[(FeIII(OH2)2)3 (A-α-PW9O34)2]9- on cationic silica nanoparticles, a new type of material and efficient heterogeneous catalyst for aerobic oxidations

Article abstract of DOI:10.1021/ja0267223

Polyoxometalates (POMs) electrostatically bind to silica nanoparticles coated with cationic aluminum oxide "(Si/AlO₂)ⁿ⁺" to form a new type of material (the anionic POMs replace Cl^- counterions associated with the cationic surface sites). Association of a new _～D_3h POM of formula [(Fe^III(OH₂)₂)₃(A-α-PW₉O₃₄)₂]^9- (1) with the cationic nanoparticles (to form "K₈1/(Si/AlO₂)") was studied in detail. Elemental analysis, particle sizes from both laser light scattering and TEM before and after association of 1, the size of 1 from X-ray crystallography, and other methods provide mutually consistent data that indicate about 58 K₈[(Fe^III(OH₂)₂)₃(A-α-PW₉O₃₄)₂]^- monoanions associate with the average nanoparticle (diameter of the K₈1/(Si/AlO₂) product = _～17 nm). While heterogeneity of the cationic sites and roughness of the (Si/AlO₂)ⁿ⁺ surfaces make the associated POMs structurally nonuniform, the equivalent of ～1 monolayer of 1 is present in K₈1/(Si/AlO₂). Remarkably, while 1, the precursor (Si/AlO₂)ⁿ⁺, and the components of 1, each alone, are inactive as catalysts for O₂/air-based oxidation of sulfides or aldehydes in solution, K₈1/(Si/AlO₂) is an active catalyst for both reactions (facile reaction with air at low temperature). Copyright

Full text of DOI:10.1021/ja0267223

Published on Web 02/22/2003
[(FeIII(OH2)2)3(A-r-PW9O34)2]9- on Cationic Silica Nanoparticles, a New Type of
Material and Efficient Heterogeneous Catalyst for Aerobic Oxidations
Nelya M. Okun, Travis M. Anderson, and Craig L. Hill*
Department of Chemistry, Emory UniVersity, Atlanta, Georgia 30322
Received April 29, 2002 ; E-mail: chill@emory.edu
The development of materials that catalyze selective O2/air-based
oxidations under mild conditions are of both considerable intel-
lectual interest and potential utility. To date there are very few
molecules or materials that catalyze rapid air-based oxidations under
mild conditions.1 We report here the preparation and characteriza-
tion of a new type of material comprised of anionic metal oxygen
clusters (polyoxometalates or “POMs”)2 that are electrostatically
bound to cationic silica nanoparticles (henceforth “(Si/AlO2)n+.”).3
Figure 1. (A) Illustration of the electrostatic association of K81 monoanions
The binding of a new sandwich-type POM4 of formula, [(FeIII-
with the cationic surfaces of the (Si/AlO2)n+ nanoparticles. Compound 1 is
shown in combination polyhedral/ball-and-stick notation. The WO6 and PO4
(OH2)2)3(A-R-PW9O34)2]9- (1), to (Si/AlO2)n+ forms an active
heterogeneous catalyst, for selective O2 oxidations under mild
conditions that is actually more reactive than the same quantity of
the same POM catalyst in solution.
polyhedra are shaded gray and pink, respectively, and the Fe atoms and
the K+ cations are shaded orange and translucent purple, respectively. (B)
A TEM image of an average-sized (∼17 nm) particle of K81/(Si/AlO2) after
catalysis. The sizing bar is 10 nm in length. The dark spots of 1 are more
visible on the lighter background of the larger Si/AlO2 nanoparticle.
The POM, 1, is prepared by the reaction of (A-Na8HPW9O34)5
and Fe(NO3)3 in water, and the K+ salt readily produces both ana-
lytically pure bulk samples and diffraction quality crystals.6 The
X-ray structure of 1 shows it to be an A-type sandwich POM of
approximate D3h symmetry.7,8 The charge on 1 was inferred from
the number of K+ counterions determined by both elemental anal-
ysis and X-ray crystallography. Bond-valence-sum calculations indi-
cate the three central metals in 1 are ferric ions. It is a trivalent
is consistent with ∼58 K81 monoanions per nanoparticle (loss of
one K+ per POM unit).9 Third, the relative sizes of 1 from X-ray
crystallography and the average K81/(Si/AlO2) particle from DLS
are also consistent with an approximate single layer of 1 on the
surface of each nanoparticle. Fourth, high-resolution TEM confirms
the diameter of the average K81/(Si/AlO2) particle from DLS (∼17
nm). Fifth, the TEM of the K81/(Si/AlO2) particles shows the
electrostatically bound electron-dense W-based POMs as dark spots;
the reactant (Si/AlO2)Cl particles show no such spots. The photo-
micrographs all show some spots the size of individual POMs and
larger spots consistent with some POMs that are proximal or on
top of one another on a surface that is rough and also one that has
a nonuniform distribution of cationic surface sites as per the manu-
facturer’s specifications (Figure 1B).3 Catalysis does not change
the appearance of the TEM images.
structural analogue of the POMs of formula, [MII (A-R-PW9O34)2]12-
,
A
3
where M ) MnII, FeII, NiII, CuII, and ZnII, reported by Knoth.4b,c
thermal ellipsoid plot (Figure S1) and all the structural data on the
K+ salt of 1 (K91) (R1 ) 4.19%) are given in the Supporting
Information.
Whereas the anionic POMs do not bind strongly to conventional
negatively charged or neutral silicas (POMs are displaced from most
silicas upon simple washing with conventional solvents), they do
bind very strongly to cationic (Si/AlO2)n+ nanoparticles. Simply
stirring aqueous solutions of the POMs with aqueous suspensions
of (Si/AlO2)Cl produces POM-bound nanoparticles.9 In the case
of 1, five lines of evidence are consistent with the chemistry and
stoichiometry in eq 1 for this process. Namely, there is a loss, on
average, of one of the nine cations of the POM (K91) upon binding
to the (Si/AlO2)n+ nanoparticles, and the POMs form an approx-
imate single layer of K81 monoanions on the surface of each nano-
particle. This translates to an average formula of {K8[(FeIII(OH2)2)3-
(PW9O34)2]}58(Si/AlO2) (henceforth K81/(Si/AlO2)) with an average
diameter of 17 nm. Figure 1A depicts this association, with 1 shown
in combination polyhedral/ball-and-stick notation.
This new type of material, K81/(Si/AlO2), is an active catalyst
for the aerobic oxygenation of sulfides and the autoxidation of
aldehydes (see Supporting Information, part S6), eqs 2 and 3.
R2S + 0.5 O2 w R2SO
(2)
(3)
RCHO + 0.5 O2 w RCOOH
Both reactions proceed rapidly using ambient air as the oxidant at
25 or 75 °C for CH3CHO or THT, respectively. Significantly, an
equimolar quantity of 1 alone, either in the same solution (CH3-
CN) or as a powder, is totally or nearly inactive catalytically.
Aerobic sulfide oxidation in the presence of metal complexes rarely
if ever proceeds by autoxidation. The marked catalysis of eq 2
(studied most extensively where R2S ) tetrahydrothiophene,
“THT”) by K81/(Si/AlO2) (Table 1) coupled with the lack of activity
of 1 dissolved in the same medium (which, in principle, represents
the highest possible total POM “surface area”) argues for some
sort of POM activation upon binding to the particles. An induction
period in the THT-O2-K81/(Si/AlO2) system indicates the initial
K81/(Si/AlO2) is a precatalyst (see Supporting Information, part
S7).11 Typically, the two molecular catalysis parameters, turnover
number and tunover frequency, are not used for heterogeneous cata-
lytic processes because the effective concentration of active sites
58 K91 + (Si/AlO2)Cl w
{K8[(FeIII(OH2)2)3(PW9O34)2]}58(Si/AlO2) + 58 KCl (1)
The first line of evidence for this formulation is dynamic laser
light scattering (DLS) which shows the average diameter of the
initial (Si/AlO2)Cl particles is 12 nm (surface area 227 m2 g-1),
while that of the product K81/(Si/AlO2) particles is 17 nm (160 m2
g-1) (see Supporting Information) as would be roughly expected if
an approximate single layer of K81 monoanions surrounds each
polycationic (Si/AlO2)n+ particle.10 Second, the elemental analysis
9
3194
J. AM. CHEM. SOC. 2003, 125, 3194-3195
10.1021/ja0267223 CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
Table 1. Selective Catalytic Aerobic Oxidation of
Tetrahydrothiophene to Tetrahydrothiophene Oxide (THTO) under
Ambient Conditionsa
Rhule, J. T.; Neiwert, W. A.; Hardcastle, K. I.; Hill, C. L. J. Am. Chem.
Soc. 2001, 123, 12101-12102.
(2) (a) Topical issue on polyoxometalates: Chem. ReV. 1998, 98, 1-389. (b)
Polyoxometalate Chemistry: From Topology Via Self-Assembly to Ap-
plications; Pope, M. T., Mu¨ller, A., Eds.; Kluwer Academic Publishers:
Dordrecht, The Netherlands, 2001.
(3) (a) Bindzil CAT (Akzo Nobel) contains spherical amorphous silica
particles with surfaces covered by cationic alumina that is charge balanced
with Cl-. The cationic coatings are not of monolayer uniformity. The
samples used were aged and contained slightly aggregated silica particles.
Positively charged silica sols were first reported in the following:
Alexander, G. B.; Bolt, G. H. U.S. Pat. 3,007,878 (Du Pont), 1961. (b)
Neumann and others have used electronic attraction to immobilize
homogeneous POM catalysts on cationic materials: (i) Kasem, K. K.;
Schultz, F. A. Can J. Chem. 1995, 73, 858-864. (ii) Neumann, R.; Miller,
H. J. Chem. Soc., Chem. Commun. 1995, 22770-2278. See also Neumann,
R.; Cohen, M. Angew. Chem., Int. Ed. Engl. 1997, 36, 1738-1740. (c)
Layered Double Hydroxide (LDH)-POM materials have also been
prepared. See: Kwon, T.; Tsigdinos, G. A.; Pinnavaia, T. J. J. Am. Chem.
Soc. 1988, 110, 3653-3654.
f
g
catalystb
POM (mmol)c % cnvd % ylde TOF
TON
(Si/AlO2)Clh
0
0
1.5
2.5
0
1.5
2.5
0
0
0
0
1.8
2.2
Fe(Si/AlO2)i
0.0008j
0.0045
TBA9Fe3(A-PW9O34
)
2
(TBA91) (homogeneous rxn)
K81/(Si/AlO2)k
0.0045
28
28
0.5
60
a General conditions: 0.99 mmol (0.397 M) of THT, catalyst (amount
given in column 2), 1 atm of air, 0.875 mmol (0.35 M) trichlorobenzene
(internal standard) were stirred in 2.5 mL of acetonitrile at 75 °C for 120
h. b No product was observed in the absence of POM, (Si/AlO2)n+ or POM/
(Si/AlO2). c mmol of total POM present in the catalyst during turnover.d %
e
conversion ) (moles of THT consumed/moles of initial THT) × 100.
%
yield ) (moles of THTO/moles of initial THT) × 100. f Turnover frequency
) turnovers/reaction time (120 h). g Turnovers ) (moles of THTO/moles
of POM). h Cationic silica (Bindzil CAT)3. i Fe(III)-coated Bindzil CAT.
j mmol of Fe2(SO4)3 (no POM present). k 1 ) [(FeIII3)(A-PW9O34)2]9-
(preparations in Supporting Information).
(4) Representative papers on sandwich-type POMs: (a) Weakley, T. J. R.;
Evans, H. T., Jr.; Showell, J. S.; Tourne´, G. F.; Tourne´, C. M. J. Chem.
Soc., Chem. Commun. 1973, 4, 139-140. (b) Knoth, W. H.; Domaille, P.
J.; Farlee, R. D. Organometallics 1985, 4, 62-68. (c) Knoth, W. H.;
Domaille, P. J.; Harlow, R. L. Inorg. Chem. 1986, 25, 1577-1584. (d)
Finke, R. G.; Droege, M. W.; Domaille, P. J. Inorg. Chem. 1987, 26,
3886-3896. (e) Tourne´, C. M.; Tourne´, G. F.; Zonnevijlle, F. J. Chem.
Soc., Dalton Trans. 1991, 1, 143-155. (f) Zhang, X. Y.; Jameson, G. B.;
O’Connor, C. J.; Pope, M. T. Polyhedron 1996, 15, 917-922. (g) Zhang,
X.; Chen, Q.; Duncan, D. C.; Lachicotte, R. J.; Hill, C. L. Inorg. Chem.
1997, 36, 4381-4386. (h) Clemente-Juan, J. M.; Coronado, E.; Gala´n-
Mascaro´s, J., J. R.; Go´mez-Garc´ıa, C. J. Inorg. Chem. 1999, 38, 55-63.
(i) Loose, I.; Droste, E.; Bo¨sing M.; Pohlmann, H.; Dickman, M. H.; Rosu,
C.; Pope, M. T.; Krebs, B. Inorg. Chem. 1999, 38, 2688-2694. (j) Zhang,
X.; Anderson, T. M.; Chen, Q.; Hill, C. L. Inorg. Chem. 2001, 40, 418-
419. (k) Anderson, T. M.; Hardcastle, K. I.; Okun, N.; Hill, C. L. Inorg.
Chem. 2001, 40, 6418-6425. (l) Anderson, T. M.; Zhang, X.; Hardcastle,
K. I.; Hill, C. L. Inorg. Chem. 2002, 41, 2477-2488.
is not known. We are including them in Table 1 because the collec-
tive lines of evidence above make a reasonable case that nearly all
the POMs in K81/(Si/AlO2) are accessible to the solvent, and thus
their “concentration” can be reasonably approximated. Quantifica-
tion of the organic reactants (CH3CHO or THT) and products (CH3-
COOH or sulfoxide, THTO) by gas chromatography and the O2
consumption by manometry confirm that the stoichiometries are
those given in eqs 2 and 3. In a control experiment, a sample of
K81/(Si/AlO2) was used as a catalyst, and the mixture was then
filtered. The recovered solid was nearly as catalytically active as
the initial sample, whereas the filtrate was totally inactive. This
indicates that the solid is the actual catalyst and these new catalytic
materials are quite stable.
(5) Domaille, P. J. In Inorganic Syntheses; Ginsberg, A. P., Ed.; John Wiley
and Sons: New York, 1990; Vol. 27, pp 96-104.
(6) Synthesis of K9[(Fe(OH2)2)3(A-R-PW9O34)2]‚20H2O (K91): Solid A-Na9-
PW9O34‚7H2O (10 g, ca. 3.7 mmol) and Fe(NO3)3‚9H2O (3.2 g, 8 mmol)
were added simultaneously to 80 mL of deionized water. The mixture
was stirred for 15 min at 50 °C to form a clear yellow solution to which
KCl (11 g) was added. The resulting precipitate (ca. 8 g) was separated
by filtration and redissolved in a minimal amount of 50 °C water, and
the solution was filtered to remove any insoluble material. The filtrate
was cooled to 5 °C overnight to afford 6 g of yellow-orange crystals (yield
60%). Diffuse-reflectance-Fourier transform-infrared (5% sample in KBr,
1200-400 cm-1): 1080 (s), 1058 (s, sh), 953 (s), 881 (m), 799 (s), 752
(s, sh), 595 (w), and 517 (w). Anal. Calcd for H46Fe3K9O94P2W18: H,
0.85; Fe, 3.08; K, 6.47; P, 1.14; W, 60.82. Found: H, 0.88; Fe, 3.13; K,
6.53; P, 1.17; W, 59.09. Magnetic susceptibility: µeff ) 6.1 µB/mol at
296K. MW: 5441.
The loss of water molecules is the only change apparent in K81/
(Si/AlO2) (as well as K91) up to 200 °C on the basis of TGA, DSC,
and DRIFT data. The dramatic increase in catalytic activity (Table
1) suggests a significant change in 1 upon binding to (Si/AlO2)n+ 12
.
To probe this change further, IR, 57Fe Mo¨ssbauer, and EPR studies
were conducted. The P-O and W-O stretches in the IR of 1 (1000
and 800 cm-1, respectively) before and after binding are the same
within experimental error, but the dominance of the peaks from
the abundant (Si/AlO2)n+ make further inferences difficult. Scat-
tering of the γ-rays by the heavy tungsten atoms in 1 rendered
Mo¨ssbauer useless, but EPR shows a high spin ferric signal at g )
4.30 that changes slightly and becomes about 15 times as intense
when 1 binds to (Si/AlO2)n+ (Figure S4). On the basis of this result,
a control experiment was conducted: Fe2(SO4)3 was deposited on
the (Si/AlO2)n+ particles and catalytic activity of this mixture for
eq 2 was assessed. It was almost inactive (Table 1), indicating that
binding of 1 on the cationic nanoparticles does not likely involve
(7) Crystal data for K9[(Fe(OH2)2)3(A-R-PW9O34)2]‚20H2O (K91): Orthorhom-
bic space group Fdd2, dark yellow efflorescent crystal, with a ) 40.376-
(4), b ) 27.904(2), and c ) 31.186(2) Å, and Z ) 16. The data were
collected on a Bru¨ker D8 SMART APEX CCD sealed-tube diffractometer
with Mo KR (0.71073 Å) radiation (temperature ) 100(2) K). At final
convergence, R1 ) 4.19% and GOF ) 1.117 based on 31,831 reflections
with Fo > 2σFo.
(8) Each of the three Fe atoms in the central unit exhibit a MO6 coordination
polyhedron; each has two oxygens from each trivacant Keggin subunit,
one exterior oxygen atom, and one interior oxygen atom. The latter two
oxygens are most likely water molecules, although the hydrogen atoms
could not be located. There is crowding of the interior water molecules,
and the O-O distances between these oxygen atoms (ranging from 1.97
to 2.07 Å) are consistent with strong internal hydrogen bonding (which
consequently leads to buckling and lowering of the symmetry to Cs).
(9) Synthesis of POM-modified cationic silica nanoparticles, K81/(Si/AlO2):
To an aqueous suspension of (Si/AlO2)Cl (Akzo Nobel Bindzil CAT; 10.0
g) was added 0.25 g of K91 dissolved in 10 mL water. This mixture was
stirred for 3 h at 25 °C and then at 80 °C until the solvent had evaporated.
The resulting powder was dried at 120 °C for 1 h, washed with three
10-mL portions of CH3CN (with no loss of POM after the first wash),
and dried again at 120 °C for 1 h. The average number of (Si/AlO2)n+
particles was calculated from the (Si/AlO2)n+ content in the sol and the
production of solvated or silica-bound Fe(III).
Acknowledgment. We thank F. Menger for use of the DLS
apparatus, P. Bergoo and J. Rise of Akzo Nobel for assistance with
silica products, K. Hardcastle and W. Neiwert for X-ray crystal-
lography, B. H. Huynh and G. Jameson for EPR, R. P. Apkarian
and M. Ritorto (Emory IM&MF) for TEM, and colleagues for
average particle diameter. Anal. Calcd for “K81(Si/AlO2)”, i.e., (K81)58
-
(SiO2)15570(Al2O3)4315(H2O)23095: Al, 11.23; Fe, 0.454; K, 0.954; P, 0.168;
Si, 21.00. Found: Al, 11.57; Fe, 0.475; K, 0.865; P, 0.175; Si, 21.34.
(10) The specific surface area was calculated from the diameters, taking the
density of silica as 2.2 g/cm3 (Iler, R. K. The Chemistry of Silica; Wiley:
New York, 1979; p 346.).
discussion.
Supporting Information Available: CIF file for K91, DLS, TEM,
EPR, and catalytic data for CH3CHO and THT oxidations. This material
is available free of charge via the Internet at http://pubs.acs.org.
(11) The reaction is first order in THT and K81/(Si/AlO2) and zero order in
O2. The reaction shows a 20 h induction period. See Supporting
Information, part S7.
References
(12) The surface charge of (Si/AlO2)n+ and K81/(Si/AlO2) was recently
determined on a Muetek particle charge detector PCD02. The results were
consistent with a reduction in charge (i.e., the surface charge is now less
positive) upon addition of K91.
(1) Representative recent papers: (a) Bosch, E.; Kochi, J. K. J. Org. Chem.
1995, 60, 3172-3183. (b) Haruta, M. Catal. SurV. Jpn. 1997, 1, 61-73.
(c) Xu, L.; Boring, E.; Hill, C. L. J. Catal. 2000, 195, 394-405. (d) Boring,
E.; Geletii, Y. V.; Hill, C. L. J. Am. Chem. Soc. 2001, 123, 1625-1635.
(e) Martin, S. E.; Rossi, L. I. Tetrahedron Lett. 2001, 42, 7147-7152. (f)
Boring, E.; Geletii, Y.; Hill, C. L. J. Mol. Catal. 2001, 176, 49-63. (g)
JA0267223
9
J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003 3195