Manganese(III) chemistry in KOH solutions in the presence of Bi- or Ba-containing compounds and its implications on the rechargeability of γ-MnO2 in alkaline cells

Article abstract of DOI:10.1149/1.1622960

The influence of Bi- or Ba-containing compounds on the rechargeability of γ-MnO₂ in alkaline electrolytes has been investigated with AA cells containing cylindrical cathodes and flooded cells containing thin-film type cathodes. In addition to the electrochemical evaluation of the cells, the discharged cathodes were analyzed by X-ray diffraction after washing and drying. The incorporation of bismuth or barium into the cathodes was found to improve the cell cyclabibility, which is partly due to the suppression of electrochemically inactive phases such as birnessite (δ-MnO₂) and hausmannite (Mn₃O₄). A series of chemical oxidation reactions of Mn(OH)₂ with H₂O₂ in KOH medium and nonredox reactions of Mn(III) acetate with KOH followed by an analysis of the solid and filtrate indicate that the Mn³⁺ ions, which were in equilibrium with the solid phases containing Mn(III), disproportionated into Mn(II) compounds and Mn(IV) oxides. Reaction mechanisms involving Mn(III) compounds in KOH solution and the role of bismuth or barium on those reactions are discussed.

Full text of DOI:10.1149/1.1622960

Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
A1651
0013-4651/2003/150͑12͒/A1651/9/$7.00 © The Electrochemical Society, Inc.
Manganese„III… Chemistry in KOH Solutions in the Presence
of Bi- or Ba-Containing Compounds and Its Implications
on the Rechargeability of ␥-MnO2 in Alkaline Cells
D. Im,a, A. Manthiram,
*
a,**,z
and B. Coffeyb
aMaterials Science and Engineering Program, The University of Texas at Austin, Austin, Texas 78712, USA
bRBC Technologies, College Station, Texas 77840, USA
The influence of Bi- or Ba-containing compounds on the rechargeability of ␥-MnO2 in alkaline electrolytes has been investigated
with AA cells containing cylindrical cathodes and flooded cells containing thin-film type cathodes. In addition to the electrochemi-
cal evaluation of the cells, the discharged cathodes were analyzed by X-ray diffraction after washing and drying. The incorporation
of bismuth or barium into the cathodes was found to improve the cell cyclability, which is partly due to the suppression of
electrochemically inactive phases such as birnessite (␦-MnO2) and hausmannite (Mn3O4). A series of chemical oxidation reac-
tions of Mn͑OH͒ with H2O2 in KOH medium and nonredox reactions of Mn͑III͒ acetate with KOH followed by an analysis of
2
the solid and filtrate indicate that the Mn3ϩ ions, which were in equilibrium with the solid phases containing Mn͑III͒, dispropor-
tionated into Mn͑II͒ compounds and Mn͑IV͒ oxides. Reaction mechanisms involving Mn͑III͒ compounds in KOH solution and the
role of bismuth or barium on those reactions are discussed.
© 2003 The Electrochemical Society. ͓DOI: 10.1149/1.1622960͔ All rights reserved.
Manuscript submitted December 26, 2002; revised manuscript received June 3, 2003. Available electronically October 17, 2003.
Alkaline batteries based on manganese oxide cathodes are widely
used for consumer applications. However, due to the poor recharge-
ability, the usage has been restricted mostly to primary cells. Be-
cause manganese is inexpensive and environmentally benign, re-
chargeable cells employing manganese oxide cathodes could replace
the currently prevailing Ni-based secondary cells. Mn-based second-
ary cells would offer substantial cost benefits particularly with re-
spect to large-cell applications such as electric vehicles.
Two approaches have been pursued in the past to render re-
chargeability to manganese oxides. One approach is with the one-
electron process,1-12 in which ␥-MnO2 is employed as cathodes.
␥-MnO2 has a tunnel structure formed by an intergrowth of pyro-
lusite (␤-MnO2) and ramsdellite domains.13 It becomes ␦-MnOOH
after discharge by the intercalation of protons into the tunnels. The
second approach is with the two-electron process,14-25 in which the
cess, there are still some issues that need to be addressed. In a recent
study, we showed that hausmannite could still be formed even in the
presence of bismuth depending on the discharge condition and mi-
crostructure of the cathode.25 While the formation of hausmannite
could be effectively suppressed by ball milling the cathode mixture,
it is not clear whether the hausmannite formation could be totally
prevented during extended storage at an intermediate depth of dis-
charge ͑DOD͒, in which the manganese valence is about ϩ3. An-
other drawback is the low discharge voltage ͑less than 1.0 V vs. Zn͒
for the two-electron process, which makes it difficult to compete
with the current alkaline primary cells in the market.28
While the two-electron process involves two-phase reactions, the
one-electron process is nearly a homogeneous reaction with a linear
voltage profile.29-32 The one-electron process involves an insertion/
extraction of protons into/from the cathode lattice during the
discharge/charge process. Although the overall crystal structure of
␥-MnO2 is maintained during the discharge, the lattice expands due
to proton insertion and a decrease in the oxidation state of manga-
nese. The end product of discharge, ␦-MnOOH, also has an inter-
growth structure composed of ␣-MnOOH and ␥-MnOOH, which are
isostructural with ramsdellite and pyrolusite respectively.
The insertion/extraction of protons into/from ␥-MnO2 is revers-
ible to some degree. It has been reported that the lattice parameters
of ␥-MnO2 are fully recovered after charge if the degree of reduc-
tion does not exceed 0.5 e/Mn.31 Commercial cells have been de-
veloped based on this idea of limited DOD, which could be achieved
by having a controlled amount of Zn at the anode.2,4 In contrast,
deep discharge (Ͼ0.5 e/Mn) has been found to cause some irrevers-
ibility. For example, a capacity loss of 10 to 20% is observed in the
first discharge/charge cycle when the electrode is discharged to 0.9
V vs. Zn, which corresponds to ϳ0.8 e/Mn. Also, the lattice param-
eters did not revert back fully to the initial state of fresh ␥-MnO2
after a discharge of 0.8 e/Mn. A large overpotential occurring during
the charge process,33 and Jahn-Teller distortion34 and
amorphization31 occurring at deep discharge have been suggested to
be the cause of the irreversible capacity loss encountered during the
first cycle. Although no method has been proposed to fully eliminate
this capacity loss, partial reduction of ␥-MnO2 with chemical reduc-
ing agents has been found to alleviate it, but with a sacrifice in the
overall capacity.3
discharged and charged phases are, respectively, Mn͑OH͒ and
2
␦-MnO2 . The latter has a layered structure and is called birnessite
MnO2 . The two-electron process involves
a
dissolution/
precipitation process rather than
deintercalation process.
a
simple intercalation/
Between the two approaches, the two-electron process provides
an important advantage in terms of energy density. An early devel-
opment in this approach was by the researchers at the Ford Motor
Company.14-16 They found that an incorporation of small amount of
bismuth into the manganese oxide cathode improves the reversibility
remarkably. It was an innovative finding since it had been believed
that the reduction product of birnessite is hausmannite (Mn3O4),
which is electrochemically inactive.26,27 Encouraged by this finding,
several groups focused on Bi-containing manganese oxides.
For example, Conway et al.17-19 investigated the reaction inter-
mediates with spectroscopic and rotating ring disk electrode
͑RRDE͒ techniques and identified dissolved Mn͑III͒ species in the
electrolyte, confirming the theory of the dissolution/precipitation
process. Bode et al.21 suggested from voltammetric studies that the
formation of Mn-Bi complexes prevent the growth of hausmannite.
Our group showed from extensive X-ray diffraction ͑XRD͒ and
chemical analysis studies that Kϩ ions are incorporated from the
electrolyte into the cathode lattice during charge, assisting the for-
mation of a well crystalline birnessite phase.24 Despite the high
energy density and reasonable cyclability for the two-electron pro-
Although such a large capacity loss is not observed in the sub-
sequent cycles, a slow but steady fade is found to continue. This
secondary fading process limits the practical life of the cells. Ac-
cording to Mondolini et al.,31 the discharge capacity drops below
50% of the initial value within 50 to 100 cycles. Cyclability data
* Electrochemical Society Student Member.
** Electrochemical Society Active Member.
z E-mail: rmanth@mail.utexas.edu
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1652
Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
from other groups9 indicate even worse performance than this, and
several reasons have been proposed. Kordesch et al.4 pointed out
that electrode thickness is important and a thinner cathode would
provide better cycle life. They suggested the use of additives such as
BaSO4 could improve the mobility of ions within the electrode and
reduce the polarization. However, the same group reported later that
the failure of Zn anode is a dominant factor.9
and pestle, followed by rolling into thin sheets. In the case of Bi-
and Ba-containing cathodes, 10% of EMD ͑7.5 wt % of the overall
cathode͒ was replaced, respectively, with a bismuth-modified man-
24,25
ganese dioxide ͑BMD͒
and BaSO4 or Ba͑OH͒2 ; the BMD con-
sisted of 28 wt % manganese oxide and 72 wt % Bi2O3 . A three-
electrode system was constructed with the cathode thus fabricated
on a nickel-mesh current collector,25 a NiOOH/Ni͑OH͒ counter
2
As mentioned earlier, bismuth enhances the reversibility of the
two-electron process. During the investigation of the role of bismuth
on the discharge/charge process of ␥-MnO2 , it was noted that it is
also beneficial to the one-electron process. Especially, bismuth was
found to suppress the secondary fading as BaSO4 did. We also found
that the chemical stability of manganese oxides in the discharged
state played a critical role in rendering rechargeability.
electrode, a Hg/HgO reference electrode, and a 9 M KOH electro-
lyte. After discharging to Ϫ0.4 V vs. Hg/HgO reference electrode at
C/2 rate, the working electrodes were removed from the cell, still
attached to a nonwoven paper separator, transferred without washing
into a sealable bag, and stored for 7 days in contact with the elec-
trolyte. The electrodes were then separated, washed with DI water,
dried in vacuum at room temperature, and analyzed by XRD.
With an aim to develop a better understanding of the role of
bismuth or barium, we present here the results of a number of de-
signed chemical reactions. Chemical methods are often used to
simulate the electrochemical reactions not only in the study of
␥-MnO2 ,34-37 but also in cases involving nonaqueous systems.38,39
Although chemical reactions cannot be identical to their electro-
chemical counterparts, they usually allow easier and more detailed
characterization of the materials. As a model system to investigate
the chemistry of Mn͑III͒ compounds in strong alkaline environment,
The oxidation reactions of Mn͑OH͒ with H2O2 were carried out
2
as described below. 1.225 g of manganese͑II͒ acetate tetrahydrate
was dissolved in 70 mL of DI water. In the case of Bi- and Ba-
containing additives, 0.060 g of Bi2O3 or BaSO4 was dispersed into
the manganese acetate solution thus prepared. The Mn͑OH͒ sus-
2
pension in 7 M KOH solution was generated in situ by adding an
appropriate amount of KOH pellets to the above solution. It took
less than 15 s to achieve the complete dissolution of KOH. Finally,
0.283 g of a 30 wt % H2O2 solution was added and the whole
mixture was agitated on a magnetic stirrer. While the same kind of
reaction can be used to produce birnessite,41 the amount of H2O2
was set to make the final average oxidation state of Mn to be ϩ3 in
this case. After a specific length of time, the solid was filtered,
washed thoroughly, dried, and analyzed by XRD. The oxidation
state and amount of Mn in the solid were analyzed, respectively,
with a redox titration employing oxalate and atomic absorption
spectroscopy ͑AAS͒. The amount of Mn in the filtrate was also
determined by AAS after appropriate dilutions and carrying out the
measurement within 30 min of filtration to suppress any error arising
from further precipitation.
we performed a series of oxidation reactions of Mn͑OH͒ with hy-
2
drogen peroxide (H2O2) in the presence and absence of barium and
bismuth compounds. After a certain amount of reaction time, the
solid and liquid phases in the reaction mixture were separated and
characterized. In order to supplement the results of oxidation reac-
tions, we also conducted an additional series of nonredox type reac-
tions involving the dissolution of manganese͑III͒ acetate in KOH
and an analysis of the chronological variation of solid phases and
Mn concentration in solution. Based on the experimental results, the
role of bismuth or barium on the rechargeability of ␥-MnO2 in the
one-electron process is discussed.
The experiments involving nonredox type reactions were carried
out in a manner similar to that described in the previous paragraph.
0.200 g of manganese͑III͒ acetate dihydrate and, where needed,
0.015 g of Bi2O3 or BaSO4 were dissolved/dispersed in 70 mL of
water. KOH was then added to make the solution to have a net
concentration of 7 M KOH. After a specific length of time, the solid
and filtrate were separated and analyzed as described in the previous
paragraph.
Experimental
Electrolytic manganese dioxide ͑EMD͒ was supplied from Che-
metals, Inc. Manganese͑II͒ acetate tetrahydrate ͑99 %, Acros Or-
ganic͒, and manganese͑III͒ acetate dihydrate ͑Alfa Aesar͒ were used
as-received. Bi2O3 was synthesized by dissolving Bi(NO3)3 in di-
lute nitric acid ͑prepared by adding several drops of nitric acid to
100 mL of deionized ͑DI͒ water͒ and then adding dilute ammonium
hydroxide until the pH reached 12. The solid formed was filtered,
washed with DI water, and fired at 600°C for 8 h to obtain Bi2O3 .
All other chemicals were purchased from commercial sources and
used without further purification. Water was triply deionized prior to
use.
AA cells were fabricated and provided by RBC Technologies
͑College Station, TX͒ and Battery Technologies, Inc. ͑Canada͒. The
layout of the cell is similar to the description elsewhere5 and draw-
ings therein. A cylindrical cathode was positioned in the outer part
of a stainless steel container having a gelled Zn anode40 in the core.
Typically 88 wt % EMD (␥-MnO2), 7.5 wt % graphite, and 4.5
wt % KOH solution ͑9 M͒ were blended and pressed into a cathode
pellet. The actual wt % of EMD in the dry cathode was slightly
higher after the removal of water. In case of cathodes having Bi- and
Ba-containing additives, the amounts of EMD and graphite were
decreased accordingly. For cycle life tests, AA cells were discharged
into a 10 ⍀ load until the voltage reached 0.9 V, and then were
recharged at 1.75 V for 12 h. Cycled cells were disassembled, the
cathodes were thoroughly washed with DI water, and dried over-
night under vacuum at room temperature to prevent any possible
aerial oxidation of the manganese oxides. The dried cathodes were
then ground and analyzed by XRD.
Results and Discussion
XRD analysis of AA cell electrodes.—Three types of cathode
were prepared for cycle life test with the AA cells. The first type
consisted of only EMD and other conventional constituents such as
graphite, and is designated as type I cathodes. The second type
consisted of EMD and 5 wt % BaSO4 in addition to the conventional
additives, and is designated as type II cathodes. The third type con-
sisted of EMD, 5 wt % BaSO4 , and 4.5 wt % BMD24,25 in addition
to the conventional additives, and is designated as type III cathodes;
as mentioned earlier, the BMD consisted of 28 wt % manganese
oxide and 72 wt % Bi2O3 . While we continue to work toward iden-
tifying improved cathode formulations and additives, the cathode
compositions used in producing the AA cells in this study are rep-
resentative of typical best known formulations at the time of this
work and are amenable for direct cross comparison and evaluation
of the role of additives. Figure 1 compares the discharge curves of
the three types of cathodes at first, fifth, and tenth cycles. While the
first discharge curves look similar for the three types of cathodes,
the presence of additives in type II and III cathodes led to better
capacity retention.
Figure 2 shows the cyclability data of the AA cells containing the
three different types of cathodes. In accordance with the literature,4,9
a sharp decrease in capacity was observed after the first discharge/
charge cycle, and a gradual capacity fade continued after that. The
incorporation of BaSO4 into the cathode ͑type II͒ led to an improve-
Long-term storage stability of discharged cathodes was evaluated
with cells fabricated with thin-film type electrodes. The thin-film
type electrodes were prepared by mixing EMD ͑75 wt %͒, graphite
͑20 wt %͒, and polytetrafluoroethylene binder ͑5 wt %͒ in a mortar
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
A1653
Figure 2. Cycle life data of AA cells containing type I, type II, and type III
cathodes, whose compositions are described in the text.
found for the type II and III cathodes in Fig. 2 could be due to the
suppression of the electrochemically inactive phases, achieved by
the incorporation of Ba- and/or Bi-containing additives into the
cathode. However, some capacity fade still exists in the type III
Figure 1. Comparison of the discharge curves of the AA cells fabricated
with type I, type II, and type III cathodes at first, fifth, and tenth cycles.
ment in rechargeability as pointed out by Kordesch et al.4 Interest-
ingly, even better rechargeability was achieved with the type III
cathode containing both Bi2O3 and BaSO4 . After the completion of
the cycle life tests ͑Fig. 2͒, the cathode material of each cell was
collected in the discharged state and analyzed with XRD. As seen in
Fig. 3, birnessite (␦-MnO2) and spinel (Mn3O4 or ZnMn2O4)
phases could be identified in type I cathode, which did not have Ba-
and/or Bi-containing additives. Spinel phases are known to be elec-
trochemically inactive in alkaline electrolyte26 unless the electrode
potential is lowered below the practical operating range of ␥-MnO2
electrode (ϽϪ0.4 V vs. Hg/HgO͒.24 Birnessite phase is also inactive
in that potential range although it is actually the active material of
the two-electron process. Because the discharge voltage of birnessite
is ϳ0.4 V vs. Hg/HgO, it cannot contribute to the capacity of the
one-electron process. In other words, the birnessite or spinel phase,
once formed, would not participate in the one-electron discharge/
charge process any more. Based on this, the data in Fig. 3a suggest
that the formation of birnessite and/or spinel phases could be at least
partly responsible for the capacity fade.
Interestingly, the amounts of both the birnessite and spinel
phases have been suppressed with the incorporation of BaSO4 ͑type
II cathode͒ as seen in Fig. 3b. In Fig. 3b, no spinel reflections are
seen and those of birnessite are weaker than that in the type I cath-
ode, indicating that BaSO4 inhibits the formation of such inactive
phases. More importantly, no reflections of either spinel or birnessite
could be seen in the type III cathode ͑Fig. 3c͒ that contains both
BaSO4 and Bi2O3 . The results suggest that the better cyclability
Figure 3. XRD patterns of ͑a͒ type I, ͑b͒ type II, and ͑c͒ type III cathodes in
*
discharged state after cycling as in Fig. 1. The reflections marked with , ϩ,
᭝, #, ᭺ refer, respectively, to hausmannite, birnessite, graphite, BMD, and
the discharged form of ␥-MnO2 .
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1654
Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
cathode even though no inactive phase formation is seen in Fig. 3c,
suggesting that some additional factors could be contributing to the
fade.
Phase analysis of stored, discharged thin-film-type elec-
trodes.—Although it is clear that inactive phases tend to be formed
during cycling, the mechanism of their formation is not fully under-
stood. Two possible causes could be envisioned for their formation.
First, the large polarization encountered with the relatively thick
electrodes of the AA cells4 may lead to an inhomogeneity in dis-
charge. Any overdischarge that occurs would result in the formation
of some manganese͑II͒ compounds, which could be reoxidized to
birnessite during charge as in the two-electron process or could form
hausmannite by a reaction with the available manganese͑III͒
compounds.24-26 Second, the discharge product of EMD, which is
␦-MnOOH, itself could be unstable in strong alkaline solutions,
causing a dissolution of Mn3ϩ ions. The dissolved Mn3ϩ or
3ϩ
Mn͑OH͒
can readily transform into Mn͑II͒ and Mn͑IV͒ com-
͓
͔
6
pounds through a disproportionation reaction.21,22,42 This mecha-
nism is supported by some literature, proposing that the capacity
retention is inversely proportional to Mn3ϩ solubility; the solubility
of Mn3ϩ is higher in stronger alkaline solutions17-19,29 while the
rechargeability of ␥-MnO2 is better in more dilute electrolytes such
as 1 M KOH.3 In order to identify which mechanism is dominant,
we carried out long-term storage tests with the discharged thin-film
cathodes. The use of thin-film cathode was intended to minimize
inhomogeneity in discharge and provide a better phase uniformity
throughout the whole electrode.
The electrodes collected at the end of first discharge (Ϫ0.4 V vs.
Hg/HgO͒ were stored for 7 days in contact with 9 M KOH electro-
lyte ͑see Experimental section for details͒. Since the average oxida-
tion state of Mn in the EMD sample before discharge was deter-
mined to be ϩ3.91 and the discharge capacity was equivalent to
about 0.8 electrons per manganese regardless of the use of additives,
the manganese valence at the end of discharge is expected to be
about ϩ3.1. Figure 4 compares the XRD patterns of the various
cathodes after first discharge and subsequent storage for 7 days.
While no spinel reflections are found in any of the diffraction pat-
terns, a reflection at ϳ12° corresponding to birnessite is recognized
only when no Bi- or Ba-containing additive is present ͑Fig. 4a͒. The
formation of birnessite in Fig. 4a implies the dissolution of Mn3ϩ
ions since birnessite is believed to be formed through a solution
path;17-19 Mn3ϩ ions in solution can be reoxidized or disproportion-
ated to form birnessite. This result is consistent with a report by
Holton et al.35 that the birnessite is generated when ␥-MnO2 is
partially reduced with a chemical reducing agent and kept in a
strong KOH solution for 2 days even if the average oxidation state
of manganese is as high as ϩ3.84. They also suggested that the
Mn͑III͒ species, which are in a dynamic equilibrium with ␥-MnO2
solid solution, are intermediates en route to birnessite. Additionally,
it was claimed that a decrease in the oxidation state of manganese
results in an increase in the solubility of Mn3ϩ ions, facilitating the
formation of birnessite. Consequently, it can be concluded that the
formation of birnessite is initiated by the release of Mn3ϩ ions from
the reduced form of ␥-MnO2 , which could be to a large extent due
to its instability.
Figure 4. XRD patterns of thin-film cathodes composed of ͑a͒ EMD, ͑b͒
EMD ϩ BMD, ͑c͒ EMD ϩ BaSO4 , and ͑d͒ EMD ϩ Ba͑OH͒ along with
2
other conventional constituents after discharging to Ϫ0.4 V vs. Hg/HgO ref-
erence electrode and subsequently storing in contact with 9 M KOH in a
sealed bag. The reflections marked with ϩ, ᭝, #, ᭺, and ᭹ refer, respec-
tively, to birnessite, graphite, BMD, ␥-MnO2 , and polytetrafluoroethylene.
reduced manganese oxides plays a critical role on the rechargeabil-
ity. With an aim to develop a better understanding of the degradation
mechanism of reduced ␥-MnO2 and the role of barium and bismuth
on the rechargeability, we decided to investigate the chemistry of
trivalent manganese species in strong KOH solutions ͑7 M͒ with two
kinds of chemical reactions. One is an oxidation reaction of
Mn͑OH͒ with H2O2 and the other is a nonredox reaction starting
2
from manganese͑III͒ acetate. The average manganese oxidation state
was controlled to be ϩ3 in either case.
As soon as the H2O2 solution was added to a suspension of
Mn͑OH͒2 , the color of the reaction mixture changed from white to
dark brown. In some cases, the color of the filtrate was also brown
indicating the existence of Mn3ϩ ions.17-19 The solid products ob-
tained after various durations of reaction in the presence and ab-
sence of various additives were analyzed by XRD and the data are
given in Fig. 5-7. In the absence of additives, hausmannite and
birnessite phases formed nearly immediately ͑Fig. 5͒. No significant
changes in the X-ray patterns could be observed up to 24 h except-
ing a gradual growth of the birnessite reflections indicating
aging.41,43 Thermodynamically, hausmannite and birnessite appear
to be the most stable phases in aqueous KOH media, and kinetically,
their formation rate appears to be very fast. It is interesting to ob-
serve the conspicuous reflections of hausmannite here as a result of
controlled degree of oxidation since only birnessite is known to be
produced when excess amount of H2O2 is added.41 It might have
formed through a solution reaction between Mn͑II͒ and Mn͑III͒
complexes as suggested previously.21,22,25
However, no birnessite phase could be detected if Ba- or Bi-
containing compounds are incorporated into the cathode, suggesting
that they either prevent the dissolution of Mn3ϩ ions or block the
subsequent reactions of Mn3ϩ ions. Furthermore, similar results
were obtained when Ba͑OH͒ instead of BaSO4 was used ͑Fig. 4d͒,
2
suggesting that it is the Ba2ϩ ions that take part in the reaction
mechanism. Also, the chemistry of dissolved species appears to be
more important than the crystal structure or other solid-state prop-
erties of the additives.
Addition of a small amount of Bi2O3 to the manganese acetate
solution prior to the oxidation reaction produces totally different
X-ray patterns ͑Fig. 6͒. The broad reflection observed around 19°
Chemical oxidation of Mn(OH)2 with H2O2.—It is clear from
the results of the previous sections that the chemical stability of the
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
A1655
Figure 5. XRD patterns of the products obtained by oxidizing Mn͑OH͒
with H2O2 in 7 M KOH solution without any additives for ͑a͒ 5 min, ͑b͒ 10
2
Figure 6. XRD patterns of the products obtained by oxidizing Mn͑OH͒
with H2O2 in 7 M KOH solution in presence of 5 wt % Bi2O3 for ͑a͒ 5 min,
͑b͒ 10 min, ͑c͒ 30 min, ͑d͒ 60 min, and ͑e͒ 24 h. The reflections marked with
2
*
min, ͑c͒ 30 min, ͑d͒ 60 min, and ͑e͒ 24 h. The reflections marked with and
ϩ refer, respectively, to hausmannite and birnessite.
ࡗ, ϩ, and # refer, respectively to Mn͑OH͒ or ␤-MnOOH, birnessite, and
2
Bi2O3 .
might be due to an overlap of the reflections of Mn͑OH͒ and
2
␤-MnOOH. Since those two phases are isostructural, the oxidation
reaction can be considered to progress via a solid-state route. This is
and birnessite in the early stages of the reaction, leaving little or no
material for oxidation by air.
similar to the electrochemical oxidation of Mn͑OH͒ in the two-
2
The higher oxidation state found with the use of BaSO4 or Bi2O3
also suggests that the solid products contain one or more forms of
manganese͑IV͒ oxide despite the absence of reflections correspond-
ing to them in the X-ray patterns. Accordingly, barium or bismuth
appears to hamper the formation of good crystalline Mn͑IV͒ oxides
rather than to completely block the reaction leading to Mn͑IV͒ ox-
ides. The formation of Bi-Mn complexes proposed in the literature
could be a plausible explanation for this; the Bi-Mn complexes
formed may delay the construction of a layered structure consisting
of edge-shared MnO6 octahedra as they inhibit the formation of
hausmannite.21,22 Although birnessite with better crystallinity is
formed in the presence of bismuth in the case of a two-electron
process,24,25 which is a heterogeneous reaction occurring only on the
graphite surface, poorly crystalline or amorphous phases are formed
in the case of a one-electron process, which is a homogeneous re-
action occurring everywhere in the solution. Therefore, the dissimi-
larity in the degree of crystallization may be attributed to the differ-
ences in the reaction mechanisms between the two processes.
The variations of the concentration of manganese in the filtrate
with reaction time are compared in Fig. 9. The concentration of
manganese was lowest in the case of no additives, indicating a quick
formation of solid phases as anticipated from the X-ray results ͑Fig.
5͒, in which the variations in the diffraction patterns with time are
not considerable. A significant increase in the concentration of man-
ganese was found in presence of the additives BaSO4 and Bi2O3 ,
and the magnitude of increase is higher for Bi2O3 compared to
electron process, in which ␤-MnOOH is a transitional oxidation
product. No clear reflections corresponding to birnessite or haus-
mannite were observed up to 60 min ͑Fig. 6a-d͒. However,
Mn͑OH͒ or ␤-MnOOH was consumed to generate birnessite after
2
24 h as seen in Fig. 6e. Although similar results are found with the
addition of BaSO4 instead of Bi2O3 , the birnessite phase appeared
earlier in the case of BaSO4 and also the formation of hausmannite
is noticed at later stages of the reaction ͑Fig. 7͒. Based on these
results, it can be concluded that bismuth or barium blocks a solution
reaction path generating hausmannite, and bismuth is more effective
than barium for the same amount of additive.
Figure 8 shows the variations of the average oxidation state of
manganese in the solid products with reaction time. The oxidation
state is above the target value of ϩ3 in all cases. It would be rea-
sonable to regard the valence of the dissolved manganese species to
be less than ϩ3 to compensate for the higher oxidation state in the
solid products. In the absence of additives, the oxidation state in the
solid remains almost constant independent of the reaction time. In
contrast, the oxidation state increases with reaction time in the pres-
ence of bismuth or barium. The increase in oxidation state with time
could be due to the aerial oxidation of Mn͑OH͒ and ␤-MnOOH; it
2
has been reported that birnessite can be formed by bubbling oxygen
into an alkaline solution containing Mn͑OH͒2 .44 The constancy of
the oxidation state with time in the absence of barium or bismuth
could be due to the formation of stable phases such as hausmannite
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1656
Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
Figure 8. Variations with reaction time of the average oxidation state of
manganese in the solid obtained by oxidizing Mn͑OH͒ with H2O2 in 7 M
2
KOH solution in the presence and absence of additives.
nonredox type reaction. In this regard, Mn͑III͒ acetate, which is one
of the most popular Mn͑III͒ salts,45 was employed as a starting
Figure 7. XRD patterns of the products obtained by oxidizing Mn͑OH͒
with H2O2 in 7 M KOH solution in presence of 5 wt % BaSO4 for ͑a͒ 5 min,
͑b͒ 10 min, ͑c͒ 30 min, ͑d͒ 60 min, and ͑e͒ 24 h. The reflections marked with
2
material. Addition of KOH into a solution of Mn͑III͒ acetate is
3Ϫ
supposed to generate Mn͑OH͒
complexes immediately. There-
͓
͔
6
*
ࡗ, ϩ, and refer, respectively to Mn͑OH͒ or ␤-MnOOH, birnessite, and
2
hausmannite.
BaSO4 . This result supports the hypothesis that bismuth or barium
has an effect in preventing the dissolved Mn͑II͒ or Mn͑III͒ species
from forming solid phases, which has also been suggested by the
higher oxidation state (Ͼϩ3) of manganese in the solid phases. The
concentration of manganese decreases gradually with time in all
cases, but it is still significant even after 24 h if bismuth is present.
If Mn͑II͒ compounds have a greater tendency to stay in solution, the
higher concentration of manganese would mean that more Mn͑II͒
species are in solution; if so, one would expect more Mn͑IV͒ in the
solid to account for the total charge, which is in accordance with the
oxidation state data in Fig. 8. The gradual decrease in the manganese
concentration with time could be partially due to the aerial oxidation
to give Mn͑IV͒ in the solid as indicated by the increase in manga-
nese valence in the solid.
Disproportionation of manganese(III) complexes.—A series of
chemical reactions with controlled degree of oxidation have shown
that dissolved manganese ions have a tendency to transform into
solid phases and the rate of such reactions could be decreased by
incorporating Bi- or Ba-containing compounds into the reaction me-
dium. However, with the chemical oxidation method described in
the previous section, only the total number of electrons can be con-
trolled; some Mn͑III͒ ions could be oxidized further to Mn͑IV͒ by
reaction with H2O2 while some Mn͑II͒ compounds could remain
unchanged due to the depletion of oxidizing agent. Notwithstanding
the average oxidation state of ϩ3, we could observe both Mn͑II͒ and
Mn͑IV͒ compounds statistically independent of the stability of
Mn͑III͒ species. With an aim to investigate the stability of Mn͑III͒
species in the absence of oxidizing or reducing agents, we devised a
Figure 9. Variations with reaction time of the concentration of manganese in
the filtrate obtained by oxidizing Mn͑OH͒ with H2O2 in 7 M KOH solution
2
in the presence and absence of additives.
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
A1657
Figure 10. XRD patterns of the products obtained by the reaction of Mn͑III͒
acetate in 7 M KOH solution in the absence of additives for ͑a͒ 5 min, ͑b͒ 10
Figure 11. XRD patterns of the products obtained by the reaction of Mn͑III͒
acetate in 7 M KOH solution in presence of 5 wt % Bi2O3 for ͑a͒ 5 min, ͑b͒
10 min, ͑c͒ 60 min, and ͑d͒ 24 h. The reflections marked with ࡗ, ϩ, and #
*
min, ͑c͒ 60 min, and ͑d͒ 24 h. The reflections marked with and ϩ refer,
respectively, to hausmannite and birnessite.
refer, respectively, to Mn͑OH͒ or ␤-MnOOH, birnessite, and Bi2O3 .
2
fore, any detection of Mn͑II͒ or Mn͑IV͒ compounds in a later stage
would confirm the occurrence of disproportionation reactions.
The evolution of the X-ray patterns of the solid formed by add-
ing KOH into Mn͑III͒ acetate in the absence of additives is shown in
Fig. 10. Reflections corresponding to both birnessite and hausman-
nite appear almost immediately, indicating clearly a disproportion-
ation of the initial Mn͑III͒ into Mn͑IV͒ and Mn͑II͒. The birnessite
reflections become sharper with increasing reaction time, but the
reflections corresponding to hausmannite did not change signifi-
cantly. In general, the X-ray patterns in Fig. 10 are nearly identical
to those in Fig. 5 excepting that the reflections corresponding to
hausmannite are less prominent in Fig. 10. Addition of small
amounts of Bi2O3 or BaSO4 into the solution suppressed the reflec-
tions corresponding to birnessite and hausmannite ͑Fig. 11 and 12͒.
The hausmannite reflections were hardly observed while the aging
process of birnessite with an increase in intensity with time could be
noticed in both the cases. Bi2O3 appeared to be more effective than
BaSO4 in suppressing the formation of such solid phases, which is
consistent with the results of oxidation reactions presented in the
previous section.
The variations of the average oxidation state of manganese in the
solid with reaction time are shown in Fig. 13. The trend in Fig. 13 is
similar to that in Fig. 8, but the aerial oxidation was less severe
compared to that in the case of oxidation reactions. This could be
due to the lack of readily oxidizable phases such as ␤-MnOOH in
the case of nonredox reactions as indicated by the X-ray data in Fig.
11 and 12. The variations of the concentration of manganese in the
filtrate ͑Fig. 14͒ also resemble those in Fig. 10. Although the specific
values are different from those obtained with the oxidation reac-
tions, the manganese concentration in the filtrate increased on going
from no additives to BaSO4 additive to Bi2O3 additive. Thus the
data in Fig. 14 also illustrate that BaSO4 was less effective than
Bi2O3 in keeping the manganese ions in solution.
Figure 12. XRD patterns of the products obtained by the reaction of Mn͑III͒
acetate in 7 M KOH solution in presence of 5 wt % BaSO4 for ͑a͒ 5 min, ͑b͒
*
10 min, ͑c͒ 60 min, and ͑d͒ 24 h. The reflections marked with and ϩ refer,
respectively, to hausmannite and birnessite.
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1658
Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
Dissolution of Mn͑III͒ from oxides or oxyhydroxides could oc-
cur as
3Ϫ
MnOOH ↔ Mn͑OH͒
͓1͔
͓
͔
6
Disproportionation of Mn͑III͒ into Mn͑II͒ in solution and
Mn͑IV͒ in solid could occur as
3Ϫ
Þ
Mn͑OH͒
↔
Mn͑OH͒ 4Ϫ ϩ MnO
͔
2
͓2͔
͓
͔
͓
͔
͓
6
6
Formation of hausmannite (Mn3O4) could occur by a reaction
between the Mn͑III͒ and Mn͑II͒ complexes in solution as
4Ϫ
2 Mn͑OH͒ 3Ϫ ϩ Mn͑OH͒
͔
6
→ Mn3O4
͓3͔
͓
͔
͓
6
which is an irreversible reaction. Formation of birnessite could oc-
cur from the Mn͑IV͒ solid as
Þ
MnO
→ birnessite ␦-MnO ͒
͓4͔
͓
͔
͑
2
2
which is also an irreversible reaction.
It has been reported that Mn3ϩ ions are soluble to some extent in
strong alkaline solutions by forming a complex as shown in Reac-
tion 1.17 Hence, a solid containing Mn͑III͒ would be in equilibrium
with the complex ions and the solubility is determined by an equi-
librium constant, the solubility product (Ksp), for each situation.
The solubility product is proportional to eϪ⌬G, where ⌬G is the
difference in molar Gibbs free energies between the solid phase and
the complex ion in solution.46 Therefore, for a given KOH concen-
tration, the solubility will be influenced by the free energy of each
solid phase. Accordingly, energetically stable compounds such as
Mn2O3 would hardly generate the dissolved ions. Although it is
difficult to quantitatively assess the free energy of ␦-MnOOH, which
is a reduced form of ␥-MnO2 , it would not be very stable consid-
ering the lattice expansion occurring with proton insertion and the
lattice distortion caused by Mn3ϩ. A study by Holton et al.35 and
our results with the thin-film electrodes have already shown that
manganese ions could be dissolved out from the reduced ␥-MnO2 in
alkaline solutions. Bismuth or barium do not seem to suppress Re-
action 1 because manganese ions are found in the filtrate irrespective
of the presence of such additives.
Figure 13. Variations with reaction time of the average oxidation state of
manganese in the solid obtained by reacting Mn͑III͒ acetate in 7 M KOH
solution in the presence and absence of additives.
Role of bismuth and barium on the rechargeability of
␥-MnO2.—Based on our results presented in the previous sections
and some other studies in the literature,17,21,22,25,31,35 the following
reaction mechanisms could be suggested for the formation of haus-
mannite and birnessite.
The characterization data of the products formed during the
nonredox reactions confirmed the occurrence of a disproportionation
reaction denoted by Reaction 2. The Mn͑IV͒ oxides produced im-
Þ
mediately after the disproportionation are labeled as MnO
,
͓
͔
2
which represents transitional oxides that are en route to birnessite.
They could be oxides with only short-range order or with small
crystallite size that are difficult to be detected with XRD. The con-
cept of formation of MnO Þ is strongly supported by the results of
͓
͔
2
the oxidation reactions in presence of Bi2O3 . In those experiments,
while no forms of Mn͑IV͒ oxides could be detected with XRD, the
average oxidation state of manganese was above ϩ3, suggesting that
Mn͑IV͒ oxides are indeed formed. The immediate formation of bir-
nessite phase in the absence of additives may seem to suggest that
the transitional MnO Þ stage is skipped. However, the increase in
͓
͔
2
the intensity of the birnessite reflection at ϳ12° with time ͑Fig. 5͒
without a significant increase in the oxidation state ͑Fig. 8͒ suggests
that the transitional phases existed even in this case. A further as-
Þ
sumption can be made that the free energy of MnO
is not too
͓
͔
2
low for the equilibrium constant of Reaction 2 to be very high,
because it would drive rapidly Reaction 1 in the forward direction
due to a fast consumption of Mn͑OH͒ 3Ϫ. Nonetheless, the stor-
͓
͔
6
age stability tests of discharged electrodes show that ␦-MnOOH is
still the dominant phase even after 7 days of storage, indicating the
rate of Reaction 1 is only moderate.
Blocking Reaction 2 would suppress the formation of birnessite
or hausmannite. In this regard, bismuth or barium could be consid-
ered to affect Reaction 2 by reducing the equilibrium constant or by
decelerating the reaction rate. However, this hypothesis is in contra-
diction with the observation that the introduction of bismuth or
Figure 14. Variations with reaction time of the concentration of manganese
in the filtrate obtained by reacting Mn͑III͒ acetate in 7 M KOH solution in
the presence and absence of additives.
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑12͒ A1651-A1659 ͑2003͒
A1659
barium increases the amount of Mn͑IV͒ oxides ͑Fig. 8 and 13͒.
References
Therefore, we believe that bismuth or barium is more likely to affect
the subsequent Reactions 3 and 4 rather than Reaction 2 ͑see below͒.
It is now well accepted that hausmannite is formed through a
reaction involving solution species as in Reaction 3.21 This reaction
can be considered to be irreversible based on the high chemical
stability of hausmannite, which is in accordance with its poor elec-
trochemical activity.24,26 According to Bode et al.,21,22 the formation
1. K. Kordesch and J. Gsellmann, U.S. Pat. 4,384,029 ͑1983͒.
2. K. Kordesch, J. Gsellmann, M. Peri, K. Tomantschger, and R. Chemelli, Electro-
chim. Acta, 26, 1495 ͑1981͒.
3. K. Kordesch, J. Gsellmann, and K. Tomantschger, U.S. Pat. 5,011,752 ͑1991͒.
4. K. Kordesch and M. Weissenbacher, J. Power Sources, 51, 61 ͑1994͒.
5. T. N. Andersen, J. M. Berry, and J. M. Derby, U.S. Pat. 5,405,594 ͑1995͒.
6. K. Tomantschger, R. J. Book, and J. Daniel-Ivad, U.S. Pat. 5,424,145 ͑1995͒.
7. W. Jantscher, L. Binder, D. A. Fiedler, R. Andreaus, and K. Kordesch, J. Power
Sources, 79, 9 ͑1999͒.
8. F. Beck and P. Ruetschi, Electrochim. Acta, 45, 2467 ͑2000͒.
9. Y. W. Shen and K. Kordesch, J. Power Sources, 87, 162 ͑2000͒.
10. V. K. Nartey, L. Binder, and A. Huber, J. Power Sources, 87, 205 ͑2000͒.
11. S. Bodoardo, N. Penazzi, P. Spinelli, and M. Arrabito, J. Power Sources, 94, 194
͑2001͒.
12. J. Daniel-Ivad, E. Daniel-Ivad, and J. R. Book, U.S. Pat. 6,361,899 ͑2002͒.
13. P. M. De Wolff, Acta Crystallogr., 12, 341 ͑1959͒.
14. Y. F. Yao, N. Gupta, and H. S. Wroblowa, J. Electroanal. Chem. Interfacial Elec-
trochem., 223, 107 ͑1987͒.
15. H. S. Wroblowa and N. Gupta, J. Electroanal. Chem. Interfacial Electrochem., 238,
98 ͑1987͒.
16. M. A. Dzieciuch, N. Gupta, and H. S. Wroblowa, J. Electrochem. Soc., 135, 2415
͑1988͒.
of Bi-Mn complexes inhibits this reaction. For example, bismuth has
6ϩ
a tendency to form a positively charged cluster, Bi ͑OH͒
,
͓
͔
12
6
which could possibly interact with the Mn͑OH͒ 3Ϫ/4Ϫ species.22,47
͓
͔
6
The Bi-Mn complex formation would be an extreme case of such an
interaction. While it is not obvious whether the mechanism involv-
ing barium would be similar, a positively charged barium complex,
Ba(H O) 2ϩ, is known in the literature.48 However, further work
͓
͔
8
2
with spectroscopic techniques is essential to analyze the nature of
the species or complexes formed by the interaction of Mn with Bi or
Ba and understand the mechanism. Reaction 4 is another end point
of the reaction sequence generating birnessite. A lot of rearrange-
ment or migration processes are necessary for Reaction 4 to occur
whether it is a solution reaction or a solid-state reaction. Therefore,
some chemical interaction or complex formation between bismuth
or barium and manganese complexes could keep Reaction 4 from
proceeding, for example, by reducing the rate of rearrangement of
manganese ions.
17. D. Y. Qu, B. E. Conway, L. Bai, Y. H. Zhou, and W. A. Adams, J. Appl. Electro-
chem., 23, 693 ͑1993͒.
18. C. G. Castledine and B. E. Conway, J. Appl. Electrochem., 25, 707 ͑1995͒.
19. B. E. Conway, D. Qu, and J. McBreen, in Synchrontron Techniques in Interfacial
Electrochemistry, C. A. Melendres and A. Tadjeddine, Editors, p. 331, Kluwer
Academic Publishers, Dordrecht, The Netherlands ͑1994͒.
20. S. W. Donne, G. A. Lawrance, and D. A. J. Swinkels, J. Electrochem. Soc., 144,
2961 ͑1997͒.
21. M. Bode, C. Cachet, S. Bach, J.-P. Pereira-Ramos, J. C. Ginoux, and L. T. Yu, J.
Electrochem. Soc., 144, 792 ͑1997͒.
22. L. T. Yu, J. Electrochem. Soc., 144, 802 ͑1997͒.
23. Kh. S. Abou-El-Sherbini, M. H. Askar, and R. Schollhorn, Solid State Ionics, 139,
121 ͑2001͒.
24. A. M. Kannan, S. Bhavaraju, F. Prado, M. Manivel Raja, and A. Manthiram, J.
Electrochem. Soc., 149, A483 ͑2002͒.
25. D. Im and A. Manthiram, J. Electrochem. Soc., 150, A68 ͑2003͒.
26. J. McBreen, in Power Sources 5: Proceedings of the 9th International Power
Sources Symposium, D. H. Collins, Editor, p. 525, Academic Press, New York
͑1975͒.
Conclusions
An XRD analysis of the cycled manganese oxide cathodes of the
AA cells has shown that part of the capacity fade is due to the
formation of electrochemically inactive phases such as birnessite
(␦-MnO2) and hausmannite (Mn3O4) during the one-electron pro-
cess. A further study with thin-film electrodes reveals that the for-
mation of such electrochemically inactive phases is due to the insta-
bility of reduced ͑discharged͒ ␥-MnO2 and the dissolution of Mn3ϩ
ions. Incorporation of Bi- or Ba-containing compounds into the
cathode is found to suppress the formation of such inactive phases
and improve cyclability, and for the same amount of additive, bis-
muth is more effective than barium.
27. J. McBreen, Electrochim. Acta, 20, 221 ͑1975͒.
28. R. Patrice, B. Gerand, J. B. Leriche, L. Seguin, E. Wang, R. Moses, K. Brandt, and
J. M. Tarascon, J. Electrochem. Soc., 148, A448 ͑2001͒.
29. A. Kozawa and R. A. Powers, J. Electrochem. Soc., 113, 870 ͑1966͒.
30. Y. Chabre and J. Pannetier, Prog. Solid State Chem., 23, 1 ͑1995͒.
31. C. Mondoloni, M. Laborde, J. Rioux, E. Andoni, and C. Levy-Clement, J. Electro-
chem. Soc., 139, 954 ͑1992͒.
32. D. Boden, C. J. Venuto, D. Wisler, and R. B. Wylie, J. Electrochem. Soc., 114, 415
͑1967͒.
33. D. Boden, C. J. Venuto, D. Wisler, and R. B. Wylie, J. Electrochem. Soc., 115, 333
͑1968͒.
34. W. C. Maskell, J. E. A. Shaw, and F. L. Tye, Electrochim. Acta, 26, 1403 ͑1981͒.
35. D. M. Holton, W. C. Maskell, and F. L. Tye, in Power Sources 10: Proceedings of
the 14th International Power Sources Symposium, L. J. Pearce, Editor, p. 247, Paul
Press, London ͑1985͒.
36. J. Fitzpatrick and F. L. Tye, J. Appl. Electrochem., 21, 130 ͑1991͒.
37. F. Jean, C. Cachet, L. T. Yu, A. Lecerf, and A. Quivy, J. Appl. Electrochem., 27,
635 ͑1997͒.
38. R. V. Chebiam, A. M. Kannan, F. Prado, and A. Manthiram, Electrochem. Com-
mun., 3, 624 ͑2001͒.
39. R. V. Chebiam, F. Prado, and A. Manthiram, J. Solid State Chem., 163, 5 ͑2002͒.
40. J. Daniel-Ivad, R. J. Book, and K. Tomantschger, U.S. Pat. 5,626,988 ͑1997͒.
41. J. Luo, Q. Zhang, and S. L. Suib, Inorg. Chem., 39, 741 ͑2000͒.
42. J. C. Hunter, J. Solid State Chem., 39, 142 ͑1981͒.
Results of some designed chemical reactions have provided in-
formation regarding the role of bismuth and barium on the chemistry
of Mn͑III͒ compounds in alkaline solutions. Incorporation of Bi- or
Ba-containing compounds helps to keep the manganese ions in so-
lution for a longer time without allowing them to form irreversibly
stable phases such as birnessite or hausmannite. A Mn͑IV͒ oxide
intermediate, denoted as MnO Þ, that lacks long-range order and
͓
͔
2
is in equilibrium with the solution species appears to transform to
birnessite rapidly in the absence of additives. The presence of Bi- or
Ba-containing compounds delays such a transformation, with Bi be-
ing more effective than Ba, which is consistent with the electro-
chemical data. Thus, the presence of Bi- or Ba-containing com-
pounds helps to suppress some irreversible reactions leading to
birnessite or hausmannite rather than blocking the dissolution or
disproportionation of Mn͑III͒ species. Optimization of the nature
and amount of additives could improve further the rechargeability of
manganese oxides and have an impact in the commercialization
prospects of rechargeable alkaline cells based on manganese oxides.
43. Y. Ma, L. Luo, and S. L. Suib, Chem. Mater., 11, 1972 ͑1999͒.
44. D. C. Golden, C. C. Chen, and J. B. Dixon, Science, 231, 171 ͑1986͒.
45. F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 4th ed., p. 742,
John Wiley & Sons, New York ͑1980͒.
Acknowledgment
Financial support by RBC Technologies, College Station, TX, is
gratefully acknowledged.
46. P. W. Atkins, Physical Chemistry, 4th ed., p. 212, Oxford University Press, Oxford,
U.K. ͑1990͒.
47. J. Naslund, I. Persson, and M. Sandstrom, Inorg. Chem., 39, 4012 ͑2000͒.
48. I. Persson, M. Sandstrom, H. Yokoyama, and M. Chaudhry, Z. Naturforsch., A:
Phys. Sci., 50, 21 ͑1995͒.
The University of Texas at Austin assisted in meeting the publication
costs of this article.
Downloaded on 2015-01-19 to IP 129.81.226.78 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).