moreover, this approach also required the
correction for the bulk susceptibility between the external reference and the sample because the
magnetic field is distorted by the capillary tube. The volume magnetic susceptibility of reference
and sample are unknown values and each needed to be measured each separately. For these
reasons, external referencing was avoided.
Xv: volume magnetic susceptibility
33
Figure 3.9: Proton spectra of two TMS peak. TMS peak at zero indicates the external reference,
and the TMS peak at -0.18 shows the TMS added into the sample in the capillary tube
The second approach was using caged lithium ion as an internal reference. 12-crown-4 was first
purchased from Sigma-Aldrich and used as a caging reagent because 12-crown-4 is known to
have high affinity towards lithium ion. To test whether the caged lithium exchange with the
solvent, molar ratio of 2:1, 1:1, 1:2 lithium ion to 12-crown-4 were prepared in three separate
NMR tubes and 7Li spectra were obtained. It was found that the caged lithium ion does exchange
with the solvent, as in case of 2:1 molar ratio of lithium to 12-crown-4 still showed one lithium
peak. In a literature search for the new complexing reagent, it was not only found that the
stabilities of the resulting 1:1 Li complexes is more stable in 15-crown-5 instead of 12-crown-4
because the ionic size of lithium (1.72A) is larger than the ring size of 12-crown-4 (1.2-1.5A)18,
but also Cryptan 211-Li complex is very stable. In a study of the exchange kinetics of the Li+ ion
complexation with cryptand complexing reagent19, it was reported that the observed rate constant
for the exchange of Li+ in C211 could not be found because there is almost no exchange taking
34 place. Therefore Cryptan 211, purchased from Sigma-Aldrich, was employed to cage the lithium
ion and when tested to confirm whether the caged lithium ion was exchanging with solvent by
adding excess lithium compared to Cryptan 211, two lithium peaks were shown as expected; the
peak at the higher frequency (downfield) indicates the caged lithium, served as the internal
standard throughout the 7Li study, and the other peak at the lower frequency (upfield) indicates
the free lithium ion. The Figure 3.10 shows two peaks as Cryptan 211 cages lithium ion.
Figure 3.10: Caged lithium ion served as an internal standard, solvent: acetone d-6 3.5 Lithium Air Pouch Cell Fabrication and Discharge Profile
To test the performance of Lithium battery employing room temperature ionic liquid, lithium air
pouch cell with dimension of 5cm x 5cm was built. An in-house built Li/O2 pouch cell is shown in
figure 3.11.
35
Figure 3.11: Picture of assembled Li-air Pouch cell (dimension: 5cm x 5cm)
Porous carbon electrodes were prepared by first preparing ink slurries by dissolving a 90 wt%
EC600JD Ketjen black (AkzoNobel) and 5 wt. % Kynar PVdF (Arkema Corporation) in N-
methyl-2-pyrrolidone (NMP). Air electrodes were prepared with a carbon loading of
approximately 20.0 mg/cm2 by hand-painting the inks onto a carbon cloth (PANEX 35, Zoltek
Corporation), which was then dried at 180◦C overnight. The total geometric area of the electrodes
was 3.93 cm2.
The Li/O2 test pouch cells were assembled in an argon-filled glove box. The cell consists of
metallic lithium anode and the air electrode as a cathode, prepared as mentioned above.
The copper current collector for anode and the aluminum current collector for cathode were used.
A Celgard 3401 separator separated the two electrodes and it was soaked in EMITFSI/0.5M
LiTFSI solution for a minimum of 24 hours. The cathode was soaked in the oxygen saturated
EMITFSI/1M LiTFSI solution for 24 hours and was put under vacuum for an hour before used for
the cell assembly.
The cell was placed in an oxygen filled glove box where oxygen pressure was maintained at 1atm.
Cell discharge was carried out with an Arbin battery cycler at the current rate of 0.1mA/cm2 at
room temperature.
Figure 3.12 shows the discharge profile of the Lithium air pouch cell employing EMITFSI, 0.5M
LiTFSI, which was also synthesized in our laboratory for electro-chemical advanced power
(LEAP), and it showed good performance. Although the discharge profile for a Li-air cell has not
36 been obtained for the newly synthesized Ionic liquid, PYR14TFSI, it is also expected to exhibit
good battery discharging performance.
Figure 3.12: Li-air pouch cell discharge profile (Electrolyte: EMITFSI/0.5M LiTFSI)
37 CHAPTER 4: CONCLUSIONS AND FUTURE DIRECTIONS
The room temperature ionic liquid, 1-butyl-1-methyl-pyrrolidinium bis-
(triflouromethanesulfonyl)imide (PYR14TFSI) has been successfully synthesized for both
electrochemical and NMR studies. Both 13C NMR and 7Li studies employing this ionic liquid
have been described in detail in this thesis. Electrochemical studies of O2 in this electrolyte
including CV, RDE, and impedance measurements are under investigation by a colleague. The
results of this this investigation have shown that 13C NMR chemical shifts of the C(=O) moiety in
PC provide a measure of the conducting salt cations’ relative acidity. The data show that in LiPF6
and NaPF6 solutions (1M) in PC, Li+ and Na+ ions are strongly solvated to the carbonyl oxygen.
K+ is weakly coordinated while TBA+ does not seem to coordinate with the organic carbonate
oxygen atoms. The cations of ionic liquids behave like TBA+ and they exhibit similar acidity as
TBA+. To gain additional information on the solvent-salt interactions in electrolytes in PC, the 13C
Longitudinal Relaxation Time (T1) was measured using 13C NMR. The T1 measurements were
conducted on all carbon atoms of propylene carbonate in the presence of the various salts whose
chemical shifts have been discussed above. The preliminary T1 results obtained are reported
here.
The 7Li NMR study of Li salts in a series of solvents selected based on their donor numbers (DN)
showed no correlation between the chemical shift and the Gutmann’s donicity of solvents in spite
of different experimental approaches taken.
The Li-air pouch cell employing the ionic liquid, EMITFSI/0.5M LiTFSI has shown good
performance when discharged at the current rate of 0.1mA/cm2. The synthesized neat PYR14TFSI
will be soon employed for battery testing as well, and from the NMR data of its purity it is
expected to exhibit good performance.
38
REFERENCE Abstract
1 Wasserscheid. P.; Welton, T., Ionic Liquids in Synthesis. Wiley-VCH
2 Zhou, Z.B.; Matsumoto, H.; Tatsumi, K., Cyclic Quaternary Ammonium Ionic Liquids with
Perfluoroalkyltrifluoroborates: Synthesis, Characterization, and Properties. Chem. Eur. J. 2006,
12, 2196 – 2212
3 K. M. Abraham and Ziping Jiang, A Polymer Electrolyte-Based Rechargeable
Lithium/Oxygen Battery , J. Electrochem. Soc., 143, 1(1996) T
4 Laoire, C.O.; Mukerjee, S.; Abraham, K.M., Influence of Nonaqueous Solvents on the
Electrochemistry of Oxygen in the Rechargeable Lithium-Air Battery. J. Phys. Chem.C. 2010,
114, 9178-9186
5 Allen, C.J. et. al., Oxygen Electrode Rechargeability in an Ionic Liquid for the Li-Air Battery.
J. Phys. Chem. Lett. 2011, 2 (19), 2420-2424
6 Appetecchi, G.B. et. al., Synthesis of Hydrophobic Ionic Liquids for Electrochemical
Applications. J.Electrochem.Soc. 2006, 153, A1685-1691
Thesis 1 Ohno, H., (2005) Electrochemical Aspects of Ionic Liquids, Wiley-Interscience, New York
2 Buzzeo, M. C.; Evans, R. G.; Compton, R. G., ChemPhysChem, 2004, 5, 1106.
3 Visser, A.E.; Swatloski, R.P.; Reichert, W.M.; Griffin, S.T.; Rogers, R.D., Ind. Eng. Chem.
Res. 2000, 39, 3596
4 Koch, V.R.; Nanjundiah, C.; Appetecchi, G.B.; Scrosati, B., J. Electrochem. Soc. 1995, 142,
L116
5 Welton, T., Chem.Rev. 1999, 99, 2071-2083
6 Zhou, Q.; Fitzgerald, K.; Boyle, P.D.; Henderson, W.A., Chem.Matter. 2010, 22, 1203-1208
7 Sakaebe, H.; Matsumoto, H.; Kobayashi, H.; Miyazaki, Y., Meeting Abstract of the 42th
Battery Symp. In Japan, 2B04 (2001)
8 Katayama, Y.; Onodera, H.; Yamagata, M.; Miura, T., J. Electrochem. Soc. 2004, 151, A59
9 Sacco, A., Chem. Soc. Rev., 1994, 23, 129-136
10 Laoire, C.O.; Mukerjee, S.; Abraham, K.M., J. Phys. Chem.C. 2010, 114, 9178-9186
11 Fawcett, W.R.; Ryan, P.J., Collect. Czech, Chem. 2009, 74, 1665-1674
12 Appetecchi, G.B. et al. Electrochimica Acta, 2009, 54, 1325–1332
13 Ohtaki, H.; Ishiguro, S., (1994) Chemistry of Nonaqueous Solutions: Current Progress (eds
G.Mamantov and A.I. Popov), Wiley-VCH Verlag GmbH, New York
39 14 Erlich, R.H.; Roach, E.; Popov. A.I., J. Am. Chem. Soc. 1970, 92, 4989
15 Appetecchi, G.B. et. al., J.Electrochem.Soc. 2006, 153, A1685-1691
16 Zhou, Z.B.; Matsumoto, H.; Tatsumi, K., Chem. Eur. J. 2006, 12, 2196-2212
17 Popov, A.I., Pure Appl. Chem. 1975, 41, 275
18 Karkhaneei, E. et al., J. incl. phenom. macrocycl. chem. 2001, 40, 209-312
19 Pasgreta, E. et al., J. incl. phenom. macrocycl. chem. 2007, 58, 81-88
20 Abraham, K. M.; Jiang, Z.; Carroll, B., Chemistry of Materials. 1997, 9, 1978-1988