Ion size approaching the Bjerrum length in solvents of low polarity by dendritic encapsulation

Article abstract of DOI:10.1021/ma402137x

The Bjerrum length is approached in a low polarity solvent by encapsulating, both, a borate anion and a phosphonium cation in a rigid lipophilic dendrimer shell. In addition the cation size is varied by 34-fold. We thus obtain superweak ions with potential applications in catalytic processes.

Full text of DOI:10.1021/ma402137x

Article
pubs.acs.org/Macromolecules
Ion Size Approaching the Bjerrum Length in Solvents of Low Polarity
by Dendritic Encapsulation
Ralf Moritz,† George Zardalidis,‡ Hans-Jurgen Butt,† Manfred Wagner,† Klaus Mullen,†,*
̈
̈
and George Floudas‡,*
†Max Planck Institute for Polymer Research, D-55128 Mainz, Germany
‡Department of Physics, University of Ioannina, 45110 Ioannina, Greece
S
* Supporting Information
ABSTRACT: The Bjerrum length is approached in a low polarity solvent by
encapsulating, both, a borate anion and a phosphonium cation in a rigid lipophilic
dendrimer shell. In addition the cation size is varied by 34-fold. We thus obtain
superweak ions with potential applications in catalytic processes.
However, in liquids of low polarity, such as toluene or even
THF, the Bjerrum length is 20.4 and 7.4 nm, respectively. Ion
dissociation in such solvents is limited unless the ion size
approaches the “escape distance” set by λB.
I. INTRODUCTION
Traditionally, charge creation and charge transport in nonpolar
liquids have been of interest to the electric power and
petroleum processing industries1 and more recently to
applications related to reflective displays.2 In the petroleum
industry for example, trace compounds present in oily liquids
give rise to the buildup of unwanted charges along pipes, a
process known in petroleum handling as flow electrification.3
To avoid electrification, electrolytes are used that increase the
conductivity of the liquid and provide the means for the
dissipation of electric charge. It is thus of interest to explore
additional possibilities of increasing the conductivity in
nonpolar liquids. Ion dissociation in nonpolar solvents4 is
limited because of the Coulomb attractive energy, which for
two monovalent charge carriers is
Two methods have been proposed that result in stabilization
of electric charges in nonpolar electrolyte solutions against
neutralization. Both aim at encapsulating charges in large
structures such as micelles or macroions. The former approach
is based on surfactants that form reverse micelles able to
stabilize dissociated ions by encapsulation in a structure with
dimensions comparable to the Bjerrum length.1,2 The second
approach, that is employed here, is based on the dendritic
encapsulation of ions. In the present study, we report on the
ion dissociation and transport properties of a series of salts
comprising the same rigidly dendronized borate anion (BF-
G1−) with cations bearing different sizes ranging from ∼0.2 nm
(Li+) to rigidly dendronized phosphonium cations of the first
(P-G1+), second (P-G2+), and third generation (P-G3+) with
respective diameters of 2.9, 4.6, and 6.3 nm in THF (ε = 7.58 at
298 K). To our knowledge, these are the largest cations
reported in literature. The systematic variation of the cation size
in conjunction with the large borate anion provides the means
of approaching λB in solvents of low polarity resulting in
superweak ions. Furthermore, we investigate the effect of
valency on ion dissociation.
e2
EC =
4πε0εSrc
(1)
Here, e is the elementary charge, ε0 the permittivity of free
space, rc the distance separating a point-like cation from a
point-like anion, and εS the dielectric permittivity of the
surrounding medium. The “escape distance” from the Coulomb
energy is set by the Bjerrum length,5 λB = e2/4πε0εSkBT, giving
the characteristic separation between two ions at which
Coulombic interactions are balanced by the thermal energy.
At distances below λB, electrostatic interactions are important,
whereas at longer distances thermal fluctuations prevail and
justify replacing discrete ion−ion Coulomb forces with a
continuum mean-field approach. In polar liquids, such as water,
the Bjerrum length is 0.7 nm, i.e., only a few molecular lengths.
Received: October 16, 2013
Revised: December 8, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ma402137x | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Figure 1 gives a schematic of the effects of solvent polarity
and ion size on the Coulomb potential. Increasing the dielectric
II. EXPERIMENTAL SECTION
Synthesis and Characterization. Figures 2 and 3 give the
synthetic root to dendronized phosphonium compounds of the
first and higher generations, respectively.
Figure 1. Schematic of the dendritic molecules as a function of the
cation size. All salts share the same anion, i.e., borate (BF-G1−).
Monovalent and divalent ions are shown, respectively, in blue and red.
(Center) Effective Coulomb energy (green solid line) showing the
effect of increasing dielectric permittivity of the medium and the effect
of increasing cation size (dashed lines).
Figure 2. Synthetic route to dendronized phosphonium compounds of
generation 1: (a) Mg, THF, PCl3 (2.0 M in DCM), room temperature,
overnight; (b) K2CO3 in MeOH, room temperature, 3 h; (c) BF3·
OEt2, tBuNO2, DCM, −20 °C; (d) toluene, 65 °C; (e)
tetraphenylcyclopentadienone and, respectively, AB2 building block,
o-xylene, 165 °C, 12 h; (f) TBAF, THF, −20 °C, 2 h.
permittivity of the solvent raises the attractive part of the
potential, making the potential less-attractive and giving rise to
ion dissociation. Another means toward ion dissociation is by
ion encapsulation. The latter method effectively shifts the
repulsive part of the potential to length scales that approach the
Bjerrum length thus promoting ion separation
The idea rests on extensive research on increasing the size
and bulkiness of molecular anions as a means of promoting ion
separation. This led to a new class of compounds known as
weakly coordinating anions (WCAs) with potential applications
in electrochemistry, catalysis, polymerization, ionic liquids, and
battery technology.6−13 A step forward in this field has been a
novel synthetic strategy that resulted in a number of very large
and rigid molecular ions based on polyphenylene dendrimers.14
This facilitated a study of ion dissociation and transport of a
series of tetrabutylammonium salts of rigidly dendronized
borate anions as a function of anion size (with a 4-fold increase,
from 0.56 to 2.28 nm), solvent polarity and concentration.15 In
the present investigation both the cation and the anion are
encapsulated and made lipophilic by dendronization. In
addition the cation size is varied by 34-fold thus approaching
the Bjerrum length in THF. This approach resulted in the first
superweak cations. Details on the synthesis, characterization
and crystal packing of these phosphonium salts will be
described elsewhere in detail.16 Apart from the present system,
the Bjerrum length is an essential feature in other more
common charged polymers, such a polyelectrolytes, mixtures of
oppositely charged polyelectrolytes, polyampholytes, and
ionomers. The idea explored herein, namely dissociation via
ion encapsulation, could be expanded in other charged
polymers.
Details on the general procedures and materials and the
synthesis and characterization of the dendritic borates and
posphonium compounds are given in the Supporting
Information. Details on crystal packing of these phosphonium
salts will be described elsewhere in detail.14
Dielectric Spectroscopy (DS). Dielectric spectroscopy is our
method of choice, because of its inherent ability to provide
both the degree of ion dissociation and transport through the
measured dc-conductivity.17−21 Dielectric measurements were
made with a Novocontrol BDS1308 liquid sample cell within
the temperature range from 73.15 to 298.15 K at atmospheric
pressure, and for frequencies in the range from 10−2 to 107 Hz
using a Novocontrol high resolution alpha analyzer (details in
the Supporting Information). In all cases, gold-coated electro-
des were used. Different salt concentrations in THF in the
range from 10−5 to 10−1 M were used that allowed establishing
the linear concentration regime. Measurements reported here
refer to 0.001 M solutions in THF where the mean distance
between molecular ions is of the order of Bjerrum length
(Figure S1, Supporting Information). The refractive index of
the neat solvent and solutions was measured with an Abbe
refractometer at 298 K (THF, n = 1.4112; 1, n = 1.4115; 2, n =
1.4117; 10, n = 1.4120; 11, n = 1.4124; 12, n = 1.4135).
B
dx.doi.org/10.1021/ma402137x | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Figure 3. Synthetic route to higher generations of dendronized phosphonium compounds: (g) tetraphenylcyclopentadienone, o-xylene, 165 °C, 12
h; (h) AB2 building block, o-xylene, 165 °C, 36 h; (i) TBAF, THF, −20 °C, 2 h.
cations (Figure 1), the dc-conductivity of the fully dissociated
ions can be expressed as:σdc = p+μ+e + p−μ−e. Measured dc-
conductivities (with values of ∼10−5 S/cm) for the different
salts in THF conform to σdc = σo exp(−B/(T − T0), where σo is
the limited conductivity, B is the activation parameter for ion
transport, and T0 is the “ideal” glass temperature. The thus
extracted dc-conductivities at 298.15 K are plotted in Figure 5
III. RESULTS AND DISCUSSION
Figure 4 gives the temperature dependence of the dc-
conductivity in some selected borate salts bearing the same
Figure 4. Temperature dependence of the dc-conductivities in three
salts comprising of the three large phosphonium cations (squares, 10;
circles, 11; triangles, 12) and the borate anion at 0.001 M
concentration in THF. The vertical line gives the conductivity values
at 298 K. The lines are the result of fitting with the VFT equation.
Figure 5. Dc-conductivity plotted as a function of cation size at 298 K.
Red and blue symbols correspond to divalent and monovalent cations,
respectively. Filled symbols give the measured conductivities at 298 K
at 0.001 M concentration in THF. Open symbols give the respective
calculated conductivities assuming complete dissociation.
borate anion (BF-G1−), but different (large) phosphonium
cations of the first (10), second (11), and third generation
(12).
The dc-conductivity of an ion-containing medium is the sum
of the individual contributions of all charge carriers: σdc
=
as a function of cation size for the monovalent (blue symbols)
and divalent (red symbols) salts (the same data are plotted in a
linear scale in Figure S3, Supporting Information, for
comparison). Ionic radii of alkali metals was employed from
the literature22 whereas ionic radii for the larger cations and the
anion have been determined with the software Spartan (Wave
Σni=1piμiqi. Here, pi, μi, and qi are the number density, the
mobility, and the charge of the ith type of charge carrier,
respectively. An underlying assumption is that all charge
carriers move independently of each other. In the case of
monovalent charge carriers, i.e., the anion and respective
C
dx.doi.org/10.1021/ma402137x | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
function, INC) with semiempirical calculations. The concen-
tration dependence the dc-conductivity (Figure S1, Supporting
Information) clearly depends on the degree of ion association
within the investigated concentration range. The contribution
to the measured conductivity from the neat solvent is very small
(10−7 S/cm).
The reason that 9 (bis(triphenylphosphine)iminium borate)
has a higher degree of dissociation than expected is not clear at
present. It could be associated with a more effective shielding of
the charge. Despite this deviation of 9, a linear extrapolation in
the representation of Figure 6 predicts that a cation of about 7
nm in size would be largely dissociated from the borate anion.
This size is within reach of the Bjerrum length (7.4 nm) at this
temperature and in the same solvent. This similarity is, at first
sight, surprising for two reasons: first borate salts are not point
charges and second the borate anion core has a delocalized
charge due to the fluorine substitution.6,15 Nevertheless, none
of these factors seems to alter substantially the escape distance.
Hence, this synthetic approach that utilizes very large cations
leads to the formation of superweak anions in solvents of low
polarity in agreement with the theoretical predictions for the
“ideal”, i.e., point-like case.
Despite all measured conductivities being around 1 × 10−5
S/cm there are some trends. For the smallest cations 1−3 (Li+
BF-G1−, Na+ BF-G1− and K+ BF-G1−), increasing cation size
lowers the ionic conductivity whereas the opposite trend is
observed in 10, 11, and 12 (P-G1+ BF-G1−, P-G2+ BF-G1−, and
P-G3+ BF-G1−, respectively), bearing the large dendronized
phosphonium cations. This reflects the balance between ion
dissociation on one hand (promoted by the bulky ions) and
charge transport on the other (inhibited by the large ions). The
dashed lines in the figure give the respective calculated dc-
conductivities. The latter, for the monovalent case, can be
obtained from the mobility, μi = e/6πηri
In the divalent case, each cation is attracted by two separate
borate anions, thus increasing the calculated conductivity as in
this case:σdc = p+μ+Z+e + 2p−μ−e for fully dissociated ions and
μ+=Z+e/6πηr+ (Z+ = 2) resulting in
p e2
⎡
⎤
1
1
S
σ
=
+
+
calc
⎢
⎥
⎦
⎣
6πη r
r−
p 4e2
(2)
⎡
⎤
1
r
1
S
σ
=
+
+
calc
⎢
⎣
⎥
⎦
6πη
2r−
where ri is the ionic radii, ps is the total ion concentration from
stoichiometry (0.001 M), and η is the solvent viscosity (η[Pa·s]
= 2.1379 × 10−5 exp(910/T)). The measured dc-conductivity
differs from the calculated one, because the paired charges do
not contribute to the conduction mechanism.
The degree of ion dissociation can be extracted from the
ratio of the measured dc-conductivity and the calculated
conductivity that assumes complete ion dissociation, as a =
σmeas/σcalc, known as the Haven ratio (HR). This ratio is plotted
as a function of the cation size in Figure 6.
(3)
This gives rise to a lower degree of dissociation as compared to
the monovalent case (Figure 6). Nevertheless, the degree of
dissociation extrapolates reasonably to the theoretically
predicted escape distance from the Coulomb energy. The
latter, calculated for a divalent cation with a monovalent anion,
gives a characteristic distance of 14.8 nm (l* = 2lB).
Additional information on the state of ions can be obtained
by examining the temperature dependence of the dielectric
permittivity plotted in Figure 7. In polar liquids the dielectric
Figure 6. Degree of ion dissociation for borate solutions in THF
(concentration 0.001 M) at 298 K plotted as a function of cation size.
Blue and red symbols refer to monovalent and divalent cations,
respectively. Blue arrow gives the estimated Bjerrum length (lB) for
monovalent ions. Red arrow is the estimated escape distance (l*) for
the divalent case. Lines represent the result of linear fits.
Figure 7. Static dielectric permittivity plotted as a function of inverse
temperature for the neat solvent (squares) and for 0.001 M solutions
of compound 2 (circles) and 11 (triangles) in THF.
permittivity decreases with increasing temperature and its value
is related to the dipole moment of relaxing units as predicted by
Onsager.23 In evaluating the dipole moments in the solutions
we employed the following equation24
The systematic variation of the cation size from Li+ (diameter
0.18 nm) to the much larger dendronized phosphonium cations
of 10−12 (respective diameters of 2.9, 4.6, and 6.3 nm) follows
a power low dependence with an exponent ∼2/3. For the
cations with r+ ≪ r−, this suggest a dependence of the
measured conductivity on cation radius as σmeas ∼ (r+)−1/3. This
is actually the case for the cations of the alkali metals (Li+, Na+,
K+, Cs+).
2
N (ε
+ 2)(nTHF + 2)C
27kBTε0
εS(T) = μ2
A
THF
+ (εTHF − nTHF2) + ε∞
(4)
D
dx.doi.org/10.1021/ma402137x | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
where, μ is the dipole moment, C is the concentration (1 mol/
m3), εTHF and nTHF, are the measured permittivity and refractive
index of the solvent (THF), ε∞ is the high frequency limit of
the dielectric permittivity (obtained from the measured
refractive index as ε∞=n2), ε0 is the permittivity of vacuum,
kB is Boltzmann’s constant and NA is Avogandro’s number.
From the fit to eq4, assuming that all ions form pairs, we obtain
dipole moments in the range from 60 to 70 D for 2 and in the
range from 30 to 40 D for 10 and 11. These are reasonable
values in accord with the degree of ion dissociation discussed
earlier. For example, in 2 the dipole moment of an ion pair (μ =
e(r+ + r−)) is estimated as 74 D, and this values is in good
agreement with the experimental values since the majority of
ions are associated (Figure 6). In the larger phosphonium ions
the measured dipole momentunder the assumption that all
ions form pairsis below the calculated one for paired ions
(∼180 D in 11) since in this case most ions are dissociated
(Figure 6).
with potential applications in catalytic processes. Despite a high
degree of ion dissociation the effect of increasing cation size on
dc-conductivity is only moderate. The latter reflects the balance
between ion dissociation, which is promoted by the bulky ions,
and charge transport, which is inhibited by the large ions.
Finally, the effect of ion valency is to increase the characteristic
escape length.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and DOSY-NMR and dielectric spectroscopy. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: gfloudas@cc.uoi.gr, muellen@mpip-mainz.mpg.de.
Because of ion association, the diffusion coefficients
measured by DOSY-NMR (Dexp) at 298 K (Table S1,
Supporting Information) are not the diffusion coefficients of
the free ions but represent some average of the fully dissociated
and paired states.15 This is shown in Figure 8 that compares the
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support by DFG Priority Program SPP 1355, SFP
1459, and the SFB 635 is acknowledged. The current work was
supported by the Research unit on Dynamics and Thermody-
namics of the UoI cofinanced by the European Union and the
Greek state under NSRF 2007-2013 (Region of Epirus, call 18).
This work was cofinanced by the E.U. European Social Fund
and the Greek Ministry of Development GSRT in the
framework of the program THALIS.
REFERENCES
■
(1) Morrison, I. D. Colloid Surf. A: Physicochem. Eng. Asp. 1993, 71,
1−37.
(2) Parent, M. E.; Yang, J.; Jeon, Y.; Toney, M. F.; Zhou, Z.-L.;
Henzer, D. Langmuir 2011, 27, 11845.
(3) van der Hoeven, P. H. C.; Lyklema, J. Adv. Colloid Interface Sci.
1992, 42, 205−277.
(4) Klinkenberg, A.; der Minne, J. L. Electrostatics in the Petroleum
Industry; Elsevier: Amsterdam, 1958.
Figure 8. Measured (T = 298 K) DOSY-NMR diffusion coefficients
for the anion (D−) (red symbols) and the cation (D+) (blue symbols)
plotted as a function of the cation size. Calculated anion (red) and
cation (blue) diffusion coefficients corresponding to unassociated ions.
(5) Bjerrum, N. Kgl. Dan. Vidensk. Selsk. Mat.-fys. Medd. 1926, 7, 1.
(6) Krossing, I.; Raabe, I. Angew. Chem., Int. Ed. 2004, 43, 2066.
(7) Barbarich, T. J.; Handy, S. T.; Miller, S. M.; Anderson, O. P.;
Grieco, P. A.; Strauss, S. H. Organometallics 1996, 15, 3776.
(8) Chen, E. Y.-X.; Marks, T. J. Chem. Rev. 2000, 100, 1391.
(9) Raabe, I.; Wagner, K.; Guttsche, K.; Wang, M.; Gratzel, M.;
̈
measured diffusion coefficients (symbols) for the different
anion sizes/generations at 298 K with the calculated ones (from
the Stokes−Einstein relation, Di = kBT/6πηri) (lines)
corresponding to unassociated ions. For the larger cations,
the measured and calculated diffusion coefficients are in
reasonable agreement, since the degree of ion association is
low. For the smaller cations, however, the NMR diffusion
coefficients underestimate/overestimate the respective cation/
anion diffusion coefficients with regard to the free ions as a
result of the stronger ion association.
Santiso-Quinones, G.; Krossing, I. Chem.Eur. J. 2009, 15, 1966.
̃
(10) Stewart, M. P.; Paradee, L. M.; Raabe, I.; Trapp, N.; Slattery, J.
S.; Krossing, I.; Geiger, W. E. J. Fluorine Chem. 2010, 131, 1091.
(11) Geiger, W. E.; Barrier
(12) Krossing, I.; Reisinger, A. Coord. Chem. Rev. 2006, 250, 2721.
(13) Bosmann, A.; Francio, G.; Janssen, E.; Solinas, M.; Leitner, W.;
Wasserscheid, P. Angew. Chem., Int. Ed. 2001, 40, 2697.
̀
e, F. Acc. Chem. Res. 2010, 43, 1030.
̀
̈
(14) Turp, D.; Wagner, M.; Enkelmann, V.; Mullen, K. Angew. Chem.,
̈
̈
Int. Ed. 2011, 50, 4962−4965.
(15) Mpoukouvalas, K.; Turp, D.; Wagner, M.; Mullen, K.; Butt, H.-
J.; Floudas, G. J. Phys. Chem. B 2011, 115, 5801−5806.
(16) Moritz, R.; Baumgarten, M.; Mullen, K. Manuscript in
̈
̈
̈
IV. CONCLUSIONS
preparation.
The synthesis of extremely large, bulky, and stable cations and
anions resulted in molecularly defined ions that provide a route
toward the goal of truly noncoordinating ions in solvents of low
polarity. Employing a large borate anion and by systematically
varying the size of the cation we approached the theoretical
Bjerrum length. This procedure results in superweak cations
(17) Broandband Dielectric Spectroscopy; Kremer, F., Schonhals, A.,
̈
Eds.; Springer-Verlag: Berlin, 2003.
(18) Floudas, G. Dielectric Spectroscopy. In Polymer Science: A
Comprehensive Reference; Matyjaszewski, K., Moller, M., Eds.; Elsevier
̈
BV: Amsterdam, 2012; Vol. 2.32, pp 825−845.
(19) Dyre, J. C. J. Appl. Phys. 1988, 64, 2456−2468.
E
dx.doi.org/10.1021/ma402137x | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(20) Fragiadakis, D.; Dou, S.; Colby, R. H.; Runt, J. J. Chem. Phys.
2009, 130, 064907.
(21) Serghei, A.; Tress, M.; Sangoro, J. R.; Kremer, F. Phys. Rev. B.
2009, 80, 184301.
(22) Shannon, R. D. Acta Crystallogr. 1976, A32, 751.
(23) Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486.
(24) Guggenheim, E. A. Trans. Faraday Soc. 1949, 45, 714.
F
dx.doi.org/10.1021/ma402137x | Macromolecules XXXX, XXX, XXX−XXX