Speciation in the AlCl3/SO2Cl2 catholyte system

Article abstract of DOI:10.1021/ic050621z

The fundamental chemical behavior of the AlCl₃/SO ₂Cl₂ catholyte system was investigated using ²⁷Al NMR spectroscopy, Raman spectroscopy, and single-crystal X-ray diffraction. Three major Al-containing species were found to be present in this catholyte system, where the ratio of each was dependent upon aging time, concentration, and/or storage temperature. The first species was identified as [Cl₂Al(μ-Cl)]₂ in equilibrium with AlCl₃. The second species results from the decomposition of SO₂Cl ₂ which forms Cl₂(g) and SO₂(g). The SO ₂(g) is readily consumed in the presence of AlCl₃ to form the crystallographically characterized species [Cl₂Al(μ-O ₂SCl)]₂ (1). For 1, each Al is tetrahedrally (T _d) bound by two terminal Cl and two μ-O ligands whereas, the S is three-coordinated by two μ-O ligands and one terminal Cl. The third molecular species also has T_d-coordinated Al metal centers but with increased oxygen coordination. Over time it was noted that a precipitate formed from the catholyte solutions. Raman spectroscopic studies show that this gel or precipitate has a component that was consistent with thionyl chloride. We have proposed a polymerization scheme that accounts for the precipitate formation. Further NMR studies indicate that the precipitate is in equilibrium with the solution.

Full text of DOI:10.1021/ic050621z

Inorg. Chem. 2005, 44, 5934−5940
Speciation in the AlCl3/SO2Cl2 Catholyte System
Timothy J. Boyle,* Nicholas L. Andrews, Todd M. Alam, David R. Tallant, Mark A. Rodriguez, and
David Ingersoll
Sandia National Laboratories, AdVanced Materials Laboratory, 1001 UniVersity BouleVard, SE,
Albuquerque, New Mexico 87106
Received April 22, 2005
The fundamental chemical behavior of the AlCl3/SO2Cl2 catholyte system was investigated using 27Al NMR
spectroscopy, Raman spectroscopy, and single-crystal X-ray diffraction. Three major Al-containing species were
found to be present in this catholyte system, where the ratio of each was dependent upon aging time, concentration,
and/or storage temperature. The first species was identified as [Cl2Al(
species results from the decomposition of SO2Cl2 which forms Cl2(g) and SO2(g). The SO2(g) is readily consumed
in the presence of AlCl3 to form the crystallographically characterized species [Cl2Al( -O2SCl)]2 (1). For 1, each Al
is tetrahedrally (Td) bound by two terminal Cl and two -O ligands whereas, the S is three-coordinated by two -O
µ-Cl)]2 in equilibrium with AlCl3. The second
µ
µ
µ
ligands and one terminal Cl. The third molecular species also has Td-coordinated Al metal centers but with increased
oxygen coordination. Over time it was noted that a precipitate formed from the catholyte solutions. Raman
spectroscopic studies show that this gel or precipitate has a component that was consistent with thionyl chloride.
We have proposed a polymerization scheme that accounts for the precipitate formation. Further NMR studies
indicate that the precipitate is in equilibrium with the solution.
Introduction
to make this a useful material. To accomplish this, it is
necessary to understand the various chemical changes that
are occurring in the reaction mixture.
A new generation of batteries is necessary to meet power
and energy requirements of the smaller electronic devices
demanded by both the public and private sectors, as well as
the ability to meet these needs after prolonged (years)
storage.1 Due to the portability of these new devices,
progressively diverse and increasingly harsh environments
are being encountered. Therefore, it is necessary that the
battery operate over a large temperature range during storage
as well as after storage under even more severe environ-
mental conditions. While numerous systems have been
investigated, batteries using the catholyte system of alumi-
num chloride (AlCl3) dissolved in sulfuryl chloride (SO2Cl2)
have maintained viability at both the higher and lower
temperatures (-50 to +125 °C).2-4 A catholyte system is
one where a liquid operates as both the cathode and
electrolyte. Enhancing the long-term stability of the inher-
ently unstable AlCl3/SO2Cl2 catholyte system is necessary
Several studies are available concerning the chemical
behavior associated with both AlCl3 and SO2Cl2 independ-
ently, but there are little data readily available for the
combined system.2-4 Glidewell reported that [SCl3]+[AlCl4]-
was the active chlorinating species in the complex SO2Cl2-
S2Cl2-AlCl3 system.4 Additional reports indicate that mo-
lecular ionic compounds are formed upon mixing thionyl
chloride (SOCl2) and AlCl3, some of which are proposed to
be [SOCl]+[AlCl4]-,5 Cl2SO-AlCl3,6 and Cl(O)SCl-AlCl3.7
However, we can find no report that establishes an interaction
between SO2Cl2 and AlCl3.2-4
Therefore, we undertook the systematic study of the
chemical behavior of the SO2Cl2/AlCl3 catholyte solution as
a stand-alone system in order to reduce the overall complex-
ity of the problem, but also because this configuration most
closely resembles a reserve battery. In this case the catholyte
is held in a separate reservoir during storage, and on battery
activation the reservoir is breached, allowing catholyte to
* To whom correspondences should be sent. E-mail: tjboyle@Sandia.gov.
Phone: (505) 272-7625. Fax: (505) 272-7336.
(1) Doughty, D. H. SAMPE J. 1996, 32, 75.
(2) Pray, A. R.; McCrosky, C. R. J. Am. Chem. Soc. 1952, 74, 4719.
(3) Geiko, V. I.; Stryfina, O. P.; Doda, S. A.; Borovikov, A. Y. J. Appl.
Chem. USSR 1990, 63, 648.
(5) Spandau, H.; Brunneck, E. Z. Anorg. Allg. Chem. 1955, 278, 197.
(6) Long, D. A.; Bailey, R. T. Trans. Faraday Soc. 1963, 59, 594.
(7) Haraguchi, H.; Fujiwara, S. J. Phys. Chem. 1969, 73, 3467.
(4) Glidewell, C. Inorg. Chim. Acta 1986, 117, L7.
5934 Inorganic Chemistry, Vol. 44, No. 16, 2005
10.1021/ic050621z CCC: $30.25
© 2005 American Chemical Society
Published on Web 07/13/2005

Speciation in the AlCl3/SO2Cl2 Catholyte System
reflections. Lattice determination and data collection were carried
out using SMART Version 5.054 software. Data reduction was
performed using SAINT Version 6.01 software. The structure
refinement was performed using XSHELL 3.0 software. The data
were corrected for absorption using the SADABS program within
the SAINT software package.
mix and flood the remaining chemical constituents of the
battery. The information garnered will assist in understanding
and improving battery performance over extended time
frames and in identifying storage protocols that extend battery
shelf life. This report details the observed chemical changes
the SO2Cl2/AlCl3 catholyte solution underwent as determined
by 27Al NMR spectroscopy, Raman spectroscopy, and single-
crystal X-ray diffraction, which identified a unique inter-
mediate species [Cl2Al(µ-O2SCl)]2 (1)8sthe first adduct ever
isolated for this catholyte system. The synthesis of this
compound and the other species characterized in this system
are discussed below.
Discussion
Due to the dearth of information concerning the species
present in the AlCl3/SO2Cl2 solution and our interest in using
this catholyte system, we undertook a series of analytical
investigations to elucidate the chemical speciation of this
system. Several key conditions were varied: (i) the amount
of water present, (ii) the AlCl3 concentration, (iii) solution
storage or aging temperature, and (iv) the solvent. The
changes in this system were analyzed using NMR and Raman
spectroscopies in conjunction with single-crystal X-ray
diffraction.
Precursor Materials. Our initial investigations focused
on the properties of the starting materials in solution. All
attempts to purify the precursors only resulted in the inclusion
of additional contaminants or no improvement in purity.
Double sublimation of an alternative source of AlCl3 (98%)
did not remove the fluorescent impurities that dominated the
region of the Raman spectrum of interest to us; however,
the as-received 99.99% pure materials did not possess these
fluorescent contaminants. During distillation of SO2Cl2,
corrosive attack by the solvent introduced organic species
(i.e., decomposition products of silicon grease, Tygon tubing)
into the distilled material. Therefore, the reagents were used
as received but opened and handled only under an argon
atmosphere.
Experimental Section
All compounds described below were handled with rigorous
exclusion of air and water using standard Schlenk line and glovebox
techniques. The SO2Cl2 (MCB Reagents), AlCl3 (99.999%, Aldrich),
and SO2(g) (TriGas, Inc.) were opened and used as received, in a
glovebox under an argon atmosphere. The various solution con-
centrations were prepared by dissolving preweighed AlCl3 into
measured amounts of SO2Cl2. Water was added via syringe to the
appropriate samples. All samples were prepared under argon
atmospheres by placing aliquots of the appropriate solutions into
Pyrex NMR tubes and then flame-sealing the tube under vacuum.
The solution 27Al NMR spectra were obtained on a Bruker
DRX400 instrument operating at 104.2 MHz using a 5 mm broad
band probe operating at 298 K. Typical acquisition conditions
included between 16 and 64 scans averaging, 1 s recycle delay, 8
K complex points using a 600 ppm spectral width. Chemical shifts
were referenced to a secondary standard 1 M Al[H2O]63+, δ ) 0.0
ppm.
The Raman spectra were obtained using a triple spectrograph
and charge-coupled-device detector for dispersing and recording
the Raman-scattered light. A 90° scattering configuration and a 785
nm excitation laser were used to obtain Raman spectra from the
solutions sealed in NMR tubes. For gelled species in sealed
containers, we employed a 458 nm excitation laser and a microscope
accessory in a 180° scattering configuration.
[Cl2Al(µ-O2SCl)]2 (1). In a glovebox, AlCl3 (2.00 g, 14.9 mmol)
was added to a vial of SO2Cl2 (∼10 mL) to form an ∼1.5 M
solution. The yellow mixture was stirred for 12 h and became
progressively darker in color. After sitting for an extended period
of time, the resulting brown solution was placed into a freezer (-25
°C) for 2 days. Clear crystals were observed at low-temperature;
however, any attempt to isolate them at room temperature resulted
in their melting, which prevented obtaining bulk analytical data
for 1. Crystals suitable for X-ray diffraction were obtained by
pouring off the mother liquor and placing the vial containing the
crystalline material in an ice bath prior to manipulating them on a
homemade, liquid nitrogen cooled coldfinger.
X-ray Crystal Structure Information.9 The crystal was mounted
onto a thin glass fiber under an atmosphere of flowing liquid N2,
and immediately placed under a liquid N2 stream, on a Bruker AXS
diffractometer. The radiation used was graphite-monochromatized
Mo KR radiation (λ ) 0.7107 Å). The lattice parameters were
optimized from a least-squares calculation on carefully centered
The 27Al NMR chemical shift is directly related to the
coordination of the Al metal centers. AlCl3 dissolved in
toluene-d8 revealed a broad 27Al singlet (250 Hz) at δ )
+102 ppm. This chemical shift is consistent with tetra-
hedrally (Td) bound Al metal centers.10-13 The large line
width suggests either changes in the electrical field gradient
or an increase in the correlation time of these field gradient
fluctuations due to solvent interactions, increased aggregation
of the molecular species, or chemical exchange with other
Al-containing species in solution. The observed signal in our
sample is consistent with a Td Al species found for AlCl3 in
nonpolar solvents and melted AlCl3, instead of the nominal
octahedral coordination Al species (δ ∼ 0 ppm) observed
for the room-temperature solid-state AlCl3 structure.10-13
Proposed Td species for this solution species include Al2Cl6
-
or AlCl4 , which are known to be relatively insensitive to
concentration and temperature but have a 27Al chemical shift
that is solvent-dependent.12,13 The 27Al chemical shift of the
-
Al2Cl7 dimer is reported to be δ ) +116 ppm.14 For
(10) Gray, J. L.; Maciel, G. E. J. Am. Chem. Soc. 1981, 103, 7147.
(11) Legrand, L.; Heintz, M.; Tranchant, A.; Messina, R. Electrochim. Acta
1995 40, 1711.
(12) Cerny´, Z.; Macha´cek, J.; Fusek, J.; Ca´sensky´, B.; Kriz, O.; Tuck, D.
G. J. Chem. Soc., Dalton Trans. 1998, 1439.
(8) Cambridge Crystallographic Data Centre, Cambridge, United Kingdom,
www.ccdc.cam.ac.uk, was searched using ConQuest v 5.25 (April
2004).
(9) The listed versions of SAINT, SMART, XSHELL, and XPOW in
SHELXTL and SADABS Software from Bruker Analytical X-ray
Systems Inc., (Madison, WI) were used in the analysis.
(13) Cerny´, Z.; Macha´cek, J.; Fusek, J.; Ca´sensky´, B.; Kriz, O.; Tuck, D.
G. Inorg. Chim. Acta 2000, 300, 556.
(14) Gray, J. L.; Maciel, G. E. J. Am. Chem. Soc. 1981, 103, 7147.
Inorganic Chemistry, Vol. 44, No. 16, 2005 5935

Boyle et al.
Figure 1. Raman spectra of SO2Cl2 prior to aging (solid line) and after
aging at 74 °C for 2 weeks (dotted line).
-
unassociated or non-exchanging AlCl4 complexes line
widths on the order of only a few hertz have been reported
ranging from +101 to +108 ppm.15 Therefore, we propose
that the observed line width is due to the equilibrium between
the monomer and Al2Cl6 dimer (eq 1), wherein the Td-
coordinated Al dimer is the major species present.12,13
Figure 2. Solution 27Al NMR spectrum of (a) catholyte solution at room
temperature and (b) the simulated spectral deconvolution.
especially at the part per million scale. Therefore, water at
up to 500 ppm does not significantly affect the aging of the
catholyte system; nor does it play a role in the reactions
associated with the active Al metal centers.
2AlCl3 a [Cl2Al(µ-Cl)]2
(1)
Because of the limited number of NMR handles available
for SO2Cl2, Raman studies were undertaken to elucidate its
aging behavior. For a flame-sealed NMR tube of SO2Cl2,
after 2 weeks at 74 °C, we observed the in-growth of a peak
at 1142 cm-1, which is consistent with SO2(g) generation.
The Raman spectra are shown in Figure 1. Observed at 500-
550 cm-1, but not shown, are bands associated with Cl2(g)
that developed in a manner consistent with the generation
of SO2(g). The Raman spectra indicate that SO2Cl2 decom-
poses as shown in eq 2.16
SO2Cl2 + 2H2O a H2SO4 + 2HCl
SO2Cl2 + H2SO4 a 2ClSO3H
(3)
(4)
Studies on the effect of AlCl3 concentration on molecular
speciation were undertaken by varying the amounts of AlCl3
added to the same volume of SO2Cl2. It was found that the
same Al species were present in each sample but at different
relative amounts. The ratio of these species was influenced
by aging time, temperature of reaction, and AlCl3 concentra-
tion. Since the two major variables of interest (humidity and
concentration) appeared to have little effect on the identity
of the molecular species present, our efforts focused on
understanding the aging of the catholyte without added water
and with the AlCl3 concentration normally used in battery
operation.
Aging of Catholyte. Over time, the solution 27Al NMR
revealed the presence of at least three different types of Td-
bound Al species present in the catholyte (see Figure 2) at
δ ) +99.0 (line width ) 1400 Hz), +82.4 (1630 Hz), and
+77.2 ppm (820 Hz). The decreasing 27Al chemical shift of
these species is consistent with a Td Al centers gaining more
oxygen character.11,13 With increased temperature and aging
time, the relative concentrations of the smaller chemical shift
(more shielded) Al species increased. The large line widths
again suggest that these different Al species are involved in
solvent complexes or exchange processes.
SO2Cl2 a SO2(g) + Cl2(g)
(2)
Catholyte Solution. Upon addition of AlCl3 to a clear
SO2Cl2 solution, the reaction mixture turns pale yellow. Of
initial interest was whether H2O (0, 100, 300, 500 ppm)
effected the molecular speciation. Based on 27Al NMR data
from AlCl3/SO2Cl2 solutions to which up to 500 ppm of water
had been added, no change in the number or types of Al
species present was observed, and none of the observed peaks
were consistent with the previously identified hydrolysis
species of AlCl3.13 The lack of significant Al hydrolysis in
the AlCl3/SO2Cl2 catholyte solution is not unexpected since
the chemistry associated with SO2Cl2 (the major component)
would necessarily dominate. Water addition to SO2Cl2 has
been found to follow the reactions shown in eqs 3 and 4,17-19
which would have little effect on the Al species present,
(15) Akitt, J. W. Ann. Rep. NMR Spectrosc. 1972, 5A, 465.
(16) Walker, C. W. J.; Wade, W. L. J.; Binder, M. J. Power Sources 1989,
25, 187.
(17) Hausladen, M. C.; Satterley, B. W.; Burger, M. J.; Lund, C. R. F.
Appl. Catal., A 1998, 166, 55.
(18) Fleischer, N. A.; Pallivathikal, M. B. M. J. Electrochem. Soc. 1987,
134, 513.
(19) Fleishcher, N. A.; Pallivathikal, M.; Babai, M. J. Electrochem. Soc.
1985, 132, C343.
Two-dimensional 27Al NMR exchange experiments were
pursued to measure the interchange between these different
Al species (an example is shown in Figure 3). The Al species
comprising the two dominant peaks at δ ∼ +101 and +80
ppm were found to be in rapid exchange (∼1 ms), supporting
the argument of an equilibrium between complexes in
5936 Inorganic Chemistry, Vol. 44, No. 16, 2005

Speciation in the AlCl3/SO2Cl2 Catholyte System
Figure 3. 2D 27Al NMR exchange spectrum of catholyte. The broad
resonance at ∼50 ppm is due to probe background.
Figure 5. Thermal ellipsoid plot of 1. Thermal ellipsoids are drawn at
30% level.
4). Further, the evolution of SO2(g) was significantly reduced
in the SO2Cl2 solution containing AlCl3 (1142 cm-1, Figure
4) compared to the solution containing SO2Cl2 only (Figure
1). In the SO2Cl2 solution containing AlCl3, AlCl3 either
consumes any SO2(g) generated or reacts with species that
otherwise would evolve SO2(g). Cooling a sample that had
been aged yielded crystalline material which was identified
as [Cl2Al(µ-O2SCl)]2 (1) and is shown in Figure 5.
Compound 1. Compound 1 consists of a cyclic dimer with
µ-O linking the S and the main group metal Al. The Td-
bound Al metal centers have two terminal Cl, whereas the
trigonal S has one terminal Cl. The equal distances noted
for 1 imply that the double-bonded oxygen is dispersed over
the two S-O bonds. The Al-Cl and Al-O bonds are
consistent with literature values.8,20-27 The constructs of 1
are unique8 and can either represent coordination of SO2Cl2
with loss of Cl2 or as an SO2 insertion into an Al-Cl bond
product. Table 1 lists the collection data parameters for
compound 1. Due to the consumption of the SO2, as
suggested by the Raman spectra, the latter explanation is
favored. Compound 1 is not stable at high temperature and
will convert to AlCl3 (as confirmed by solid-state NMR
spectroscopy experiments) upon washing with any solvent
or drying in vacuo. This implies a reversible insertion of
SO2(g) and suggests the equilibrium shown in eq 5. The
solid-state 27Al MAS NMR spectrum (not shown) of 1
showed two peaks at δ +101 and +96.5 ppm. Due to the
rapidity of decomposition (gas evolution) it is not unexpected
Figure 4. Raman spectra of SO2Cl2/1.5 M AlCl3 prior to aging (solid
line) and after aging at 74 °C for 2 weeks (dotted line).
solution. Exchange with the minor species at ∼+85 ppm is
not readily apparent in the 2D 27Al exchange experiment due
to the low concentration of this species and the large line
widths of the other dominant resonances.
In an effort to identify some of these Al species, we
investigated the low-temperature behavior of AlCl3 in SO2Cl2
soon after its addition. This was undertaken by generating a
sample of the catholyte, sealing it in a glass tube, and freezing
it in liquid nitrogen. The sample was transferred to a cold
NMR probe and was slowly warmed to room temperature.
The single 27Al NMR resonance observed at δ +101 ppm
(identical to AlCl3 in tol-d8) is assigned to the Td dinuclear
species Al2Cl6 in eq 1.
In the freshly prepared SO2Cl2/AlCl3 solution Raman
spectroscopy (Figure 4) revealed the presence of a species
with Raman bands at 1129 and 1375 cm-1. These bands are
not present in Raman spectra of SO2Cl2 solutions without
AlCl3 (Figure 1) and are believed to be due to an adduct
between SO2Cl2 and AlCl3. This is in contrast to Walker
and co-workers who report that the spectrum remained
essentially unchanged from pure SO2Cl2; however, due to
the very high noise produced from a large amount of
background fluorescence,16 the authors may not have been
able to observe the subtle changes noted in this report. After
2 weeks of aging at 74 °C, these bands disappeared from
the Raman spectrum of the SO2Cl2/AlCl3 solution (Figure
(20) Alexander, M. R.; Mair, F. S.; Pritchard, R. G.; Warren, J. E. Appl.
Organomet. Chem. 2003, 17, 730.
(21) Bissinger, P.; Mikulcik, P.; Riede, J.; Schier, A.; Schmidbaur, H. J.
Organomet. Chem. 1993, 446, 37.
(22) Gelbrich, T.; Dumichen, U.; Jorchel, P. Acta Crystallogr., Sect C:
Cryst. Struct. Commun. 1999, 55, 856.
(23) Jorchel, P.; Sieler, J. Z. Kristallogr. 1997, 212, 884.
(24) Kouvetakis, J.; Ritter, C.; Groy, T. L. Acta Crystallogr., Sect C: Cryst.
Struct. Commun. 2000, 56, e564.
(25) Sangokoya, S. A.; Pennington, W. T.; Robinson, G. H.; Hmcir, D. C.
J. Organomet. Chem. 1990, 385, 23.
(26) Schumann, H.; Kaufmann, J.; Deschert, S.; Schmalz, H.-G. Tetrahe-
dron Lett. 2002, 43, 3507.
(27) Szumacher, S.; Kunicki, A. R.; Madura, I.; Zachara, J. J. Organomet.
Chem. 2003, 682, 196.
Inorganic Chemistry, Vol. 44, No. 16, 2005 5937

Boyle et al.
Table 1. Data Collection Parameter for 1
compd
chem formula
fw
1
Al2Cl6O4S2
394.78
temp (K)
space group
a (Å)
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
168(2)
triclinic P(1h)
6.188(7)
7.351(8)
7.816(8)
94.202(17)
106.616(16)
109.251(15)
316.0(6)
1
2.074
1.810
2.94 (3.71)
7.32 (7.68)
D
calcd (Mg/m3)
µ(Mo KR) (mm-1
)
R1a (%) (all data)
wR2b (%) (all data)
Figure 6. Raman spectrum of thionyl chloride (SOCl2, top) and the
difference spectrum (gelled solid minus SO2Cl2, bottom) from a battery
reservoir.
a R1 ) ∑||Fo| - |Fc||/∑|Fo| × 100. b wR2 ) [∑w(Fo - Fc )2/
2
2
2
∑(w|Fo| )2]1/2 × 100.
Proposed Scheme of Catholyte Behavior. On the basis
of the above data, we have proposed a general reactive
pathway (Scheme 1) for the catholye AlCl3/SO2Cl2 system.
The initial mixture has several equilibria as shown in Scheme
1a,b. From this mixture a number of adducts could result.
We have identified 1 (Scheme 1b′). Raman and NMR data
indicate a great deal of ligand exchange, and molecular
transformations are occurring for this system. An O-S bond
in a compound like 1 may be broken followed by rearrange-
ment of the freed O atoms (Scheme 1c). A simple Cl transfer
forms a small dinuclear species with more oxo character and
the SOCl2 moiety noted above. Once formed, this precursor
can start to cross-link (Scheme 2c) forming larger species
(Scheme 2d). For this process a slight rearrangement of
Al-O bonds results in larger oligomers. The degree that this
rearrangement occurs will determine the solubility. For this
mechanism, it should be noted that the Al atoms all remain
Td and adopt more oxo character. These changes are
consistent with the analytical data collected on the catholyte
system.
To study the precipitate, it was first necessary to generate
a significant volume of material. Therefore, a sample of 1.5
M AlCl3 in SO2Cl2 in a glass ampule under vacuum was
sealed and heated to 120 °C overnight. A large volume of
precipitate was observed; however, the sample was found
to be under significant pressure even at room temperature
(Caution: Great care must be taken in handling these
samples due to the potential of explosion from the gas
buildup). Therefore, after cooling to room temperature, the
sample was frozen in liquid nitrogen, transferred into the
glovebox, opened, and allowed to slowly melt and de-gas.
The mother liquor was decanted off from the off-white solid
precipitate that formed. From the 27Al solid-state NMR data
(not shown) for this precipitate, it is readily apparent that
the Td coordination is no longer the dominant feature. The
major species present now have octahedral (Oh) and trigonal-
bipyramidal (tbp) geometries. The cross-linking of the
smaller molecules would lead to polymers and eventually a
precipitate (Scheme 2).
to see the peak associated with the Al2Cl6 equilibrium
present. The +96.5 ppm resonance is similar to the +94 ppm
resonance assigned to the AlCl3‚xSO2 observed in dialkyl
sulfone mixtures.11
(xs) SO2(g) + Al2Cl6 a [Cl2Al(µ-O2SCl)]2
(5)
To verify the insertion/deinsertion of SO2(g), the synthesis
of 1 was undertaken by bubbling SO2(g) through AlCl3 in
toluene. Undissolved AlCl3 was instantly solubilized, and
the resulting crystals isolated proved to be 1. Again, studying
the early species formed by freezing an NMR tube of AlCl3
with condensed SO2(g) and allowing it to slowly warm to
room temperature allowed some insight into the reaction
behavior of the AlCl3. Caution: Extreme caution must
be used during condensation of SO2(g) to aVoid oVer-
pressurization of the NMR tube. Initially the Al2Cl6 resonance
at +102 ppm was observed; however an SO2(g) adduct,
which grows in over time, was also observed at a resonance
of ∼+80 ppm. This 27Al NMR feature is consistent with
the assignment discussed for the original crystal of 1. The
27Al NMR spectrum of this compound has resonances that
are similar but not identical to what was observed for the
original catholyte mixture. Therefore, the SO2Cl2 leads to
additional alternative reactivity (i.e., there is more to the
reaction than the interaction of SO2(g) and AlCl3).
Precipitate. As the catholyte solution ages, 27Al NMR
studies indicate that the soluble species become more oxo
in nature until eventually a precipitate forms. This gellike
precipitate, present in aged solutions in both NMR tubes and
battery reservoirs, was analyzed by micro-Raman spectros-
copy. Bands due to SO2Cl2 entrapped in the precipitate were
subtracted from the as-obtained spectra, and with a typical
difference Raman spectrum being shown in Figure 6. The
derivative-like feature(s) in the difference spectrum of Figure
6 are related to incomplete removal of the SO2Cl2 bands.
Bands due to dissolved Cl2 are also present. Interestingly,
the remaining features in the difference spectrum are similar
to Raman bands from thionyl chloride (SOCl2, top, Figure
6), indicating that the precipitate has features consistent with
SOCl2.
Investigation of the mother liquor lends some insight into
the behavior of this precipitate. For the 24 h mother liquor
5938 Inorganic Chemistry, Vol. 44, No. 16, 2005

Speciation in the AlCl3/SO2Cl2 Catholyte System
Scheme 1. Observed Chemistry of SO2Cl2 Catholyte Solution: (a) Initial Mixture, (b) Intermediate Solution, (b′) One Possible Adduct Isolated at
Low Temperature (1), and (c) Precipitate Precursor
Scheme 2. Conversion from (c) Precipitate Precursor to (d) Precipitate or Gel
solution, the 27Al NMR spectrum is similar to what was
previously discussed. However, when compared to a sample
that was heated for 168 h, Al2Cl6 was found to dominate
the latter 27Al NMR spectrum. This can only be accounted
for if the precipitate is in equilibrium with the solution. That
is, the Al2Cl6 should not spontaneously regenerate if the
reaction is slowly and irreversibly forming precipitate;
however, if the precipitate is being re-dissolved, the equi-
librium would shift to the left and form the observed Al2Cl6
species.
Finally, it should be mentioned that the experimental
approach used in these studies, namely, the (1) hermetically
sealed vials, (2) temperatures, (3) concentrations, and (4)
minor impurities and their relative amounts, closely ap-
proximate those used for storage of both the catholyte and
the reserve battery itself, where the catholyte is held in a
hermetically sealed container prior to battery activation.
Consequently, the results of these studies are directly
applicable to the problem of understanding and improving
battery performance over extended time frames and in
identifying storage protocols that extend battery shelf life.
Summary and Conclusion
The fundamental chemical behavior of the catholyte AlCl3/
SO2Cl2 system was investigated using a variety of analytical
tools including Raman and 27Al NMR spectroscopies and
single-crystal X-ray diffraction. From these investigations,
it was determined that all catholyte samples possessed the
same chemical species in solution. Hydrolysis has been
eliminated as a dominant mechanism in the precipitation
formation since both the 27Al NMR and Raman investigations
reveal that water does not significantly alter the number, or
concentration, of the catholyte’s chemical species. Three
species that were identified include a starting material
(Al2Cl6), an SO2 adduct 1, and a rearranged oxo-like species
with a SOCl2 moiety. Scheme 1 shows some of the potential
pathways to these species. Both time and temperature control
the kinetics of this complex reaction system. However, over
time a precipitate will form that appears to be in equilibrium
with the mother liquor. Therefore, based on these data, low-
temperature storage will prevent precipitate formation.
Nevertheless, even after prolonged high-temperature storage,
Inorganic Chemistry, Vol. 44, No. 16, 2005 5939

Boyle et al.
the catholyte is expected to remain active due to the inter-
action of the solution, precipitate, and the relative amounts
of unreacted species remaining. Hence, the reserve battery
is expected to function normally, although formation of
excessive amounts of precipitate may impair the mechanical
aspects of the battery activation process. Additional studies
on the extent of reactivity that this catholyte can display
under more extreme conditions as well as the electrochemical
behavior of the stored material remain to be investigated.
Acknowledgment. Sandia is a multiprogram laboratory
operated by Sandia Corp., a Lockheed Martin Co., for the
United States Department of Energy’s National Nuclear
Security Administration under Contract DE-AC04-94AL85000.
Supporting Information Available: CIF data are given.
This material is available free of charge via the Internet at
http://pubs.acs.org.
IC050621Z
5940 Inorganic Chemistry, Vol. 44, No. 16, 2005