Iodide-conducting polymer electrolytes based on poly-ethylene glycol and MgI2: Synthesis and structural characterization

Article abstract of DOI:10.1016/j.electacta.2011.07.133

A major obstacle for a viable technological development of dye sensitized solar cells (DSSCs) is still the synthesis of a high performance iodide-conducting polymer electrolyte. Here we present a series of eight electrolytic complexes with formula PEG1000/(MgI₂)_x(I ₂)_y (0.0038 ≤ x ≤ 0.5801, 0 ≤ y ≤ 0.0636). The synthesis involves the preparation of a disordered form of MgI₂ by a metallorganic route, which enables us to dissolve high amounts of salt in the chosen polymer host. The thermal analysis of the resulting polymer electrolytes was performed using modulated differential scanning calorimetry measurements. Vibrational studies were carried out using medium FT-IR, far FT-IR and FT-Raman. The variation of the CO and OH stretching modes in the medium infrared, as a function of the mole-to-mole ratio nMg/nO, was investigated by Gaussian decomposition to provide insight into the polymer-polymer and salt-polymer interactions in these materials. The FT-Raman spectra confirmed and complemented the vibrational assignment. The conductivity study of these systems was performed by electrical spectroscopy in the frequency interval 10 mHz-10 MHz. The direct current conductivity (σDC) profiles versus the reciprocal temperature exhibited a Vo?gel-Tamman-Fu?lcher (VTF) behavior. The best σDC at 50°C was 5 × 10^-5 S cm^-1. The overall results indicate the presence of bivalent, monovalent and neutral species, Mg2+, [MgI]+ and MgI₂, respectively, which participate in the conduction process. These results are consistent with what was previously observed in PEG400-based systems doped with δ-MgCl2. The presence of at least one Mg site containing a distribution in parameters was observed using ²⁵Mg solid state magic angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy. The site has been assigned to a Mg complex involving the coordination by oxygen atoms of the polymer backbone.

Full text of DOI:10.1016/j.electacta.2011.07.133

Electrochimica Acta 57 (2011) 112–122
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Iodide-conducting polymer electrolytes based on poly-ethylene glycol and MgI2:
Synthesis and structural characterization
Michele Vittadelloa,∗ , David I. Waxmanb , Paul J. Siderisc , Zhehong Gand , Keti Vezzùe , Enrico Negrof ,
Ahmad Safarib, Steve G. Greenbaumc, Vito Di Notof
a PECS Department, Medgar Evers College and Graduate Center of the City University of New York, 1638 Bedford Avenue, Brooklyn, NY 11225, USA
b Department of Materials Science and Engineering, Rutgers, The State University of New Jersey, 607 Taylor Road, Piscataway, NJ 08854-8065, USA
c Department of Physics, Hunter College and Graduate Center of the City University of New York, 695 Park Avenue, New York, NY 10021, USA
d National High Magnetic Field Laboratory, 1800 East Paul Dirac Drive, Tallahassee, FL 32310, USA
e Dipartimento di Chimica, Università Ca’ Foscari di Venezia, Dorsoduro, Calle Larga S. Marta 2137, 30123 Venezia, Italy
f
Dipartimento di Scienze Chimiche, Università di Padova, Via Marzolo 1, 35131 Padova (Pd), Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A major obstacle for a viable technological development of dye sensitized solar cells (DSSCs) is still
the synthesis of a high performance iodide-conducting polymer electrolyte. Here we present a series
of eight electrolytic complexes with formula PEG1000/(MgI2)x(I2)y (0.0038 ≤ x ≤ 0.5801, 0 ≤ y ≤ 0.0636).
The synthesis involves the preparation of a disordered form of MgI2 by a metallorganic route, which
enables us to dissolve high amounts of salt in the chosen polymer host. The thermal analysis of the
resulting polymer electrolytes was performed using modulated differential scanning calorimetry mea-
surements. Vibrational studies were carried out using medium FT-IR, far FT-IR and FT-Raman. The
variation of the CO and OH stretching modes in the medium infrared, as a function of the mole-to-mole
ratio nMg/nO, was investigated by Gaussian decomposition to provide insight into the polymer–polymer
and salt–polymer interactions in these materials. The FT-Raman spectra confirmed and complemented
the vibrational assignment. The conductivity study of these systems was performed by electrical spec-
troscopy in the frequency interval 10 mHz–10 MHz. The direct current conductivity (ꢀDC) profiles versus
the reciprocal temperature exhibited a Vögel-Tamman-Fülcher (VTF) behavior. The best ꢀDC at 50 ◦C was
5 × 10−5 S cm−1. The overall results indicate the presence of bivalent, monovalent and neutral species,
Mg2+, [MgI]+ and MgI2, respectively, which participate in the conduction process. These results are con-
sistent with what was previously observed in PEG400-based systems doped with ␦-MgCl2. The presence
of at least one Mg site containing a distribution in parameters was observed using 25Mg solid state magic
angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy. The site has been assigned to a Mg
complex involving the coordination by oxygen atoms of the polymer backbone.
Received 17 December 2010
Received in revised form 27 July 2011
Accepted 29 July 2011
Available online 5 August 2011
Keywords:
Polymer electrolytes
Iodide-conducting polymers
MgI2
DSSCs
Electrical spectroscopy
Published by Elsevier Ltd.
1. Introduction
dye-sensitized photovoltaic devices [4]. Most of the research
has been devoted to iodide-conducting gel electrolytes [5]. Gel
Dye-sensitized solar cells (DSSCs) are photoelectrochemical
devices with the potential of offering a low-cost alternative to
silicon-based and thin-film solar cells [1,2]. DSSCS typically consist
of nanocrystalline anatase (TiO2) coated with a dye, a ruthenium
electrolytes systems usually afford a higher conductivity, but also
present the drawback of potential solvent leakage issues. For
instance, quasi-solid state dye-sensitized TiO2 solar cells have
been proposed, based on gel electrolytes containing iodide and
iodine impregnated in mesoporous TiO2 [6]. It was shown that
the formation of less-mobile polyiodide ions in the electrolyte
increased the conductivity in the mesoporous phase. This behavior
was rationalized as due to the Grotthuss-type electron exchange
mechanism caused by packed polyiodide species in the electrolyte.
An alternative strategy to improve the conductivity has
been the synthesis of nanocomposite polymer electrolytes. For
instance, iodide-conducting poly(ethylene oxide)/titania polymer
electrolytes were synthesized [7]. It was shown that the intro-
duction of the titania filler into the poly(ethylene oxide) matrix
complex sensitizer, an electrolyte with I−/I3 redox couple, and
−
a Pt-coated counter-electrode. Conversion efficiencies up to 12%
have been reported [3]. The quest for an iodide-conducting poly-
mer electrolyte represents a crucial issue toward a widespread
use of the DSSC technology. It has been shown that polymer elec-
trolytes do indeed have the potential to yield efficient solid-state
∗
Corresponding author. Tel.: +1 718 270 6442; fax: +1 718 270 6197.
E-mail address: mvittadello@mec.cuny.edu (M. Vittadello).
0013-4686/$ – see front matter. Published by Elsevier Ltd.
doi:10.1016/j.electacta.2011.07.133

M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
113
Table 1
reduces the crystallinity of the polymer and enhances the mobility
Composition of the PEG1000/(␦-MgI2)x(I2)y polymer electrolytes.
of the I−/I3− redox couple. The DSSCs fabricated with these materi-
als, Ru(dcbpy)2(NCS)2 complex as the sensitizer and nanoporous
TiO2 films as photoanodes, resulted in outstanding overall con-
version efficiency (4.2% under direct sunlight illumination), one of
the best efficiencies reported to date for an all solid-state device.
Another example is given by alkyloxy-imidazolium iodide ionic
polymer/SiO2 nanocomposite electrolyte [8]. By optimizing the
content of I2, 1,2-dimethyl-3-propylimidazolium iodide, and SiO2
nanoparticles in the electrolyte, a considerable ionic conductivity
of 1.51 × 10−3 S cm−1 was achieved due to the formation of high-
efficiency electron exchange tunnels.
a
b
Sample
%
MgI
nMg/nO
nI /nO
xa
yb
2
2
PEG1000
0.0000
0.1043
0.3020
0.5536
1.1393
2.5874
4.3495
7.8937
13.7025
0.0000
0.0002
0.0005
0.0009
0.0018
0.0041
0.0071
0.0133
0.0243
0.0000
0.0027
0.0026
0.0026
0.0023
0.0019
0.0013
0.0003
0.0000
0.0000
0.0038
0.0111
0.0203
0.0420
0.0966
0.1647
0.3088
0.5668
0.0000
0.0636
0.0605
0.0595
0.0534
0.0445
0.0295
0.0078
0.0000
7
6
5
4
3
2
1
0
a
Determined on the basis of ICP-AES measurements of Mg and the amount of
PEG1000 used in the synthesis.
The influence of different iodide salts on the per-
b
Determined by difference on the basis of the amount of PEG1000 used in the
formance
of
dye-sensitized
solar
cells
containing
synthesis and Mg ICP-AES analysis.
non-volatile electrolytes was investigated in phosphazene-
based electrolytic complexes [9]. Electrolytes based on
hexa[methoxyethoxyethoxycyclotriphosphazene] (MEE trimer)
with dissolved LiI, NaI, NH4I, and 1-methyl-3-propylimidazolium
(PMII) and I2 were examined. The anionic conductivities
were highest in the PMII system and decreased in the order
PMII > NH4I > NaI > LiI. The effect of cation size on solid polymer
electrolytes for DSSCs has also been investigated. Solid poly-
mer electrolyte comprised of a PEO:PEG blend and LiI, NaI, KI,
NH4I, 1-ethyl-3-methylimidazolium iodide (EMII), and 1-hexyl-3-
methylimidazolium iodide (HMII) with the same ion-to-monomer
ratio were investigated [10]. It was shown that the cell perfor-
mance of DSSCs fabricated with these systems showed strong
dependence on the cation radii.
To date, the research carried out in iodide-conducting polymer
electrolytes has been limited mostly to monovalent iodide salts.
In general, the possibility of preparing PEO-based polymer elec-
trolytes with a MX2 salt (M = alkaline earth metal, X = halide) had
been previously ruled out in the literature. After that, the synthesis
of PEG-based MgCl2 electrolytes with high Mg-ion conductivity was
shown to be possible [11]. The breakthrough was due to the use of
a highly disordered special form of MgCl2 (␦-MgCl2) [11–13]. It was
shown that the conduction mechanism of these systems involves
bivalent, monovalent and neutral species (Mg2+, [MgCl]+, MgCl2)
[14].
complete dissolution of MgI2. By diluting this mother solution with
PEG1000, eight electrolytic complexes with decreasing concentra-
tion of MgI2 were obtained. Molecular iodine (I2) was subsequently
added to each solution and mixed at 90 ◦C. The resulting solid
solutions were dark red in color. Mg content in the samples was
determined by ICP-AES using the method of standard additions and
the emission line at 279.079 nm. The argon reference line was set at
404.305 nm. A Spectroflame Modula sequential and simultaneous
ICP-AES spectrometer equipped with a capillary cross-flow nebu-
lizer was used (Spectro Analytical, Kleve, Germany). Mineralization
of PEG1000/(␦-MgI2)x(I2)y polymer electrolytes was carried out
by treating ca. 15 mg of each sample with a 3 wt.% water solution
of HCl. The composition of each of the PEG1000/(␦-MgI2)x(I2)y
polymer electrolytes thus prepared is reported in Table 1.
2.4. Instruments and methods
2.4.1. Thermal analysis
Modulated differential scanning calorimetry (MDSC) measure-
ments were carried out with a MDSC 2920 Differential Scanning
Calorimeter (TA Instruments) equipped with the LNCA low-
temperature attachment operating under a helium flux of 30
cm3·min-1. Measurements were collected with a heating rate of
3 ◦C min−1 in the −60 < T < 60 ◦C temperature range on ca. 7 mg of
sample hermetically sealed in an aluminum pan under a dry Ar
atmosphere in a glove box.
In this paper we propose the synthesis of a new polymer elec-
trolyte family with general formula PEG1000/(MgI2)x(I2)y. These
materials were characterized by thermal analysis, vibrational spec-
troscopy, electrical spectroscopy and solid state NMR.
2.4.2. Infrared spectroscopy
FT-IR measurements in both the medium (MIR) and in the far-IR
(FIR) were collected in transmission mode in a Nicolet FT-IR Nexus
spectrometer equipped with a triglycine sulphate (TGS) detector
at a resolution of 4 cm−1. MIR measurements were derived by
averaging 500 scans, after sandwiching the samples between two
KBr windows. FIR spectra were derived by averaging 400 scans;
measurements were carried out on samples spread as thin films
between two polyethylene windows. The samples for both MIR and
FIR measurements were placed in a cell hermetically closed under
argon. Baseline correction was performed with Nicolet FT-IR Nexus
spectrometer software.
2. Experimental
2.1. Reagents
Metallic magnesium, anhydrous 1-iodobutane and PEG1000
were Aldrich reagent grade. PEG1000 was stirred under vacuum for
a week at 10−3 mbar at 90 ◦C to ensure dehydration. All the other
reagents were used as received. The manipulation and the storage
of the samples were carried out under an argon atmosphere.
2.2. Synthesis of ı-MgI2
2.4.3. Electrical measurements
Excess anhydrous 1-iodobutane (∼500 g) was added to 32.5 g of
Mg metal, and the mixture was refluxed for 72 h. After refluxing,
the residual solvent was removed and the salt was dried at 90 ◦C
under vacuum for 120 h. The identity of the salt was verified by
powder X-ray diffraction measurements (data not shown)..
The measurement of complex conductivity spectra was carried
out in the frequency range 10 mHz–10 MHz, using a Novocontrol
Alpha-A analyzer. The temperature range from 20 to 90 ◦C was
explored by using a home-made cryostat operating with a N2 gas
jet heating and cooling system. The measurements were carried out
using a closed home-made cell. The geometrical constant of the cell
was determined by measuring the electrode–electrolyte contact
surface and the distance between electrodes with a micrometer. No
corrections for thermal expansion of the cell were carried out. The
2.3. Synthesis of PEG1000/(ı-MgI2)x(I2)y polymer electrolytes
In a 250 mL Pyrex bottle, 7.9714 g MgI2 was added to 58.8320 g
of PEG1000. The mixture was heated to 90 ◦C and stirred until

114
M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
40 ◦C is expected. In case of a leak a polymer electrolyte that is
based on a polymer which melts at about 40 ◦C would readily solid-
ify and minimize toxic effects on the environment. Furthermore
PEG1000 is a hydroxyl-terminated polyethylene oxide oligomer
endowed with good salt-solvation properties and known vibra-
tional spectra characteristics (as shown below). Our original plan
was to prepare a series of polymer electrolytes with general formula
PEG1000/(MgI2)x(I2)y, having a decreasing amount of MgI2 and a
constant amount of I2. Unfortunately, the presence of trace quanti-
ties of Mg in ␦-MgI2 produced an undesired variation of molecular
iodine. This can be explained by the reaction between Mg traces
and I2 to yield quantitatively MgI2, especially in the samples at the
higher concentration of salt. The final non-constant value of I2 can
be seen in Table 1. We should mention that the compositional anal-
ysis presented some challenges. The analysis of the amount of Mg
was straightforward but a direct determination of the iodine con-
tent proved to be unreliable. It is well known that iodine is volatile
and the loss of iodine from the mineralized samples was unavoid-
able. Indirectly, the volatility of iodine affected also the accuracy
of the elemental analysis of C and H. Therefore we decided to rely
on the ICP analysis of Mg and the PEG1000 experimental quantities
used in the synthesis, in order to determine the iodine content by
difference (Table 1). As shown below, it turns out that the fact that
I2 is not constant is actually an advantage in the interpretation of
the equivalent conductivity spectra of the proposed materials (see
below).
Scheme 1. (a) Synthesis of ␦-MgI2. (b) Synthesis of the PEG1000/(MgI2)x(I2)y poly-
mer electrolytes.
temperature was measured with an accuracy higher than 0.05 ◦C.
The complex impedance Z*(ω) = Zꢀ(ω) + iZꢀꢀ(ω) was converted into
the complex permittivity spectra (ε*(ω) = εꢀ(ω) − iεꢀꢀ(ω)), where
ω = 2␲f (f is the frequency in Hz).
2.4.4. NMR measurements
25Mg MAS NMR experiments were performed at the National
High Magnetic Field Laboratory at 19.6 T (ꢁ0Mg = 55.99 MHz) using
a single channel home-built 4 mm probe. The radiofrequency field
strength used for all measurements was approximately 42 kHz.
Two samples were run:pure MgI2 (Aldrich 98%) and a polymer elec-
trolyte (sample 0, see Table 1). A total of 32 and 819,200 transients
were collected for MgI2 and sample 0, respectively. 25Mg data for
MgI2 was acquired using a standard single-pulse sequence, a 1 s
recycle delay, and a spinning speed of 10 kHz. A rotor-synchronized
Hahn echo sequence was used for sample 0. The polymer elec-
trolyte was spun at 5 kHz. A single rotor period was used for
the delay between the two pulses. The recycle time was 0.5 s.
The samples were referenced to an aqueous solution of MgSO4 at
0 ppm. An exponential smoothing function was applied to the sig-
nals (LB = 20 Hz and 200 Hz for MgI2 and sample 0, respectively)
prior to Fourier Transformation. NMR simulations were performed
using the WinSolids software package. The simulated spectrum was
smoothed using a LB = 350 Hz.
3.2. Thermal analysis of PEG1000/(MgI2)x(I2)y polymer
electrolytes
In Fig.
1
the MDSC profiles of the electrolytes
PEG1000/(MgI2)x(I2)y are shown. The main spectral feature
that is visible is the endothermic peak at about 40 ◦C. This peak
was attributed to the melting (Tm) of PEG1000. The overall area of
the peak was used for calculating the crystallinity of PEG1000 in
the samples. This was possible on the basis of the tabulated value
for the enthalpy of fusion per number of moles of ethyleneoxydic
units for 100% crystalline PEO (polyethyleneoxide), which is equal
to 9.41 kJ [14]. The results of the analysis, carried out as previously
reported, are presented in Fig. 2. The glass transition temperature
(Tg) of the materials could be measured by slowly collecting the
spectra below 0 ◦C (data not shown). The Tg of pristine PEG1000
was about −44 ◦C. Minor variations of 1–2 ◦C occurred with
increasing salt concentration.
3. Results and discussion
3.1. Syntheses of ı-MgI2 and PEG1000/(MgI2)x(I2)y polymer
electrolytes
3.3. Vibrational characterization of PEG1000/(MgI2)x(I2)y
polymer electrolytes
The overarching goal of this paper is to explore the use of alka-
line earth halides in the synthesis of iodide-conducting polymer
electrolytes. A highly amorphous form of MgI2 was synthesized
from the metal and excess iodobutane (Scheme 1a). This salt has
been called “␦-MgI2” by analogy with ␦-MgCl2, previously synthe-
sized from Mg metal and chlorobutane [11,12]. The strategy of
using ␦-MgI2 was undertaken because it proved to be successful
in the synthesis of Mg-ion conducting polymer electrolytes such
as PEG400/(MgCl2)x, based on PEG400 and ␦-MgCl2. While in the
case of ␦-MgCl2, the amorphous nature of this salt was clearly
revealed by broadened peaks in XRD spectra, no clear difference
was observed between the commercial MgI2 and ␦-MgI2 (data not
shown). However it is also a fact that the dissolution of the salt in
PEG1000 (Scheme 1b) yields a MgI2 percentage as high as 13.7% in
weight. It needs to be recalled that the possibility to dissolve alka-
line earth halides in high quantities had been ruled out in earlier
literature [11]. Further studies are necessary to clarify the nature
of ␦-MgI2, which will be carried out in the future.
The far FT-IR spectra (FIR) of the PEG1000/(MgI2)x(I2)y are
shown in Fig. 3. The far FT-IR spectra and the FT-Raman spectra are
shown in Figs. 4 and 5, respectively. The full assignment of the MIR,
FIR and Raman spectra is given in Table 2. All vibrational spectra are
in agreement with those of other PEG-based polymer electrolytes
previously studied in our group [15,16]. The inspection of the MIR
and Raman “fingerprint” region (700–1500 cm−1) allowed us to
study the conformational morphology of the polyether chains in
PEG1000. Doping of PEG1000 with I2 and ␦-MgI2 did not alter the
secondary structure of the polyether chains of PEG1000. The con-
formation is of a TGT type (T = trans; G = gauche) and the vibrational
assignment is possible under the group D(4␲/7) [15,16]. Main char-
acteristic vibrations can be immediately recognized: CO stretching
bands between 800 and 1500 cm−1, CH stretching bands between
2700 and 3000 cm−1 and OH stretching bands between 3000 and
3700 cm−1
.
The choice of PEG1000 as the polymer host was motivated by
the consideration that in a typical DSSC a temperature of at least
Doping of PEG1000 with I2 and ␦-MgI2 affects the CO and
OH stretching regions. The FT-Raman spectra of the electrolytic

M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
115
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
0.0
0.0243
0.0000
0.0243
0.0000
0.0133
0.0001
-1.0
-2.0
0.0133
0.0003
-3.0
0
-1
-2
-3
0.0071
0.0013
0.0071
0.0013
-4
0.0
-1.0
-2.0
0.0042
0.0019
0.0042
0.0019
-3.0
0.0
-1.0
-2.0
0.0018
0.0023
0.0018
0.0023
-3.0
0
-1
-2
-3
0.0009
0.0026
0.0009
0.0026
-4
0
-1
-2
-3
0.0005
0.0026
-4
0.0005
0.0026
0
-1
-2
-3
0.0002
0.0027
0.0002
0.0027
-4
0.0
-1.0
-2.0
-3.0
nMg/nO: 0.0000
nI /nO: 0.0000
2
nMg/nO: 0.0000
nI /nO: 0.0000
2
-60
-40
-20
0
20
40
60
Temperature / °C
500
1000
1500
2000
2500
3000
3500
4000
Wavenumber / cm-1
Fig. 1. MDSC profiles of PEG1000/(MgI2)x(I2)y polymer electrolytes.
Fig. 3. Medium FT-IR absorption spectra of PEG1000/(MgI2)x(I2)y polymer elec-
trolytes.
complexes, as shown in Fig. 5, confirm the tentative assignments
and complete the vibrational analysis. Information about the
metal-to-ligand vibrations in the samples was provided by the
FIR spectra, as shown in Fig. 4. CCO and COC bending vibra-
tions and CC and CO rotations were detected. Symmetric and
anti-symmetric Mg–I stretching vibrations were identified
along with Mg–O stretching vibrations. These results reveal the
coordination of the possible Mg species, namely Mg2+, [MgI]+,
MgI2, by the oxygen of the polyether moieties. This was expected
by analogy with what observed in polyether-based polymers
−
doped with MgCl2. No clear vibrational pro−of of the presence of I3
anionic species was found at 110 cm−1. I3 is most likely present
in these systems but its vibrational signature cannot be resolved
from other vibrations.
220
200
180
160
140
120
100
80
100
80
60
40
20
0
The salt–polymer interactions were further studied by Gaus-
sian decomposition of the MIR CO and OH regions, as shown in
Figs. 6 and 7. The changes in the percentage area and wavenum-
ber of the individual OH and CO peaks with respect to the molar
ratio nMg/nO are shown in Figs. 8 and 9. The observed increase
or decrease in wavenumber with nMg/nO can be related to the
change in electronic density among the vibrating atoms caused by
the salt–polymer and polymer–polymer interactions, as previously
demonstrated. These changes are useful in the vibrational assign-
ment as previously reported. The Gaussian analysis confirms the
assignment given in Table 2. In the stretching OH region the sym-
metric and anti-symmetric OH vibration were detected at 3332
and 3491 cm−1, respectively [16]. The vibration at 3121 cm−1 was
attributed to stretching of PEG1000 OH groups which are involved
in hydrogen-bonding cages with other OH groups and PEG1000
-3
0
5
10
15
20x10
OH groups interacting with the anions I− or I3 [ꢁHy(OH)]. It can
−
nMg/nO
be seen that while both the symmetric and anti-symmetric OH
decrease with increasing nMg/nO ratio, the intensity of ꢁHy(OH)
increases. The crossing occurs at about nMg/nO = 0.0133. In the
Fig. 2. Enthalpy of fusion per gram of PEG1000 and PEG1000 crystallinity with
respect to nMg/nO of PEG1000/(MgI2)x(I2)y polymer electrolytes.

116
M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
Table 2
FT-IR and Raman band assignments for PEG1000/(␦-MgI2)x(I2)y polymer electrolytes.a.
(A) Polyether vibrations
Observed wavenumbers (cm-1)
Calculated d
Assignment and potential energy distribution d, e
References
Speciesb
IR c
Raman c
νa(OH)
νs(OH)
[11-18]
[11-18]
[11-18]
3491 (vw)
3332 (m)
3121 (w)
3444
3269
3114
ν
Hy(OH)
νs(CH2) (101)
νs(CH2) (101)
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
A1
inactive
2941 (s sh)
2887 (vs)
1482 (m)
1259 (vw)
887 (vw)
2940
2874
1479
1252
866
inactive
inactive
inactive
inactive
sr(CH2) (98)
t(CH2) (81)
ν(CO) (50), r(CH2) (46)
δ(CCO) (24), ϕi(CC) (19), δ(COC) (17)
274
νs(CH2) (100)
sr(CH2) (100)
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
A2
2871 (s)
1454 (w), 1472 (w)
1326 (vw sh),1350 (m)
1251 (m)
inactive
inactive
inactive
inactive
inactive
inactive
2883
1470
1344
1264
1087
964
w(CH2) (107)
t(CH2) (81)
ν(CO) (94)
r(CH2) (49), t(CH2) (18)
δ(CCO) (89), r(CH2) (33)
1088 (vs)
996 (vw)
511 (vw sh), 530 (w)
533
νa(CH2) (101, -162°)
νa(CH2) (101, 18°)
νs(CH2) (100, -5°)
2943
2940
E1
2931 (s sh)
2832 (w)
2941 (s sh)
[11-18]
[11-18]
{
2694 (vw), 2796 (vw),
2752 (vw), 2856 (s sh)
1448 (w sh)
2873
1487 (vw sh)
1438 (vw sh)
1359 (w sh)
1285 (w), 1303 (w)
1237 (w)
1471
1401
1353
1286
1234
1142
1112
sr(CH2) (100, -132°)
w(CH2) (95, 53°)
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
[11-18]
1396 (vw)
1363 (vw)
w(CH2) (107, -129°)
1280 (m)
t(CH2) (73, -40°)
1234 (w)
t(CH2) (87, 146°)
1148 (w)
1126 (s)
1142 (w)
ν(CO) (37, -27°), r(CH2) (29, -66°)
ν(CO) (81, 136°), ν(CC) (21)
ν(CO) (36, -102°),
1036 (m sh)
1065 (w)
1060
[11-18]
r(CH2) (35, 70°), ?(CC) (17)
ν(C-O-I)
r(CH2) (34, 42°), ν(CC) (27), ν(CO) (14, 167°)
1063 (vw)
935 (m), 953 (m)
806 (w), 844 (m), 871 (w)
935 (vw)
845 (m), 804 (vw)
941
847
[11-18]
[11-18]
r(CH2) (58, -149°), ν(CO) (39,-9°)
δ(CCO) (49, -142°), δ(COC) (21),
r(CH2) (17, 152°)
511 (w)
536 (vw)
524
[11-18]
301 (vw)
108 (w), 168 (w)
366
164
[11-18]
[11-18]
δ(CCO) (42, -73°), δ(COC) (38)
ϕi(CC) (42), ϕi(CO) (38, -161°), δ(CCO) (18, 16°)
(B) Inorganic vibrations
Observed wavenumbers (cm-1)
Raman c
Assignmente
References
IR c
[11-16]
[11-16]
[11-16]
351 (vw), 391 (vw), 426 (w)
301 (w)
ν(Mg-O)
νa(Mg-I)
νs(Mg-I)
137 (vw)
aPEG1000 in the polymer electrolytic complex was assumed under the D(4␲/7) symmetry group.
bSymmetry species.
cRelative intensities of observed bands are reported in parentheses; vs, very strong; s, strong; m, medium; w, weak; vw, very weak; sh, shoulder.
dCalculated wavenumber values of the normal coordinate analysis were derived from the references reported in the last column.
eꢁ, stretching; ı, bending; w, wagging, t, twisting, r, rocking; sr, scissoring; ϕi, internal rotations; a, antisymmetric mode; s, symmetric.
stretching CO region the vibrations that are most affected by the
salt–polymer interactions are the bands at 1126 and 1088 cm−1
The difference in crossing point of the OH and CO vibrations indi-
cates that these functional groups have a different role in the
dissolution of the salt at different salt concentrations. This lack
of syncronism with respect to the metal-to-oxygen molar ratio
is in contrast with the synchronism observed in other systems
such as PEG400/(LiCl)x [16], PEG400/(MgCl2)x [11], and PEG400-
alt-DEOS/(MgCl2)x [17,18]. Both OH and CO vibrations measured
for the PEG1000/(MgI2)x(I2)y complexes have a direct correspon-
dence with those detected in the polymer electrolytes mentioned
,
which were attributed to [ꢁ(CO) (81, 136◦), ꢁ(CC) (21)] and [ꢁ(CO)
(94)], respectively. The former vibration [ꢁ(CO) (81, 136◦), ꢁ(CC)
(21)] is diagnostic of the breaking and forming of hydrogen bonds
among PEG1000 molecules, the latter vibration is indicative of ter-
minal CO groups that are influenced by anion coordination. It can
be seen that the crossing between the percentage area of these
two bands occurs between nMg/nO = 0.0018 and nMg/nO = 0.0041.

M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
117
0.0243
0.0000
0.0243
0.0000
0.0133
0.0003
0.0133
0.0003
0.0071
0.0013
0.0071
0.0013
0.0042
0.0019
0.0042
0.0019
0.0018
0.0023
0.0018
0.0023
0.0009
0.0026
0.0009
0.0026
0.0005
0.0026
0.0005
0.0026
0.0002
0.0027
0.0002
0.0027
nMg/nO: 0.0000
nI /nO: 0.0000
nMg /nO: 0.0000
2
nI /nO: 0.0000
2
500
1000
1500
2000
2500
3000
3500
50 100 150 200 250 300 350 400 450 500 550 600
Wavenumber / cm-1
Wavenumber / cm-1
Fig. 5. FT-Raman spectra of PEG1000/(MgI2)x(I2)y polymer electrolytes.
Fig. 4. Far FT-IR absorption spectra of the PEG1000/(MgI2)x(I2)y polymer elec-
trolytes.
of the kth relaxation peak, ꢃk = 1/2ꢄfk is the dielectric relaxation
time (with fk in Hz the frequency of the kth peak position), ꢀel,n
is the conductivity associated to the relaxation ꢃel,n, the relaxation
time associated with the electrode polarization phenomena.
Spectroscopic data were fitted in the following representations:
(a) real dielectric spectrum (εꢀ versus f) and (b) imaginary dielec-
tric spectrum (εꢀꢀ versus f). Selected spectra are shown in Fig. 10.
One electrodic polarization term (ꢀel) and two-to-three dielectric
relaxation terms (ꢃk) were used for the spectra measured below
the melting point (T < Tm) of the polymer electrolytes. Up to 3
polarization terms and 2 dielectric relaxation terms were used for
T > Tm. The presence of more polarization terms is evident from the
impedance spectra representation, which shows two semicircles
(data not shown).
above. However, there is one exception. It should be mentioned
that the band detected at 1063 cm−1 was attributed to ꢁ(C–O–I). We
think that a small fraction of molecular iodine reacts with terminal
C–O–H groups forming C–O–I groups. It is interesting to notice that
the percentage area of this vibration decreases slightly with the
nMg/nO ratio. This decrease in intensity is accompanied by a small
increase in wavenumber. This latter observation indicates that the
C–O–I functional groups are involved in the coordination of the
cationic species Mg2+, and [MgI]+.
3.4. Electrical characterization of PEG1000/(MgI2)x(I2)y polymer
electrolytes
In Fig. 11 the profiles of ꢀel,n (n = 1, 2, 3) for the
PEG1000/(MgI2)x(I2)y polymer electrolytes log ꢀ versus recip-
rocal temperature are shown. Three temperature regions were
The
polymer
electrolytes
PEG1000/(MgI2)x(I2)y
with
0 ≤ x ≤ 0.5801, 0 ≤ y ≤ 0.0636 were analyzed by electric spec-
troscopy between 0.1 Hz and 1 MHz, 293.15 K and 363.15 K.
Electrical spectra were analyzed by using the following equation
[19]:
identified: the first (I) at low T (T < Tm), the second (II) for T ∼ Tm
,
and the third for T > Tm. In the first two regions the fitting was
performed by using the following equation [20]:
n
ꢀ
ꢀ
Ea
ꢀel,n
iω
(iωꢃel,n)
ꢂεk
ε∗(ω) = ε
+
+
, (1)
ln(ꢀel,n(T)) = ln(Aꢀel,n ) +
(2)
ꢁ
ꢂ
∞
ꢁ
ꢂ
ˇ
n
RT
1 + (iωꢃk)˛
k
1 + (iωꢃel,n
)
k
n
k
while for T > Tm through the Vögel-Tamman-Fülcher (VTF) [21]:
ꢃ
ꢄ
where ε is the permittivity of the material, the second and third
∞
−Ea
terms in the equation describe the contributions of the dielectric
relaxations and of electrodic and interfacial polarization phenom-
ena, respectively. ꢂεk is the relaxation strength, ˛k and ˇk are shape
parameters describing the symmetric and asymmetric broadening
ꢀel,n(T) = Aꢀel,n T−1/2exp
(3)
R(T − T0)
where Aꢀ
is proportional to the number of charge carriers, R
is the ideeal,ln gas constant, Ea is the pseudo-activation energy for

118
M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
0.05
0.00
200
0
a
a
-200
-0.05
4x10-3
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
3
2
1
0
2600
2800
2800
2800
3000
3200
3400
3400
3400
3600
3600
3600
3800
3800
3800
400
400
400
600
600
600
800
1000
1200
1200
1200
1400
1400
1400
Wavenumber / cm-1
-1
Wavenumber / cm
200
0
0.04
0.00
b
b
-200
-0.04
2.0
5x10-3
4
3
2
1
0
1.5
1.0
0.5
0.0
-0.5
800
1000
2600
3000
3200
Wavenumber / cm-1
-1
Wavenumber / cm
0.02
0.00
200
0
c
c
-0.02
0.6
-200
5x10-3
0.5
0.4
0.3
0.2
0.1
0.0
-0.1
4
3
2
1
0
800
1000
2600
3000
3200
-1
Wavenumber / cm-1
Wavenumber / cm
Fig. 7. Decomposition by Gaussian functions of the FT-IR spectral region
Fig. 6. Decomposition by Gaussian functions of the FT-IR spectral region
400–1540 cm−1 measured on selected PEG1000/(MgI2)x(I2)y polymer electrolytes.
2450–3800 cm−1 measured on selected PEG1000/(MgI2)x(I2)y polymer electrolytes.
conduction and T0 is the ideal thermodynamic glass transition
temperature. The initial value for T0 was determined from the
experimental value of Tg measured by thermal analysis, using the
equation T0 = Tg − 50 ◦C. The Tg of all the materials does not change
significantly in the polymer electrolytes and is equal to −44 2 ◦C.
It can be observed that the conductivity increases significantly
above the onset of the melting of the polymer which occurs within
region II (between 30 and 40 ◦C, see Fig. 2). The highest value among
the ꢀel,n is in most cases ꢀel,1. This indicates that ꢀel,1 is mostly
responsible for the overall direct current conductivity ꢀDC. There-
fore, in the following we reasonably assume that ꢀel,1 = ꢀDC. The
presence of three types of interfacial conductivities ꢀel,n (n = 1, 2,
3) indicates that there are three types of interfaces in the systems.
This implies the existence of two types of domains in the materials
under investigation: domains A and B. The conduction can occur
within domain A (corresponding to ꢀel,2), within domain B (cor-

M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
119
80
80
60
40
20
0
a
1151 cm-1
1126 cm-1
1088 cm-1
1063 cm-1
1036 cm-1
3121 cm-1
a
3332 cm-1
80
80
60
40
20
0
3491 cm-1
60
fit
60
996 cm-1
fit
40
40
20
20
0
0
0
5
10
15
20x10-3
nMg/nO
0
5
10
15
20x10-3
nMg/nO
1050
b
b
1151 cm-1
1126 cm-1
1088 cm-1
1063 cm-1
1036 cm-1
1040
1030
1020
1010
1000
990
3500
3400
3300
3200
3100
3500
3400
3300
3200
3100
1140
1120
1100
1080
1060
996 cm-1
fit
3121 cm-1
3332 cm-1
3491 cm-1
fit
0
5
10
15
20x10-3
nMg/nO
0
5
10
15
20x10-3
Fig. 9. Dependence on nMg/nO molar ratio of the fractional percentage (%)i of each
CO band revealed by the CO stretching region (Fig. 7). The semiquantitative values
of the percentages were determined with the formula (%)i = Ai100/˙, where Ai is the
nMg/nO
ꢅ
n
Fig. 8. Dependence on nMg/nO molar ratio of the fractional percentage (%)i of each
OH band revealed by the OH stretching region (Fig. 6). The semiquantitative values
of the percentages were determined with the formula (%)i = Ai100/˙, where Ai is the
band area of peak i-th, ˙ =
i=1Ai, and n is the number of bands considered.
ꢅ
n
band area of peak i-th, ˙ =
i=1Ai, and n is the number of bands considered.
Fig. 10. Real and imaginary relative permittivity spectra of selected PEG1000/(MgI2)x(I2)y polymer electrolytes at different temperatures. The molar ratios nMg/nO and nI /nO
2
(underlined) are indicated for each sample.

120
M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
-4
-6
12
10
8
12
10
8
σel,1
σel,2
σel,3
-8
6
6
4
4
-10
2
2
0.0133
0.0003
0.0071
0.0013
0.0243
0.0000
0
0
-12
-4
12
10
8
12
10
8
-6
-8
6
6
4
4
-10
2
2
0.0041
0.0020
0.0009
0
0.0018
0.0023
0
-12
-4
0.0026
12
10
8
12
10
8
-6
-8
6
6
4
4
-10
-12
2
2
0
0
0.0000
0.0000
0.0005
0.0026
0.0002
0.0027
0
5
10
15
20
25x10-3
-3
2.8 3.0 3.2 3.4x10 2.8 3.0 3.2 3.4 2.8 3.0 3.2 3.4
nMg / nO
(1000/T) / K-1
Fig. 12. Activation energy for the conduction process of the PEG1000/(MgI2)x(I2)y
polymer electrolytes as a function of the nMg/nO molar ratio.
Fig. 11. Dependence on the reciprocal absolute temperature of the parameters ꢀel,n
(n = 1, 2, 3) for the PEG1000/(MgI2)x(I2)y polymer electrolytes. ꢀel,1 corresponds to
the overall direct current conductivity ꢀ. The fitting was carried out with an Arrhe-
nius equation in regions I and II, while the VTF was used in region III. The molar
ratios nMg/nO and nI /nO (underlined) are indicated for each sample.
2
domains B would be smaller sites comprised of empty PEG1000
chains (with or without I−, I3 or I2) with no MgI2 salt. The bound-
−
ary sites are comprised of A and B sites coordinated to each other
by means of dynamic cross-links via bridging anionic species such
responding to ꢀel,3) and across domains from A to B or vice versa
(corresponding to ꢀel,1). This latter phenomenon results in long-
range conductivity. It can be seen in Fig. 11 that for nMg/nO = 0.0005
and nI /nO = 0.0026 ꢀel,2 partially prevails over ꢀel,1, indicating
that a 2localized conduction phenomenon is taking place in region
III. The conductivity in pristine PEG1000 is accounted for by a single
electrode polarization phenomenon (a single domain is present).
Further insight about the conductivity mechanism is given
by the activation energy for the conduction process of the
PEG1000/(MgI2)x(I2)y polymer electrolytes as a function of the
nMg/nO molar ratio, as shown in Fig. 12. The barrier for conduc-
tion of ꢀel,1 is mostly lower than or similar to both ꢀel,2 and ꢀel,3 at
the higher concentrations of salt. This finding indicates that while
it is relatively easy for the free ion carriers to move across domains
from A to B in order to have long-range conductivity (the crossings
A–B or B–A correspond to ꢀel,1), motion is more difficult within
domain A (which corresponds to ꢀel,2) and B (which corresponds
to ꢀel,3).
as I−, I3 or via I2. Conducting species can escape both A and B
−
domains, giving rise to long range conductivity.
Information about the conducting species in the systems was
gathered by means of equivalent conductivity spectra, as shown in
1/2
Mg
Fig. 13. The profiles of ꢅ with respect to m
exhibit an overall
rapid decrease with increasing salt concentration with the excep-
tion of a small local maximum at 0.14 (mol/kg)1/2. The fitting of the
data was conducted by using the following generalized equation
[21,22]:
ꢅ = H0 + K1exp(−˛1m1/2) + K2exp(−˛2m1/2
)
Mg
Mg
+ K3exp(−˛3(m1/2 − x0)2)
(4)
Mg
This equation reduces to the Onsager equation (ꢅ = ꢅ0
−
Am1M/g2) in the limit of zero concentration, under the conditions that
ꢅ0 = H0 + K1 + K2 + K3 and A = K1˛1 + K2˛2 + 2K3˛3:
We hypothesize, in analogy with previous studies [14], that
domains A are comprised of one ore more PEG1000 chains
coordinated to bivalent, monovalent and neutral species, Mg2+
,
[MgI]+, and MgI2 via direct Mg–O complexation. These species can
migrate locally by intra-chain hopping. We hypothesize also that
ꢅ = (H0 + K1 + K2 + K3) − (K1˛1 + K2˛2 + 2K3˛3) · m1/2
.
(5)
Mg

M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
121
and outside the typical range of octahedrally coordinated Mg sites.
This may be attributed to the electron-rich iodide atoms in the first
coordination shell. Considering the electronegativities of Mg and
I, the Mg–I bond is expected to be fairly covalent. Previous work
has shown that the magnitude of the quadrupole coupling con-
stant (CQ) increases with an increase in the degree of covalency
of the bonds to the quadrupole nucleus. Surprisingly, the narrow
peak implies a very small quadrupole coupling constant found in
symmetric environments.
In contrast to the MgI2 spectrum, PEG1000/(MgI2)0.5668(I2)0.0000
displays asymmetrically broad resonances whose intensities lie
between 0 and −125 ppm. The sharp resonance present in MgI2
is no longer visible. However, spectral intensity distributed about
−80 ppm is present, implying that there may be some MgI2
nanoparticles within the polymer matrix. The peak broadening is
attributed to the decrease in particle size and an increase in the
amount of disorder present in the remaining magnesium iodide
crystallites. A relatively large amount of signal intensity lies at
approximately −25 ppm. This shift is to be expected for Mg atoms
which are no longer part of the MgI2 lattice, but have coordinated
to the oxygen atoms in the polymer backbone. Bonds to oxygen
atoms would withdraw electron density at the Mg site and shift the
resonance downfield. Inherent distribution in the bonding geom-
etry of the Mg atom to the polymer backbone will consequently
lead to a distribution in both the quadrupole coupling constant and
asymmetry parameter. These distributions induce a characteristic
asymmetric broadening of the second order line shape extending
upfield (lower ppm). The breadth of the line shape was used to esti-
mate the NMR parameters of this characteristic site. The simulated
parameters are ıiso = 6 ppm, Á = 0.5, CQ = 3.5 MHz.
800
600
400
200
0
293 K
303
313
323
333
343
353
363
fit
298 K
308
318
328
338
348
358
0.1
0.2
0.3
½
0.4
½
0.5
0.6
0.7
c _Mg/mol kg-½
Fig. 13. Equivalent conductivity as a function of m1M/g2 of PEG1000/(MgI2)x(I2)y poly-
mer electrolytes.
1/2
The initial decrement (0 ≤ m
< 0.10 mol/kg1/2
) can be
Mg
explained in terms of the presence of ionic species such as Mg2+
,
[MgI]+, I− and I3−. In this regime both cationic and anionic
1/2
conductivity is possible. At the local maximum (0.10 ≤ m
<
Mg
0.20 mol/kg1/2), which is accounted for by the fourth term in Eq.
(4), magnesium is present mostly as [MgI]+ and MgI2, opening the
possibility of dominant anionic migration. This possible mechanism
is expected on the basis of previous results on PEG-based MgCl2
electrolytes [14]. The formation of triple ions is also a possibility:
I−· · ·[MgI]+· · ·I− or I3−· · ·[MgI]+· · ·I3−. The formation of triple ions
The 25Mg NMR study has revealed the presence of at least two
different Mg sites in the polymer complex. The first site, whose
shift is centered at approximately −80 ppm, has been assigned to
residual MgI2 in the polymer matrix. The second site is present
at relatively higher ppm values and assigned to Mg coordinated
to the oxygen atoms of the PEG backbone. It is characterized by
a second-order line shape and distributions in NMR parameters.
The site could be adequately simulated with ıiso = 6 ppm, Á = 0.5,
CQ = 3.5 MHz. The relative intensity of the second site is much larger
than the first site, implying the successful, albeit incomplete, dis-
solution of the MgI2 salt into the polymer matrix. Further studies,
preferably at different magnetic field strengths, are needed to ver-
ify the amount of sites as well as determine the effects of chemical
shift anisotropy on the line shapes. The description emerging from
the NMR study is compatible with the results obtained from the
equivalent conductivity study: the coordination Mg2+, [MgI]+, and
MgI2 species by the polymer backbone.
will further facilitate anionic conductivity. The equivalent conduc-
1/2
tivity decreases further (0.20 ≤ m
≤ 0.70 mol/kg1/2) with the
Mg
widespread formation of neutral species such as [MgI]+· · ·I− or
[MgI]+· · ·I3−. Thus, both cationic and anionic conductivity is dra-
matically reduced.
Fig. 14 shows the 25Mg MAS NMR of MgI2 and sample 0
(see Table 1), along with a preliminary simulation of the dom-
inant Mg site in the composite. Sample 0 indicates the system
PEG1000/(MgI2)0.5668(I2)0.0000. MgI2 is isostructural with CdI2,
¯
crystallizing in the P3m1 space group with a Mg atom lying on the
3m site. The 25Mg MAS NMR spectrum of MgI2 contains a single,
sharp resonance at −81.1 ppm. This is consistent with the high sym-
metry of the Mg site. The shift of the resonance is highly shielded
4. Conclusions
A new series of polymer electrolytes with formula PEG1000/
(MgI2)x(I2)y was synthesized with a magnesium iodide content as
high as 13.7% in weight. This achievement was possible on the basis
of a highly amorphous form of MgI2 synthesized via a metallorganic
route. Thermal analysis has shown that PEG1000 oligomers par-
tially crystallize below Tm = 39–41 ◦C, with a crystallinity varying
parabolically between 46 and 76%. Based on our vibrational inves-
tigations we have shown that: (1) full band assignment is possible
by assuming a D(4␲/7) symmetry for the PEG1000 polymer host;
(2) terminal OH groups dominate the salt–polymer interactions
at low salt concentrations while both OH and CO groups are
relevant at higher concentrations; (3) Mg atoms are coordinated
by oxygen atoms of the polymer host while retaining iodide anions
in their first sphere of coordination. The presence of molecular
iodine does not seem to significantly alter the distribution of
ionic species for PEG1000/(MgI2)x systems, as expected from our
Fig. 14. 25Mg MAS NMR of MgI2, a polymer composite (see text) and a Win-
Solids simulation of a Mg site. The simulated parameters are ıiso = 6 ppm, Á = 0.5,
CQ = 3.5 MHz.

122
M. Vittadello et al. / Electrochimica Acta 57 (2011) 112–122
previous findings on PEG400/(MgCl2)x electrolytic complexes.
This indicates that molecular iodine may be present as I3−, with
a substitutional role with respect to iodides. The electric mea-
surements highlighted three electrode polarization phenomena
in PEG1000/(MgI2)x(I2)y at temperatures higher than the melting
point of the polymer host. This evidence has been attributed to
the presence of two distinct domains (A and B) characterized by
different ion concentrations [7,8]. For nMg/nO < 7 × 10−3, ꢀDC is
determined by the mobility of the charge carriers within each
domain; at higher concentrations of the salt, ꢀDC is bottlenecked
by the charge transfer between the domains. The overall study
indicates that at least three types of Mg-containing species are
present: Mg2+, MgI+ and MgI2. Bivalent, monovalent and neutral
species participate in the conduction process. Residual nanoparti-
cles of MgI2 may be present. 25Mg MAS NMR measurements on the
References
[1] M. Grätzel, Nature 414 (2001) 338.
[2] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Chem. Rev. 110 (2010)
6595.
[3] S. Yanagida, Y. Yu, K. Manseki, Acc. Chem. Res. 42 (2009) 1827.
[4] A.F. Nogueira, J.R. Durrant, M.A. De Paoli, Adv. Mater. 13 (2001) 826.
[5] J. Wu, Z. Lan, D. Wang, S. Hao, J. Lin, Y. Huang, S. Yin, T. Sato, Electrochim. Acta
51 (2006) 4243.
[6] W. Kubo, K. Murakoshi, T. Kitamura, S. Yoshida, M. Haruki, K. Hanabusa, H.
Shirai, Y. Wada, S. Yanagida, J. Phys. Chem. B 105 (2001) 12809.
[7] T. Stergiopoulos, I.M. Arabatzis, G. Katsaros, P. Falaras, Nano Lett. 2 (2002)
1259.
[8] J. Shi, L. Wang, Y. Liang, S. Peng, F. Cheng, J. Chen, J. Phys. Chem. C 114 (2010)
6814.
[9] S.H.A. Lee, A.M.S. Jackson, A. Hess, S.T. Fei, S.M. Pursel, J. Basham, C.A. Grimes,
M.W. Horn, H.R. Allcock, T.E. Mallouk, J. Phys. Chem. C 114 (2010) 15234.
[10] B. Bhattacharya, J.Y. Lee, J. Geng, H.T. Jung, J.K. Park, Langmuir 25 (2009)
3276.
[11] V. Di Noto, S. Lavina, D. Longo, M. Vidali, Electrochim. Acta 43 (1998)
1225.
highest salt concentration sample, PEG1000/(MgI2)0.5668(I2)0.0000
,
confirm such a description. Depending on the salt concentration,
both cationic and anionic hopping mechanisms may be possible.
[12] V. Di Noto, S. Bresadola, Macromol. Chem. Phys. 197 (1996) 3827.
[13] M. Vittadello, P.E. Stallworth, F. Alamgir, S. Suarez, S. Abbrent, C.M. Drain, V. Di
Noto, S.G. Greenbaum, Inorg. Chim. Acta 359 (2006) 2513.
[14] V. Di Noto, M. Vittadello, Solid State Ionics 147 (2002) 309.
[15] V. Di Noto, M. Bettinelli, M. Furlani, S. Lavina, M. Vidali, Macromol. Chem. Phys.
197 (1996) 375.
[16] V. Di Noto, D. Longo, V. Münchow, J. Phys. Chem. B 103 (1999) 2636.
[17] M. Vittadello, S. Biscazzo, S. Lavina, M. Fauri, V. Di Noto, Solid State Ionics 147
(2002) 341.
[18] V. Di Noto, V. Münchow, M. Vittadello, J.C. Collet, S. Lavina, Macromol. Chem.
Phys. 203 (2002) 1211.
[19] V. Di Noto, E. Negro, J.Y. Sanchez, C. Iojoiu, J. Am. Chem. Soc. 132 (2010)
2183.
1/2
Mg
In particular, for 0 ≤ c
< 0.10 mol/kg1/2 both cationic and
anionic species (Mg2+, [MgI]+, I− and I3−) are likely to participate
1/2
in the conduction process, while for 0 ≤ c
< 0.20 mol/kg1/2
Mg
mostly iodide hopping is possible among the species [MgI]+ and
MgI2 and the triple ions I−· · ·[MgI]+· · ·I− or I3−· · ·[MgI]+· · ·I3−. For
1/2
0.20 ≤ c
≤ 0.70 mol/kg1/2 low residual cationic and anionic con-
Mg
ductivity is present due to neutralization caused by extensive ionic
coupling.
[20] V. Di Noto, J. Phys. Chem. B 106 (2002) 11154.
[21] V. Di Noto, S. Lavina, E. Negro, M. Vittadello, in: D. Santos, C. Sequeira (Eds.),
Polymer Electrolytes: Fundamentals and Applications, Woodhead Publishing
Ltd., Cambridge, UK, 2010, pp. 219–277.
Acknowledgements
[22] V. Di Noto, M. Vittadello, S. Lavina, M. Fauri, S. Biscazzo, J. Phys. Chem. B 105
(2001) 4584.
The National High Magnetic Field Laboratory is supported by
the National Science Foundation and the State of Florida through
Cooperative Agreement (DMR-0084173).