Nanoscaled bismuth oxido clusters: Probing factors of structure formation and photocatalytic activity

Article abstract of DOI:10.1002/ejic.201300889

Hydrolysis of bismuth(III) nitrate in the presence of sodium benzoate (NaOBz) in dimethyl sulfoxide (DMSO) gave the bismuth oxido cluster [{Bi₃₈O₄₅(NO₃)₂₀(OBz)₄(DMSO)₂₄}·4DMSO][{Bi₃₈O₄₅(NO₃)₂₄(DMSO)₂₆}·4DMSO] (1), which crystallizes in the monoclinic space group P2₁/c with cell parameters of a = 35.8728(12) ?, b = 20.2750(5) ?, c = 34.8624(11) ?, and β = 111.797(4)°. The asymmetric unit contains two crystallographically independent bismuth oxido clusters of different composition. Because of the distinct structural relationship of this type of bismuth oxido clusters with the photocatalytically active polymorph β-Bi₂O₃, colloidal suspensions of nanoscaled 1 as well as colloids composed of the recently reported clusters [Bi₆O₄(OH)₄(NO₃)₆]·H₂O (2), [{Bi₃₈O₄₅(NO₃)₂₄(DMSO)₂₆}·2DMSO][{Bi₃₈O₄₅(NO₃)₂₄(DMSO)₂₄}·0.5DMSO] (3), and [Bi₃₈O₄₅(OMc)₂₄(DMSO)₉]·2DMSO·7H₂O (4) were probed for their photocatalytic activity. The bismuth oxido cluster 3 decomposes more than 85 % of a 1 × 10^-5 M aqueous solution of rhodamine B (RhB) under visible-light irradiation within 150 min. Colloids of such clusters might serve as model compounds for indirect photodegradation processes of organic contaminants at the metal oxide surface. The bismuth oxido cluster [{Bi₃₈O₄₅(NO₃)₂₀(OBz)₄(DMSO)₂₄}·4DMSO][{Bi₃₈O₄₅(NO₃)₂₄(DMSO)₂₆}·4DMSO] (DMSO = dimethyl sulfoxide, OBz = benzoate) is presented. We show that additives significantly influence the crystallization behavior of high-nuclearity bismuth oxido clusters. Colloidal suspensions of such clusters show high photocatalytic activity as demonstrated by degradation of rhodamine B (RhB).

Full text of DOI:10.1002/ejic.201300889

FULL PAPER
DOI:10.1002/ejic.201300889
Nanoscaled Bismuth Oxido Clusters: Probing Factors of
Structure Formation and Photocatalytic Activity
Maik Schlesinger,[a] Marcus Weber,[a] Tobias Rüffer,[b]
Heinrich Lang,[b] and Michael Mehring*[a]
Dedicated to Professor Werner Uhl on the occasion of his 60th birthday
Keywords: Bismuth / Nanostructures / Cluster compounds / Hydrolysis / Photocatalysis
Hydrolysis of bismuth(III) nitrate in the presence of sodium
benzoate (NaOBz) in dimethyl sulfoxide (DMSO) gave the
bismuth oxido cluster [{Bi38O45(NO3)20(OBz)4(DMSO)24}·
4DMSO][{Bi38O45(NO3)24(DMSO)26}·4DMSO] (1), which
crystallizes in the monoclinic space group P21/c with cell
parameters of a = 35.8728(12) Å, b = 20.2750(5) Å, c =
34.8624(11) Å, and β = 111.797(4)°. The asymmetric unit con-
tains two crystallographically independent bismuth oxido
clusters of different composition. Because of the distinct
structural relationship of this type of bismuth oxido clusters
dal suspensions of nanoscaled 1 as well as colloids composed
of the recently reported clusters [Bi6O4(OH)4(NO3)6]·H2O (2),
[{Bi38O45(NO3)24(DMSO)26}·2DMSO][{Bi38O45(NO3)24
(DMSO)24}·0.5DMSO] (3), and [Bi38O45(OMc)24
-
-
(DMSO)9]·2DMSO·7H2O (4) were probed for their photocata-
lytic activity. The bismuth oxido cluster 3 decomposes more
than 85% of a 1ϫ10–5
M aqueous solution of rhodamine B
(RhB) under visible-light irradiation within 150 min. Colloids
of such clusters might serve as model compounds for indirect
photodegradation processes of organic contaminants at the
with the photocatalytically active polymorph β-Bi2O3, colloi- metal oxide surface.
solvents such as dimethyl sulfoxide (DMSO), acetone,
and CH2Cl2 gave access to several ligand-stabilized
[Bi38O44+x
(26–x)+ (x = 0, 1) clusters.[5,6] Such bismuth oxido
clusters are composed of 13 [Bi6O8+x
2(1–x)+ units (x = 0, 1)
Introduction
It is well known that bismuth(III) nitrate tends to hy-
drolyze quickly and forms various condensation products
(at low pH values), so-called bismuth subnitrates. The phe-
nomenon has been known since the Middle Ages, when
“bismutum subnitricum” was administered for the treat-
ment of dyspepsia. Remarkably, such formulations are still
in use for wound-healing agents.[1] However, the composi-
tion of bismuth subnitrate as obtained from aqueous solu-
tion is ill defined and is often described as a mixture of
BiONO3, Bi(NO3)(OH)2, and BiOOH.[2] Recently it was
shown that, when starting from acidic aqueous solutions of
]
]
and represent an additional intermediate within the nucle-
ation process of bismuth oxido/hydroxido nitrates to finally
result in Bi2O3. It is worth noting that the hexanuclear mo-
tifs are present as structural units in the solid-state struc-
tures of β- and δ-Bi2O3.[6a,7] Thus, the bismuth oxido clus-
ters are excellent precursors for the synthesis of β-Bi2O3
nanoparticles, which show high photocatalytic activity un-
der visible-light irradiation as reported earlier.[8] However,
metal oxido clusters themselves are also well known for
their photocatalytic activity,[9] and recently the efficient
photodegradation of dyes by using the heterobimetallic bis-
muth-containing cluster H3[{Bi(DMSO)3}4V13O40] was re-
ported.[10] This prompted us to study the photocatalytic
performance of homometallic bismuth oxido clusters, which
has not been reported so far. Thus, colloidal suspensions of
the following nanoscaled clusters were tested:
bismuth(III) nitrate, the formation of hexanuclear bismuth
(6–x)+
oxido clusters of the type [Bi6O4+x(OH)4–x
]
(x = 0, 1)
is favored.[3] These complex ions result from intramolecular
condensation of hydrated Bi3+ ions[4] and represent the
largest bismuth oxido clusters that have been crystallized
from aqueous media so far.[5] However, the use of polar
[a] Technische Universität Chemnitz, Fakultät für Naturwissen-
schaften, Institut für Chemie, Professur Koordinationschemie,
09107 Chemnitz, Germany
[{Bi38O45(NO3)20(OBz)4(DMSO)24}·4DMSO][{Bi38O45
(NO3)24(DMSO)26}·4DMSO] (1), [Bi6O4(OH)4(NO3)6]·
H2O [{Bi38O45(NO3)24(DMSO)26}·2DMSO]-
(2),[3a]
and
-
E-mail: michael.mehring@chemie.tu-chemnitz.de
http://www.tu-chemnitz.de/chemie/koord/index.php
[b] Technische Universität Chemnitz, Fakultät für Naturwissen-
schaften, Institut für Chemie, Professur Anorganische Chemie,
09107 Chemnitz, Germany
[{Bi38O45(NO3)24(DMSO)24}·0.5DMSO][5]
(3),
[Bi38O45(OMc)24(DMSO)9]·2DMSO·7H2O[6b] (4). The
photodegradation of rhodamine B (RhB) using visible-light
irradiation under the same conditions as reported for β-
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300889.
Eur. J. Inorg. Chem. 2014, 302–309
302
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Bi2O3 nanoparticles was chosen as a model reaction.[8] For
comparison, the photocatalytic activities of 3 in solution
(DMF/H2O) and of suspensions of 1–4 in water under vis-
ible-light irradiation and for cluster 3 also under UV/Vis
conditions were determined. The syntheses and characteri-
zation of compounds 2–4 were reported recently; the bis-
muth oxido cluster 1 is reported here for the first time. The
latter was obtained upon testing the influence of a variety
of crystallization modulators on the hydrolysis behavior of
bismuth nitrate.
Results and Discussion
Studies on the Hydrolysis of Bismuth(III) Nitrate
The addition of sodium methacrylate (NaOMc) to a
solution of [Bi6O4(OH)4(NO3)6]·H2O[3a] (2) in DMSO re-
sulted in the formation of [Bi38O45(OMc)24(DMSO)9]·
Figure 1. Powder X-ray diffraction patterns obtained for crystals
after slow evaporation of DMSO from the reaction mixture:
Bi(NO3)3·5H2O with (a) sodium trichloroacetate, (b) sodium
dichloroacetate, and (c) sodium chloroacetate, respectively. (d) Cal-
culated powder pattern on the basis of single-crystal X-ray data of
[Bi(NO3)3(DMSO)3].[11] (e) The measured PXRD pattern of crys-
tals of the novel bismuth oxido cluster 1.
2DMSO·7H2O
(4),[6b]
whereas
[{Bi38O45(NO3)24-
(DMSO)26}·2DMSO][{Bi38O45(NO3)24(DMSO)24}·
0.5DMSO][5] (3) is obtained when Bi(NO3)3·5H2O was used
instead of the hexanuclear bismuth oxido cluster 2. In the
latter case, the additive NaOMc modulates the crystalli-
zation behavior but is not incorporated into the final mate-
rial. Without addition of NaOMc, the bismuth complex
[Bi(NO3)3(DMSO)3][11] was observed.[5] These previously
reported results demonstrate that additives are essential to
obtain bismuth oxido clusters of high nuclearity from bis-
muth nitrate solutions, but the role of the modulators is still
unclear. The basicity of additives such as NaOMc might be
one of the main aspects that influences condensation reac-
tions to give [Bi38O45]24+ species. We have thus investigated
the influence of additives with different pKb values[12] such
as sodium acetate, sodium benzoate, sodium formiate, so-
dium chloroacetate, sodium dichloroacetate, and sodium
trichloroacetate on the condensation reaction of bismuth
nitrate in DMSO. After slow evaporation of the solvent,
crystalline materials were observed in all cases that were
shown to be mainly [Bi(NO3)3(DMSO)3] if the sodium salts
of chloroacetate, dichloroacetate, and trichloroacetate are
added (Figure 1), whereas the addition of sodium formiate
and sodium acetate gave [{Bi38O45(NO3)24(DMSO)26}·
2DMSO][{Bi38O45(NO3)24(DMSO)24}·0.5DMSO] (3), sim-
ilar to the results observed upon addition of NaOMc.[5] The
identity of compound 3 was confirmed by single-crystal and
powder X-ray diffraction. In contrast, the data obtained for
crystals formed upon addition of sodium benzoate did not
fit to any reported data on bismuth compounds (Figure 1),
and single-crystal X-ray diffraction analysis revealed the
formation of a novel cluster with a mixed ligand shell,
(DMSO)24]·4DMSO (1a·4DMSO) and (2) [Bi38O45(NO3)24-
(DMSO)26]·4DMSO (1b·4DMSO) (Figure 2). The crystal-
line material is a cocrystal of a {Bi38} nitrate cluster, as
observed previously, and a {Bi38} cluster in which four
nitrates are substituted by benzoate.[5] As reported earlier,
a μ6-coordinated oxygen atom occupies an octahedral void
in the center of the [Bi38O45]24+ core of both clusters [Bi–O
1a: 2.5205(8)–2.6477(11) Å; 1b: 2.5422(9)–2.5846(9) Å], and
thus a [Bi6O9] unit represents the central motif, which is
surrounded by twelve edge-sharing octahedral [Bi6O8]2+
units that form a nearly face-centered cubic lattice in which
the tetrahedral voids are filled with oxygen atoms.
[{Bi38O45(NO3)20(OBz)4(DMSO)24}·4DMSO][{Bi38O45
(NO3)24(DMSO)26}·4DMSO] (1) (OBz = benzoate).
Compound 1 crystallizes in the monoclinic space group
P21/c with cell parameters of a = 35.8728(12) Å, b =
20.2750(5) Å, c = 34.8624(11) Å, and β = 111.797(4)°, and
two crystallographically independent molecules in the
-
Figure 2. Cutout of the monoclinic unit cell of 1 displaying the
packing of the two centrosymmetric bismuth oxido clusters
[Bi38O45(NO3)20(OBz)4(DMSO)24]·4DMSO (1a·4DMSO) and
[Bi38O45(NO3)24(DMSO)26]·4DMSO (1b·4DMSO). Hydrogen
atoms are omitted for clarity.
In cluster 1a, the [Bi38O45]24+ core is built up by 20 μ3-
asymmetric unit, both with different substitution patterns and 24 μ4-bridging oxygen atoms with Bi–O bond lengths
at the [Bi38O45]24+ core: (1) [Bi38O45(NO3)20(OBz)4- in the range from 2.084(13) to 2.520(12) Å and 2.065(13) to
Eur. J. Inorg. Chem. 2014, 302–309
303
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
2.964(12) Å, respectively. In cluster 1b, 16 oxygen atoms DMSO molecules are noncoordinated, eight coordinate in
show a μ3-coordination, whereas 28 oxygen atoms build a a monodentate fashion [2.719(15)–2.946(14) Å], and 18 bi-
μ4-coordination with Bi–O bond lengths of 2.089(12)– dentate [2.530(15)–2.938(15) Å] through their oxygen atoms
2.415(13) Å and 2.067(13)–2.905(11) Å, respectively. The to the [Bi38O45]24+ core. Of the 24 coordinating nitrate li-
coordination spheres of the bismuth atoms in 1 are built up gands, four show a monodentate [2.738(14)–2.979(17) Å],
by a varying number of anionic ligands and DMSO mole- ten a μ-1κO:2κO-type [2.716(15)–2.976(19) Å], six a μ-
cules. Thus, the coordination numbers of the bismuth atoms 1κO:2κOЈ-type [2.694(19)–2.824(19) Å], and four
a
vary from 5 to 9 to result in triangular-based [3+x] (x = 2– 1κ2O,OЈ-type coordination [2.670(16)–2.954(13) Å].
6) and square-based [4+x] (x = 1–5) pyramids capped by a
The [Bi38O45]24+ cores of the cocrystal 1 composed of 1a
varying number of oxygen atoms (x). Primary bonds with and 1b are quite similar but not identical, as it is shown by
Bi–O bond lengths in the range from 2.065(13) to a comparison of the atom positions of both bismuth oxido
2.489(11) Å constitute the base of the pyramids. The sec- cores (Figure S1 in the Supporting Information). The only
ondary bonds, which are suggested to result from the inter- identical position is found for the μ6-coordinated oxygen
action of the HOMOs of the ligands with the δ*(Bi–O) or- atom in the center of both bismuth oxido clusters. A
bitals, range from 2.520(12) to 2.991(12) Å.[13] The distor- weighted root mean square deviation of 0.317 Å containing
tion of the polyhedra is assigned to a stereochemically a maximum deviation of 0.847 Å is obtained for the differ-
active lone pair at the bismuth atoms. The manifold coordi- ences of the oxygen atom positions in clusters 1a and 1b, the
nation numbers of the bismuth atoms are also expressed by bismuth atom positions show a weighted root mean square
bismuth bond-valence parameters that vary from 2.48 to deviation of 0.288 Å with a minimum deviation of 0.042 Å
3.54.[14] In 1a·4DMSO, the coordination sphere of the bis- and a maximum deviation of 0.524 Å.
muth atoms is built up by the coordination of four benzo-
ates, 20 nitrates, and 24 DMSO molecules, and in addition
four noncoordinated DMSO molecules are present in the
crystal lattice (Figure 3). Two benzoates coordinate in a μ4-
Photocatalytic Investigations
1κO:2κO:3κOЈ:4κOЈ mode [2.523(13)–2.97(3) Å] and two in
Recently, heterobimetallic bismuth-containing clusters
a μ3-1κO:2κO:3κOЈ mode [2.584(14)–2.742(12) Å], whereas such as H3[{Bi(DMSO)3}4V13O40], defined as “molecular
four DMSO molecules show a monodentate [2.711(15)– bismuth vanadate”, have been demonstrated to be efficient
2.897(14) Å] and 20 a bidentate [2.542(14)–2.946(13) Å] co- photocatalysts in the visible-light region even though their
ordination through their oxygen atoms. Ten nitrate ligands absorption maximum is located below 300 nm and the ab-
exhibit a monodentate [2.77(2)–2.991(12) Å], six a μ- sorption intensity clearly diminishes at wavelengths higher
1κO:2κO-type [2.730(19)–2.967(17) Å], two a μ-1κO:2κOЈ- than 400 nm.[10] This report encouraged us to probe
type [2.868(16)–2.923(18) Å], and two a 1κ2O,OЈ-type coor- whether homometallic molecular bismuth oxido clusters,
dination [2.679(17)–2.847(17) Å]. Similar coordination which we have been studying in recent years, exhibit photo-
modes of the ligands are observed in 1b·4DMSO: Four catalytic activity as well. Here we focus on the novel bis-
muth oxido cluster [{Bi38O45(NO3)20(OBz)4(DMSO)24}·
4DMSO][{Bi38O45(NO3)24(DMSO)26}·4DMSO] (1) and the
previously reported clusters [Bi6O4(OH)4(NO3)6]·H2O[3a]
(2), [{Bi38O45(NO3)24(DMSO)26}·2DMSO][{Bi38O45(NO3)24-
(DMSO)24}·0.5DMSO][5] (3), and [Bi38O45(OMc)24-
(DMSO)9]·2DMSO·7H2O[6b] (4). The bismuth oxido clus-
ters were tested as photocatalysts for rhodamine B (RhB)
degradation under visible-light irradiation to evaluate the
influence of the cluster size and the nature of the ligand
shell. The latter aspect is of importance in developing
model compounds to mimic bismuth oxide surfaces.
First, the bandgaps of clusters 1–4 were analyzed. On the
basis of UV/Vis measurements in diffuse reflectance mode,
the absorption edges and the bandgaps, which were calcu-
lated using Tauc plots,[15] were determined as 351 nm
(3.61 eV) for 1, 335 nm (3.76 eV) for 2, and 346 nm
(3.64 eV) for 3 and 4, respectively (Figure 4 and Figure S2
in the Supporting Information). For comparison, the ab-
sorption edges were also determined for solutions in
DMSO and DMF. As shown in Figure S3 of the Support-
ing Information, smaller absorption edges relative to the
solid state were obtained, whereas a small shift to longer
wavelengths was observed by changing the solvent from
DMSO (1: 295 nm; 2: 309 nm; 4: 306 nm) to DMF (1:
Figure 3. Ball-and-stick model of [Bi38O45(NO3)20(OBz)4-
(DMSO)24]·4DMSO (1a·4DMSO) and typical coordination modes
(arrows show the coordination of the four benzoate ligands): (A)
μ4-1κO:2κO:3κOЈ:4κOЈ mode, (B) μ3-1κO:2κO:3κOЈ mode. The
lower part represents the coordination of the nitrate ligands: (C)
Monodentate, (D) μ-1κO:2κO mode, (E) 1κ2O,OЈ mode, and (F)
μ-1κO:2κOЈ mode. Hydrogen atoms are omitted for clarity.
Eur. J. Inorg. Chem. 2014, 302–309
304
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
301 nm; 3: 301 nm; 4: 311 nm). Cluster 2 was measured so-
lely in DMSO and cluster 3 in DMF due to low solubility
in other solvents. As observed from the experiments, a re-
sponse of the clusters with regard to photocatalysis is ex-
pected under UV light rather than under visible light irradi-
ation. However, heterobimetallic bismuth-containing clus-
ters, as mentioned above, also exhibit absorption band max-
ima in the UV/Vis region but show auspicious visible-light-
induced photodegradation, for example, of patent blue V.[10]
However, in most cases, photosensitizers such as [Ru-
(bpy)3]2+ (bpy = 2,2Ј-bipyridine) are added to improve the
visible-light photoactivity of the metal oxido clusters.[9a]
Additionally, photocatalytic activity also depends on the
model dyes that are used. RhB, used as the model dye in
this study, shows a distinct chromophoric system that facili-
tates light absorption in the visible-light region. By interac-
tion of RhB with a semiconductor, electron-transfer pro-
cesses can be induced by irradiation with visible light, which
results in N-deethylation processes of RhB even if the
bandgap of the semiconductor is larger than 3.0 eV.[16] In
addition to such photosensitization processes RhB can also
be degraded with complete destruction of the chromo-
phoric system as a result of the interaction with electron–
hole pairs generated by excitation of the semiconductor
with light irradiation.[16b]
One interesting aspect for photocatalysts based on mo-
lecular materials relative to bulk metal oxides is their solu-
bility in organic solvents, which might offer the possibility
to perform photocatalysis in homogeneous solution. To elu-
cidate this possibility, the highly soluble bismuth oxido clus-
ter 3 (40 mg) was dissolved in dimethylformamide (20 mL)
and was mixed with an aqueous rhodamine B (RhB) solu-
tion (20 mL, 2ϫ10–5 m). These conditions were chosen be-
Figure 4. Normalized UV/Vis diffuse reflectance spectra (top) and
(αhυ)2 versus hυ curves (bottom) of bismuth oxido clusters 1, 2, 3,
and 4. The single UV/Vis diffuse reflectance spectra are given in
Figure S2 of the Supporting Information.
cause they are comparable to those in our previously re- sition), which is still low-level. The low performance as a
ported heterogeneous photocatalytic experiments using β- result of decomposition of the cluster is excluded. We have
Bi2O3 nanoparticles.[8] After dissolution, the homogeneous demonstrated by several methods (including ESI-MS) that
mixture
was
illuminated
with
visible
light cluster 3 is stable in diverse solvents.[6a,17] In addition, clus-
(400 nmՅλՅ700 nm), and the photocatalytic activity was ter 3 was dissolved in a DMF/H2O mixture (1:1) and after
monitored by in situ UV/Vis measurements. However, in evaporation of the solvent a crystalline material was ob-
this experiment,
a
negligible degradation rate of tained, which indicates that the bismuth oxido core is still
1.21ϫ10–6 s–1 (1.3% decomposition after 150 min) was ob- intact as demonstrated by typical diffraction peaks centered
served, and we conclude that our molecular bismuth oxido at 4.93 and 5.73° (2θ) (Figure S7 in the Supporting Infor-
clusters are almost inactive, as expected from the size of the mation).[6c,17] The main reason for the low activity seems
bandgaps (Figure S4 in the Supporting Information). The to be that RhB does not adsorb effectively to the cluster
use of UV/Vis light gave a slightly higher decomposition surface.
rate of 4.10ϫ10–6 s–1 (3.5% decomposition after 150 min;
In a typical heterogeneous photocatalytic experiment, the
Figure S5 in the Supporting Information). The negligible bismuth oxido clusters (40 mg) were dispersed in an aque-
increase in the decomposition rate might be attributed to ous RhB solution (40 mL, 10–5 m). Thus, dynamic light
the instrumental light source, which only provides light with scattering (DLS) experiments in water were made for all
wavelengths longer than 370 nm. Thus, the bismuth oxido tested bismuth oxido clusters to investigate their behavior
cluster is not directly excited and the higher decomposition in aqueous suspensions. Values of (2.2Ϯ0.8) and
rate is a result of the higher energy of the light source, (17.1Ϯ14.3) nm (1), (2.3Ϯ0.5) and (14.4Ϯ4.8) nm (2),
which is provided as a result of the removal of the UV cut- (2.5Ϯ0.5) and (10.9Ϯ4.4) nm (3), as well as (3.1Ϯ0.7) and
off filter. For further investigations, another photoreactor (11.5Ϯ5.3) nm (4) were obtained (Figures S6–S9 in the
with a 160 W mercury-vapor lamp that provided light above Supporting Information). As a result of agglomeration, bi-
250 nm was used (Figure S6 in the Supporting Infor- modal distributions are obtained for all bismuth oxido clus-
mation). In this experiment, a decomposition rate of ters, but the tendency to form agglomerates is less distinct
4.71ϫ10–6 s–1 after 150 min was obtained (4.3% decompo- for cluster 2. Prior to irradiation, the samples were stirred
Eur. J. Inorg. Chem. 2014, 302–309
305
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
in the dark to establish the adsorption/desorption equilib- 1.92ϫ10–4 s–1. Under identical conditions, the samples 1,
rium (–30 minՅtՅ0 min). A change in intensity of the 2, and 4 decompose 79 (1.67ϫ10–4 s–1), 31 (3.89ϫ10–5 s–1),
characteristic RhB absorption band during this period and 80% (1.38ϫ10–4 s–1) of RhB, respectively. The use of
shows that RhB is partially adsorbed at the bismuth oxido commercially available Bi(NO3)3·5H2O shows decomposi-
cluster surfaces; the large clusters 1 (13%), 3 (16%), and 4 tion of approximately 30% of RhB with a reaction rate con-
(27%) adsorb a significant amount of the initially present stant of 3.95ϫ10–5 s–1, which is comparable to the perform-
RhB (Figure 5), whereas adsorption by cluster 2 and the ance of the hexanuclear cluster 2. However, all of the tested
commercially available Bi(NO3)3·5H2O is negligible (Ͻ1%). compounds show a lower photocatalytic activity than β-
Bi2O3 nanoparticles as prepared starting from bismuth ox-
ido clusters.[8] A closer look at the chemical composition of
the bismuth oxido clusters tested show that, in the case of
cluster 2, DMSO is missing as coordinated solvent. Thus,
the photocatalytic experiment using cluster 2 as photocata-
lyst was repeated, and 9 mg of DMSO was added to give a
similar amount of DMSO molecules per bismuth atom rela-
tive to the other bismuth oxido clusters tested. However, no
increase in the photocatalytic performance was observed (k
= 3.81ϫ10–5 s–1; Figure S12 in the Supporting Infor-
mation) and thus any influence of DMSO is excluded. A
higher nuclearity of bismuth oxido clusters is associated
with an increasing photocatalytic activity, and hence higher
decomposition rates compared to cluster 2 and Bi(NO3)3·
5H2O are obtained, respectively. This is most likely attrib-
uted to agglomeration in aqueous media, as shown by DLS
experiments, as well as to significant RhB adsorption,
which is much more distinct for bismuth oxido clusters 1,
3, and 4. Furthermore, higher-nuclearity bismuth oxido
2(1–x)+
clusters are composed of a larger number of [Bi6O8+x
]
units (x = 0, 1), which results in a closer structural rela-
tionship to the photocatalytically active β-Bi2O3.[7,8] Powder
X-ray diffraction (PXRD) analyses after photocatalytic de-
gradation experiments have shown that the crystalline mate-
rials were transformed to give powders of lower crystal-
linity, but they are not fully amorphous (Figure S13 in the
Supporting Information). However, the loss of crystallinity
during photodegradation is attributed not to photocor-
rosion but to ligand exchange in aqueous media, which was
similarly observed by simply stirring the crystalline materi-
als in water in control experiments. Nevertheless, the PXRD
patterns indicate that the bismuth oxido cluster cores are
not destroyed. This assumption was additionally proven by
solid-state UV/Vis measurements, which show no change
for the absorption edges of clusters 1–4 after their use in a
photocatalytic process (Figure S14 in the Supporting Infor-
mation).
Figure 5. Time-dependent conversion (top) and semilogarithmic
plots (bottom) of the photodegradation of an aqueous RhB solu-
tion (10–5 m) under visible light irradiation (tϾ0 min) using bis-
muth oxido clusters 1–4 and Bi(NO3)3·5H2O as photocatalysts;
tϽ0 min shows the adsorption behavior of the samples towards
RhB.
Starting the irradiation using visible light (tϾ0 min) im-
To shed some more light on the degradation process of
mediately initiates photodegradation processes. All of the RhB, further photocatalytic experiments were performed.
tested bismuth oxido clusters show photocatalytic activity, First, a RhB solution that contained the best photocatalyst,
whereas in the absence of a photocatalyst a negligible de- cluster 3, was irradiated with UV/Vis light to probe the
composition of RhB within 150 min (Ͻ5%) is observed. As photocatalytic performance under standard procedures. Af-
shown in Figure 5, the photocatalytic degradation in the ter 150 min of irradiation, approximately 95% of the initial
presence of the bismuth oxido clusters follows pseudo-first- RhB amount was degraded (k = 3.29ϫ10–4 s–1; Figure S15
order reaction kinetics and hence the reaction rate con- in the Supporting Information). Thus, the photocatalytic
stants were determined from the slope of the linear func- performance was approximately doubled (a factor of 1.7)
tions obtained by semilogarithmic plots of ln(c/c0) versus relative to the previous investigations in which visible light
time. The highest photocatalytic activity is observed for the is used exclusively. The irradiation with UV/Vis light allows
bismuth oxido cluster 3, which decomposes 87% of RhB direct excitation of the bismuth oxido cluster, which results
within 150 min with
a
reaction rate constant of in higher decomposition rates. Thus, direct and indirect
Eur. J. Inorg. Chem. 2014, 302–309
306
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
photocatalytic processes might be present. Further UV/Vis clusters or pure RhB did result in very low photodecompo-
experiments were performed to identify crucial processes in sition rates without any blueshift (Figures S16 and S17 in
RhB degradation on the surface of bismuth oxido clusters. the Supporting Information). Thus we conclude that the
As shown by way of example in Figure 6 for a suspension photodegradation process of RhB occurs preferentially at
of cluster 3, no significant blueshift of the absorption band heterogeneous surfaces of the bismuth-containing colloids
of RhB is observed in all degradation experiments, which is under UV/Vis as well as visible conditions and is compar-
indicative of a fast decomposition of the chromophore sys- able to the process as reported for bismuth oxide nanopar-
tem to finally give CO2 and water.[18] Another possible de- ticles.[8] This explains the low activity of 2 and also of 3 in
gradation process that occurs by means of photooxidation solution, both of which lack strong adsorption of RhB.
induces stepwise deethylation of the N,N,NЈ,NЈ-tetraeth-
ylrhodamine, which results in a sequential blueshift of the
characteristic RhB absorption band (λmax = 554 nm). This
is a result of the stepwise formation of the degradation
products N,N,NЈ-triethylrhodamine (λmax = 539 nm), N,NЈ-
diethylrhodamine (λmax = 522 nm), N-ethylrhodamine (λmax
= 510 nm), and rhodamine (λmax = 498 nm).[16b] However,
the absence of the blueshift in solution during RhB decom-
position might not fully reflect the degradation process of
RhB that typically occurs at the surface of a heterogeneous
photocatalyst. To study the photooxidation process at the
cluster surface, the photodegradation of adsorbed RhB was
studied under visible light and for comparison also under
UV/Vis conditions. A small amount of RhB (1 μm) was ad-
sorbed on the surface of cluster 3 (Figure 7), and UV/Vis
Figure 6. UV/Vis spectra of the photodegradation of RhB using
bismuth oxido cluster 3 as dispersed photocatalyst in water (t =
–30 min: starting stirring in the dark; t = 0 min: starting irradiation
with visible light).
measurements in diffuse reflectance mode were recorded. A
decrease in intensity of the RhB absorption band along
with a sequential blueshift to 510 nm within 60 min during
continuous irradiation with the instrument light source
(100 W) was observed. A similar behavior was obtained un-
der visible-light conditions, but a smaller amount of RhB
was decomposed than in the experiment under UV/Vis con-
ditions. The lower activity results from the lower energy
that is available for the photodegradation process due to the
absorption of light by the UV cutoff filter. Additionally, the
bismuth oxido cluster is not directly excited under visible-
light conditions as a result of the bandgap of 3.64 eV, there-
fore the degradation of RhB must be induced by photosen-
sitization. In this process, one electron of the chromophoric
system of RhB is excited by absorption of visible light,
which is then transferred to the conduction band of the
photocatalyst.[16a] The electron is removed from the con-
duction band by interaction with oxygen to give a superox-
ide radical, and a double bond at RhB is formed by separa-
tion of a proton. Several repetitions of these processes lead
to N-deethylation of RhB, which results in ongoing degra-
dation. Thus, direct interactions of RhB with the surface of
a photocatalyst is important prior to inducing degradation.
Direct excitation of the photocatalyst by light to generate
electron holes is not necessary. These processes have been
reported previously.[16] The experiments presented here
show that N,N,NЈ-triethylrhodamine and N,NЈ-diethylrhod-
amine are continuously formed (as a result of an ongoing
deethylation process of RhB at the surface of the bismuth
oxido cluster 3) regardless of whether UV/Vis or visible
light is used. The results therefore indicate that photosensi-
tization processes are predominant in degradation of RhB
with bismuth oxido clusters as photocatalysts. It is worth
Figure 7. Diffuse reflectance spectra of solid bismuth oxido cluster
3 with adsorbed RhB under UV/Vis (top) and visible light irradia-
noting that similar experiments using MgO instead of the tion (bottom).
Eur. J. Inorg. Chem. 2014, 302–309
307
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
of 830 nm. In every DLS experiment a concentration of 1 gL–1 was
used.
Conclusion
The investigations presented here have shown that addi-
tives play an important role in the formation of bismuth
oxido clusters starting from solutions of bismuth(III)
nitrate in DMSO. We have demonstrated the influence of
sodium acetate, sodium benzoate, sodium formiate, sodium
chloroacetate, sodium dichloroacetate, and sodium tri-
chloroacetate. Depending on the pKb values of the addi-
tives, the formation of [Bi(NO3)3(DMSO)3] (pKb Ն11.2)
Bismuth oxido clusters 2,[3a] 3,[5] and 4[6b] were synthesized accord-
ing to the literature.
General Procedure for Hydrolysis of Bi(NO3)3 in the Presence of
Additives: A suspension of Bi(NO3)3·5H2O (2.1 g, 4.25 mmol) in
DMSO (10 mL) was stirred at 80 °C to give a colorless solution.
An appropriate amount (4.25 mmol) of the investigated additive
was added in one portion, then the reaction mixture was stirred at
80 °C overnight and filtered. The filtrate was kept under ambient
conditions for crystallization.
and
(NO3)24(DMSO)24}·0.5DMSO][5] (pKb Յ10.3) was ob-
served. However, in the case of sodium benzoate (pKb 9.8), Synthesis
[{Bi38O45(NO3)24(DMSO)26}·2DMSO][{Bi38O45-
of
[{Bi38O45(NO3)20(OBz)4(DMSO)24}·4DMSO]-
[{Bi38O45(NO3)24(DMSO)26}·4DMSO] (1): By using the above-de-
scribed procedure and sodium benzoate (0.6 g, 4.25 mmol) as addi-
tive, colorless crystals of compound 1 (0.4 g, 28%) were obtained,
four nitrate ligands were substituted for four benzoates,
which resulted in the formation of a cocrystal of similar
{Bi38}
clusters,
[{Bi38O45(NO3)20(OBz)4(DMSO)24}·
m.p. Ͼ200 °C decomp. IR: ν = 3436 (w, br.) , 3008 (w), 2923 (w),
˜
4DMSO][{Bi38O45(NO3)24(DMSO)26}·4DMSO] (1). The
photocatalytic investigations on the degradation of rhod-
amine B under visible-light irradiation have shown the
promising activity of colloidal suspensions of the bismuth
oxido clusters 1, [{Bi38O45(NO3)24(DMSO)26}·2DMSO]-
1603 (w), 1509 (m) , 1472 (s), 1441 (s), 1401 (s), 1312 (m), 1268 (s),
1172 (m), 1072 (w), 1018 (s), 998 (s), 978 (s), 925 (s), 858 (s), 812
(m), 718 (s), 685 (s), 431 (w), 400 (s) cm–1. Bi76C144H368O288N44S58
(25066.78 gmol–1): calcd. C 6.9, H 1.5, N 2.5; found C 7.0, H 1.6,
N 2.5.
[{Bi38O45(NO3)24(DMSO)24}·0.5DMSO][5]
(3),
and
Crystal Data: C138H358Bi76N44O289S58, Mr = 24904.53 gmol–1,
crystal size 0.3ϫ0.2ϫ0.2 mm3, monoclinic, space group P21/c, Z
= 2, a = 35.8728(12) Å, b = 20.2750(5) Å, c = 34.8624(11) Å, β =
111.797(4)°, V = 23543.3(12) Å3, Dcalcd. = 3.536 gcm3, F(000) =
22160, μ(Mo-Kα) = 28.629 mm–1, T = 293 K, 2.92Ͻ2θϽ25.00,
completeness to 2θ: 99.4%. Max./min. residual electron density:
3.852/–3.337 eÅ–3. A total of 111370 reflections collected, 41216
reflections were independent (Rint = 0.0878). Final R1 = 0.0524
[Bi38O45(OMc)24(DMSO)9]·2DMSO·7H2O[6b] (4), but only
minor activity for [Bi6O4(OH)4(NO3)6]·H2O[3a] (2). The
highest degradation rate using visible light irradation was
obtained for compound 3, which decomposes approxi-
mately 87% of a 10–5 m aqueous solution of RhB within
150 min with a reaction rate constant of 1.92ϫ10–4 s–1,
whereas under homogeneous conditions a negligible photo-
catalytic activity was observed. Furthermore, the studies [IϾ2σ(I)] and wR2 = 0.1067 (all data).
have shown that the photocatalytic degradation of RhB un-
CCDC-947073 contains the supplementary crystallographic data
der visible-light irradiation by using bismuth oxido clusters
as photocatalysts preferentially occurs through indirect
photocatalytic processes and that adsorption properties
play an important role. Finally, we postulate that bismuth
oxido clusters might serve as model compounds to mimic
bismuth oxide surfaces in photocatalytic processes, espe-
cially when the indirect photocatalytic process is favored.
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
Photocatalytic Tests: In a typical photodegradation experiment, an
aqueous solution (40 mL) of 1ϫ10–5 m rhodamine B dye (RhB)
and the as-prepared samples (40 mg) in a water-cooled glass reactor
(15 °C) were used. After the suspension was stirred in the dark for
30 min to establish the adsorption/desorption equilibrium, irradia-
tion with a 300 W xenon lamp (Cermax VQTM ME300BF, Co.
Perkin–Elmer) equipped with a hot mirror filter (λՅ700 nm) and
a UV cutoff filter (λՆ420 nm, GG420, Co. Schott) to provide vis-
ible light was started. The effective irradiation area of the suspen-
sion was 4.52 cm2 (approx. 25% of the reactor area). The photo-
catalytic degradation was investigated by in situ UV/Vis measure-
ments. Therefore stirring was stopped for 10 s and darkening of the
light beam was achieved by the addition of a cover. The measure-
ments were carried out up to a maximum time of 150 min (up to
30 min with a 5 min interval; up to 120 min with a 10 min interval).
The photocatalytic degradation rates were determined by calculat-
ing the mathematical area under the characteristic UV/Vis absorp-
tion bands of RhB. In some extra reference experiments, the above-
mentioned UV cutoff filter was removed and a 160 W mercury
vapor lamp (TQ 150 EXKLUSIV UVCP, Co. Heraeus Noblelight
GmbH) was used as light source.
Experimental Section
General: PXRD patterns were measured with a STOE-STADI-P
diffractometer using Cu-Kα radiation (40 kV, 40 mA) and
a
Ge(111) monochromator. The decomposition points were deter-
mined with a Büchi melting apparatus “Melting Point B-540” and
are uncorrected. The data set for the single-crystal X-ray diffraction
study was collected by using Mo-Kα radiation (0.71073 Å) with an
Oxford Gemini S diffractometer at 110 K. All calculations were
performed using the SHELXTL program.[19] The structures were
solved by direct methods and refined by full-matrix least-squares
on F2. Diffuse reflectance UV/Vis spectroscopy was performed with
a single-beam simultaneous spectrometer MCS 400 (Co. Carl Zeiss
Jena GmbH). The UV and visible-light radiation were generated
with a CLD 300 deuterium lamp and a CLX 11 xenon lamp,
respectively. For in situ UV/Vis measurements, which were carried
out to examine the photocatalytic activity, an Agilent Cary 60 UV/
Vis spectrometer (Co. Agilent Technologies) equipped with fiber
optics was used. DLS was performed with a Viscotek model 802
equipped with a laser diode that provided light with a wavelength
Supporting Information (see footnote on the first page of this arti-
cle): It contains DLS data, UV/Vis spectra, data of photocatalytic
degradation experiments, powder X-ray diffraction patterns, and a
structural comparison of the two [Bi38O45]24+ cores in 1.
Eur. J. Inorg. Chem. 2014, 302–309
308
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Zhang, B. Li, Angew. Chem. 2006, 118, 5574; Angew. Chem.
Int. Ed. 2006, 45, 5448–5451; h) V. Andre, A. Hardeman, I.
Halasz, R. S. Stein, G. J. Jackson, D. G. Reid, M. J. Duer, C.
Curfs, M. T. Duarte, T. Friscic, Angew. Chem. 2011, 123, 8004;
Angew. Chem. Int. Ed. 2011, 50, 7858–7861.
Acknowledgments
The authors thank Prof. Dr. Stefan Spange for access to diffuse
reflectance UV/Vis spectroscopy and Janine Fritzsch for per-
forming CHN analyses. We gratefully acknowledge the financial
support of this work by the Deutsche Forschungsgemeinschaft
(DFG) (SPP1415) and the Fonds der Chemischen Industrie (FCI).
[7] M. Mehring, Coord. Chem. Rev. 2007, 251, 974–1006.
[8] a) M. Schlesinger, S. Schulze, M. Hietschold, M. Mehring, Dal-
ton Trans. 2013, 42, 1047–1056; b) M. Schlesinger, M. Weber,
S. Schulze, M. Hietschold, M. Mehring, ChemistryOpen 2013,
2, 146–155.
[1] a) S. Serefis, Arch. Dermatol. Syph. 1934, 171, 1–98; b) P. J. [9] a) C. Streb, Dalton Trans. 2012, 41, 1651–1659; b) C. L. Hill,
Sadler, H. Y. Li, H. Z. Sun, Coord. Chem. Rev. 1999, 185–6,
689–709; c) N. Yang, H. Sun, Coord. Chem. Rev. 2007, 251,
2354–2366; d) H. Ippen, Hautarzt 1997, 48, 424–424; e) M. W.
Whitehead, R. H. Phillips, C. E. Sieniawska, H. T. Delves, P. T.
Seed, R. P. H. Thompson, J. J. Powell, Helicobacter 2000, 5,
169–175.
Comprehensive Coord. Chem. II 2003, 4, 679–759; c) A. Hiskia,
A. Mylonas, E. Papaconstantinou, Chem. Soc. Rev. 2001, 30,
62–69.
[10] J. Tucher, L. C. Nye, I. Ivanovic-Burmazovic, A. Notarnicola,
C. Streb, Chem. Eur. J. 2012, 18, 10949–10953.
[11] L. J. Barbour, S. J. Belfield, P. C. Junk, M. K. Smith, Aust. J.
Chem. 1998, 51, 337–342.
[12] pKb values in water: sodium methacrylate 9.4; sodium acetate
9.4; sodium bezoate 9.8; sodium formiate 10.3; sodium chloro-
acetate 11.2; sodium dichloroacetate 12.47; sodium trichloro-
acetate 13.3.
[13] N. C. Norman, Chemistry of Arsenic, Antimony and Bismuth,
1. ed., Blackie Academic and Professional, London, 1998.
[14] N. E. Brese, M. Okeeffe, Acta Crystallogr., Sect. B 1991, 47,
192–197.
[2] a) Y. M. Yukhin, T. V. Daminova, L. I. Afonina, B. B. Bok-
honov, O. A. Logutenko, A. I. Aparnev, K. Y. Mikhailov, T. A.
Udalova, V. I. Evseenko, Chem. Sustainable Dev. 2004, 12, 395–
401; b) G. Gattow, D. Schott, Z. Anorg. Allg. Chem. 1963, 324,
31–47; c) B. S. Brcic, D. Kolar, F. Lazarini, M. Malesic, Mon-
atsh. Chem. 1973, 104, 365–375.
[3] a) F. Lazarini, Acta Crystallogr., Sect. B 1979, 35, 448–450; b)
F. Lazarini, Cryst. Struct. Commun. 1979, 8, 69–74; c) F. Lazar-
ini, Acta Crystallogr., Sect. B 1978, 34, 3169–3173; d) B. Liu,
W. Zhou, Z. Zhou, X. Zhang, Inorg. Chem. Commun. 2007, 10,
1145–1148; e) A. N. Christensen, B. Lebech, Dalton Trans.
2012, 41, 1971–1980.
[15] a) J. Tauc, R. Grigorov, A. Vancu, Phys. Status Solidi 1966, 15,
627–637; b) A. B. Murphy, Sol. Energy Mater. Sol. Cells 2007,
91, 1326–1337.
[4] K. H. Tytko, Chem. Unserer Zeit 1979, 13, 184–194.
[5] L. Miersch, T. Rüffer, M. Schlesinger, H. Lang, M. Mehring,
Inorg. Chem. 2012, 51, 9376–9384.
[6] a) L. Miersch, M. Schlesinger, R. W. Troff, C. A. Schalley, T.
Rüffer, H. Lang, D. Zahn, M. Mehring, Chem. Eur. J. 2011,
17, 6985–6990; b) L. Miersch, T. Rüffer, M. Mehring, Chem.
Commun. 2011, 47, 6353–6355; c) D. Mansfeld, L. Miersch, T.
Rüffer, D. Schaarschmidt, H. Lang, T. Böhle, R. W. Troff, C. A.
Schalley, J. Müller, M. Mehring, Chem. Eur. J. 2011, 17, 14805–
14810; d) M. Schlesinger, L. Miersch, T. Rüffer, H. Lang, M.
[16] a) O. Merka, V. Yarovyi, D. W. Bahnemann, M. Wark, J. Phys.
Chem. C 2011, 115, 8014–8023; b) T. Watanabe, T. Takizawa,
K. Honda, J. Phys. Chem. 1977, 81, 1845–1851; c) F. Chen,
J. C. Zhao, H. Hidaka, Int. J. Photoenergy 2003, 5, 209–217.
[17] D. Sattler, M. Schlesinger, M. Mehring, C. A. Schalley, Chem-
PlusChem 2013, 78, 1005–1014.
[18] a) K. Yu, S. Yang, H. He, C. Sun, C. Gu, Y. Ju, J. Phys. Chem.
A 2009, 113, 10024–10032; b) Q. Wang, C. Chen, D. Zhao, W.
Ma, J. Zhao, Langmuir 2008, 24, 7338–7345; c) H. Park, W.
Choi, J. Phys. Chem. B 2005, 109, 11667–11674.
Mehring, Main Group Met. Chem. 2013, 36, 11–19; e) P. C. [19] a) G. M. Sheldrick, in: An Integrated System for Solving, Re-
Andrews, G. B. Deacon, C. M. Forsyth, P. C. Junk, I. Kumar,
M. Maguire, Angew. Chem. 2006, 118, 5766; Angew. Chem. Int.
Ed. 2006, 45, 5638–5642; f) P. C. Andrews, M. Busse, G. B.
Deacon, R. L. Ferrero, P. C. Junk, J. G. Maclellan, A. Vom,
Dalton Trans. 2012, 41, 11798–11806; g) E. V. Dikarev, H.
fining and Displaying Crystal Structures from Diffraction Data,
Siemens Analytical X-ray Instruments, Madison, WI, 1990; b)
G. M. Sheldrick, Acta Crystallogr., Sect. A 1990, 46, 467–473.
Received: July 12, 2013
Published Online: November 20, 2013
Eur. J. Inorg. Chem. 2014, 302–309
309
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim