High-activity, single-site mesoporous Pd/Al2O3 catalysts for selective aerobic oxidation of allylic alcohols

Article abstract of DOI:10.1002/anie.200702534

(Chemical Equation Presented) Pd does it alone: Tailored heterogeneous catalysts offer exciting, alternative, clean technologies for regioselective molecular transformations. A mesoporous alumina support stabilizes atomically dispersed Pd^II surface sites (see picture, C light gray, O red, Pd dark gray, Al purple, H white), thereby dramatically enhancing catalytic performance in the aerobic selective oxidation of alcohols.

Full text of DOI:10.1002/anie.200702534

Angewandte
Chemie
DOI: 10.1002/anie.200702534
Heterogeneous Catalysis
High-Activity, Single-Site Mesoporous Pd/Al2O3 Catalysts for Selective
Aerobic Oxidation of Allylic Alcohols**
Simon F. J. Hackett, Rik M. Brydson, Mhairi H. Gass, Ian Harvey, Andrew D. Newman,
Karen Wilson, and Adam F. Lee*
The heterogeneously catalyzed aerobic selective oxidation of
hydrocarbons offers new, environmentally benign routes to a
diverse range of valuable intermediates for the pharmaceut-
ical, fine chemical, and agrochemical sectors.[1] Such powerful
catalytic technologies circumvent the use of stoichiometric
reagents and expensive homogeneous complexes along with
the associated process disadvantages and safety issues.[2]
While there has been recent interest in the potential of gold
as a partial oxidation catalyst,[3] such systems typically offer
low oxygenate yields and require either radical initiators,[4]
high temperatures, or high O2 partial pressures.[5] Supported
ruthenium,[6] palladium, and platinum clusters are also
promising candidates to catalyze the selective oxidation
(selox) of primary alcohols to their corresponding aldehydes
under mild conditions.[7] These reactions are highly regiose-
lective in the presence of diverse functionalities, including
allylic groups. Most studies in this area employing Pd and Pt
have utilized commercial formulations based upon amor-
phous carbon or oxide supports. However, the poor intrinsic
performance of these materials relative to their homogeneous
counterparts has often necessitated ad hoc promotion of the
reaction by non-noble metals to achieve even moderate
yields.[8] Two factors presently constrain the wider adoption of
heterogeneous Pd selox catalysts amongst both academic and
industrial communities: first, uncertainty over the active site
responsible for the rate-limiting oxidative dehydrogenation
step,[9] and second, the use of conventional supports with
restrictive pore dimensions that inhibit efficient mass and
heat transfer to and from reaction sites and limit the range of
viable solvents and substrates.[10]
nated surfaces with decreasing cluster size,[12] we hypothe-
sized that a mesoporous high-area support would serve to
both stabilize highly dispersed palladium oxide nanoparticles
and would facilitate efficient alcohol and aldehyde diffusion.
Herein, we report the successful synthesis of tailored Pd–
Al2O3 catalysts and demonstrate that extremely low palla-
dium loadings generate atomically dispersed PdII surface
species that confer exceptional selox activity of allylic
alcohols.
Mesoporous alumina-supported palladium catalysts (Pd/
meso-Al2O3) were prepared by a modified surfactant-tem-
plated route through hydrolysis of aluminum sec-butoxide
and subsequent aging in the presence of lauric acid.[13] The
organic template was removed by calcination prior to
incipient wetness impregnation with tetraammine palladiu-
m(II) nitrate solution (see the Supporting Information).
Samples were then calcined, reduced, and stored in air.
Porosimetry and powder X-ray diffraction confirmed that the
final processed materials possessed well-defined, hexagonal
pore structures, with surface areas of 350 m2 gꢀ1 and uniform
3-nm-diameter thick-walled pores. X-ray analysis provides
clear evidence that the mesoporous support stabilizes palla-
dium nanoparticles that are smaller than can be achieved
using conventional amorphous aluminas. Furthermore,
shrinking palladium nanoparticle dimensions from 4.5-nm to
below 1.7-nm diameter induces a striking transition from
surface-bound Pd0 aggregates to isolated PdII sites. Palladium
K-edge EXAFS analysis (Figure 1, EXAFS = extended X-ray
ꢀ
absorption fine structure) reveals a strong signal from a Pd
Pd coordination shell at 2.75 Š for B (4.7 wt% Pd), indicative
of large metallic Pd clusters. In sharp contrast, the ultradilute
sample C (containing only 0.03 wt% Pd) possesses only four-
Recent studies have implicated surface palladium oxide as
the active center in allylic alcohol selox.[11] Since the
energetics of metal clusters increasingly favor oxide-termi-
ꢀ
coordinate Pd O scatterers at 2.00 Š, akin to a truncated
PdO environment.
Linear combination fitting of the corresponding X-ray
absorption near-edge structure (XANES) data with oxide and
metal standards verifies that mesoporous alumina stabilizes
palladium in the + 2oxidation state. A weak shoulder in the
edge threshold of C further indicates that palladium adopts a
four-coordinate, pseudo-square-planar geometry similar to
that in PdO;[14] this shoulder is absent for isotropic symme-
tries at PdIII and PdIV centers.[15] X-ray photoelectron spec-
troscopy (XPS) corroborates this gradual evolution from
metallic to oxidic surface palladium with decreasing cluster
size (i.e. Pd loading).
[*] S. F. J. Hackett, A. D. Newman, Dr. K. Wilson, Dr. A. F. Lee
Department of Chemistry
University of York, York YO10 5DD (UK)
Fax: (+44)1904-432-516
E-mail: afl2@york.ac.uk
Homepage: http://www-users.york.ac.uk/%7Eafl2/index.htm
Dr. M. H. Gass, Dr. I. Harvey
Daresbury Laboratory
Daresbury, Cheshire WA4 4AD (UK)
Prof. R. M. Brydson
Institute for Materials Research
University of Leeds, Leeds LS2 9JT (UK)
The core-level chemical shift and near-edge absorbance of
the atomically dispersed PdII species are indistinguishable
from those of PdII residing within extended palladium oxide
phases, despite their different coordination spheres. However,
[**] We thank the EPSRC (EP/E046754/1, EP/P500575/1) and CCLRC
(beamtime award 46106) for financial support of this work.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or fromthe author.
Angew. Chem. Int. Ed. 2007, 46, 8593 –8596
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8593
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Take advantage of blue reference links
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&

Communications
distributions are presented in the Supporting Information. It
is impossible to identify such small particles or single atoms in
low-Pd-loaded samples using standard TEM or in aberration-
corrected bright-field imaging (see the Supporting
Information for comparison).
The catalytic performance of mesoporous alumina-sup-
ported palladium clusters was subsequently explored in the
aerobic selox of cinnamyl and crotyl alcohols to their
corresponding aldehydes. Figure 3 shows the resultant turn-
Figure 1. Left: Pd EXAFS of Pd foil (A), 4.7 wt% Pd/meso-Al2O3 (B),
0.03 wt% Pd/meso-Al2O3 (C), and PdO (D). Right: Fitted Pd XANES
(theory corresponds to 15% A and 85% D) and schematic depiction of
Pd bound in fourfold hollow sites on a (100)-terminated g-alumina
surface.
using high-resolution high-angle annular dark-field
(HAADF) scanning transmission electron microscopy
(STEM), it was possible to directly image these isolated Pd
centers. Figure 2shows a smoothed HAADF STEM image of
the 0.03 wt% Pd/meso-Al2O3 sample; two pores of approx-
imately 2.5 nm are apparent in the support in the center of the
image. HAADF imaging is particularly sensitive to atomic
number, with heavy atoms appearing brighter.[16] A large
number of such bright spots are distributed across the alumina
framework, with common diameters of approximately
0.13 nm, consistent with individual heavy atoms either on or
within the support.[17]
Figure 3. Catalytic performance of Pd clusters dispersed across con-
ventional versus mesoporous alumina supports as a function of size
and oxidation state in aerobic crotyl alcohol selox.
Time-lapse imaging reveals that these bright spots are
mobile, indicating movement across the pore walls of the
support as a result of excitation from the focused electron
beam and thus confirming the atomic nature of the spots. No
larger aggregates were observed, although larger particles
were seen at higher Pd loadings. Representative particle-size
over frequencies (TOF) per exposed surface Pd site for crotyl
alcohol as a function of cluster size (loading) and total (oxide
and single-site) PdII content. Metal dispersions were mea-
sured by H2 chemisorption, while the total PdII concentration
was derived from the fitted XANES spectra. The latter was in
excellent agreement with the corresponding surface PdII
concentration from XPS (Table 1), thus indicating that
electron-deficient palladium was always surface-segregated
(either as a capping surface PdOx layer around metal cores, as
small oxide particles, or as single sites, depending on the
loading).
The impact of using a high-area, mesoporous alumina
support compared with conventional g-Al2O3 is dramatic. As
the Pd loading and cluster size fall, the ability of the
mesoporous support to stabilize atomically dispersed PdII
sites confers a tenfold rate enhancement over the conven-
tional system. The higher support surface area undoubtedly
boosts palladium dispersion, resulting in clusters for similar
loadings that are smaller than those achieved over the
amorphous alumina (see the Supporting Information). How-
ever, another key factor may be a higher defect density within
the transitional alumina phase formed in our mesoporous
support.[13] Such defects are likely nucleation centers for
palladium, and indeed, alumina surfaces have been recently
shown to inhibit adatom diffusion and sintering of small Au
particles.[18]
Figure 2. HAADF STEM image of a 0.03 wt% Pd/meso-Al2O3 sample.
8594 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8593 –8596
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Take advantage of blue reference links
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&

Angewandte
Chemie
Table 1: Structural and catalytic properties of Pd/meso-Al2O3.
kilogram scale, that they exhibit
excellent longevity and recyclabil-
ity, and that maximum activity is
attained concomitant with vast
reductions in precious-metal con-
tent. Highly dispersed PdII has been
postulated as the active species in
Wacker-type partial oxidation in
doped MoVNbOx catalysts.[20]
Catalyst
[wt%]
Particle size[a] Total PdII
[nm]
Surface PdII TON[d]
[%][c]
Yield
Selectivity
[%][f]
ꢀ1 [e]
[%][b]
[mmolhꢀ1 gcat.
]
0.03
0.06
1.03
2.37
4.70
0.13
0.8
1.3
2.0
4.4
85
59
43.4
23.6
17.8
76.6
57.2
44.1
22.2
10.2
1305 (822) 3.1 (1.9)
1908 (678) 5.1 (1.8)
163 (50) 29.9 (9.0)
55 (18) 24.8 (8.2)
36 (16) 17.1 (7.2)
91 (92)
88 (92)
91 (91)
93 (90)
94 (89)
[a] Average fromEXAFS 1st shell CN, X-ray line-broadening, and TEM. [b] Fromfitted XANES Pd K-edge
spectra. [c] From fitted Pd 3d XP spectra. [d] Turnover number based on crotyl alcohol turnover per mole
Pd after 1 h; cinnamyl alcohol values are given in parentheses. [e] Cinnamyl alcohol values are given in
parentheses. [f] Selectivity for aldehyde; cinnamyl alcohol values are given in parentheses.
In conclusion, we have synthe-
sized new mesoporous catalysts that
contain
atomically
dispersed
PdII centers. These materials show
exceptional catalytic performance
It is interesting to note the nonlinear dependence of TOF
on the total PdII concentration, which reflects the structural
evolution of supported palladium. We have previously shown
that metallic Pd (dominant at high loadings) exhibits poor
activity towards allylic alcohol selox.[11b] The slow initial rise
in TOF in Figure 3 over both alumina supports is associated
with the transition from large metal to small oxidic clusters
that expose moderately active surface PdOx. To confirm that
in the aerobic selox of allylic alcohols under mild conditions.
The dramatic dependence of cinnamaldehyde production on
electron-deficient palladium highlights the parallel between
homogeneous and heterogeneous alcohol selox chemistry and
the importance of isolated PdII sites in the catalytic cycle—a
discovery that opens new avenues for tailored selox catalysts.
the dramatic rate enhancement in the ultradilute regime for Experimental Section
Alcohol oxidation was performed on a Radleys carousel on a 10-mL
the Pd/meso-Al2O3 catalyst is attributable to the genesis of
isolated PdII centers, apparent activation energies were
measured for different loadings. These measurements
indeed reveal two distinct regimes associated with the switch-
over at approximately 0.06 wt% palladium as given in
Equation (1):
scale at 608C. Individual reactors were charged with benzyl, cinnamyl,
or crotyl alcohol (0.84 mmol, Aldrich, 99%) in toluene (10 mL) with
mesitylene as an internal standard. Reactions were performed on air
with catalyst (50 mg) and analyzed by GC using a DB5 capillary
column. Cinnamaldehyde, crotonaldehyde, and benzaldehyde were
the principal reaction products in all cases (87–94% selectivity).
There was no evidence for solvent oxidation or alcohol isomerization
or polymerization, reflecting our mild operating conditions. Reac-
tions were run for up to 24 h with initial rates determined from the
linear portion of the reaction profile. Catalyst selectivity and mass
balances (closure was greater than 98%) were determined using
reactant and product response factors with quoted conversions and
selectivities given to Æ 2and Æ 3%, respectively. All catalyst samples
could be recycled by simple filtration, washing with toluene, and air
drying without significant loss of activity over three consecutive runs
(TOFs remaining within 10% of initial values). Exemplary kinetic
profiles are shown in the Supporting Information.
surface PdIIOx phase !atomically dispersed PdII species
ð1Þ
DEact ꢁ 45 Æ 5 kJ molꢀ1 DEact ꢁ 10 Æ 5 kJ molꢀ1
With TOF = 4400 hꢀ1, our 0.03 wt% Pd/meso-Al2O3 cata-
lyst is the most active ever reported for heterogeneous
cinnamyl alcohol oxidation. This result compares favorably
with the best literature values of 538 hꢀ1 over Au/CeO2[3c] (at
1208C) and 27 hꢀ1 over Ru/Al2O3.[6b] The performance in
crotyl alcohol oxidation is no less exceptional, with TOF =
7080 hꢀ1 achievable at 0.21 atmospheres O2 and 608C, second
only to a recently reported AuPd/TiO2 catalyst that attained
12600 hꢀ1, albeit under 5 atm O2 and at 1608C.[5] In contrast,
the best literature report for crotyl alcohol selox under
comparable reaction conditions, over Ru/Faujisite, achieved a
TOF of only 3 hꢀ1.[19] For both alcohols, aldehyde selectivities
with our system were better than 91% (Table 1). In absolute
terms, redistributing palladium from large metal clusters
(4.7 wt%) to PdII species decorating the mesoporous alumina
(0.03 wt%) results in a 30-fold increase in the activity per
Pd atom. We also observed an excellent high TOF of 4096 hꢀ1
for benzyl alcohol oxidation, which was more than 99%
selective for benzaldehyde, thus indicating the applicability of
the catalyst beyond allylic systems.
Received: June 11, 2007
Revised: August 30, 2007
Published online: October 5, 2007
Keywords: heterogeneous catalysis · mesoporous materials ·
.
oxidation · palladium· X-ray absorption spectroscopy
[1] a) T. Mallat, A. Baiker, Chem. Rev. 2004, 104, 3037 – 3058; b) M.
Besson, P. Gallezot, Catal. Today 2000, 57, 127 – 141.
[2] a) R. A. Sheldon, Pure Appl. Chem. 2000, 72, 1233 – 1246;
b) R. A. Sheldon, I. W. C. E. Arends, G.-J. ten Brink, A. Dijks-
man, Acc. Chem. Res. 2002, 35, 774 – 781.
[3] a) G. J. Hutchings, S. Carrettin, P. Landon, J. K. Edwards, D.
Enache, D. W. Knight, Y.-J. Xu, A. F. Carley, Top. Catal. 2006, 38,
223 – 230; b) A. Abad, P. Concepción, A. Corma, H. García,
Angew. Chem. 2005, 117, 4134 – 4137; Angew. Chem. Int. Ed.
2005, 44, 4066 – 4069; c) A. Abad, C. Almela, A. Corma, H.
Operando EXAFS measurements confirm that these PdII
centers remain unchanged during the course of reaction over
periods of days (unlike other supported Pd systems, for which
reduction and sintering occur in situ[7a,9a,11a]) and are therefore
unequivocally the active catalytic species. It is important to
note that these catalysts are simple to synthesize on a
´
Garcyꢀa, Chem. Commun. 2006, 3178 – 3180.
[4] M. D. Hughes, Y.-J. Xu1, P. Jenkins, P. McMorn, P. Landon, D. I.
Enache, A. F. Carley, G. A. Attard, G. J. Hutchings, F. King,
Angew. Chem. Int. Ed. 2007, 46, 8593 –8596
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8595
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Take advantage of blue reference links
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&

Communications
E. H. Stitt, P. Johnston, K. Griffin, C. J. Kiely, Nature 2005, 437,
1132– 1135.
2006, 8, 549 – 555; c) J. D. Grunwaldt, M. Caravati, M. Ramin, A.
Baiker, Catal. Lett. 2003, 90, 221 – 229.
[5] D. I. Enache, J. K. Edwards, P. Landon, B. Solsona-Espriu, A. F.
Carley, A. A. Herzing, M. Watanabe, C. J. Kiely, D. W. Knight,
G. J. Hutchings, Science 2006, 311, 362– 365.
[12] C. T. Campbell, Surf. Sci. Rep. 1997, 27, 1 – 111.
[13] F. Vaudry, S. Khodabandeh, M. E. Davis, Chem. Mater. 1996, 8,
1451 – 1464; see the Supporting Information.
[6] a) K. Yamaguchi, N. Mizuno, Angew. Chem. 2002, 114, 4720 –
4724; Angew. Chem. Int. Ed. 2002, 41, 4538 – 4542; b) K.
Yamaguchi, N. Mizuno, Chem. Eur. J. 2003, 9, 4353 – 4361.
[7] a) K. Mori, K. Yamaguchi, T. Hara, T. Mizugaki, K. Ebitani, K.
Kaneda, J. Am. Chem. Soc. 2002, 124, 11572– 11573; b) A. F.
Lee, J. J. Gee, H. J. Theyers, Green Chem. 2000, 2, 279 – 282.
[8] a) T. Mallat, Z. Bodnar, A. Baiker, Stud. Surf. Sci. Catal. 1993,
78, 377 – 384; b) F. Alardin, B. Delmon, P. Ruiz, M. Devillers,
Catal. Today 2000, 61, 255 – 262.
[14] K.-T. Park, D. L. Novikov, V. A. Gubanov, A. J. Freeman, Phys.
Rev. B 1994, 49, 4425 – 4431.
[15] S.-J. Kim, S. Lemaux, G. Demazeau, J.-Y. Kim, J.-H. Choy, J.
Mater. Chem. 2002, 12, 995 – 1000.
[16] S. J. Pennycook, L. A. Boatner, Nature 1988, 336, 565 – 567.
[17] M. D. Shannon, C. M. Lok, J. L. Casci, J. Catal. 2007, 249, 41 – 51.
[18] Z. Ma, S. H. Overbury, S. Dai, J. Mol. Catal. A 2007, 273, 186 –
197.
[19] B.-Z. Zhan, M. A. White, T.-K. Sham, J. A. Pincock, R. J.
Doucet, K. V. R. Rao, K. N. Robertson, T. S. Cameron, J. Am.
Chem. Soc. 2003, 125, 2195 – 2199.
[9] a) J. D. Grunwaldt, M. Caravati, A. Baiker, J. Phys. Chem. B
2006, 110, 9916 – 9922; b) J. D. Grunwaldt, M. Caravati, A.
Baiker, J. Phys. Chem. B 2006, 110, 25586 – 25589.
[20] D. Linke, D. Wolf, M. Baerns, O. Timpe, R. Schlögl, S. Zeyß, U.
Dingerdissen, J. Catal. 2002, 205, 16 – 31.
[10] A. Corma, Chem. Rev. 1997, 97, 2373 – 2419.
[11] a) A. F. Lee, K. Wilson, Green Chem. 2004, 6, 37 – 42 ; b) A. F.
Lee, S. F. J. Hackett, J. S. J. Hargreaves, K. Wilson, Green Chem.
8596 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8593 –8596
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&
&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
&
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Take advantage of blue reference links
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&