Tetraalkylphosphonium-based ionic liquids

Article abstract of DOI:10.1016/j.jorganchem.2004.09.060

Ionic liquids are salts that are liquid at or near room temperature. Their wide liquid range, good thermal stability, and very low vapor pressure make them attractive for numerous applications. The general approach to creating ionic liquids is to employ a large, unreactive, low symmetry cation with and an anion that largely controls the physical and chemical properties. The most common cations used in ionic liquids are N-alkylpyridinium and N,N′- dialkylimidazolium. Another very effective cation for the creation of ionic liquids is tetraalkylphosphonium, [PR₁R₂R ₃R₄]⁺. The alkyl groups, R_n, generally are large and not all the same. The halide salts of several phosphonium cations are available as starting materials for metathesis reactions used to prepare ionic liquids. The large phosphonium cations can combine with relatively large anions to make viscous but free flowing liquids with formula mass greater than 1000 g mol^-1. Some other more massive salts are waxes and glasses. The synthesis and the physical, chemical, and optical properties of phosphonium-ionic liquids having anions with a wide range of masses were measured and are reported here.

Full text of DOI:10.1016/j.jorganchem.2004.09.060

Journal of Organometallic Chemistry 690 (2005) 2536–2542
www.elsevier.com/locate/jorganchem
Tetraalkylphosphonium-based ionic liquids
a
Rico E. Del Sesto , Cynthia Corley , Al Robertson , John S. Wilkes
a
b
a,
*
a
Department of Chemistry, US Air Force Academy, 2355 Fairchild Dr., Suite 2N225, Colorado, USAF Academy 80840-6230, USA
b
Cytec Canada Inc., 9061 Garner Road, Niagara Falls, Ont., Canada L2E6T4
Received 29 June 2004; accepted 20 September 2004
Available online 28 October 2004
Abstract
Ionic liquids are salts that are liquid at or near room temperature. Their wide liquid range, good thermal stability, and very low
vapor pressure make them attractive for numerous applications. The general approach to creating ionic liquids is to employ a large,
unreactive, low symmetry cation with and an anion that largely controls the physical and chemical properties. The most common
0
cations used in ionic liquids are N-alkylpyridinium and N,N -dialkylimidazolium. Another very effective cation for the creation of
ionic liquids is tetraalkylphosphonium, [PR
+
] . The alkyl groups, R
1
R
2
R
3
R
4
n
, generally are large and not all the same. The halide
salts of several phosphonium cations are available as starting materials for metathesis reactions used to prepare ionic liquids. The
large phosphonium cations can combine with relatively large anions to make viscous but free flowing liquids with formula mass
ꢀ1
greater than 1000 g mol . Some other more massive salts are waxes and glasses. The synthesis and the physical, chemical, and opti-
cal properties of phosphonium-ionic liquids having anions with a wide range of masses were measured and are reported here.
ꢀ
2004 Elsevier B.V. All rights reserved.
Keywords: Ionic liquids; Molten salt; Physical properties; Phosphonium salts; Viscosity; Room temperature ionic liquids
1
. Introduction
materials have led us to RTILs with large phosphonium
+
cations, [PR R R R ] . The anions in the phosphonium
1
2
3
4
The diverse nature of room temperature ionic liquids
RTILs) continues to expand significantly as more re-
sources are dedicated to understanding these materials
as solvents, electrolytes, thermal fluids, and much more
salts contain structural features which have been pre-
dicted to display significant third order nonlinear optical
properties, particularly nonlinear indices or refraction
(
(often designated as n ) and nonlinear absorptions
2
[
1–3]. RTILs have numerous advantages over molecular
(a ), – such as multiphoton absorption – both of which
2
organic materials, particularly in their stability due to
their negligible vapor pressure. Other advantages in-
clude easy functionalization, large liquid ranges, high
effective concentrations as compared to solutions, and
the ability to flow – an interesting and attractive quality
from a materials perspective.
The structural tunability of both the cations and ani-
ons suggest great flexibility in the potential applications
of ionic liquids. Our interests in optical and magnetic
are intensity dependent functions, and therefore useful
for optical limiting applications. These include conju-
gated thiolate and dithioacetate compounds [4], Fig. 1,
such as methylxanthate (xan), diethyldithiocarbamate
(dtc), nitrodithioacetate – also known as the K-salt,
and dithiomaleonitrile (dtmn). Other conjugated anions
of interest for their physical properties are bistrifylamide
(NTf ), nonafluorobutylsulfonate (C F SO ), tricya-
2
4
9
3
nomethanide (C(CN) ), and dicyanamide (N(CN) ).
3
2
Some further potentially magnetically and/or optically
interesting anions are those such as thiocyanate, cya-
nide, or dicyanamide complexes of Ni(II), Co(II), and
Fe(II), bis-dicarbollylcobalt(III) (CoCB) – which is
*
Corresponding author. Tel.: +1 719 333 6005; fax: +1 719 333
2
947.
E-mail address: john.wilkes@usafa.af.mil (J.S. Wilkes).
0
022-328X/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.09.060

R.E. Del Sesto et al. / Journal of Organometallic Chemistry 690 (2005) 2536–2542
2537
H
S
S
S
C
S
C
S
S
Et
Et
S
S
C
C
N
C
O
C
C
C
O
N
H3C
O
N
N
[
dtc] -
[xan] -
[K-salt]2-
[dtmn]2-
N
C
C
O
O
O
N
C
N
S
S
F3C
CF2
S
O
F3C
CF3
3
C
N
N
O
O
C
C
O
N
N
N(Tf)2]-
[C F SO ]
-
[C(CN) ]-
[N(CN) ]-
[
4
9
3
3
2
O3S
SO3
N
Co
= BH
CH
N
N
N
N
N
N
N
NNH
=
Cu
N
S
S
N
H3C
CH3
SO3
SO3
N
O3S
SO3
[
CoCB]-
[CuPc]4-
[ThY]2-
Fig. 1. Anion structures for new phosphonium RTILs.
interesting due to its r-aromaticity, as well as copper
phthalocyaninetetrasulfonate (CuPc) and thiazolium
yellow G (ThY), both of which are common dyes.
the large side chains around the phosphonium cation
interferes with electrostatic interactions between the ani-
ons and the cation – which is generally localized on the
phosphorus – allowing for greater impact of the proper-
ties of the anion itself, as compared to salts with more
strongly interacting smaller cations. Naturally, there
are disadvantages as well, such as very high viscosities,
poorly understood solubility properties, etc. – all of
which may potentially be optimized through functional-
ization of the cations alkyl/aryl components.
Although the quaternary phosphonium ILs have not
received the attention and extensive resources that the
more common imidazolium and other nitrogen-contain-
ing ILs do, they have been used extensively in several
areas of chemistry for many years, particularly catalysis
and phase transfer reactions and extractions [6]. Addi-
tionally, they have been applied to materials as well,
such as liquid crystals [7].
+
The tetraalkylphosphonium salts, [PR R R R ] ,
4
1
2
3
were chosen for a number of features which differentiate
them from other common RTIL cations. The numerous
possible permutations that can be synthesized using var-
ious alkyl and aryl side chains increases the fine tunabil-
ity of the resulting properties, Table 1, such as their
melting/glass transitions or viscosities. The phosphoni-
ums lack the weakly acidic ring protons of the more
common imidazolium cations – a limitation when
strongly basic anions are used, such as the thiolates
mentioned above. Also, the quaternary phosphonium
salts have been reported to be more stable to increased
temperatures for longer periods of time than nitrogen-
based cations [5]. With materials aspects in mind, they
are additionally attractive because the steric bulk of
Table 1
Physical properties of [PR
= R = R
Ethyl
1 2 3 4
R R R ][X] salts
ꢀ
a
b
R
1
2
3
R
4
X
Solid/liquid
Trade name
Ethyl
n-Butyl
Br
Cl
Cl
Cl
Cl
Br
Cl
I
Solid (Tmp > 300 ꢁC)
Solid (Tmp = 87 ꢁC)
Solid (Tmp = 56 ꢁC)
Liquid
CYPHOS 422
n-Butyl
n-Butyl
n-Pentyl
n-Hexyl
n-Hexyl
n-Octyl
Isobutyl
a
CYPHOS IL 164
CYPHOS IL 167
CYPHOS IL 127
CYPHOS IL 101
CYPHOS IL 102
CYPHOS IL 128
CYPHOS IL 205
n-Tetradecyl
n-Tetradecyl
n-Tetradecyl
n-Tetradecyl
n-Tetradecyl
n-Octyl
Liquid (T
Liquid
g
= ꢀ56 ꢁC)
Liquid
Liquid
State at room temperature.
Cytec specialty chemicals.
b

2538
R.E. Del Sesto et al. / Journal of Organometallic Chemistry 690 (2005) 2536–2542
A number of these quaternary phosphonium salts
solved in chloroform, and added to an aqueous solution
of the anion salt. In cases where the anion salt was not
water-soluble, e.g., the Na (dtmn), the metathesis reac-
with desired anions were synthesized, with nearly all
being liquid at and well below room temperature, and
stable up to 400 ꢁC in some cases. Their chemical and
physical properties were measured, and while the anions
do dictate the major physical properties such as optical
and magnetic features, there appears to be slight corre-
lations between the cation size/structure with their vis-
cosity and glass transition temperatures.
2
tion was carried out in pure EtOH, the resulting solid
byproducts (NaCl) filtered, the solvent evaporated,
and then the product redissolved in chloroform for
washing and purifying. Any remaining solvent and
traces of water could be removed by heating the sample
in vacuo to 90–120 ꢁC. All of the anions discussed here
were incorporated into salts with the trihexyltetradecyl-
+
phosphonium cation, [PC C C C ] . The dtmn and
6 6 14
6
NTf2 anions were also made as salts with
+
+
+
2
. Results and discussion
[PC C C C ] , [PC C C C ] , [PC C C C ] , and
4 4 4 14 4 4 4 4 8 8 8 8
+
[
PPh ] cations.
4
2
.1. Synthesis of RTILs
The amount of residual alkali metal and chloride in
the resulting liquids was determined through ICP–OES
measurements. As most of the [PR R R R ] RTILs are
The sources of the anionic components of the RTILs
1
2
3
4
were alkali metal salts, and those which are not commer-
cially available could generally be made by known syn-
theses. The salts of dtmn and the K-salt, could be made
not water-soluble, they were dissolved in minimal
amount of chloroform and extracted with 20 ml of
20% nitric acid to try to dissolve as much of the RTIL
components in the aqueous layer as possible, and in
doing so should extract any alkali or chloride salts pre-
sent in the RTIL. In all cases, less than 100 ppm of
either impurity was found. The minute chloride impu-
rities are also confirmed through chloride elemental
analyses on a few selected ILs, which included both
the purely organic ILs and those with inorganic com-
plexes. In all samples, the chloride levels were below
the detection limits of the instrument – less than
0.2% by mass, which even for the lowest molecular
weight IL calculates to be less than 5 mol% (5 mol
Cl per 100 mol IL). This indicates that the samples
are pure, and are not solutions or binary salts of the
two starting materials.
via simple nucleophilic additions to CS , as seen in
2
Scheme 1. The metal complexes are also easily produced
from metal chloride salts reacted with an excess of the
salts of the desired ligands. These syntheses are generally
higher yield reactions and use fairly inexpensive starting
materials – desirable qualities should a large scale-up
process be considered.
Simple metathesis reactions of these alkali metal salts
of the desired anions with the appropriate phospho-
nium halide, [PR R R R ]X, in an organic solvent or
1
2
3
4
organic/aqueous system, results in the alkali halide pre-
cipitating out or extracted into the aqueous layer, and
xꢀ
the [PR R R R ] [anion] left in the organic layer, as
2
1
3
4 x
with the [PR R R R ] [dtmn] salt shown in Scheme 2.
1
2
3
4 2
Washing the organic layer several times with water gen-
erally removes the alkali halide and excess starting
materials. In most cases, the [PR R R R ]X was dis-
Water content of the RTILs was determined using
Karl–Fischer titrations. Not all of the salts could be
tested, however, due to low solubilities in the methanol
solvent used in the titrations. Most of the thiol-contain-
ing ILs were tested before and after drying, including the
xan, dtc, dtmn, and K-salt salts. Before drying, these
RTILs had water content ranging from 0.75 to 1.5 water
molecules per formula unit – fairly high presumably due
1
2
3
4
S
C
S
S
S
Nu
+
Nu
C
ꢀ
to strong [R–S ]–[H–OH] interactions. After drying at
9
0 ꢁC under vacuum for several hours, these dropped
to less than 0.05 waters per formula unit (less than
.1% by mass).
0
Scheme 1.
Na
S
S
C
Na
S
C
S
C
[
PR R R R ]+X-
C
C
C C
EtOH
1
2
3
4
+
[PR R R R ]
1 2 3 4 2
+ NaCl
C
N
N
N
N
[
dtmn]
Scheme 2.

R.E. Del Sesto et al. / Journal of Organometallic Chemistry 690 (2005) 2536–2542
2539
2
.2. Thermal properties
All of the materials presented exhibit glass transitions
occurring well below room temperature as seen in DSC
measurements, Table 2, with the exception of the ThY
salt, which is a glass a room temperature with its transi-
tion at +65 ꢁC. Some do show significant transition
enthalpies as seen in a true freezing/melting point, e.g.
the [PC C C C ][xan], but most only show changes in
6
6 6 14
their enthalpy at their transitions, Fig. 2. Correlations
of the T Õs amongst the [PC C C C ] salts with varying
g
6 6 6 14
anions is difficult to establish due to the charges of the
anions, and thus different number of cations necessary
for charge balance. Other factors that most likely con-
tribute to the T Õs and physical properties are morphol-
Fig. 2. DSC for: (a) [PC C C C ][xan], and (b) [PC C C C14][CoCB]
6
g
6
6
14
6 6 6
RTILs.
ogy (planar versus nonplanar) and electrostatic
distributions of the anions, which also contribute to
packing efficiency of the salts.
In a series of common anions with the different phos-
phonium cations, a slight trend can be observed in the
T versus size of the cation, Table 3. This trend is more
the [PC C C C ], and less symmetry, such as
8 8 8 8
[PC C C C ], results in RTILs with much lower T Õs.
4
4
4
14
g
As salts, they have no readily apparent vapor pres-
sure, thus their upper temperature limit is typically
determined by their thermal decomposition tempera-
tures, which generally occurs well below boiling. The
carbamate and xanthate materials, [PC C C C ][dtc]
g
easily seen in the [PR R R R ] [dtmn] salts, where
1
2
3
4 2
cation size is [PC C C C ] = [PC C C C ] > [PC C -
14
6
6
6
8
8
8
8
4 4
C C ] > [PC C C C ]. This trend is not necessarily lin-
4
14
4 4 4 4
ear, again due to the morphology of the cation, as well
as the symmetry and packing efficiency. The cation
6
6 6 14
and [PC C C C ][xan], tend to decompose at lower
6 6 14
6
[
PC C C C ] approaches the limit for creating RTILs
4
temperatures – around 250 ꢁC – whereas most others
are stable up to 400–450 ꢁC based on TGA experiments,
Table 2 (decomposition in these cases are the onset tem-
4 4 4
due to its much smaller size and higher symmetry than
the other cations. Introducing larger cations, such as
Table 2
xꢀ
Physical properties of [PC
6
C
6
C
6
C
14
]
x
[anion] RTILs
ꢀ1
a
b
Density (g cm
ꢀ3
b
c
[
Anion]
FW (g mol
)
T
g
(ꢁC)
T
dec (ꢁC)
)
Viscosity (g, cP)
kmax (nm)
ꢀ
ꢀ
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
dtc]
xan]
K-salt]
631
590
ꢀ77
ꢀ70
ꢀ79
ꢀ71
ꢀ65
ꢀ70
ꢀ76
ꢀ68
ꢀ65
ꢀ67
ꢀ70
ꢀ72
ꢀ70
ꢀ60
ꢀ70
ꢀ71
ꢀ45
+65
255
290
350
360
370
380
400
390
415
395
375
0.942
0.920
0.960
1470
1480
560
305
345
d
2
ꢀ
1101
1106
1198
1198
782
396
394
2
ꢀ
dtmn]
Co(dtmn)
0.942
e
4780
e
2
ꢀ
2
]
460
2
ꢀ
Ni(dtmn)
ꢀ
2
]
0.994
1.080
1.079
0.901
0.904
0.907
0.963
1.02
6480
450
805
320
490
310
2436
660
760
5790
475
NTf
C
2
]
NA
NA
NA
NA
618
ꢀ
4
F
9
SO
3
]
763
ꢀ
C(CN)
N(CN)
3
]
585
549
ꢀ
2
]
2ꢀ
Co(N(CN)
]
2
2
)
4
]
1289
1257
1447
2228
2144
806
ꢀ
Co(NCS)
4
405
f
626
628
2
ꢀ
Co(NCSe)
4
]
2
ꢀ
Ni(NCS)
6
]
4
380
f
0.902
0.942
346
355/410
450
ꢀ
Fe(CN)
6
]
ꢀ
CoCB]
CuPc(SO
420
400
395
1.00
e
3702
e
4
ꢀ
3
)
4
]
2816
1615
598/698
405
2ꢀ
e
e
ThY]
a
T
dec is decomposition onset temperature.
Density and viscosity reported for room temperature (20 ꢁC).
UV/Vis from neat samples with pathlength ꢁ10 lm.
b
c
d
e
f
T
g
mp, not T .
Viscosity too large to measure at RT.
Not measured due to toxicity.

2540
R.E. Del Sesto et al. / Journal of Organometallic Chemistry 690 (2005) 2536–2542
Table 3
more robust anions tend to decompose at much higher
temperatures, at which it appears the cation stability be-
comes an issue near 400 ꢁC.
Physical properties of [PR
RTILs
1
R
2
R
3
R
4
][NTf
2
1 2 3 4 2
] and [PR R R R ] [dtmn]
a
ꢀ3
a
Cation
[PC
T
g
(ꢁC) Density (g cm
)
Viscosity (g, cP)
NTf
2
6 6
C
C
6
C
14
]
ꢀ76
ꢀ60
ꢀ60
+65
1.080
1.071
1.112
450
418
2.3. Physical properties
[
[
[
[
PC
PC
PC
PPh
8
C
8
C
8
C
8
]
4
C
4
C
4
C
14
]
464
b
b
The density of all of the materials show linear behav-
ior over changing temperature as expected, Fig. 4. All of
the RTILs, with the exception of [PC C C C ] [Co-
4
C
4
C
4
C
4
]
b
b
4
]
+160
6
6 6 14 2
dtmn [PC
C
6 6
C
6
C
14
]
ꢀ71
ꢀ69
ꢀ46
ꢀ35
+154
0.942
0.946
0.950
4780
5590
7480
(NCSe) ], [PC C C C ][C F SO ], and the NTf salts,
4 6 6 6 14 4 9 3 2
[
[
[
[
PC
PC
PC
PPh
8
C
8
C
8
C
8
]
show densities below 1.0, Table 2, presumably due to
a significant contribution from large alkyl chains on
the cation.
4
C
4
C
4
C
14
]
b
c
4
C
4
C
4
C
4
]
16,150
b
b
4
]
a
b
c
Density and viscosity reported for room temperature (20 ꢁC).
Not measured.
All RTILs presented show very high viscosities at
room temperature, ranging from 310 cP for
Measured at 30 ꢁC.
[
C ] [dtmn], Tables 2 and 3. The CuPc, ThY, and some
PC C C C ] [Co(N(CN) ) ] to >7000 cP [PC C C
6 6 6 14 2 2 4 4 4 4
1
4 2
dtmn salts could not be measured due to their extremely
viscous and waxy/glassy state at room temperature and
up to near 90 ꢁC – the temperature limit for the viscom-
eter. As with the glass transitions, varying cation size
with common anions also correlates with the viscosity
of the RTILs. This effect is shown for the dtmn and
peratures, which is the intersection of the baseline below
their decomposition temperature with the tangent to the
mass loss versus temperature plots in the TGA
experiment).
Thermal stability for some of the RTILs was deter-
mined through isothermal TGA experiments. An exam-
ple of the thiolates is the [PC C C C ][xan] salt, which
NTf salts in Table 3. Arrhenius plots of the viscosities
2
6
6 6 14
versus temperature of the RTILs show some deviation
from two-parameter exponential behavior, Fig. 5.
Therefore the viscosities were modeled with the Vogel–
Tammann–Fulcher (VTF) formula, Eq. (1)
shows moderate stability below 200 ꢁC, but decomposes
fairly rapidly at higher temperatures, Fig. 3(a). The
[
PC C C C ][CoCB] on the other hand, is fairly stable
6 6 6 14
up to 375 ꢁC, showing only 0.1%/min mass loss, Fig.
(b). This sample rapidly decomposes and outgases
ꢀ
ꢁ
3
B
ln g ¼ ln A þ
;
ð1Þ
around 400 ꢁC. The lack of stability of the thiolate based
salts seems to be due to decomposition of the anion,
T ꢀ T
where A is the frequency factor, B is the activation en-
g
most likely though facile loss of CS . The salts with
2
ergy, and T is the thermodynamic glass transition tem-
g
perature. The three-parameter fits to this model are in
Table 4 [8].
No trend is seen with viscosity or T with varying the
(
a)
1
11 55 00 ˚˚ CC
2 00 00 ˚˚ CC
g
2
anions with common cations, as expected due to various
8
6
2
1.1
4
1
.05
3
2
(
a)
3
0
1
1
(b)
c)
d)
(
9
0.95
(
9
(e)
0.9
4
(f)
8
(g)
0.85
(
b)
0
20
40
60
80
100
8
0
10
20
30
Temperature, ˚C
2
14] RTILs with anions: (a) NTf ,
Time, minutes
Fig. 4. Densities of [PC
(b) [Co(NCSe)
(g) [N(CN) ].
6
C
6
C
6
C
Fig. 3. Isothermal TGA for: (a) [PC
PC 14][CoCB] RTILs.
6
C
6
C
6
C
14][xan], and (b)
4
], (c) CoCB, (d) [Co(NCS) ], (e) dtmn, (f) xan, and
4
[
6
C
6
C
6
C
2

R.E. Del Sesto et al. / Journal of Organometallic Chemistry 690 (2005) 2536–2542
2541
9
8
7
6
5
4
3
2
a Z-scan setup to determine nonlinear index of refrac-
tion, n , and nonlinear absorptivity, a , with a laser
wavelength of 910 nm. Most of the samples did show
(
(
a)
b)
2
2
(c)
nonlinear behavior, with the largest nonlinear refraction
ꢀ14
(
d)
seen in the [PC C C C ] [dtmn] salt, n = ꢀ6.1 · 10
6
6
6
14 2
2
(
e)
2
ꢀ1
m W . This value is slightly less than values observed
for common NLO polymers and semiconductors
ꢀ12
2
m W ) [9], but greater than previously re-
ꢀ1
(ꢁ10
ported metal cluster compounds [10] and molecular or-
ꢀ
R.T.
16
2
m W ) [11]. The dtmn salt
ꢀ1
ganic compounds (<10
is approximately 20,000 times greater than the typical
carbon disulfide standard (CS2), which has an
0
.0027
0.0029
0.0031
0.0033
0.0035
1
/T, K-1
ꢀ18
2
m W . These results are only prelimi-
ꢀ1
n ꢁ 3 · 10
2
nary, as there appears to be a thermal contribution to
the nonlinear behavior at the 910 nm wavelength. Non-
linear absorptions were also observed with several of the
samples, but again linear absorptions – and therefore
thermal effects – appear to contribute to the behavior.
Fig. 5. Viscosities of [PC
b) dtc, (c) [Ni(NCS) ], (d) [N(CN)
fit (solid lines mostly obscured by the data points).
6 6 6
C C C14] RTILs with anions: (a) dtmn,
(
6
2
], and (e) NTf
2
, with VTF model
Table 4
xꢀ
VTF parameters for viscosities of [PC
using Eq. (1)
6
C
6
C
6
C
14
]
x
[anion]
RTILs
2
3. Conclusion
[
Anion]
A
B (K)
T
g
(K)
R
ꢀ
ꢀ
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
dtc]
xan]
K-salt]
0.0271
0.0227
0.0477
1531.1
1646.9
1294.8
1815.8
149.0
143.6
154.5
140.8
0.99992
0.99931
0.99988
0.99985
A number of phosphonium based RTILs have been
synthesized, with optical and magnetic applications of
these materials being the focus. The larger anions uti-
lized along with their greater charges allow for develop-
ment of massive ionic liquids which can exceed 2000
2
ꢀ
2
ꢀ
dtmn]
0.0210
a
2
ꢀ
Co(dtmn) ]
2
2
ꢀ
a
Ni(dtmn)
ꢀ
2
]
NTf
C
2
]
0.0427
a
1327.1
144.9
0.99981
ꢀ
1
ꢀ
g mol formula weight. Most of the materials are liquid
well below room temperature and stable up to 400 ꢁC,
and those which are glass at RT are still potentially
4
F
9
SO
3
]
ꢀ
C(CN)
N(CN)
3
]
0.0741
0.0423
0.0734
0.0156
0.0627
0.0347
0.0024
1108.0
1374.4
1132.7
1983.8
1311.3
1467.5
2704.4
1600.6
161.5
145.8
158.5
127.3
152.3
145.8
110.9
155.8
0.99985
0.99992
0.99992
0.99994
0.99995
0.99993
0.99932
0.99961
ꢀ
2
]
2ꢀ
Co(N(CN) ) ]
2 4
interesting due to their glassy state. In fact, the phospho-
+
nium cation [PC C C C ] actually results in an RTIL
2ꢀ
Co(NCS)
4
]
2
ꢀ
6 6 6 14
Co(NCSe)
4
]
2
ꢀ
in conjunction with every anion attempted, with the
exception of the ThY salt. Although viscosities are
rather high compared to the more common RTILs, they
still have flow, and the viscosities can be tuned through
modification of the cation size and symmetry. Other
properties such as their color (absorption) can be ad-
justed through the use of the various conjugated organic
anions as well as the inorganic metal–ligand complex
anions. Additionally, preliminary experiments do show
nonlinear refractive indices in most of the samples,
and nonlinear absorption processes in several samples,
suggesting these new ILs as potential novel optical
materials.
Ni(NCS) ]
6
4
ꢀ
Fe(CN) ]
6
ꢀ
CoCB]
ꢀ
3 4
CuPc(SO ) ]
0.0308
b
4
2ꢀ
b
ThY]
a
Could not fit data with VTF or other models.
Wax/solid – did not obtain viscosity data.
b
other influences involved. However, the anions do dic-
tate other properties of the salts, particularly the color.
While some are colorless, such as the NTf2 and
C F SO , the conjugated thiolates are deep orange or
4
9
3
red in color, Table 2. The colors of the metal complexes
are determined by the spectrochemical series, which is a
result of the metals and ligands used. It is therefore pos-
sible to develop these RTILs to absorb in any region of
the UV, Visible and NIR region based on the anions
utilized.
4. Experimental
4.1. General techniques and reagents
The magnetic properties of several of the Co(II) con-
taining ILs were measured with an Evans balance. All of
the Co(II) salts were paramagnetic, with room tempera-
ture (293 K) magnetic moments ranging from 3.3. to 3.7
lB – typical of high-spin (S = 3/2) Co(II) ions. Prelimi-
nary nonlinear optical measurements were done using
The alkali metal salts Na(dtc) (Fluka), Li(NTf ) (Ald-
2
rich), K(C F SO ) (Aldrich), Na[N(CN) ] (Fluka),
4
9
3
2
K[C(CN) ] (Strem), K [Fe(CN) ] (Strem), Na (CuPc)
3
4
6
4
(Aldrich), and Na (ThY) (ACROS) are all commer-
2
cially available, and were used as received. The

2542
R.E. Del Sesto et al. / Journal of Organometallic Chemistry 690 (2005) 2536–2542
[
PC C C C ]Cl, [PC C C C ]Cl, [PC C C C ]Cl, and
14
Acknowledgments
6
6
6
14
4
4
4
4 4 4 4
[
PC C C C ]Br salts were prepared by Cytec Specialty
8
8 8 8
Chemicals, and dried under vacuum at 90 ꢁC before
being used.
This work was performed while R.E.D.S. held a
National Research Council Associateship Award at
USAFA. This research was funded by the Air Force Of-
fice of Scientific Research and the Air Force Research
Laboratory. The authors thank Rebecca Chamberlin
at LANL for the Cs[CoCB] salt.
For those that were not available, the K-salt [12],
Na(xan) [13], Na (dtmn) and Na [M(dtmn) ] [14] were
2
2
2
synthesized by known methods. The Cs(CoCB) salt
was provided by Los Alamos National Laboratory.
The metal complexes K [Co(NCS/Se) ], K [Ni-
2
4
4
(
substitution by reacting the MCl salt with excess so-
dium or potassium salts of the desired ligands in water.
The aqueous solutions were used without isolation of
the salt.
NCS) ], and Na [Co(N(CN) ) ] were made by simple
6 2 2 4
2
References
[1] P. Wasserscheid, T. Welton (Eds.), Ionic Liquids in Synthesis,
Wiley–VCH, 2003.
[2] J.S. Wilkes, J. Mol. Catal. A. 214 (2004) 11.
Viscosities of the samples were measured on a Cam-
bridge Applied Systems ViscoLab 4000 Viscometer
and measured over the range of 20–90 ꢁC. The densities
were determined by a Mettler–Toledo DE40 Density
Meter at 20, 45, 60 and 80 ꢁC. Absorptions were meas-
ured on an HP 8452A diode array spectrophotometer,
using neat samples in a 0.01 mm (10 lm) etched quartz
cuvette. Thermal characterization was carried out using
a TA Instruments Q100 DSC and SDT 2960 SDT-TGA.
Alakli metal and halide impurities were determined
using a Varian VistaPro ICP–OES with CCD detector.
Water content of the RTILs was measured with a
Mettler Toledo DL38 Karl–Fischer titrator.
[
3] M.E. Van Valkenburg, R.L. Vaughn, M. Williams, J.S. Wilkes, in:
H.C. Delong, R.W. Bradshaw, M. Matsunaga, G.R. Stafford,
P.C. Trulove (Eds.), Molten Salts XIII, Proc. Electrochem. Soc.,
PV 2002–19 (2002) 112–123.
[4] (a) R.E. Del Sesto, D.S. Dudis, F. Ghebremichael, N.E. Heimer,
T.K.C. Low, J.S. Wilkes, A.T. Yeates, Proc. SPIE 5212 (2003)
2
92;
b) A.T. Yeates, D.S. Dudis, J.S. Wilkes, Proc. SPIE 4106 (2000)
34.
(
3
[
[
5] M.O. Wolff, K.M. Alexander, G. Belder, Chim. Oggi (2000) 9.
6] (a) N. Karodia, S. Guise, C. Newlands, J. Andersen, Chem.
Commun. (1998) 2341;
(
(
(
b) D.E. Kaufmann, M. Nouroozian, H. Henze, Chem. Syn. Lett.
1996) 1091;
c) J. McNulty, A. Capretta, J. Wilson, J. Dyck, G. Adjabeng, A.
Robertson, Chem. Commun. (2002) 1986.
7] (a) H. Chen, D.C. Kwait, S. G o¨ nen, B.T. Weslowski, D.J.
Abdallah, R.G. Weiss, Chem. Mater. 14 (2002) 4063;
4
.2. Metathesis reactions to synthesize RTILs
[
(
b) D.J. Abdallah, A. Robertson, H. Hsiu-Fu, R.G. Weiss, J.
Am. Chem. Soc. 122 (2000) 3053.
To make the [PR R R R ] phosphonium RTILs
1
2
3
4
from the above alkali metal salts (all except the dtc,
dtmn, [C F SO ], and CoCB RTILs), metathesis reac-
tions were carried out in which the [PR R R R ]Cl (ex-
[
[
8] Viscosities were fit with SigmaPlot 8.0 Regression Wizard.
9] U. Gubler, C. Bosshard, Adv. Polym. Sci. 158 (2002) 123.
4
9
3
1
2
3
4
[10] (a) H. Hou, X. Meng, Y. Song, Y. Fan, Y. Zhu, H. Lu, C. Du,
W. Shao, Inorg. Chem. 41 (2002) 4068;
cept in the case of the [PC C C C ], which is only
8
8 8 8
(
b) H. Hou, H.G. Ang, S.G. Ang, Y. Fan, M.K.M. Low, W. Ji,
available as the Br salt) was dissolved in excess CHCl3
and added to an aqueous solution of the above salts.
The solution was rapidly stirred for 24 h, and the organ-
ic layer containing the RTIL product was separated and
washed several times with water. The solvent was re-
moved under vacuum, and the resulting RTIL dried at
Y.W. Lee, Phys Chem. Chem Phys. 1 (1999) 3145.
11] (a) e.g. C. Zhan, W. Xu, D. Zhang, D. Li, Z. Lu, Y. Nie, D. Zhu,
J. Mater. Chem. 12 (2002) 2945;
[
(
b) E.M. Garc ´ı a-Frutos, S.M. OÕFlaherty, E.M. Maya, G. de la
Torre, W. Blau, P. V a´ zquez, T. Torres, J. Mater. Chem. 13 (2003)
49;
c) Z. Dai, X. Yue, B. Peng, Q. Yang, X. Liu, P. Ye, Chem. Phys.
Lett. 317 (2000) 9;
d) H.I. Elim, W. Ji, G.C. Meng, J. Ouyang, S.H. Goh, Chem.
Phys. Lett. 366 (2002) 224.
[12] (a) K.A. Jensen, O. Buchardt, C. Lohse, Acta Chem. Scand. 21
7
(
9
0–120 ꢁC, typically for 12 h.
The dtc and dtmn RTILs were synthesized by mixing
(
the [PR R R R ]Cl and respective alkali metal salt in
1
2
3
4
ethanol. The NaCl precipitate was filtered, the ethanol
evaporated, and the resulting crude RTIL purified by
redissolving in CHCl and washing with water as above.
(
(
1967) 2797;
b) E. Freund, Ber. B 52 (1919) 542.
3
[
13] The salt Na(xan) was carried out in a procedure similar to that of
The [C F SO ] and CoCB RTILs were made in a similar
4
9
3
the K-salt, in which NaOH and CS
2
were added in a 1:1 ratio to a
fashion except CHCl and THF were the reaction sol-
3
methanol solvent.
[14] A. Davison, R.H. Holm, Inorg. Synth. 10 (1967) 8.
vents, respectively.