Solvothermal morphology studies: Alkali and alkaline earth molybdates

Article abstract of DOI:10.1002/hlca.200490095

A convenient and systematic solvothermal pathway towards alkali and alkaline-earth molybdates has been established. The solvothermal treatment of a molybdenum-based precursor material (yellow molybdic acid, MoO ₃·2H₂O) with ionic additives (alkali or alkaline-earth halides) provides access to a spectrum of molybdates. Their particle morphology can further be addressed by optimizing the reaction conditions. The resulting products cover a wide scope of sizes and morphologies, ranging from molybdenum oxide fibres with high aspect ratios and nanoscale diameters to millimeter-sized crystals of novel alkali molybdates. Both anionic and cationic additives exhibit certain synthetic profiles that offer the perspective of turning this approach into a 'toolbox' for the tailoring of molybdate-based materials.

Full text of DOI:10.1002/hlca.200490095

Helvetica Chimica Acta ± Vol. 87 (2004)
1029
Solvothermal Morphology Studies: Alkali and Alkaline Earth Molybdates
by Alexej Michailovski, Frank Krumeich, and Greta R.Patzke*
Laboratory of Inorganic Chemistry, ETH Hˆnggerberg, Wolfgang-Pauli-Strasse 10, CH-8093 Zurich
(
Phone: ‡41-1-6326743; fax: ‡41-1-6321149, e-mail: patzke@inorg.chem.ethz.ch)
A convenient and systematic solvothermal pathway towards alkali and alkaline-earth molybdates has been
established. The solvothermal treatment of a molybdenum-based precursor material (yellow molybdic acid,
MoO ¥ 2 H O) with ionic additives (alkali or alkaline-earth halides) provides access to a spectrum of
3 2
molybdates. Their particle morphology can further be addressed by optimizing the reaction conditions. The
resulting products cover a wide scope of sizes and morphologies, ranging from molybdenum oxide fibres with
high aspect ratios and nanoscale diameters to millimeter-sized crystals of novel alkali molybdates. Both anionic
and cationic additives exhibit certain synthetic profiles that offer the perspective of turning this approach into a
−
toolbox× for the tailoring of molybdate-based materials.
1.Introduction. ± The outstanding catalytic properties of MoO3 and their
mechanistic foundations [1 ± 3] have been in the focus of research activities over the
past decades. Consequently, the crystal growth of this versatile material has been
studied as well [4][5]. To fully explore the potential of MoO and related molybdates, it
3
would be highly desirable to downscale these substances to the nano regime so that they
are accessible as anisotropic nanotools in large quantities via simple preparative
routines [6 ± 8]. Recently, the first successful formation of anisotropic MoO by means
3
of solvothermal methods has been reported [9][10]. Solvothermal reactions offer a
high degree of preparative flexibility, and this synthetic route is predestined to facilitate
the development of a future nanotechnology, because it provides control of particle size
and gives access to kinetically stable phases [11 ± 13]. Morphology control of inorganic
materials [14] is a further preparative challenge that could be tackled by means of
solvothermal methods. A multitude of solvothermal syntheses have been published, but
there is still plenty of room for experimental work, and a particular need for systematic
investigations. The standard autoclave setup remains a −black box×, and, therefore, most
key steps of solvothermal reactions are not fully understood up to now [15]. The direct
monitoring of a solvothermal processes can only be performed with rather sophisti-
cated in situ techniques.
The standard orthorhombic form of MoO serves as a basis for highly efficient
3
catalysts that are widely applied, e.g., in alcohol or methane oxidation. However, there
is another modification of MoO that might even bear superior technical potential:
3
hexagonal MoO (h-MoO ), which displays a porous (3 ± 3.5 ä) framework structure.
3
3
Considerable preparative effort has been undertaken since the 1970s to synthesize the
hexagonal channel structure of h-MoO in pure form. All those approaches, however,
3
had to overcome a crucial problem: the hexagonal channels are highly dependent on
the stabilizing presence of large cations, which are extremely difficult to remove [16 ±
1
9]. Even the most-successful routes towards pure h-MoO have not afforded the
3

1
030
Helvetica Chimica Acta ± Vol. 87 (2004)
entirely dehydrated material [20][21]. The cation exchange (alkali vs. ammonium ions)
and subsequent thermal decomposition of hexagonal molybdates has been established
as a route to h-MoO [22]. The preparation of hexagonal molybdates, however, has
3
never been explored beyond standard procedures, with no further parameter varia-
tions such as the dissolution of M MoO (M ˆ alkali-metal cation) phases in acids,
2
4
followed by either heating at atmospheric pressure or a few days of solvothermal
treatment.
During our systematic investigations, we have already developed highly efficient
solvothermal syntheses of (NH ) (W,Mo)O nanorods [23] and fibrous MoO [24],
4
x
3
3
starting from yellow molybdic acid, MoO ¥ 2 H O. Preliminary results [11] pointed out
3
2
that the aspect ratio of the MoO fibres can be tuned by standard laboratory acids and
3
ionic additives. The solvothermal reaction of MoO ¥ 2 H O in the presence of ionic
3
2
additives is in the focus of the present study. In order to keep the systems as concise and
straightforward as possible, we have opted for alkali and alkaline-earth halides that are
subjected to an optimized solvothermal routine.
2
.Experimental. ± In a standard experiment, 180mg of MoO
of distilled H O, and 1 mmol of the appropriate alkali or alkaline-earth halide were added to a Teflon-lined
stainless-steel autoclave with a capacity of 23 ml. The autoclave was then sealed, heated at 1808 for 2 d, and
subsequently cooled to r.t. The resulting precipitate was collected by filtration, washed with distilled H O,
EtOH, and Et O, and dried in air. X-ray powder-diffraction analyses were conducted on a STOE STADI-P2
diffractometer in transmission mode (flat sample holders, graphite monochromated CuK radiation) equipped
with a position-sensitive detector (resolution ca. 0.018 in 2q). Transmission-electron-microscopy (TEM)
investigations were performed on a Philips CM30ST microscope operated at 300 kV (LaB cathode). The
3 2
¥ 2 H O (synthesized according to [25]), 2 ml
2
2
2
a
6
material was deposited on a perforated carbon foil supported on a copper grid. For scanning-electron
microscopy (SEM), performed on a LEO 1530 (FEG) microscope with 1 kV electrons, samples were dispersed
in EtOH and subsequently deposited on a silicon wafer. A CamScan CS-44 electron microscope with an EDAX-
Phoenix energy-dispersive X-ray spectrometer (EDXS) was employed for approximate elemental analyses.
Differential-thermal-analysis/thermogravimetry (DTA/TG) measurements were performed on a NETZSCH
À1
STA 409-C apparatus between 20and 1200 8 at a heating rate of 10K min . Laser-ablation-inductively-coupled-
mass-spectrometry (LA-ICP-MS) investigations were performed on an Elan 6100-DRC-plus-ICP-MS
apparatus (Perkin Elmer/Sciex).
3
.The Interaction of MoO ¥ 2 H O and Alkali Halides: Key Observations. ± 3.1.
3 2
Solvothermal Reaction Pathways. 3.1.1. Additive-Free Reactions. In the course of our
previous investigations on the formation of MoO nanorods [24], we have observed
3
that the reaction pathway depends on the solvent. Neutral, unpolar organic solvents
such as toluene exert the same effect as a solvent-free environment and give rise to
bundles of agglomerated MoO rods. This indicates that the topotactic dehydration of
3
MoO ¥ 2 H O must be triggered by a polar solvent system such as H O or diluted acid
3
2
2
to produce well-defined, isolated rods.
The solvothermal treatment of MoO ¥ 2 H O in H O at 1808 afforded highly
3
2
2
anisotropic rods: their average diameter was in the range of 90± 130nm, and the rod
length varied from 2 to 4 mm (Fig. 1,a). At 808, fibrous a-MoO ¥ H O was isolated as an
3
2
intermediate product of the topotactic reaction. Concerning the reaction timescale, the
rods were already present after 3 h of treatment at 1808, and extended reaction times
(
2 d) induced the formation of tiny holes on the rod surface (Fig. 1,b). When enhanced
aspect ratios were required (such as diameters of 200 ± 500 nm and lengths up to
0 mm), the reaction temperature had to be raised to 2208 (Fig. 1,c and d).
5

Helvetica Chimica Acta ± Vol. 87 (2004)
1031
Fig. 1. Scanning-electron-microscope (SEM) images of MoO
3
nanorods grown from MoO
3
¥ 2 H
2
O in H
2
O for 2
days at different temperatures. a) and b) Tˆ 1808; scale barˆ 1 mm each; the rod surface is becoming porous.
c) and d) Tˆ 2208, scale barˆ 4 mm; the aspect ratio of the rods is enhanced considerably.
3
.1.2. The Influence of Ionic Additives. Changing the solvent from H O to 25%
2
aqueous AcOH considerably altered the course of the reaction: no intermediates could
hitherto be isolated, so that a dissolution/precipitation process appears more
reasonable than the topotactic-dehydration route. Further options for tuning the
shape and aspect ratio of MoO nanorods are accessible by varying the acidic solvent.
3
In the following, the solvothermal potential of neutral additives was explored. The
halide anions were expected to resemble ideal −spectator× anions with minimum
influence on the reaction, which would be predestined for studying the impact of alkali
and alkaline-earth cations. This assumption, however, was quickly disproved, because
the halide ions turned out to exert a remarkable influence on the course of the
solvothermal reaction. Among them, only the chlorides and bromides exhibited
analogous behavior (cf. Tables 1 and 2), so that they were selected for further
‡
investigations. The products emerging from the according MoO ¥ 2 H O/MX (M ˆ Li
3
2
‡
À
À
through Cs , X ˆ Cl , Br ) systems can be subdivided into three main categories
(
Table 1): 1) LiCl and LiBr afford MoO nanorods; 2) The reaction of MoO ¥ 2 H O
3 3 2
‡
‡
with chlorides and bromides of the heavier alkali cations (Na through Cs ) leads to
the formation of hexagonal molybdates; and 3) the fluorides/iodides of Rb and Cs give
rise to the formation of new alkali-molybdate and -fluoromolybdate modifications
(
Table 1, footnote), as will be discussed in follow-up publications.

1
032
Helvetica Chimica Acta ± Vol. 87 (2004)
O (M ˆ Li‡ through Cs , X ˆ FÀ
‡
Table 1. Solvothermal-Reaction Products in the System MX/MoO
3
¥ 2 H
2
À
through I )
‡
‡
‡
‡
‡
Li
Dissolution Dissolution
MoO
nanorods
MoO
nanorods
Amorphous Hexagonal
material phase
Na
K
K
Rb
Rb
Cs
À
2 2 6 8 26 2 2 4
¥ 6 H Mo O F /Cs Mo O
Oa) Cs
13a)
F
6
Mo
8
O
26
F
2
¥ 6 H
2
O
6
Mo
8
O
26
F
À
Cl
3
Hexagonal Hexagonal
phase phase (17% K)
Hexagonal Hexagonal
phase phase (16% K)
Mo 13 plus
hexagonal phase
18% K)
Hexagonal
phase (13% Rb)
Hexagonal
Hexagonal phase (21% Cs)
À
Br
3
Hexagonal phase (21% Cs)
phase (16% Rb)
À
a
2 4
Cs Mo O
13a) plus
hexagonal phase
(23% Cs)
I
K
2
4
O
Rb
2
Mo
4
O
13 ) plus
hexagonal phase
(
a) Unknown alkali-molybdate or -fluoromolybdate modification.
3
.2. Adjustment of Solvothermal-Reaction Parameters: MoO Nanorods vs. Hex-
3
agonal Phases. 3.2.1. Characterization of MoO Nanorods. The MoO nanorods grown
3
3
in the presence of LiCl (Fig. 2) were slightly shorter than those produced in plain H O
2
(diameters ranging from 80to 120nm, combined with lengths between 750nm and
2
.2 mm; Fig. 1,a, see also Fig. 11,a). Their lattice constants (a ˆ 3.960(1), b ˆ 13.853(3),
c ˆ 3.696(1) ä) corresponded well with those reported for a-MoO (a ˆ 3.962, b ˆ
3
‡
1
3.855, c ˆ 3.701 ä) [26]. The absence of Li in the rods was confirmed by LA-ICP-MS
analyses.
Fig. 2. X-Ray-diffraction (XRD) patterns of MoO
3
rods grown in the presence of a) LiCl compared to
b) literature data for a-MoO
3
[26]. Inset: SEM image of the rods.
3
.2.2. Characterization of Rubidium-Based Hexagonal Molybdates. A representa-
tive characterization of the hexagonal molybdates was performed on a sample of

Helvetica Chimica Acta ± Vol. 87 (2004)
1033
hexagonal rubidium molybdate grown in the presence of RbBr. The Rb content was ca.
1
1
6% ) (EDX analysis referring to the overall metal content, cf. Table 1). This
corresponds well to the expected amount of 15.8% based on the empirical formula
RbH6 Mo O (x ˆ 2/3) [27]. The experimental lattice constants (a ˆ 10.556(4) and
xÀ1
6Àx 18
c ˆ 3.734(1) ä, cf. Fig. 3) were only slightly larger than the literature data (a ˆ
1
0.573(1), c ˆ 3.720(2) ä) [27]. All columnar hexagonal microcrystals generally
displayed diameters ranging from 0.8 ± 4 mm, and lengths from 10to 40 mm
Fig. 4).
(
Fig. 3. XRD Pattern of hexagonal rubidium molybdate a) before and b) after heating to 4008, together with a
representative structural model for hexagonal phases (KMo 15OH ¥ 2 H O) [22]
5
O
2
Fig. 4. SEM Images of hexagonal rubidium molybdate a) before and b) after heating to 4008 (scale barˆ 20 mm)
1)
Unless specified, elemental compositions are expressed in atomic percent.

1
034
Helvetica Chimica Acta ± Vol. 87 (2004)
The thermal behavior of the hexagonal rubidium-molybdate microcrystals was
characterized by DTA/TG investigations. Heating to 4008 left the columnar shape of
the microcrystals intact (Fig. 4) and did not affect their XRD pattern (Fig. 3). A mass
loss of 2.7% was observed. The RbH6 Mo O (x ˆ 2/3) model leads to a theoretical
xÀ1
6Àx 18
mass loss of 3.1%, attributable to H O, and the according experimental value has been
2
reported as 3.4% [27]. Concerning the accuracy of the EDX and DTA/TG measure-
ments, the composition of the rubidium-molybdate microcrystals should, thus, be in the
range of the literature data [27].
Furthermore, the microcrystals were investigated by means of electron diffraction
(
Fig. 5), which confirmed the hexagonal symmetry. Additional diffuse reflections along
the c*-axis (Fig. 5,a) indicated the presence of a fivefold superstructure with a c-axis of
ca. 18.5 ä. This might be due to an ordering of the partially occupied Rb positions
inside the channels. HR-TEM Images (Fig. 6) showed that the channels were arranged
without defects in the ab-plane.
Fig. 5. Electron-diffraction patterns of hexagonal rubidium molybdate a) along [100] and b) along [001].
Parameters: d100 ˆ 9.07, d001 ˆ 3.70ä.
Fig. 6. a) High-resolution transmission-electron-microscope (HR-TEM) image and b) Fourier-filtered HR-
TEM image of hexagonal rubidium molybdate along [001] (scale bar ˆ 10nm). In b, the Rb positions are
revealed by grey shades that can be seen in the centers of the channels.

Helvetica Chimica Acta ± Vol. 87 (2004)
1035
3
.2.3. Reaction Pathway. The formation of MoO nanorods can either proceed via
3
topotactic dehydration in H O or via a dissolution/precipitation sequence in AcOH. In
2
order to garner information about the reaction pathway in the presence of neutral ionic
additives, we have employed glass ampoules to monitor the solvothermal reaction of
MoO ¥ 2 H O in LiBr (Fig. 7) and RbBr solutions, respectively. A reaction period of
3
2
48 h was sufficient for quantitative conversion, and in both experiments, a precipitate
was present at any stage of the reaction. The only visible phenomena were marginal
color changes of the precipitate during the first 90min. This time interval corresponds
well to our previous observations concerning the timescale of rod formation in H O:
2
after 3 h of autoclave treatment, the reaction was basically finished [24]. The resulting
MoO nanorods formed with LiBr (Fig. 7) displayed the same dimensions as their
3
autoclave-grown analogs, and so did the hexagonal rubidium molybdate microcrystals
arising from the reaction of MoO ¥ 2 H O and RbBr in a glass ampoule. Both reactions
3
2
exclusively proceeded in the presence of a solid phase, so that they may be described as
heterogeneous× solvothermal processes. Accordingly, reactions based on the precip-
−
itation from a previously clear solution could then be specified as −homogeneous×
syntheses. As a consequence, heterogeneous solvothermal reactions do not even require
the most elementary pre-treatments such as mixing or dissolution of starting materials,
which might render them convenient for technical applications.
Fig. 7. SEM Images of MoO
3
nanorods formed in the presence of LiBr as an additive (scale barˆ 1 mm). The
products grown in a glass ampoule (b) are compared to those grown in a standard autoclave (a).
3
.2.4. Concentration of Additives. Both MoO rods and hexagonal molybdates
3
displayed morphological changes, when the additive concentration was varied. LiBr,
for example, was most suitable for MoO -rod formation in the 0.5 ± 1m concentration
3
range, whereas 2 ± 4m solutions gave rise to rod bundles that displayed an increasing
degree of agglomeration (Fig. 8,a and b). Even in the presence of drastically increased
‡
Li concentrations, there were no hexagonal molybdates formed at all. The hexagonal
rubidium molybdates tended to intergrow as well when the additive concentration was
raised: whereas a 1m RbBr solution favored the growth of isolated hexagonal
microcrystals (Fig. 4), flower-like morphologies (Fig. 8,c and d) prevailed at higher
‡
concentrations (2 ± 4m RbBr). The degree of Rb incorporation, however, remained
unchanged (13 ± 16%, cf. Table 1). As a result, the production of particles with a uniform
morphology requires carefully adjusted additive concentrations.

1
036
Helvetica Chimica Acta ± Vol. 87 (2004)
Fig. 8. SEM Images demonstrating the effect of the concentration of additives on the morphology of a), b) MoO
3
nanorods (0.5 vs. 2m LiBr; scale bar ˆ 1 mm) and c), d) hexagonal rubidium molybdate (2 vs. 4m RbBr; scale
barˆ 6 mm)
3
.2.5. Reaction Temperature. When MoO nanorods were synthesized in AcOH via
3
a homogeneous process, the temperature range of 90± 120 8 gave rise to fine rods with
minimum diameters of 40nm. Their maximum diameter of 150nm was reached when
the temperature was raised to 1508 [24]. Rod formation in the presence of an ionic
additive, however, required higher reaction temperatures (ca. 1808) to prevent rod
agglomeration. As shown in Fig. 9, treatment of MoO ¥ 2 H O in 1m LiCl solution at
3
2
1008 partially afforded bundle-like products that finally separated into individual rods
Fig. 9. SEM Images of MoO
3
nanorods formed in the presence of LiCl at a) 1008 and b) 1808. Scale bar ˆ 1 mm.

Helvetica Chimica Acta ± Vol. 87 (2004)
1037
at 1808. Hence, heterogeneous reactions might require higher temperatures than
dissolution/precipitation routes to generate well-separated nanoparticles.
3
.2.6. Reactions in Acidic Media. Reference experiments concerning the reaction of
MoO ¥ 2 H O with RbCl in acidic media (1%, 5%, and 25% aqueous AcOH) were
3
2
performed to reveal whether the heavier alkali cations could provide MoO fibres,
3
when subjected to altered reaction conditions. Detailed SEM investigations showed
that no MoO nanorods were formed at all. In other words: the presence of larger alkali
3
‡
‡
cations (Rb , Cs ) completely inhibits MoO -rod formation.
3
3
.3. Sodium-Based Additives: the Interplay of MoO Nanorods and Hexagonal
3
Molybdates. Depending on the appropriate alkali cation, hydrothermal treatment of
‡
‡
MoO ¥ 2 H O either provides MoO nanorods (Li ) or hexagonal molybdates (K
3
2
‡
3
‡
through Cs ). Na Ions, however, represent a borderline case, leading to mixtures of
both phases; the small ionic radius of Na does not suffice to stabilize the hexagonal
framework. As a result, sodium is incorporated as Na ¥ 2 H O [27].
‡
‡
2
3
.3.1. Sodium-Halide (NaX) Concentration (X ˆ Cl, Br). Fig. 10,a ± f illustrates
how the NaX concentration affects the ratio of MoO nanorods to hexagonal sodium
3
molybdate in the reaction products. Lower NaCl concentrations (ca. 0.25m) led to the
formation of almost phase-pure MoO3 nanorods, containing ca. 5% hexagonal
microcrystals as a side product (Fig. 10,a). When the concentration was raised to
0.5m NaCl, a mixture containing equal amounts of both phases was obtained. The
standard concentration of 1m NaCl yielded hexagonal sodium molybdate as the main
product, and the fraction of fibrous MoO side products was just above the XRD
3
detection limit (Fig. 10,b). Finally, 2 ± 4m NaCl solutions yielded pure hexagonal
sodium molybdate (Fig. 10,c). With NaBr as an additive, the above trend was
confirmed for concentrations of 0.25m (Fig. 10,d) and 1m (Fig. 10,e) up to 2m NaBr
(
pure hexagonal phase). Raising the NaBr concentration to 4m, however, re-
introduced MoO fibres into the solvothermal system (Fig. 10,f). Hence, the NaCl or
3
NaBr concentration must be carefully adjusted in the 0.25 ± 2m range to generate phase-
pure products.
3
.3.2. Reaction Temperature. The influence of the temperature on the reaction of
MoO ¥ 2 H O in 1m NaCl solution is summed up in Fig. 10 ,g ± i. At 10 08 , well-shaped
3
2
and phase-pure microcrystals were obtained (Fig. 10,g), with no traces of MoO3
nanorods being present. Raising the reaction temperature to 1408 (Fig. 10,h) did not
affect the phase purity, but the morphology of the hexagonal phase was beginning to
deteriorate. Under standard conditions (1808), MoO nanorods were beginning to
3
emerge (Fig. 10,b), and the reaction at 2208 finally led to a noticeable amount of coarse
MoO microfibers (Fig. 10,i). They were smaller and less homogeneous than the fibers
3
generated in an additive-free environment at 2208 (Fig. 1,c ± d).
In summary, low NaX (X ˆ Cl, Br) concentrations and high reaction temperatures
favor the formation of MoO nanorods, and the overall solvothermal process is quite
3
sensitive towards parameter shifts. The heavier alkali chlorides and bromides exclusively
provide phase-pure products.
4
.The Interaction of MoO 3 ¥ 2 H O with Alkaline-Earth-Based Additives. ± 4.1.
2
General Trends. Table 2 provides a survey of key results concerning the interaction of
alkaline-earth additives and MoO ¥ 2 H O. They resemble the behavior of the
3
2

1
038
Helvetica Chimica Acta ± Vol. 87 (2004)
Fig. 10. SEM Images of MoO
3
nanorods produced in concentration- (a ± f) and temperature-dependent (g ± i)
reactions from MoO ¥ 2 H O in the presence of sodium-based additives. a) 0.25m NaCl, b) 1m NaCl, c) 4m NaCl,
3
2
d) 0.25m NaBr, e) 1m NaBr, f) 4m NaBr. The reaction was also performed at constant NaCl concentration (1m) at
different temperatures: g) 10 08 , h) 1408, and i) 2208. The scale bars are 10 mm (a ± c) or 6 mm (d ± i), resp.
3 2
Table 2. Solvothermal-Reaction Products of the Alkaline-Earth Hydroxides and Halides with MoO ¥ 2 H O
under standard conditions
Mg2
Phase unknown
MoO nanorods
plus MgF
‡
Ca
CaMoO
MoO nanorods
plus CaF
2‡
Sr
2‡
Ba
BaMoO
Ba MoO
2‡
À
OH
F
4
SrMoO
SrMo
4
4
À
3
3
4
O
13 ¥ 2 H
2
O
2
3 4
F
2
2
À
Cl
MoO
MoO
3
nanorods
nanorods
MoO
MoO
3
nanorods
nanorods
SrMo
SrMo
4
4
O
O
13 ¥ 2 H
2
O
O
Hexagonal phase (14% Ba)
Ba Mo
plus hexagonal phase (13% Ba)
Ba Mo
plus amorphous material
À
Br
3
3
13 ¥ 2 H
13 ¥ 2 H
2
2
4 2
O13 ¥ 2 H O
À
I
Amorphous
material
Amorphous material
SrMo
4
O
2
O
2
4 2
O13 ¥ 2 H O
corresponding alkali-based systems (Table 1): 1) The choice of the counteranion is
crucial for the course of the reaction. 2) Fluorides and iodides trigger the formation of
either microcrystalline molybdates or amorphous material. MgF and CaF give rise to
2
2
MoO -rod formation, but due to their low solubility under the reaction conditions,
3
these results were confirmed by additional experiments under more-drastic conditions

Helvetica Chimica Acta ± Vol. 87 (2004)
1039
(1m HCl, 2208). 3) The alkaline-earth bromides and chlorides again display analogous
solvothermal behavior, and they are the only anions supporting the quantitative
formation of MoO nanorods. 4) The alkaline-earth hydroxides give rise to the
3
formation of MMoO phases (M ˆ Ca, Sr, Ba).
4
4
.2. Trends in the Formation of MoO Nanorods. MoO Rods are accessible in the
3 3
‡
2‡
2‡
presence of the smallest alkali and alkaline-earth cations only: Li , Ca , and Mg .
.2.1. Influence of Cations. The MoO rods obtained with LiCl (Fig. 11,a) and
4
3
MgCl (Fig. 11,d) exhibited similar dimensions (ca. 80-nm diameters and lengths
2
ranging from 600 nm to 2 mm). This might be due to the related electrostatic
‡
2‡
parameters of Li and Mg . The presence of additive ions slightly diminished the
aspect ratio with respect to the particles grown in H O (Fig. 1,a and b). This trend was
2
continued with LiBr (Fig. 11,b) and MgBr (Fig. 11,e), and the latter already induced
2
rod agglomeration. Finally, CaCl and CaBr (Fig. 11,g and h) quantitatively yielded
2
2
MoO rod bundles. The various MoO -rod dimensions determined are summed up in
3
3
Table 3. Three preparative guidelines can be derived from these results: 1) The most
‡
2‡
straightforward additive-based formation of MoO rods proceeds with Li and Mg
3
‡
ions. Na Ions represent a borderline case and require optimized reaction conditions. 2)
Compared to the reference reaction in bulk H O, Li - and Mg -based additives slightly
downsize the aspect ratio of the rods, whereas Na cations yield considerably longer
‡
2‡
2
‡
fibres. 3) Maximum rod sizes are achieved at elevated temperatures (2208) in pure H O.
2
Fig. 11. SEM Images of rod-shaped solvothermal-reaction products of MoO
3 2
¥ 2 H O under standard conditions
with a) ±c) lithium-based, d) ±f) magnesium-based, and g) ±i) calcium-based additives. Scale bars ˆ 1 mm.

1
040
Helvetica Chimica Acta ± Vol. 87 (2004)
3
Table 3. Size Trends among MoO Nanorods
Growth Conditions
m LiCl, 1808
m Mg .C6 l± ,2 1808
O, 1808
Diameter [nm]
Length [mm]
1
1
H
80± 120
800
90± 130
1705 ± 30
175 ± 35012
200 ± 500
0.75 ± 2.2
2
2
1.8 ± 4
b)
a
0
0
.25m NaCl, 1808 )
.25m NaBr, 1808 )
a
b)
H
2
O, 2208
ꢀ 44
a) Small amounts of hexagonal sodium molybdate formed as a side product. b) Partly also longer rods.
4
.2.2. Influence of Anions. The iodide series of experiments with LiI, MgI , CaI2
2
(
Fig. 11,c,f, and i, resp.) revealed a deteriorating effect of the lighter iodides on the
overall morphology. The aforementioned trend among the cations, however, appears to
be inverted: LiI afforded only unstructured, fibrous material, and MgI2 yielded
intergrown rods. Whereas CaCl and CaBr favored rod-bundle formation, CaI led to
2
2
2
the formation of well-defined fibres with diameters of ca. 20nm and typical lengths of
00 nm.
All three products exhibited hitherto unidentified XRD patterns (Fig. 12): the less-
4
defined the morphology of the product, the broader the reflections were. Certain
analogies in the XRD patterns might point to the formation of similar products. The
redox activity of the iodide ion obviously prevents a clear-cut topotactical reaction and
VI
might give rise to a reduction of the Mo -based parent structure instead. Consequently,
2 2
Fig. 12. XRD Patterns of the fibrous materials grown with a) LiI, b) MgI , and c) CaI . See also Figs. 11,c, f, and i,
resp.

Helvetica Chimica Acta ± Vol. 87 (2004)
1041
neutral and chemically −inert× counteranions are preferable for morphology studies and
nanoparticle production.
4
.3. The Role of Counteranions: Sr- and Ba-Based Additives. Strontium and barium
halides exhibited major differences in their interaction with MoO ¥ 2 H O (Table 2).
3
2
Among them, BaCl and BaBr exclusively provided alkaline-earth-based hexagonal
2
2
molybdates.
.3.1. Hexagonal Molybdates ± an Interesting Class of Channel-Structured Materials.
4
‡
‡
‡
‡
‡
‡
Only NH , Na ¥ 2 H O, K , Rb , Cs , and Ag ¥ 2 H O have been found to stabilize the
4
2
2
‡
channels of hexagonal molybdates [27]. Li Ions have been considered too small to
exert a stabilizing effect. To the best of our knowledge, no general and comprehensive
study has been performed to investigate the precise criteria required for cation
incorporation into hexagonal molybdates. Furthermore, hexagonal molybdates have
often been regarded as mere precursor materials for h-MoO , although they may well
3
give rise to interesting properties. They provide cation-exchange facilities, and,
moreover, structural investigations on molybdate-channel systems are quite challeng-
ing, because there are at least two alternative models for rationalizing the charge
balance in fully oxidized hexagonal molybdates: electroneutrality might either be
2À
achieved in terms of Mo vacancies or by the incorporation of extra O ions along with
H O in the tunnels [20]. The precise interplay of molybdenum vacancies, and of oxo-
2
and hydroxo-groups can only be enlightened with the help of neutron-diffraction
methods [20]. Substitution reactions of the molybdenum framework are feasible as
‡
well: although Li is too small to stabilize the common hexagonal molybdate structures,
it can be incorporated into a vanadium-substituted hexagonal framework resulting in
phases of the type (Li H )0.13V0.13Mo0.87O ¥ n H O [18]. These substances undergo a
x
1Àx
3
2
‡
−
hopping× transition upon heating that shifts the Li ions on trigonal prismatic sites
within the tunnel walls.
4
.3.2. Formation of a Novel Hexagonal Barium Molybdate. BaCl quantitatively
2
1
yielded a phase-pure hexagonal molybdate containing ca. 14% of Ba ). This is a key
result, because, to the best of our knowledge, no alkaline-earth-based hexagonal
molybdates have been reported up to now. BaBr afforded BaMo O ¥ 2 H O as a side
2
4
13
2
product, but the hexagonal barium molybdate could clearly be identified and separated
due to its characteristic columnar crystal shape. The aspect ratio of the Ba-containing
hexagonal molybdates was considerably smaller with respect to their alkali-based
analogs: BaCl afforded hexagonal crystals (Fig. 13,a) with diameters between 1.5 and
2
6
.5 mm, and lengths ranging from 6 to 20 mm, and BaBr (Fig. 13,b) further reduced the
2
average crystal dimension to diameters of 1 ± 4 mm, and lengths of 4 ± 13 mm. The lattice
constants for the BaCl -based product (Fig. 14) were determined as a ˆ 10.514(2) and
2
c ˆ 3.733(1) ä.
The newly synthesized hexagonal barium molybdate fits into the established
correlation between the ionic radius of an incorporated cation and the resulting volume
2‡
‡
of the hexagonal unit-cell (Table 4) [27]. Ba and K display almost the same ionic
radii so that their unit-cell volumes are closely related as well. All other alkaline-earth
cations failed to yield hexagonal phases under standard reaction conditions: whereas
2‡
2‡
Ba (149 pm) is sufficiently large to stabilize the hexagonal framework, Sr (132 pm)
is obviously too small, and so SrMo O ¥ 2 H O was formed instead. This points to a
4
13
2
critical threshold ionic radius of 132 ± 150pm for the formation of hexagonal phases.

1
042
Helvetica Chimica Acta ± Vol. 87 (2004)
Fig. 13. SEM Images of hexagonal barium molybdate crystals grown in the presence of a) BaCl
together with the EDX spectrum of the latter (scale barˆ 6 mm)
2 2
and b) BaBr ,
Table 4. Ionic Radii and Unit-Cell Volumes of Known Alkali-Based Hexagonal Phases Compared to the Newly
Synthesized Barium-Containing Phase
Cation
Unit-cell volume [ä3]
Ionic radius [pm]a)
2
‡
Ba (new)
K
357.37(13)
356.97(17)
360.14(18)
362.35(6)
149
152
166
181
‡
[27]
‡
Rb [27]
‡
Cs [27]
a) According to [28], for a sixfold coordination sphere.
To investigate whether Sr2 could still be incorporated into the hexagonal phases via
a combined solvothermal reaction, a series of experiments, starting from mixtures of
SrCl and BaCl (Sr/Ba ratios of 1:4, 2 :3, 3 :2, and 4 :1), were carried out. Surprisingly,
‡
2
2
the formation of hexagonal phases broke down altogether, and all main products
adopted the MMo O structure.
4
13
4
.3.3. Strontium- vs. Barium-Based Additives. Whereas Ba-based additives afforded
a variety of products, their Sr-based counterparts showed a distinct preference for
SrMo O ¥ 2 H O (Table 2), which might point to an outstanding thermodynamic
4
13
2
stability of this phase. Even SrI generated well-developed crystals of SrMo O ¥ 2 H O
2
4
13
2
(
Fig. 15,a), whereas all other alkali and alkaline-earth iodides mainly gave rise to

Helvetica Chimica Acta ± Vol. 87 (2004)
1043
3 2 2
Fig. 14. XRD Pattern of hexagonal barium molybdate grown from MoO ¥ 2 H O in the presence of BaCl
amorphous materials. The size of the SrMo O ¥ 2 H O platelets could be adjusted by
4
13
2
choosing between SrBr and SrI : the latter afforded crystals with edges ranging from
2
2
2
5 to 100 mm, and sides between 500 nm and 1 mm (Fig. 15,a). SrBr downsized the
2
platelet dimensions (Fig. 15,b): the smallest crystals were thin squares (200 Â 200 nm),
and the maximum edge lengths reached 20 mm (max. thickness: 200 nm). Up to the
1
990s, the Sr/Mo/O/H system was far from explored, and SrMoO and SrMoO had
4 3
been the only well-characterized phases until SrMo O ¥ 2 H O was found to be
4
13
2
Fig. 15. SEM Images of platelets of SrMo
4
O
13 ¥ 2 H
2
O grown in the presence of a) SrI
0 mm.
2
and b) SrBr
2
. Scale barˆ
1

1
044
Helvetica Chimica Acta ± Vol. 87 (2004)
isostructural with BaMo O ¥ 2 H O [29]. Infinite sheets of distorted MoO and MoO
5
4
13
2
6
moieties, sharing edges and vertices, represent the main structural element of the
MMo O ¥ 2 H O (M ˆ Sr, Ba) compounds [30]. The original hydrothermal synthesis
4
13
2
of single crystals of SrMo O ¥ 2 H O starts from a mixture of SeO , SrCO , and MoO .
4
13
2
2
3
3
Hence, our new bromide/iodide route represents a selenium-free alternative.
The existence of strontium fluoromolybdates has not been reported hitherto.
Ba MoO F , however, is a well-known fluoromolybdate that has attracted interest due
2
3
4
to its nonlinear optical properties [31]. Furthermore, it has been in the focus of NMR
studies concerning the O/F ordering in the chains of corner-sharing MoO/F octahedra
separated by Ba cations [32]. The synthesis of Ba MoO F single crystals requires
2
3
4
harsh conditions (BaF , MoO , 3 months of heating at 6808) [33], and even a newly
2
3
developed hydrothermal approach still requires HF and rather high temperatures
BaF /MoO , 15 ml of 3.2% aqueous HF solution, 3 d, 2278) [32].
(
2
3
The solvothermal route presented here provides phase-pure microcrystals of
Ba MoO F via a novel and fairly mild approach. Flower-like aggregates of intergrown
2
3
4
microcolumns exhibiting lengths from 1.6 to 6 mm and diameters of ca. 0.85 mm
Fig. 16) are formed. The constants of the monoclinic lattice were determined as a ˆ
1.497(8), b ˆ 9.352(6), c ˆ 7.258(6) ä, and b ˆ 126.23(7)8, and they agree well with the
literature data (a ˆ 11.457, b ˆ 9.365, c ˆ 7.237 ä, b ˆ 126.308) [34].
(
1
Fig. 16. SEM Image of flower-like intergrown microcrystals of Ba
2
MoO
3
F
4
. Scale barˆ 2 mm.
5
.Discussion. ± The solvothermal behavior of MoO ¥ 2 H O in the presence of
3 2
alkali- and alkaline-earth-based halides is summed up graphically in Fig. 17 and can be
discussed with respect to the following criteria.
1
) The resulting products can be classified according to the role of the cation: a)
‡
‡
2‡
MoO nanorods are preferably formed with the smallest cations (Li , Na , Mg ), and
3
their aspect ratios can be changed accordingly with respect to the reference route in
‡
‡
pure H O; b) The heavier and larger alkali (K through Cs ) and alkaline-earth cations
(
2
2‡
2‡
Ba ) are capable of stabilizing the hexagonal molybdate channel structure; c) Sr
Ions exhibit a remarkable tendency towards the formation of the tetramolybdate phase,
SrMo O ¥ 2 H O, while all remaining additive cations fail to provide phase-pure
4
13
2
tetramolybdates.

Helvetica Chimica Acta ± Vol. 87 (2004)
1045
3 2
Fig. 17. Solvothermal-reaction products of MoO ¥ 2 H O with alkali and alkaline-earth halides
2
) The anionic part of an additive exerts a distinct influence upon the products: a)
Fluorides tend to be incorporated into the products, thus providing a chemically mild
access to barium fluoromolybdate, and convenient routes to novel rubidium and
caesium molybdate and fluoromolybdate modifications. b) Chlorides and bromides
display an almost entirely identical behavior, and their interference with the
solvothermal reaction system is kept to a minimum, which renders them suitable for
competitive experiments and morphology studies. c) Iodides often exert a destructive
influence on both phase and morphology of the reaction products due to their redox
activity; only tetramolybdates and hexagonal molybdates withstand this tendency.
3
) The product morphologies cover a wide size range: a) The smallest MoO rods
3
exhibit nanoscale diameters, and the larger ones provide enhanced aspect ratios (up to
:200 ).b) Hexagonal molybdates usually form well-shaped microcrystals. c) Especially
1
the novel tetra- and fluoromolybdate modifications tend to form large crystals on the
millimeter scale.
In summary, the synthetic strategy employed here provides reasonable access to a
wide spectrum of products via a simple and straightforward solvothermal routine.
The underlying mechanisms remain to be enlightened: the ionic radii of the
additives only provide a useful preparative guideline for addressing certain molybdates,
and considerable effort is necessary to fully understand the solvothermal process. Lab-
scale quenching experiments provide basic insights, but sophisticated in situ techniques
are the method of choice. We are currently working on EXAFS experiments in situ
concerning the formation of MoO nanorods from MoO ¥ 2 H O. The interaction of
3
3
2
molybdenum-based oxidic precursor materials with alkali iodides and fluorides has
been investigated in another systematic study that continues this work and will be
presented in a follow-up publication.

1
046
Helvetica Chimica Acta ± Vol. 87 (2004)
6
.Conclusions. ± Four main trends finally emerge from the present study. 1) A
spectrum of molybdates can be selectively addressed by the appropriate choice of an
alkali or alkaline-earth halide. Further morphology tuning is feasible in terms of
parameter optimization. 2) Pathways to novel phases, such as the first Ba-containing
hexagonal molybdates, have been established. 3) The preparative potential of ionic
additives is currently explored with respect to other anions/cations and expanded upon
quaternary systems. Especially the newly formed Rb and Cs molybdates provide a rich
crystal chemistry that serves as a basis for manifold structural studies, including the
formation of solid solutions. 4) The precursor-additive approach may finally be turned
into a tailored additive −toolbox× for the straightforward synthesis and morphology
adjustment of molybdates.
All in all, this study might be regarded as a contribution to the challenging task of
introducing a −design× approach into solid-state chemistry.
The authors thank Prof. R. Nesper (Laboratory of Inorganic Chemistry, ETH Zurich) for his steady interest
and continuous support of this work. Our research was funded by the ETH Zurich, by the Swiss National Science
Foundation (−MaNEP ± Materials with Novel Electronic Properties×), and by the National Research Program
(−Supramolecular Functional Materials×). The authors thank Kathrin Hametner (group of Prof. D. G¸nther,
ETH Zurich) for LA-ICP-MS analyses, and Christian Mensing for DTA/TG measurements.
REFERENCES
[
[
[
[
[
[
[
[
[
1] A. Baiker, D. Gasser, Z. Phys. Chem. 1986, 149, 119.
2] H.-F. Liu, R.-S. Liu, K. Y. Liew, R. E. Johnson, J. H. Lunsford, J. Am. Chem. Soc. 1984, 106, 4117.
3] A. Baiker, P. Dollenmeier, A. Reller, J. Catal. 1987, 103, 394.
4] J. L. Callahan, R. H. Petrucci, C. A. Brown, Science 1958, 128, 841.
5] J. Li, P. Wei, W. Ruigang, J. Chen, Key Eng. Mat. 2002, 224 ± 226, 367.
6] C. N. R. Rao, A. K. Cheetham, J. Mater. Chem. 2001, 11, 2887.
7] C. N. R. Rao, B. C. Satishkumar, A. Govindaraj, M. Nath, ChemPhysChem 2001, 2, 79.
8] R. Tenne, Prog. Inorg. Chem. 2001, 50, 269.
9] X. W. Lou, H. C. Zeng, Chem. Mater. 2002, 14, 4781.
[
[
[
[
[
10] M. Niederberger, F. Krumeich, H.-J. Muhr, M. M¸ller, R. Nesper, J. Mater. Chem. 2001, 11, 1941.
11] G. R. Patzke, F. Krumeich, R. Nesper, Angew. Chem., Int. Ed. 2002, 41, 2446.
12] R. I. Walton, Chem. Soc. Rev. 2002, 31, 230.
13] K. Byrappa, M. Yoshimura, − Handbook of Hydrothermal Technology×, Noyes, Park Ridge, NJ, 2001.
14] S. Mann, Angew. Chem., Int. Ed. 2000, 39, 3392; G. A. Ozin, H. Yang, N. Coombs, Nature (London) 1997,
386, 692; G. R. Patzke, Angew. Chem., Int. Ed. 2003, 42, 972.
[
[
[
[
[
[
[
[
[
[
15] R. I. Walton, F. Millange, R. I. Smith, T. C. Hansen, D. O× Hare, J. Am. Chem. Soc. 2001, 123, 12547.
16] N. Sotani, Bull. Chem. Soc. Jpn. 1975, 48, 1820.
17] T. P. Feist, P. K. Davies, Chem. Mater. 1991, 3, 1011.
18] Y. Hu, P. K. Davies, T. P. Feist, Solid State Ionics 1992, 53 ± 56, 539.
19] O. Mougin, J.-L. Dubois, F. Mathieu, A. Rousset, J. Solid State Chem. 2000, 152, 353.
20] J. Guo, P. Zavalij, M. S. Whittingham, J. Solid State Chem. 1995, 117, 323.
21] Y. Muraoka, J.-C. Grenier, S. Petit, M. Pouchard, Solid State Sci. 1999, 1, 133.
22] B. Krebs, I. Paulat-Bˆschen, Acta Crystallogr., Sect. B 1976, 32, 1697.
23] A. Michailovski, F. Krumeich, G. R. Patzke, Chem. Mater. 2004, 16, 1433.
24] G. R. Patzke, A. Michailovski, F. Krumeich, R. Nesper, J.-D. Grunwaldt, A. Baiker, Chem. Mater. 2004, 16,
1126.
[
[
[
[
[
25] J. J. Cruywagen, J. B. B. Heyns, S. Afr. J. Chem. 1981, 34, 118.
26] G. Andersson, A. Magneli, Acta Chem. Scand. 1950, 4, 793.
27] E. M. McCarron III, D. M. Thomas, J. C. Calabrese, Inorg. Chem. 1987, 26, 370.
28] R. D. Shannon, Acta Crystallogr., Sect A 1976, 32, 751.
29] W. T. A. Harrison, L. L. Dussack, A. J. Jacobson, J. Solid State Chem. 1995, 116, 95.

Helvetica Chimica Acta ± Vol. 87 (2004)
30] W. T. A. Harrison, Acta Crystallogr., Sect C 1999, 55, 485.
31] M. Wiegel, G. Blasse, Solid State Commun. 1993, 86, 239.
32] L.-S. Du, F. Wang, C. P. Grey, J. Solid State Chem. 1998, 140, 285.
33] G. Wingefeld, R. Hoppe, Z. Anorg. Allg. Chem. 1984, 518, 149.
34] R. Domesl, R. Hoppe, Z. Anorg. Allg. Chem. 1982, 492, 63.
1047
[
[
[
[
[
Received November 26, 2003