Synthesis of a series of fluorinated boronate compounds and their use as additives in lithium battery electrolytes

Article abstract of DOI:10.1149/1.1779407

A new series of anion receptors based on boronate compounds have been synthesized. These compounds can be used as anion receptors in lithium battery electrolytes. The so-called boronate means that the compounds contain a boron bonded with two oxygen atoms and one carbon atom. This series includes various boronate compounds with different fluorinated aryl and fluorinated alkyl groups. When these anion receptors are used as additives in 1,2-dimethoxyethane (DME) solutions containing various lithium salts, the ionic conductivities of these solutions are greatly increased. The electrolytes tested in this study were DME solutions containing the following lithium salts: LiF, CF₃COOLi, and C₂F₅COOLi. Without the additive, the solubility of LiF in DME (and all other nonaqueous solvents) is very low. With some of these boronate compounds as additives, LiF solutions in DME with concentration as high as 1 M were obtained. The solubilities of the other salts were also increased by these additives. Near-edge X-ray absorption fine structure (NEXAFS) spectroscopy studies show that I^- anions are complexed with these compounds in DME solutions containing LiI salts. The degree of complexation is also closely related to the structures of the fluorinated aryl and alkyl groups which act as electron-withdrawing groups. The NEXAFS results are in good agreement with ionic conductivity studies.

Full text of DOI:10.1149/1.1779407

Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
A1429
0013-4651/2004/151͑9͒/A1429/7/$7.00 © The Electrochemical Society, Inc.
Synthesis of a Series of Fluorinated Boronate Compounds
and Their Use as Additives in Lithium Battery Electrolytes
a, ,z
,c
H. S. Lee,a Z. F. Ma,b, X. Q. Yang,
*
*
*
a,**
X. Sun,a, and J. McBreen
aBrookhaven National Laboratory, Upton, New York 11973, USA
bDepartment of Chemical Engineering, Shanghai Jiaotong University, Shanghai 200240, China
A new series of anion receptors based on boronate compounds have been synthesized. These compounds can be used as anion
receptors in lithium battery electrolytes. The so-called boronate means that the compounds contain a boron bonded with two
oxygen atoms and one carbon atom. This series includes various boronate compounds with different fluorinated aryl and fluori-
nated alkyl groups. When these anion receptors are used as additives in 1,2-dimethoxyethane ͑DME͒ solutions containing various
lithium salts, the ionic conductivities of these solutions are greatly increased. The electrolytes tested in this study were DME
solutions containing the following lithium salts: LiF, CF3COOLi, and C2F5COOLi. Without the additive, the solubility of LiF in
DME ͑and all other nonaqueous solvents͒ is very low. With some of these boronate compounds as additives, LiF solutions in DME
with concentration as high as 1 M were obtained. The solubilities of the other salts were also increased by these additives.
Near-edge X-ray absorption fine structure ͑NEXAFS͒ spectroscopy studies show that IϪ anions are complexed with these com-
pounds in DME solutions containing LiI salts. The degree of complexation is also closely related to the structures of the
fluorinated aryl and alkyl groups which act as electron-withdrawing groups. The NEXAFS results are in good agreement with
ionic conductivity studies.
© 2004 The Electrochemical Society. ͓DOI: 10.1149/1.1779407͔ All rights reserved.
Manuscript submitted October 8, 2003; revised manuscript received January 12, 2004. Available electronically August 18, 2004.
Numerous studies on developing new cathode and anode mate-
rials for lithium batteries have been reported. In contrast, studies on
developing new electrolyte systems somewhat lag behind. Recent
studies1,2 have shown that the low thermal stability of LiPF6-based
electrolyte plays an important role in the capacity fading of lithium-
ion batteries during cycling. Therefore, the development of new
electrolyte systems with high ionic conductivity, and good chemical
and electrochemical stability will help to improve the cycling life of
lithium-ion batteries. Studies on a new electrolyte using a boron-
compounds were shown in Fig. 1. The starting materials were pur-
chased from Aldrich Chemical Co. All the reactions and processes
were performed under an argon atmosphere or under vacuum. Pro-
ton nuclear magnetic resonance ͑HNMR͒ spectra were recorded on a
Hitachi R-1200 ͑60 MHz͒ spectrometer. The synthesis procedures of
compounds ͑1͒, ͑2͒, and ͑3͒ are sketched in Fig. 2. The following are
the outlines of these procedures:
Compound (1).—2-͑2.4-Difluorophenyl͒-4-fluoro-1,3,2-benzodioxa-
borole, or (C6H3F͒O2B͑C6H3F2). 3-fluorocatechol ͑3.84 g, 0.03
mol͒ and 2,4-difluoroboronic acid ͑4.73 g, 0.03 mol͒ were mixed in
40 mL toluene. The mixture was heated to reflux. Water from con-
densation reaction was removed by azeotropic distillation. After 4 h
of reaction, the solvent was evaporated from the reaction mixture
under reduced pressure. The residue was sublimed at 98-100°C/0.1
mm Hg. The final product ͑7.2 g of white crystal͒ was obtained in
96% yield. Melting point: 112°C. HNMR (CDCl3 ppm͒ ␦: 6.7-7.2
͑m, 5H͒, 7.8-8.3 ͑m, 1H͒. IR ͑neat cmϪ1͒, ␯ 3082.7, 1613.0, 1499.4,
1462.4, 1422.6, 1393.3, 1368.5, 1331.9, 1268.9, 1173.6, 1140.3,
1103.5, 1057.7, 1025.0, 968.7, 853.3, 774.7, 725.5, 653.4.
3,4
containing lithium salt bis͑oxalato͒ borate ͑LiBOB͒ have shown
the advantages of the higher electrochemical stability of this new
system. The potential of using boron-containing compounds in non-
aqueous electrolytes is attracting more and more attention from both
the academic and industrial communities. In our previous
publications,5,6 it has been demonstrated that using anion receptors
as additives can significantly reduce ion pairing in nonaqueous elec-
trolytes. These types of anion receptors can be used in two ways.
One is to use them as additives to improve the properties of cur-
rently used electrolytes;1 the other is to use them in combination
with various salts to develop new electrolyte systems. Here we re-
port the synthesis of a series of boronate compounds. The effect of
these new compounds on conductivity of lithium salts in nonaque-
ous solution was studied. The molecular weights of most of these
new boronate compounds are lower than our previously reported
boron compounds, and many of them are liquids at room tempera-
ture. Therefore, their effects on conductivity enhancement are supe-
rior. The effects on anion complexation were investigated by near-
edge X-ray absorption spectroscopy. Electrolytes using some
compounds in this series display high electrochemical stability up to
5 V. Although the molecular structures and the synthesis of some
compounds in this new series of boronate compounds had been
briefly described in our previous publication,7,8 detailed and system-
atic study of the series reported here provides more complete infor-
mation about these compounds.
Compound
(2).—2-͑3-Trifluoromethyl
phenyl͒-4-fluoro-1,3,2-
benzodioxaborole, or (C6H3F͒O2B͑C7H4F3). The same synthesis
procedure as for compound ͑1͒ was used. 3-Fluorocatechol reacted
Experimental
Synthesis and characterization.—Fourteen fluorinated 1,3,2-
benzodioxaborole and 1,3,2-dioxaborolane compounds synthesized
in our laboratory are reported here. The chemical structures of these
* Electrochemical Society Active Member.
** Electrochemical Society Fellow.
c Present address: DuPont Fluoroproducts, Wilmington, Delaware 19803, USA.
z E-mail: xyang@bnl.gov
Figure 1. Chemical structure of boronate-based anion receptors. The num-
bers under the structures correspond to the compound number in the text.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1430
Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
Figure 2. Scheme of synthesis steps of compounds ͑1-3͒.
with equivalent 3-trifluoromethylboronic acid to yield 95% pure
product. Melting point: 81-82°C. HNMR (CDCl3 ppm͒ ␦ 6.8-7.4
͑m, 3H͒, 7.5-8.1 ͑m, 2H͒, 8.2-8.5 ͑m, 2H͒. IR ͑neat cmϪ1͒, ␯ 3059.5,
1632.7, 1499.7, 1462.5, 1376.1, 1301.9, 1171.3, 1131.0, 1080.0,
1026.9, 917.7, 847.7, 808.7, 774.7, 721.7, 701.1, 605.7.
Figure 4. Scheme of synthesis steps of compounds ͑9-11͒.
2,5-Bis͑trifluoromethyl͒phenylboronic acid, an intermediate
compound. 160 mL of 2.5 M butyl lithium in hexane was mixed
with 400 mL of anhydrous ether. 77.2 g ͑0.27 M͒ of 1,4-
bis͑trifluoromethyl͒benzene was added dropwise to this solution at
room temperature. After refluxing for 6 h, the solution was added
dropwise, over a 1 h period, from a dropping funnel, to a flask in an
ice bath containing 100 mL trimethylborate mixed with 400 mL
ether. The mixture was further stirred for 2 h at room temperature.
Then the mixture was poured into an aqueous ammonium chloride
solution. The ether layer was separated and washed with water. The
product was dried over magnesium sulfate. The solvent was evapo-
rated and a crystalline of the product was formed. Pentane was
added to precipitate out all the crystals. The crude product was fil-
tered out and further purified through recrystallization. The yield
was 21 g, and the melting point was 133-135°C.
Compound (5).—2-͑2,4-Difluorophenyl͒-tetrafluoro-1,3,2-benzodi-
oxaborole ͑5͒ or (C6F4)O2B͑C6H3F2). The crude product was ob-
tained from 3.64 g of tetrafluorocatechol and 3.16 g of 2,4-
difluorophenylboronic acid. The crude product was sublimed at
125°C/0.2 mm Hg to obtain 5.7 g of pure compound. The melting
point of the compound was 112-114°C. HNMR ͓dimethyl sulfoxide
͑DMSO͒-d6 ppm͔ ␦: 6.7-7.2 ͑m, 2H͒, 7.3-7.7 ͑m, 1H͒.
Compound (6).—2-͑Pentafluorophenyl͒-tetrafluoro-1,3,2-benzodi-
oxaborole or (C6F4)O2B͑C6F5). 3.64 g of tetrafluorocatechol was
condensed with 4.2 g pentafluoroboronic acid. The crude product
was then sublimed at 110°C/0.1 mm Hg to obtain 6.2 g of pure
compound. The melting point of the compound was 131-132°C.
Compound (7).—2-͑2-Trifluoromethyl phenyl͒-tetrafluoro-1,3,2-
benzodioxaborole or (C6F4)O2B͑C7H4F3). 3.3 g of tetrafluorocat-
echol was reacted with 3.42 g of 2-trifluoromethylphenylboronic
acid to obtain 5.9 g of crude product. The crude product was then
sublimed at 110°C/0.1 mm Hg to obtain 5.2 g of pure compound.
The melting point of the compound was 102-103°C. HNMR
͑acetone-d6 ppm͒ ␦: 7.8-8.1 ͑m, 3H͒, 8.2-8.4 ͑m, 1H͒.
Compound
(3).—2,5-Bis͑trifluoromethyl͒phenyl-4-fluoro-1,3,2-
benzodioxaborole, or (C6H3F͒O2B͑C8H3F6). The same synthesis
procedure as for compound ͑1͒ was used. 2.56 g of 3-fluorocatechol
was condensed with 5.16 g of 2,5-bis͑trifluoromethyl͒phenylboronic
acid to obtain 6.8 g of pure product ͑sublimed at 120°C/0.3 mm Hg͒.
The yield was 92%. The melting point of the compound was 81-
83°C. HNMR ͑Acetone-d6 ppm͒ ␦: 7-7.4 ͑m, 3H͒, 8.1-8.25 ͑m, 2H͒,
8.6 ͑s, 1H͒.
Compound (8).—2,5-Bis͑trifluoromethyl phenyl͒-tetrafluoro-1,3,2-
benzodioxaborole or (C6F4)O2B͑C8H3F6). 3.2 g of tetrafluorocat-
The synthesis procedure of compounds ͑4-8͒ was similar as de-
scribed for compound ͑1͒ and is sketched in Fig. 3.
echol
was
condensed
with
5.16
g
of
2,5-
bis͑trifluoromethyl͒phenylboronic acid to obtain the crude
compound. The crude product was distilled using a Kugelrohr at
105-110°C/0.3 mm Hg. After cooling, crystals were formed and the
yield was 7.4 g. The melting point of the compound was 81-83°C.
HNMR ͑Acetone-d6 ppm͒ ␦: 8.05-8.2 ͑m, 2H͒, 8.6 ͑s, 1H͒.
The synthesis procedure of compounds ͑9-11͒ are sketched in Fig.
4. The following are the outlines of these procedures:
Compound
(4).—2-͑4-Fluorophenyl͒-tetrafluoro-1,3,2-benzodi-
oxaborole or (C6F4)O2B͑C6H4F). 3.3 g of tetrafluorocatechol was
condensed with 2.5 g of 4-fluoroboronic acid to obtain 5 g of crude
product. It was sublimed at 135°C/͑0.15 mm Hg͒ twice to obtain 4 g
of pure product. The melting point of the compound was 128-130°C.
Figure 3. Scheme of synthesis steps of compounds ͑4-8͒.
Figure 5. Scheme of synthesis steps of compounds ͑12-14͒.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
A1431
Table I. Ionic conductivity of 0.8 M boronate compounds ϩ lithium salts in DME.
Conductivity
CF3COOLi
͑S/cm͒
Conductivity
C2F5COOLi
͑S/cm͒
Conductivity
LiF ͑S/cm͒
Group number
of compounds
Boronate compounds
a
a
a
a
3.3 ϫ 10Ϫ5
4.20 ϫ 10Ϫ3
4.59 ϫ 10Ϫ3
4.67 ϫ 10Ϫ3
4.20 ϫ 10Ϫ3
7.50 ϫ 10Ϫ3
8.24 ϫ 10Ϫ3
6.56 ϫ 10Ϫ3
6.60 ϫ 10Ϫ3
5.65 ϫ 10Ϫ3
7.37 ϫ 10Ϫ3
6.74 ϫ 10Ϫ3
6.45 ϫ 10Ϫ3
7.80 ϫ 10Ϫ3
8.33 ϫ 10Ϫ3
2.1 ϫ 10Ϫ5
4.07 ϫ 10Ϫ3
3.57 ϫ 10Ϫ3
4.57 ϫ 10Ϫ3
4.07 ϫ 10Ϫ3
7.09 ϫ 10Ϫ3
8.10 ϫ 10Ϫ3
6.14 ϫ 10Ϫ3
6.00 ϫ 10Ϫ3
5.79 ϫ 10Ϫ3
6.44 ϫ 10Ϫ3
1
1
1
2
2
2
2
2
3
3
3
4
4
4
1. (C6H3F͒O2B͑C6H3F2)
2. (C6H3F͒O2B͑C7H4F3)
3. (C6H3F͒O2B͑C8H3F6)
4. (C6F4)O2B͑C6H4F)
5. (C6F4)O2B͑C6H3F2)
6. (C6F4)O2B͑C6F5)
7. (C6F4)O2B͑C7H4F3)
8. (C6F4)O2B͑C8H3F6)
9. (C6F12)O2B͑C6H5)
10. (C6F12)O2B͑C6H3F2)
11. (C6F12)O2B͑C6F5)
12. (C3HF6O)2B͑C6H5)
13. (C3HF6O)2B͑C6H3F2)
14. (C3HF6O)2B͑C6F5)
1.00 ϫ 10Ϫ3
2.75 ϫ 10Ϫ3
9.54 ϫ 10Ϫ3
3.06 ϫ 10Ϫ3
3.42 ϫ 10Ϫ3
2.31 ϫ 10Ϫ3
6.97 ϫ 10Ϫ3
7.48 ϫ 10Ϫ3
4.78 ϫ 10Ϫ3
4.75 ϫ 10Ϫ3
4.03 ϫ 10Ϫ3
6.27 ϫ 10Ϫ3
b
b
b
a The solubility of LiF was low; therefore, the conductivity was not measured.
b Conductivity was not measured.
Compound (9).—2-Phenyl-4,4,5,5-tetrakis͑trifluoromethyl͒-1,3,2-
dioxaborolane or (C6F12)O2B͑C6H5). The synthesis of this com-
pound followed the procedure described by Allan et al.9 By reacting
perfluoropinacol with dichlorophenylborane, the product was ob-
tained with a yield of 78%. The boiling point of the compound was
68°C/18 mm Hg.
1,1,1,3,3,3-hexafluoro-2-propanol was added dropwise at room tem-
perature. After the addition was completed, the ether was evaporated
from the solution. The residue was dried under vacuum to obtain the
solid alkoxide. 200 mL of pentane was added to the alkoxide, then
15.9 g ͑0.1 M͒ of dichlorophenylborane ͑purchased from Aldrich͒
was dropped into the mixture. After stirring the mixture for 20 h, the
sodium chloride was filtered off. The ether was removed by evapo-
ration under atmospheric pressure from the filtrate. The liquid resi-
due was distilled under vacuum using a Kugelrohr. 22 g of the pure
product was obtained. The boiling point of the compound was 90-
95°C/15 mm Hg. HNMR (CDCl3 ppm͒ ␦: 5.2 ͑q, 2H͒, 7.55 ͑s, 5H͒.
IR ͑neat cmϪ1͒, ␯ 3061.1, 2966.9, 1603.9, 1344.9, 1267.9, 1110.4,
905.8, 874.4, 693.2.
Compound (10).—2-͑3,5-Difluorophenyl-4,4,5,5-tetrakis͑trifluoro-
methyl͒-1,3,2-dioxaborolane or (C6F12)O2B͑C6H3F2). 1-bromo-
3,5-difluorobenzene ͑38.6 g, 0.2 mol͒ was added slowly to a mixture
of magnesium turnings ͑4.8 g, 0.2 mol͒ in 200 mL anhydrous ethyl
ether. After the reaction mixture was refluxed for 2 h, trimethyltin
chloride ͑40 g, 0.2 mol͒ was added dropwise to the solution and
refluxing continued for one more hour. Then the reaction mixture
was hydrolyzed by saturated ammonium chloride. The organic layer
of 3,5-difluorophenyltrimethyltin was separated and purified by dis-
tillation with a 62% yield. Then the 3,5-difluorophenyltrimethyltin
͑11 g, 0.04 mol͒ was reacted with boron trichloride ͑11 g, 0.093 M͒
in a sealed tube at 0°C for 1 h and then at room temperature for 12
h. Dichloro-3,5-difluorophenylborane ͑6.3 g, 0.032 mol͒ was ob-
tained from the reaction with a yield of 81%. Dichloro-3,5-
difluorophenylborane ͑5.84 g, 0.03 mol͒ was added to a solution
containing perfluoropinacol ͑10 g, 0.03 M͒ in 50 mL of anhydrous
chloroform at Ϫ40°C. Then the reaction mixture was stirred con-
tinuously at room temperature for 4 h. The insoluble material was
filtered off from the solution. The chloroform solvent was removed
by evaporation. The pure product was obtained by distillation of the
residue with a yield of 73%. The boiling point of the compound was
80°C/14 mm Hg. HNMR (CDCl3 ppm͒ ␦: 6.9-7.4 ͑m, 1H͒, 7.35-7.7
͑m, 2H͒. IR ͑neat cmϪ1͒, ␯ 3092.0, 1593.8, 1480.6, 1432.0, 1388.2,
1248.9, 1109.8, 1080.5, 986.9, 952.8, 889.6, 870.4, 748.2, 721.8.
Compound (13).—Bis͑1,1,1,3,3,3-hexafluoroisopropyl͒-3,5-difluoro
phenylboronate, or (C3HF6O)2B͑C6H3F2). This compound was
synthesized with the same procedure as described for compound
͑12͒, but using dichloro-3,5-difluorophenylborane to replace the
dichlorophenylborane. The yield was 62%. The boiling point of the
compound was 70-73°C/15 mm Hg. HNMR (CDCl3 ppm͒ ␦: 5.1 ͑q,
2H͒, 6.9-7.35 ͑m, 3H͒. IR ͑neat cmϪ1͒, ␯ 3092.3, 1590.6, 1348.7,
1267.7, 1204.4, 1117.0, 987.2, 872.4, 698.2.
Compound (14).—Bis͑1,1,1,3,3,3-hexafluoroisopropyl͒pentafluoro-
phenylboronate, or (C3HF6O)2B͑C6F5). This compound was syn-
thesized with the same procedure as described for compound ͑12͒,
Table II. Ionic conductivity of compound „1…¿ lithium salt in
DME solutions.
Conductivity Conductivity
Concentration
of lithium
salt ͑M͒
Conductivity
LiF
͑S/cm͒
Compound
(11).—2-pentafluorophenyl-4,4,5,5-tetrakis͑trisfluoro-
Compound
͑1͒ ͑M͒
CF3COOLi
C2F5COOLi
methyl͒-1,3,2-dioxaborolane or (C6F12)O2B͑C6F5). This compound
was synthesized with the same procedure for compound ͑10͒. Bro-
mopentafluorobenzene and perfluoropinacol were used as starting
materials. The final product had a boiling point: 87-90°C/12 mm Hg.
IR ͑neat cmϪ1͒, ␯ 1657.6, 1493.5, 1426.4, 1368.4, 1246.5, 1083.7,
989.9, 952.2, 889.0, 749.2, 718.2.
͑S/cm͒
͑S/cm͒
a
a
a
a
a
a
a
3.3 ϫ 10Ϫ5 2.1 ϫ 10Ϫ5
1.24 ϫ 10Ϫ3 1.1 ϫ 10Ϫ3
2.83 ϫ 10Ϫ3 2.73 ϫ 10Ϫ3
3.83 ϫ 10Ϫ3 3.69 ϫ 10Ϫ3
4.20 ϫ 10Ϫ3 4.07 ϫ 10Ϫ3
3.96 ϫ 10Ϫ3 3.86 ϫ 10Ϫ3
3.40 ϫ 10Ϫ3 3.35 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
0.2
0.4
0.6
0.8
1.0
1.2
The synthesis procedure of compounds ͑12-14͒ are sketched in
Fig. 5. Following are the outlines of these procedures.
Compound
(12).—Bis͑1,1,1,3,3,3-hexafluoroisopropyl͒phenyl-
a The solubility of LiF was low; therefore, the conductivity was not
measured.
boronate, or (C3HF6O)2B͑C6H5). To a mixture of 4.8 g ͑0.2 M͒ of
sodium hydride in 200 mL of anhydrous ether, 33.6 g ͑0.2 M͒ of
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1432
Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
Table III. Ionic conductivity of compound „2…¿ lithium salt in DME solutions.
Conductivity
CF3COOLi
͑S/cm͒
Conductivity
C2F5COOLi
͑S/cm͒
Conductivity
LiF
͑S/cm͒
Concentration of
compound ͑2͒ ͑M͒
Concentration of
lithium salt ͑M͒
a
a
a
a
a
a
a
3.3 ϫ 10Ϫ5
1.93 ϫ 10Ϫ3
3.57 ϫ 10Ϫ3
4.53 ϫ 10Ϫ3
4.59 ϫ 10Ϫ3
3.95 ϫ 10Ϫ3
3.00 ϫ 10Ϫ3
2.1 ϫ 10Ϫ5
7.60 ϫ 10Ϫ4
1.94 ϫ 10Ϫ3
2.90 ϫ 10Ϫ3
3.57 ϫ 10Ϫ3
3.79 ϫ 10Ϫ3
3.68 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
0.2
0.4
0.6
0.8
1.0
1.2
a The solubility of LiF was low; therefore, the conductivity was not measured.
Table IV. Ionic conductivity of compound „6…¿ lithium salt in DME solutions.
Conductivity
CF3COOLi
͑S/cm͒
Conductivity
C2F5COOLi
͑S/cm͒
Conductivity
LiF
͑S/cm͒
Concentration of
compound ͑6͒ ͑M͒
Concentration of
lithium salt ͑M͒
a
3.3 ϫ 10Ϫ5
4.11 ϫ 10Ϫ3
7.51 ϫ 10Ϫ3
9.00 ϫ 10Ϫ3
8.24 ϫ 10Ϫ3
7.76 ϫ 10Ϫ3
6.62 ϫ 10Ϫ3
2.1 ϫ 10Ϫ5
4.52 ϫ 10Ϫ3
7.67 ϫ 10Ϫ3
8.92 ϫ 10Ϫ3
8.10 ϫ 10Ϫ3
6.49 ϫ 10Ϫ3
4.67 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
3.58 ϫ 10Ϫ3
6.85 ϫ 10Ϫ3
8.98 ϫ 10Ϫ3
9.54 ϫ 10Ϫ3
9.00 ϫ 10Ϫ3
7.68 ϫ 10Ϫ3
0.2
0.4
0.6
0.8
1.0
1.2
a The solubility of LiF was low; therefore, the conductivity was not measured.
Table V. Ionic conductivity of compound „8…¿ lithium salt in DME solutions.
Conductivity
CF3COOLi
͑S/cm͒
Conductivity
C2F5COOLi
͑S/cm͒
Conductivity
LiF
͑S/cm͒
Concentration of
Concentration of
compound ͑8͒ ͑M͒
lithium salt ͑M͒
a
a
a
a
3.3 ϫ 10Ϫ5
3.60 ϫ 10Ϫ3
6.16 ϫ 10Ϫ3
6.95 ϫ 10Ϫ3
6.60 ϫ 10Ϫ3
5.22 ϫ 10Ϫ3
3.47 ϫ 10Ϫ3
2.1 ϫ 10Ϫ5
3.92 ϫ 10Ϫ3
6.56 ϫ 10Ϫ3
7.15 ϫ 10Ϫ3
6.00 ϫ 10Ϫ3
4.11 ϫ 10Ϫ3
2.23 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
0.2
0.4
0.6
0.8
1.0
1.2
3.42 ϫ 10Ϫ3
3.78 ϫ 10Ϫ3
3.46 ϫ 10Ϫ3
a LiF was only partially dissolved in these solutions; therefore, conductivity was not measured.
Table VI. Ionic conductivity of compound „9…¿ lithium salt in DME solutions.
Conductivity
CF3COOLi
͑S/cm͒
Conductivity
C2F5COOLi
͑S/cm͒
Conductivity
LiF
͑S/cm͒
Concentration of
Concentration of
compound ͑9͒ ͑M͒
lithium salt ͑M͒
a
3.3 ϫ 10Ϫ5
3.15 ϫ 10Ϫ3
4.95 ϫ 10Ϫ3
5.89 ϫ 10Ϫ3
5.65 ϫ 10Ϫ3
4.59 ϫ 10Ϫ3
3.02 ϫ 10Ϫ3
2.1 ϫ 10Ϫ5
3.46 ϫ 10Ϫ3
5.33 ϫ 10Ϫ3
6.14 ϫ 10Ϫ3
5.79 ϫ 10Ϫ3
4.46 ϫ 10Ϫ3
2.78 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
7.40 ϫ 10Ϫ4
1.38 ϫ 10Ϫ3
1.90 ϫ 10Ϫ3
2.31 ϫ 10Ϫ3
2.47 ϫ 10Ϫ3
2.40 ϫ 10Ϫ3
0.2
0.4
0.6
0.8
1.0
1.2
a The solubility of LiF was low; therefore, the conductivity was not measured.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
A1433
Table VII. Ionic conductivity of compound „11…¿ lithium salt in DME solutions.
Conductivity
CF3COOLi
͑S/cm͒
Conductivity
C2F5COOLi
͑S/cm͒
Conductivity
LiF
͑S/cm͒
Concentration of
compound ͑11͒ ͑M͒
Concentration of
lithium salt ͑M͒
a
3.3 ϫ 10Ϫ5
4.30 ϫ 10Ϫ3
7.07 ϫ 10Ϫ3
7.83 ϫ 10Ϫ3
6.74 ϫ 10Ϫ3
4.69 ϫ 10Ϫ3
2.86 ϫ 10Ϫ3
2.1 ϫ 10Ϫ5
4.57 ϫ 10Ϫ3
7.16 ϫ 10Ϫ3
7.72 ϫ 10Ϫ3
6.27 ϫ 10Ϫ3
3.97 ϫ 10Ϫ3
1.91 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
3.47 ϫ 10Ϫ3
6.13 ϫ 10Ϫ3
4.53 ϫ 10Ϫ3
7.48 ϫ 10Ϫ3
6.55 ϫ 10Ϫ3
5.10 ϫ 10Ϫ3
0.2
0.4
0.6
0.8
1.0
1.2
a The solubility of LiF was low; therefore, the conductivity was not measured.
but using dichloro-pentafluorophenylborane to replace the dichlo-
rophenylborane. The yield was 43%. The boiling point of the com-
pound was 73-75°C/15 mm Hg. HNMR (CDCl3 ppm͒ ␦: 4.9 ͑q,
CH͒. IR ͑neat cmϪ1͒, ␯ 2972.8, 1654.8, 1491.3, 1379.6, 1271.7,
1209.0, 1113.7, 982.5, 874.5, 693.4.
increasing number of the electron-withdrawing groups.6,7 The larg-
est conductivity enhancement in group 2 is from compound ͑6͒,
which has fully fluorinated rings on both sides. Compound 6 has the
strongest effect in group 2. This tells us that the conductivity en-
hancement ͑obtained from the anion complexation effect͒ strongly
depends on the electron-withdrawing effect of the substituting
group, but it is also affected by steric hindrance effects. Compound
͑8͒ has a greater conductivity enhancement effect than compound
͑5͒, due to the two stronger CF3 electron-withdrawing groups as
compared to the two single fluorine atoms in compound ͑5͒. It has a
weaker effect than compound ͑6͒ due to the steric hindrance effect
of the bulky CF3 structure. All compounds in groups ͑3͒ and ͑4͒ are
liquids at room temperature due to their partial branch structures.
This liquid nature makes them much easier to be dissolved in non-
aqueous solvents. In group 3, the conductivity enhancement in-
creases with the increased number of substituting fluorine atoms.
However, in group 4, the conductivity enhancement does not have
clear correlation with the number of fluorine atoms. This may be due
to the more flexible branch structure of group 4. Although we have
measured conductivity data for all 14 compounds at various concen-
trations, only selected data are presented here for each group. The
ionic conductivity data of compound 1 and compound 2 as additives
in DME with various concentrations are listed in Tables II and III.
The corresponding data for compound ͑6͒ and compound ͑8͒ are
listed in Tables IV and V. Table VI contains data for compound ͑9͒
and Table VII contains data for compound ͑11͒. The results for
compound ͑14͒ are listed in Table VIII. For most compounds, the
conductivity has a maximum at around 0.8 M concentration. For
compounds with CF3 substituting groups attached to phenol rings,
the conductivity decreases rapidly with increasing concentration af-
ter the maximum was reached, while the others with fluorine substi-
tuting atoms only decrease slightly. We would like to point out that
Conductivity measurements.—Conductivity measurements were
made at 25°C using a Hewlett-Packard 4129A impedance analyzer
in the frequency range from 5 Hz to 10 MHz. Cells with Pt elec-
trodes were used for the measurements. The cell constants were
calibrated by measuring a 0.05 N KCl standard aqueous solution.
The ionic conductivities of dimethoxyethane ͑DME͒ solutions con-
taining various boron-based compounds and lithium salts were com-
pared with the conductivities of DME solutions containing lithium
salts only.
Near-edge absorption X-ray fine structure measurements.—
Near-edge absorption X-ray fine structure ͑NEXAFS͒ measurements
were made at the iodine LI edge for LiI/DME solutions with and
without boron-based anion receptors. The measurements were done
at Beam Line X19A of the National Synchrotron Light Source
͑NSLS͒. The data were collected using a large solid angle ionization
chamber as the fluorescence detector. The solutions containing LiI
salt and various boronate compounds were poured into cells with
thin Mylar windows for NEXAFS studies.
Electrochemical stability measurements.—Electrochemical sta-
bility measurements were performed using a Solatron SI 1287 elec-
trochemical interface in the potential dynamic mode.
Results and Discussion
The ionic conductivity of electrolytes containing lithium salts
only, such as CF3CO2Li and C2F5CO2Li, dissolved in DME are
low, with respective conductivities of only 3.3 and 2.1 ϫ 10Ϫ5 S/cm
at concentrations of 0.2 M. LiF is insoluble in this solvent. The
conductivity enhancement effects of the boronate compounds are
summarized in Table I for 0.8 M concentrations of both the lithium
salts and the anion receptors. All measurements were done at 25°C.
According to the chemical structures in Fig. 1, the 14 compounds
can be divided into four groups: compounds ͑1-3͒, compounds ͑4-8͒,
compounds ͑9-11͒, and compounds ͑12-14͒. The 0.8 M concentra-
tion data are given because this is close to the optimized concentra-
tion. With the addition of stoichiometric amounts of compounds
͑1-3͒ as additives, the conductivities of 0.8 M solutions of
CF3CO2Li and C2F5CO2Li salts in DME were significantly in-
creased to a range around 4 ϫ 10Ϫ3 S/cm, as shown in Table I.
However, their anion complexation effects of these compounds were
still rather weak and the solubility of LiF remained low. When the
second group compounds ͓compounds ͑4-8͔͒ were used as additives,
their effects were strong enough to dissolve LiF salt to concentra-
tions as high as 1.2 M. The main difference between the first group
and the second is the fully fluorinated left ring in group 2. This is
consistent with our previous results on the increased conductivity
enhancement effects of other boron-based anion receptors with the
Table VIII. Ionic conductivity of compound „14…¿ lithium salt
in DME solutions.
Concentration
Concentration
of
lithium salt
͑M͒
Conductivity
CF3COOLi
͑S/cm͒
of
compound ͑11͒
͑M͒
Conductivity
LiF
͑S/cm͒
a
3.30 ϫ 10Ϫ5
3.67 ϫ 10Ϫ3
4.49 ϫ 10Ϫ3
8.96 ϫ 10Ϫ3
8.33 ϫ 10Ϫ3
6.30 ϫ 10Ϫ3
3.68 ϫ 10Ϫ3
0
0.2
0.2
0.4
0.6
0.8
1.0
1.2
1.45 ϫ 10Ϫ3
2.70 ϫ 10Ϫ3
3.56 ϫ 10Ϫ3
4.03 ϫ 10Ϫ3
4.09 ϫ 10Ϫ3
3.77 ϫ 10Ϫ3
0.2
0.4
0.6
0.8
1.0
1.2
a The solubility of LiF was low; therefore, the conductivity was not
measured.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1434
Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
Figure 7. Electrochemical stability of electrolyte containing 0.2 M LIF and
compound ͑14͒ in EC-DMC solution ͑v/v 1:1͒.
these spectra. This technique has been used to study the complex-
ation between ClϪ and aza-based anion receptors10 and other types
of boron-based anion receptors.5 The white line in curve ͑a͒ for LiI
in DME is a broad peak. When LiI salt was dissolved in solvent, the
IϪ ion experiencing a spherical symmetric environment, the 5p final
states of the IϪ are close to being degenerate, and the transition
feature is a broad peak. White-line peak splitting can occur if there
is an asymmetric distribution of atoms ͑or molecules͒ surrounding
the IϪ with strong bonding. The more asymmetric the distribution is,
the stronger the splitting effect will be. The asymmetric local field
experienced by each of the degenerate 5p final states of IϪ is dif-
ferent and the degeneracy is lifted. This results in a splitting of the
absorption peak. In curve ͑a͒, when the LiI salt was dissolved in
DME, the white-line splitting is almost undetectable, because the
distribution of the solvent molecules has a spherical symmetry sur-
rounding the IϪ ions, and the interaction between the solvent and IϪ
is weak. We classify these anions as uncomplexed anions. In curve
͑b͒, white-line splitting was observed when compound ͑4͒ was
added into the solution. This gives strong evidence that the IϪ is
indeed complexed with the boron atom in compound ͑4͒. In curves
͑c-f͒, for solutions containing LiI and compounds ͑7-10͒, respec-
tively, the white-line splitting becomes more and more obvious, in-
dicating stronger and stronger anion complexing effects of these
compounds. For easy comparison, the chemical structures and the
ionic conductivities of these compounds used as additives for LiF
salts are also shown in Fig. 6. It is quite clear that the stronger the
electron-withdrawing power of the substituting groups in the bor-
onate compounds, the higher the conductivity enhancement, and
also the larger white-line splitting of the NEXAFS spectra.
Figure 6. NEXAFS at the I LI edge for DME solutions of: ͑a͒ 0.2 M LiI salt,
͑b͒ 0.2 M LiI ϩ compound ͑4͔͒, ͑c͒ 0.2 M LiI ϩ compound ͑7͔͒, ͑d͒ 0.2
͓
͓
M
LiI ϩ compound ͑8͔͒, ͑e͒ 0.2 M LiI ϩ compound ͑10͔͒, and ͑f͒ 0.2 M
͓
͓
LiI ϩ compound ͑11͔͒.
͓
for the practical lithium battery electrolyte, DME is not a good
choice due to its poor electrochemical stability compared with other
solvents such as dimethyl carbonate ͑DMC͒.
In our previous publication,5 NEXAFS was used to study the
anion complexation of boron compounds. This technique is used
here to study the anion complexation effect of boronate compounds.
NEXAFS spectra at the iodine LI edge were used to study the
coordination symmetry of the IϪ ion in solutions. It would be better
if we could get spectroscopic evidence for complex formation in LiF
solutions directly by studying the fluorine K edge. Unfortunately, the
energy of fluorine K edge is too low and difficult to apply to solution
studies. Therefore, the spectroscopy was done at the chlorine K edge
and iodine L edges. We report our iodine L edge spectroscopy results
here. In an iodine L edge NEXAFS experiment, as the incident
X-ray energy is scanned through the iodine L edge, the ejected 2s
photoelectrons can undergo bound state transitions to empty 5p
states. Coordination symmetry can affect the energy levels of the
final states. Therefore, the fine structure of the absorption spectra
near the edge provides information about the coordination symmetry
of the IϪ anion.
The electrochemical stability studies were carried out using a
Solatron SI 1287 electrochemical interface in potential dynamic
mode at a scan rate of 20 mV/s. A three-electrode cell with a glassy
carbon as working electrode ͑7.0 mm2͒, platinum wire as counter
electrode, and metal lithium foil as reference electrode was used for
this study. The resulting curve is plotted in Fig. 7. The electrolyte
made of compound ͑14͒, LiF salt, and ethylene carbonate/diethyl
carbonate ͑EC/DMC͒ solvent has excellent electrochemical stability
at voltage vs. Li as high as 5 V.
Figure 6 shows how the boron-based additives with different
electron-withdrawing groups have a dramatic effect on fine structure
of the absorption spectra. NEXAFS data at the iodine K edge are
given for DME solutions of ͑a͒ 0.2 M LiI, ͑b͒ 0.2 M LiI ϩ 0.2 M
compound ͑4͒, ͑c͒ 0.2 M LiI ϩ 0.2 M compound ͑7͒, ͑d͒ 0.2 M
LiI ϩ 0.2 M compound ͑8͒, ͑e͒ 0.2 M LiI ϩ 0.2 M compound ͑10͒,
͑f͒ 0.2 M LiI ϩ 0.2 M compound ͑11͒. The ‘‘white line’’ peak,
between 5185 and 5195 eV, above the edge, is due to dipole-allowed
transitions to final states of p symmetry. The structure of the white
line is sensitive to the coordination of the absorbing atom, IϪ, in
Acknowledgment
This work was supported by the U.S. Department of Energy,
Division of Materials Science of the Office Basic Energy Sciences,
and the Office of Energy Research, Laboratory Technology Re-
search Program, under contract no. DE-AC02-98CH10886. The
NEXAFS experiments were done at Beam Line X19A at NSLS.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑9͒ A1429-A1435 ͑2004͒
A1435
Brookhaven National Laboratory assisted in meeting the publication
costs of this article.
5. H. S. Lee, X. Q. Yang, C. L. Xiang, J. McBreen, and L. S. Choi, J. Electrochem.
Soc., 145, 2813 ͑1998͒.
6. X. Sun, H. S. Lee, S. Lee, X. Q. Yang, and J. McBreen, Electrochem. Solid-State
Lett., 1, 239 ͑1998͒.
7. H. S. Lee, X. Q. Yang, X. Sun, and J. McBreen, J. Power Sources, 97-98, 4188
͑2001͒.
8. H. S. Lee, X. Sun, X. Q. Yang, and J. McBreen, J. Electrochem. Soc., 149, A1460
͑2002͒.
References
1. X. Sun, H. S. Lee, X. Q. Yang, and J. McBreen, Electrochem. Solid-State Lett., 5,
A248 ͑2002͒.
2. S. E. Sloop, J. K. Pugh, S. Wang, J. B. Kerr, and K. Kinoshita, Electrochem.
Solid-State Lett., 4, A42 ͑2001͒.
9. M. Allan, A. F. Janzen, and C. J. Willis, Can. J. Chem., 46, 3671 ͑1968͒.
10. H. S. Lee, X. Q. Yang, J. McBreen, L. S. Choi, and Y. Okamoto, J. Electrochem.
Soc., 143, 3825 ͑1996͒.
3. K. Xu, S. Zhang, T. R. Jow, W. Xu, and C. A. Angell, Electrochem. Solid-State
Lett., 5, A26 ͑2002͒.
4. W. Xu and C. A. Angell, Electrochem. Solid-State Lett., 4, E1 ͑2001͒.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).