The role of outer-sphere surface acidity in alkene epoxidation catalyzed by calixarene-Ti(IV) complexes

Article abstract of DOI:10.1021/ja074614g

Cooperativity between Bronsted acidic defect sites on oxide surfaces and Lewis acid catalyst sites consisting of grafted calixarene-Ti(IV) complexes is investigated for controlling epoxidation catalysis. Materials are synthesized that, regardless of the surface or calixarene substituent, demonstrate nearly identical UV-visible ligand-to-metal charge-transfer bands and Ti K-edge X-ray absorption near edge spectral features consistent with site-isolated, coordinatively unsaturated Ti(IV) atoms. Despite similar Ti frontier orbital energies demonstrated by these spectra, replacing a homogeneous triphenylsilanol ligand with a silanol on a SiO₂ surface increases cyclohexene epoxidation rates with tert-butyl hydroperoxide 20-fold per Ti site. Supporting calixarene-Ti active sites on fully hydroxylated Al ₂O₃ or TiO₂, which possess lower average surface hydroxyl pK_a than that of SiO₂, reduces catalytic rates 50-fold relative to SiO₂. These effects are consistent with SiO₂ surfaces balancing two competing factors that control epoxidation rates-equilibrated hydroperoxide binding at Ti, disfavored by stronger surface Bronsted acidity, and rate-limiting oxygen transfer from this intermediate to alkenes, favored by strongly H-bonding intermediates. These observations also imply that Ti-OSi rather than Ti-OCalix bonds are broken upon hydroperoxide binding to Ti in kinetically relevant steps, which is verified by the lack of a calixarene upper-rim substituent effect on epoxidation rate. The pronounced sensitivity of observed epoxidation rates to the support oxide, in the absence of changes to the Ti coordination environment, provides experimental evidence for the importance of outer-sphere H-bonding interactions for the exceptional epoxidation reactivity of titanium silicalite and related catalysts.

Full text of DOI:10.1021/ja074614g

Published on Web 11/22/2007
The Role of Outer-Sphere Surface Acidity in Alkene
Epoxidation Catalyzed by Calixarene-Ti(IV) Complexes
Justin M. Notestein,†,§ Andrew Solovyov,† Leandro R. Andrini,‡ Felix G. Requejo,‡
Alexander Katz,*,† and Enrique Iglesia*,†
Contribution from the Department of Chemical Engineering, UniVersity of California at
Berkeley, Berkeley, California 94720, and INIFTA e IFLP (CONICET), Facultad de Ciencias
Exactas, UniVersidad Nacional de La Plata, 1900 La Plata, Argentina
Received June 23, 2007; E-mail: askatz@berkeley.edu; iglesia@berkeley.edu
Abstract: Cooperativity between Brønsted acidic defect sites on oxide surfaces and Lewis acid catalyst
sites consisting of grafted calixarene-Ti(IV) complexes is investigated for controlling epoxidation catalysis.
Materials are synthesized that, regardless of the surface or calixarene substituent, demonstrate nearly
identical UV-visible ligand-to-metal charge-transfer bands and Ti K-edge X-ray absorption near edge
spectral features consistent with site-isolated, coordinatively unsaturated Ti(IV) atoms. Despite similar Ti
frontier orbital energies demonstrated by these spectra, replacing a homogeneous triphenylsilanol ligand
with a silanol on a SiO2 surface increases cyclohexene epoxidation rates with tert-butyl hydroperoxide
20-fold per Ti site. Supporting calixarene-Ti active sites on fully hydroxylated Al2O3 or TiO2, which possess
lower average surface hydroxyl pKa than that of SiO2, reduces catalytic rates 50-fold relative to SiO2. These
effects are consistent with SiO2 surfaces balancing two competing factors that control epoxidation rates-
equilibrated hydroperoxide binding at Ti, disfavored by stronger surface Brønsted acidity, and rate-limiting
oxygen transfer from this intermediate to alkenes, favored by strongly H-bonding intermediates. These
observations also imply that Ti-OSi rather than Ti-OCalix bonds are broken upon hydroperoxide binding
to Ti in kinetically relevant steps, which is verified by the lack of a calixarene upper-rim substituent effect
on epoxidation rate. The pronounced sensitivity of observed epoxidation rates to the support oxide, in the
absence of changes to the Ti coordination environment, provides experimental evidence for the importance
of outer-sphere H-bonding interactions for the exceptional epoxidation reactivity of titanium silicalite and
related catalysts.
Introduction
on all Ti-SiO2 materials involves concerted oxygen transfer
from an electrophilic oxygen on Ti-bound hydroperoxides to
The design and characterization of solid catalysts at the atomic
level is challenging, at least in part because of the difficulty in
deconvoluting inner- and outer-sphere effects on their catalytic
function.1,2 Inner-sphere effects arise from ligands coordinated
directly to active atoms and consequent changes to the ground-
state electronic structure of these atoms. The outer-sphere is
less well-defined for inorganic solid catalysts and consists of
noncoordinated moieties, such as vicinal Brønsted acid sites,
coadsorbed species, and the fluid phase.
the alkene substrate.4-7 In other epoxidation systems, reaction
rates have been shown to increase with increasing electrophi-
licity of stoichiometric oxidants8 and homogeneous catalysts.9
Therefore, the superior alkene epoxidation properties of Ti-
SiO2 relative to other molecular and supported metal oxides,10
have been primarily attributed to an inner-sphere effect of
coordinatively unsaturated tetrahedral Ti atoms stabilized within
an electronegative SiO2 framework. This assignment of the
active structure is consistent with theoretical calculations,11-17
Titanosilicates, such as TS-1, represent a class of materials
for which atomic-level details are only gradually emerging,
despite their broad application as epoxidation catalysts.3 The
accepted mechanism for epoxidation of unfunctionalized alkenes
(4) Sheldon, R. A.; van Doorn, J. A. J. Catal. 1973, 31, 427-437.
(5) Sheldon, R. A.; van Doorn, J. A.; Schram, C. W. A.; De Jong, A. J. J.
Catal. 1973, 31, 438-443. Khouw, C. B.; Dartt, C. B.; Labinger, J. A.;
Davis, M. E. J. Catal. 1994, 149, 195-205; Corma, A.; Garcia, H. Chem.
ReV. 2002, 102, 3837-3892.
(6) Sheldon, R. A. J. Mol. Catal. 1980, 7, 107-126.
(7) Oldroyd, R. D.; Thomas, J. M.; Maschmeyer, T.; MacFaul, P. A.; Snelgrove,
D. W.; Ingold, K. U.; Wayner, D. D. M. Angew. Chem., Int. Ed. Engl.
1996, 35, 2787-2790.
† University of California.
‡ Universidad Nacional de La Plata.
§ Current address: Department of Chemical and Biological Engineering,
Northwestern University, Evanston IL 60208.
(8) Lynch, B. M.; Pausacker, K. H. J. Chem. Soc. 1955, 1525-1531. Medvedev,
S.; Blokh, O. J. Phys. Chem. USSR 1933, 4, 721-730.
(1) Corma, A. Catal. ReV. Sci. Eng. 2004, 46, 369-417.
(2) Notestein, J. M.; Katz, A. Chem. Eur. J. 2006, 12, 3954-3965.
(3) Clerici, M. G.; Bellussi, G.; Romano, U. J. Catal. 1991, 129, 159-167.
Notari, B. AdV. Catal. 1996, 41, 253-334. Sheldon, R. A.; Wallau, M.;
Arends, I. W. C. E.; Schuchardt, U. Acc. Chem. Res. 1998, 31, 485-493.
Saxton, R. J. Top. Catal. 1999, 9, 43-57. Corma, A. J. Catal. 2003, 216,
298-312.
(9) Palucki, M.; Finney, N. S.; Pospisil, P. J.; Gu¨ler, M. L.; Ishida, T.; Jacobsen,
E. N. J. Am. Chem. Soc. 1998, 120, 948-954. Cavallo, L.; Jacobsen, H. J.
Org. Chem. 2003, 68, 6202-6207.
(10) Wulff, H. P.; Wattimena, F. (Shell Oil Company) 1978. Nemeth, L. T.;
Malloy, T. P.; Jones, R. R. (UOP) 1995.
(11) Neurock, M.; Manzer, L. E. Chem. Commun. 1996, 1133-1134.
(12) Sinclair, P. E.; Catlow, C. R. A. J. Phys. Chem. B 1999, 103, 1084-1095.
9
10.1021/ja074614g CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 15585-15595
15585

A R T I C L E S
Notestein et al.
Scheme 1. Overview of Surface-Grafted and Homogeneous
UV-visible spectroscopy studies,18,19 and X-ray absorption
spectroscopy studies.20,21 The basis for the marked differences
in epoxidation rates, optimal conditions, and preferred oxidants
between TS-1 and other titanosilicates,22 Ti-SiO2 materials,23
and homogeneous complexes,24 however, remains controver-
sial,25,26 particularly because all effective catalysts show similar
Ti inner-sphere environments by the aforementioned electronic
spectroscopies.
Calixarene-Ti Complexes
H-bonding assemblies of outer-sphere Brønsted acids have
been proposed as intermediates in epoxidation catalysis on TS-1
to explain the strong effects of protic solvents on rates.27 These
rate increases could not be explained solely on the basis of
adsorption and partitioning.26,28 The presence of Brønsted acidic
silanol nests15 or silanols generated by hydroperoxide binding,11
alone or in combination with protic solvents, has been suggested
to explain differences between the reactivity of TS-1 and other
Ti catalysts. However, experimental proof of an outer-sphere
effect in TS-1 catalysis has remained elusive, at least in part,
because of the difficulty in identifying which bond-breaking
processes form Brønsted acids in the kinetically relevant
steps,29,30 and because systematic and proven variations of the
inner- and outer-spheres are difficult to achieve in heterogeneous
catalysts without unintended changes in structural features of
uncertain importance to catalytic function.31
Here, electron-donating and electron-withdrawing calixarenes
(macrocyclic phenols 1a-4a) are synthesized and molecular
calixarene-Ti complexes are grafted covalently to oxide
surfaces (Scheme 1: SiO2, TiO2, Al2O3, and triphenylsilanol
as a model) to address the kinetic relevance of outer-sphere
Brønsted acidity on Ti-catalyzed cyclohexene epoxidation with
tert-butylhydroperoxide. The Ti inner-sphere environment re-
mains constant, as probed by ligand-to-metal charge transitions
(LMCT) in the UV-visible spectrum and by Ti 1s-3d
electronic transitions that give rise to near-edge features in the
Ti K-edge absorption spectrum. In contrast, calixarene phenols
and oxide support OH defects are chosen to span a broad pKa
range. Therefore, the sensitivity of the observed epoxidation
rate to ligand or support variations probes the kinetic relevance
of Brønsted acids formed from Ti-O-support or Ti-O-
calixarene bond cleavage.
We have previously demonstrated that calixarene-Ti com-
plexes grafted covalently on SiO2 are single-site heterogeneous
catalysts for epoxidation of alkenes with organic hydroperox-
ides.32 The single-site nature of these catalysts eliminates the
confounding effects of ensemble-averaging in structural and
functional characterization, in constrast with the nonuniform
active sites that can arise from post-synthesis modifications of
Ti centers33 or doping of support oxides.34 Grafted calixarene-
Ti catalysts are also more active than samples prepared by
removing their organic ligands by thermal treatment in air
(material 1-SiO2-c) or Ti-SiO2 samples prepared by alkoxide
grafting at similar Ti surface densities.35 The multidentate and
bulky calixarene-Ti precursors enforce site isolation during
synthesis and during reaction and organize phenolate ligands
to create Ti Lewis centers stronger than those formed
(13) Yudanov, I. V.; Gisdakis, P.; Di, Valentin, C.; Rosch, N. Eur. J. Inorg.
Chem. 1999, 12, 2135-2145.
(14) Sever, R. R.; Root, T. W. J. Phys. Chem. B 2003, 107, 4090-4099.
(15) Wells, D. H.; Delgass, W. N.; Thomson, K. T. J. Am. Chem. Soc. 2004,
126, 2956-2962. Wells, D. H.; Joshi, A. M.; Delgass, W. N.; Thomson,
K. T. J. Phys. Chem. B 2006, 110, 14627-14639.
(16) Vayssilov, G. N.; van Santen, R. A. J. Catal. 1998, 175, 170-174.
(17) Urakawa, A.; Burgi, T.; Skrabal, P.; Bangerter, F.; Baiker, A. J. Phys. Chem.
B 2005, 109, 2212-2221.
(18) Geobaldo, F.; Bordiga, S.; Zecchina, A.; Giamello, E.; Leofanti, G.; Petrini,
G. Catal. Lett. 1992, 16, 109-115. Bellussi, G.; Rigutto, M. S. Stud. Surf.
Sci. Catal. 1994, 85, 177-213. Klein, S.; Thorimbert, S.; Maier, W. F. J.
Catal. 1996, 163, 476-488. Marchese, L.; Maschmeyer, T.; Gianotti, E.;
Coluccia, S.; Thomas, J. M. J. Phys. Chem. B 1997, 101, 8836-8838. Soult,
A. S.; Poore´, D. D.; Mayo, E. I.; Stiegman, A. E. J. Phys. Chem. B 2001,
105, 2687-2693. Bonino, F.; Damin, A.; Ricchiardi, G.; Ricci, M.; Spano`,
G.; D’Aloisio, R.; Zecchina, A.; Lamberti, C.; Prestipino, C.; Bordiga, S.
J. Phys. Chem. B 2004, 108, 3573-3583. Fraile, J. M.; Garcia, J. I.;
Mayoral, J. A.; Vispe, E. J. Catal. 2005, 233, 90-99.
(19) Gao, X.; Bare, S. R.; Fierro, J. L. G.; Banares, M. A.; Wachs, I. E. J.
Phys. Chem. B 1998, 102, 5653-5666.
(20) Blanco-Brieva, G.; Capel-Sanchez, M. C.; Campos-Martin, J. M.; Fierro,
J. L. G.; Lede, E. J.; Adrini, L.; Requejo, F. G. AdV. Synth. Catal. 2003,
345, 1314-1320. Bordiga, S.; Coluccia, S.; Lamberti, C.; Marchese, L.;
Zecchina, A.; Boscherini, F.; Buffa, F.; Genoni, F.; Leofanti, G.; Petrini,
G.; Vlaic, G. J. Phys. Chem. 1994, 98, 4125-4132. Davis, R. J.; Liu, Z.;
Tabora, J. E.; Wieland, W. S. Catal. Lett. 1995, 34, 101. Fraile, J. M.;
Garc´ıa, J. I.; Mayoral, J. A.; Proietti, M. G.; Sa´nchez, M. C. J. Phys. Chem.
1996, 100, 19484-19488. Imamura, S.; Nakai, T.; Kanai, H.; Shiono, T.;
Utano, K. Catal. Lett. 1996, 39, 79-82.
(21) Thomas, J. M.; Sankar, G. Acc. Chem. Res. 2001, 34, 571-581.
(22) Blasco, T.; Camblor, M. A.; Corma, A.; Pariente, J. P. J. Am. Chem. Soc.
1993, 115, 11806-11813; Reddy, J. S.; Kumar, R.; Ratnasamy, P. Appl.
Catal. 1990, 58, L1-L4.
(23) Maschmeyer, T.; Rey, F.; Sankar, G.; Thomas, J. M. Nature 1995, 378,
159-162. Blasco, T.; Corma, A.; Navarro, M. T.; Pariente, J. P. J. Catal.
1995, 156, 65-74. Cativiela, C.; Fraile, J. M.; Garc´ıa, J. I.; Mayoral, J. A.
J. Mol. Catal. 1996, 112, 259-267. Corma, A.; D´ıaz, U.; Forne´s, V.; Jorda´,
J. L.; Domine, M.; Rey, F. Chem. Commun. 1999, 779-780. Deng, Y.;
Maier, W. F. J. Catal. 2001, 199, 115-122.
(24) Johnson, B. F. G.; Klunduk, M. C.; Martin, C. M.; Sankar, G.; Teate, S.
J.; Thomas, J. M. J. Organomet. Chem. 2000, 596, 221-225. Beck, C.;
Mallat, T.; Baiker, A. New J. Chem. 2003, 27, 1284-1289.
(25) Sastre, G.; Corma, A. Chem. Phys. Lett. 1999, 302, 447-453. Adam, W.;
Corma, A.; Garc´ıa, H.; Weichold, O. J. Catal. 2000, 196, 339-344.
(26) Clerici, M. G. Top. Catal. 2001, 15, 257-263. Clerici, M. G. Oil Gas
2006, 32, 77-82.
(27) Bellussi, G.; Carati, A.; Clerici, M. G.; Maddinelli, G.; Millini, R. J. Catal.
1992, 133, 220-230. Clerici, M. G.; Ingallina, P. J. Catal. 1993, 140, 71-
83. Corma, A.; Esteve, P.; Mart´ınez, A. J. Catal. 1996, 161, 11-19.
(28) Langhendries, G.; De Vos, D. E.; Baron, G. V.; Jacobs, P. A. J. Catal.
1999, 187, 453-463.
(29) Crocker, M.; Herold, R. H. M.; Orpen, A. G.; Overgaag, M. T. A. J. Chem.
Soc., Dalton Trans. 1999, 3791-3804.
(30) Ikeue, K.; Ikeda, S.; Watanabe, A.; Ohtani, B. Phys. Chem. Chem. Phys.
2004, 6, 2523-2528. Lin, W. Y.; Frei, H. J. Am. Chem. Soc. 2002, 124,
9292-9298.
(31) We do not discuss here the role of surface Brønsted acidity in leading to
undesired epoxide ring opening in the presence of water or other primary
alcohols. This side reaction is eliminated here by strictly anhydrous process
conditions and does not arise from cooperative interactions between the Ti
active site and surface OH defects.
(32) Notestein, J. M.; Iglesia, E.; Katz, A. J. Am. Chem. Soc. 2004, 126, 16478-
16486.
(33) Fraile, J. M.; Garc´ıa, J. I.; Mayoral, J. A.; Vispe, E. J. Catal. 2000, 189,
40-51. Fraile, J. M.; Garc´ıa, J. I.; Mayoral, J. A.; Salvatella, L.; Vispe,
E.; Brown, D. R.; Fuller, G. J. Phys. Chem. B 2003, 107, 519-526.
Brutchey, R. L.; Ruddy, D. A.; Anderson, L. K.; Tilley, T. D. Langmuir
2005, 21, 9576-9583.
(34) Oldroyd, R. D.; Sankar, G.; Thomas, J. M.; O¨ zkaya, D. J. Phys. Chem. B
1998, 102, 1849-1855. Thomas, J. M.; Sankar, G.; Klunduk, M. C.;
Attfield, M. P.; Maschmeyer, T.; Johnson, B. F. G.; Bell, R. G. J. Phys.
Chem. B 1999, 103, 8809-8813.
(35) See Supporting Information for further discussion.
9
15586 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Alkene Epoxidation Catalyzed by Calizarene-Ti(IV)
A R T I C L E S
from alkoxy precursors.36,37 We have recently shown that the
calixarene macrocycle controls the oxygen coordination at the
grafted Ti atom without perturbing the isolated nature of active
Ti atoms.38
of four Ti neighbors at 3.04 Å, while the intensity of A2 and A3 peaks
is assigned to the transitions due to perturbation by the second set of
Ti neighbors at 3.78 Å.42 The Ti-3d atomic orbitals influence each other
via hybridization with oxygen orbitals located between them. Transition
B has a predominantly Ti-4p character, but is hybridized with the Ti-
4s and/or O-2p orbitals.43
Experimental Methods
The relative area of the combined A2 and A3 features is reported
here as a measure of the availability of unoccupied electronic states at
the Ti 3d level, which decreases with the Ti average coordination
number. In particular, it is greatest for isolated tetrahedral Ti centers.
This area has been shown to be a useful descriptor of epoxidation
turnover rates for a wide range of electronic environments in Ti-SiO2
materials.38 Spectra were normalized by the average X-ray absorption
intensity of the sample between 5050 and 5200 eV. Detailed procedures
and analyses have been reported elsewhere.38 All samples were treated
in situ at 383 K in a vacuum of ∼10-6 Torr during X-ray absorption
experiments.
General Analytical. The mass loss of all grafted calixarene materials
was determined by thermogravimetry (TGA; TA Instruments 2950) in
a flow of dry synthetic air (2.0 cm3 s-1 3:1 N2:O2 from vaporized liquid
O2 and N2 at a heating rate of 0.083 K s-1) after correction for hydroxyl
condensation losses (determined from mass losses in unmodified
supports). Calixarene contents were estimated by assuming that mass
losses reflect combustion of a molecular weight of 612 g/mol site
(C45H55O) in material 1 and were adjusted accordingly for other upper-
rim substituted grafted calixarenes. Ti inductively coupled plasma
emission mass spectrometry (ICP/MS) was performed by Quantitative
Technologies, Inc.
Calixarene Synthesis Protocols. The syntheses of calixarene ligands
and calixarene-Ti complexes (Scheme 2) were conducted using
standard Schlenk-line conditions. The solvents and reagents used were
the highest available purity (Aldrich); they were dried and distilled
using standard methods before use.44 1a and 2a were synthesized from
iodomethane and tert-butylcalix[4]arene 1 and calix[4]arene 2, respec-
tively.45 Calix[4]arene 2 was synthesized from 1 via reaction with
phenol in the presence of AlCl3 in toluene.46 New compounds 3a and
4a were synthesized from 2a via modified Friedel-Crafts procedures
described in the Supporting Information. These procedures led to
selective substitution at the para position of free phenols and isolation
of the desired “cone” conformation shown in Scheme 2. This “cone”
conformation was confirmed from its methylene 13C NMR resonances
at ∼31.5 ppm.47
UV-visible spectra were measured at ambient conditions using a
Cary 400 Bio UV-visible spectrophotometer (Varian) equipped with
a Harrick Praying Mantis diffuse reflectance accessory. Pressed poly-
(tetrafluoroethylene) powders were used as perfect reflector standard
in calculating Kubelka-Munk pseudoabsorbances. The edge energies
were calculated by assuming that the lowest energy transitions
(calixarene-Ti LMCT) were indirect.39 Tangent lines as shown in
Figure 2 were calculated from the first derivatives of the transformed
1
sample spectra. Solid-state H magic angle spinning (MAS) and 13C
cross-polarized magic angle spinning (CP/MAS) NMR spectra were
collected using a Bruker DSX500 spectrometer operating at 500 MHz
at the Caltech Solid-State NMR facility. Solution NMR spectra were
measured using a Bruker AMX400 spectrometer operating at 400 MHz
at the UC-Berkeley NMR facility.
Ti Complex and Grafted Materials Synthesis. Complex 1d
(Scheme 1) was synthesized from calixarene 1a and bis(tetrahydrofu-
ran)-TiCl4, as reported previously.48 Complexes 1b,49 2b, 3b, and 4b
(Scheme 2) were synthesized by adding 1 equiv Ti(OiPr)4 (Aldrich
99.999%) to a 0.1 M solution of 1a-4a in toluene and then stirring
under a N2 atmosphere for 48 h at ambient temperature. Complexes
1b and 2b were isolated by evaporating solvents to form red glassy
and dull orange solids, respectively. Complexes 3b and 4b could not
be isolated successfully as analytically pure species because of their
low solubility. Complexes 1c and 2c were prepared by adding 0.1 M
solutions of 1b or 2b in toluene to 1 equiv triphenylsilanol and refluxing
24 h in N2 flow. The evaporation of volatile species at ambient
temperature produced analytically pure 1c as an orange solid. Analyti-
cally pure yellow powders of 2c were obtained after evaporation of
the synthesis solution and washing with anhydrous hexane. Detailed
characterization data for these complexes are provided in the Supporting
Information.
Mesoporous SiO2 (Selecto Scientific) consists of 250-500 µm
diameter aggregates with ∼6.0 nm average diameter pores, possesses
a N2 BET surface area of 498 m2/g-1, and produces pH ∼7 as a 5%
suspension in H2O. Hombikat UV100 TiO2 (Sachtleben) consists of
12 nm anatase crystallites embedded within 50 nm primary particles,
which further form ∼15 µm agglomerates possessing only macropores
and a N2 BET surface area of 343 m2 g-1. Al2O3 (Brockman I; CAMAG
Ti K-Edge X-ray Absorption Near Edge Spectroscopy (XANES).
Ti K-edge XANES were measured using a Si(111) single-crystal
monochromator and a multielement fluorescence detector held within
a sample holder with pressure and temperature control (incident beam
energy of 1.37 MeV, < 1% harmonics; 0.8 eV resolution; Laboratorio
Nacional de Luz Sincrotro´n; D04B-XAS beamline; LNLS, Campinas,
Brazil). Photon energies were calibrated in transmission mode using a
Ti foil by setting the first inflection point in its spectrum at its known
absorption edge (4966 eV).40 Background contributions were removed
by fitting the region below 4960 eV to a fourth-order polynomial
(Victoreen). Pre-edge features with maxima at 4968-4970 eV were
fitted using four Gaussian peaks (A1, A2, A3, and B in order of
increasing energy).
The four pre-peaks, assigned as A1, A2, A3, and B, correspond to
transitions from the 1s core electron to Ti 3d-4p-4s hybridized states.
The A1 pre-peak arises from a t2g bandlike state, whereas A2 and A3
features reflect eg bandlike states. The t2g and eg states represent the
components of the crystal-field split Ti-3d atomic orbitals, which are
shifted in energy relative to each other as a result of the oxygen
environment surrounding the Ti center.41 In the particular case of anatase
TiO2, the A1 feature mainly arises from perturbations by the first shell
(36) Nielson, A. J.; Schwerdtfeger, P.; Waters, J. M. Dalton Trans. 2000, 529-
537.
(37) Fantacci, S.; Sgamellotti, A.; Re, N.; Floriani, C. Inorg. Chem. 2001, 40,
1544-1549.
(42) Wu, Z. Y.; Ouvrard, G.; Gressier, P.; Natoli, C. R. Phys. ReV. B 1997, 55,
10382-10391.
(38) Notestein, J. M.; Andrini, L. R.; Kalchenko, V. I.; Requejo, F. G.; Katz,
A.; Iglesia, E. J. Am. Chem. Soc. 2007, 129, 1122-1131.
(39) Delgass, W. N.; Haller, G. L.; Kellerman, R.; Lunsford, J. H. Spectroscopy
in Heterogeneous Catalysis; Academic Press: New York, 1979. Barton,
D. G.; Shtein, M.; Wilson, R. D.; Soled, S. L.; Iglesia, E. J. Phys. Chem.
B 1999, 103, 630-640. Argyle, M. D.; Chen, K. D.; Resini, C.; Krebs, C.;
Bell, A. T.; Iglesia, E. J. Phys. Chem. B 2004, 108, 2345-2353.
(40) Waychunas, G. A. Am. Miner. 1987, 72, 89-101. Farges, F.; Brown, G.
E.; Rehr, J. J. Phys. ReV. B 1997, 56, 1809-1819.
(43) Grunes, L. A. Phys. ReV. B 1983, 27, 2111-2131.
(44) Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory Chemicals;
4th ed.; Butterworth-Heinemann: Boston, MA, 1996.
(45) Groenen, L. C.; Rue¨l, B. H. M.; Casnati, A.; Verboom, W.; Pochini, A.;
Ungaro, R.; Reinhoudt, D. N. Tetrahedron 1991, 47, 8379-8384.
(46) Rathore, R.; Abdelwahed, S. H.; Guzei, I. A. J. Am. Chem. Soc. 2004,
126, 13582-13583.
(47) Jaime, C.; Demendoza, J.; Prados, P.; Nieto, P. M.; Sanchez, C. J. Org.
Chem. 1991, 56, 3372-3376.
(41) Orgel, L. An Introduction to Transition-Metal Chemistry. Ligand-field
theory; Methuen & Co. LTD.: London, 1964. Cotton, F. A. Chemical
Applications of Group Theory; John Wiley and Sons, Inc.: New York,
1967.
(48) Zanotti-Gerosa, A.; Solari, E.; Giannini, L.; Floriani, C.; Re, N.; Chiesi-
Villa, A.; Rizzoli, C. Inorg. Chim. Acta 1998, 270, 298-311.
(49) Friedrich, A.; Radius, U. Eur. J. Inorg. Chem. 2004, 4300-4316.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15587

A R T I C L E S
Notestein et al.
Scheme 2. Synthesis of Grafted Calixarene-Ti Materials
(i) 1.2 equiv CsF and 12 equiv iodomethane, anhydrous DMF, 333 K, 18 h; (ii) 4.8 equiv phenol and 5.3 equiv AlCl3, toluene, 1 h room temp; (iii) 4.5
equiv AlCl3 and 4 equiv benzyl chloride, toluene/nitrobenzene 10:1, 273 K, 0.75 h; (iv) 4.5 equiv KNO3 and 6.8 equiv AlCl3, acetonitrile, 273 K, 2 h; (v)
1 equiv TiOPr4, toluene, 48 h room temp; (vi) 1 equiv Ph3SiOH (1c or 2c) or oxide surface, toluene, 24 h reflux. Compounds 3b and 4b were used directly
in the synthesis of surface-grafted materials without intervening isolation.
Scientific) is a pseudo-boehmite phase material treated to 773 K during
manufacture to remove 30% of its constitutional water resulting in a
material with a N2 BET surface area of 155 m2 g-1 and pH ∼4.5 in a
5% suspension in H2O. SiO2 was treated in dynamic vacuum at 773 K
for 24 h before grafting calixarene-Ti complexes. TiO2 and Al2O3 were
treated in dynamic vacuum at 393 K for 4 h before grafting calixarene-
Ti species.
Surface-grafted complexes were synthesized by treating 0.1 M
solutions or suspensions of 1b-4b in toluene with the corresponding
oxide support (0.25 mmol 1b-4b per gram oxide) in sufficient
additional toluene to fully suspend particles during magnetic stirring.
Complexes 1b-4b were used as synthesized without intervening
purification steps. After adding 1b-4b, suspensions were refluxed for
24 h, and the reactor was held at 388 K as N2 was bubbled through the
suspension until all volatiles were removed. The dry solids were washed
with boiling toluene until calixarenes were no longer visible in the
effluent and treated in dynamic vacuum for 4 h at ambient temperature.
Material 1-TiO2 was protected from ambient light during synthesis and
storage; all other materials were stored without specific precautions
until use. Material 1-SiO2-c was prepared by treating 1-SiO2 at 823 K
for 1 h in flowing N2/O2 (2.0 cm3 s-1 3:1 N2:O2) to remove all organic
ligands by combustion. The UV-visible spectrum of catalyst 1-SiO2-c
is similar to that of isolated Ti-oxo species on SiO2 prepared by other
means.19
sieves (previously dehydrated at 573 K under dynamic vacuum, 12 h)
in a 25 mL sidearm flask and heating under dynamic vacuum for 1 h
to the prescribed temperature indicated in the text (393-523 K). The
flask was then backfilled with Ar and 20 mL of anhydrous n-octane
(Aldrich, 99.8%, passed over SiO2 and Al2O3 columns, freshly distilled
off CaH2) and 0.3 mL of cyclohexene (Aldrich, 99%, passed over Al2O3
column, freshly distilled off CaH2) were added, and the reactor was
sealed, magnetically stirred, and heated to 333 K. A solution of tert-
butyl hydroperoxide (TBHP) in nonane (0.125 mL, 4.8 M solution in
nonane with added 4A sieves, Aldrich) was added, and liquid aliquots
were periodically removed and analyzed for reactants and products by
gas chromatography (Agilent 6890, HP-1 methylsilicone capillary
column) to measure catalytic reaction rates and selectivities. Solvents
and reagents were kept rigorously dry to eliminate ring opening of
epoxides catalyzed by Brønsted or Lewis acids. Neither diols nor other
cyclohexene oxidation products were detected on SiO2- and TiO2-
supported catalysts; less than one turnover to cyclohexane diol after
24 h was detected in reactions catalyzed by 1-Al2O3. TBHP was used
as the oxidant instead of the more reactive cumene hydroperoxide so
as to minimize hydroperoxide decomposition (to acetone and phenol)
in the presence of Brønsted acids.50 Homogeneous catalysts were used
at a concentration of ∼0.2 mM, which led to the same total Ti atom
(50) Bewley, T.; Bowen, B. E. V.; Bramwyche, P. L.; Jackson, G. W. Hercules
Powder Company: U.S.A., 1955. Wirth, M. M.; BP Chemicals Limited:
U.S.A., 1979. Knifton, J. F.; Sanderson, J. R. Appl. Catal. A 1997, 161,
199-211.
Catalysis. Cyclohexene epoxidation rates and selectivities were
measured by combining ∼30 mg of catalyst and ∼200 mg 4A molecular
9
15588 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Alkene Epoxidation Catalyzed by Calizarene-Ti(IV)
A R T I C L E S
Table 1. Physicochemical and Spectroscopic Data for Synthesized Materials and Complexes
site density
Ti K-edge XANES
pre-edged
2
b
materiala
(nm-
)
LMCT edgec
(eV)
R2
X
Ti
calix
(eV)
(AA2 + AA3)/AT
1b
1c
tBu
tBu
tBu
-OiPr
-OSiPh3
SiO2
SiO2
TiO2
Al2O3
SiO2
SiO2
n/a
n/a
n/a
n/a
0.24
2.34e
2.37e
2.21
3.82f
2.18
2.18
2.31
2.26
2.26
4968.8
4968.7
4968.9
n/a
0.77
0.83
0.78
n/a
1-SiO2
1-SiO2-c
1-TiO2
1-Al2O3
2-SiO2
3-SiO2
4-SiO2
0.24
0.24
n/a
0.17
0.26
0.19
0.18
tBu
0.19
0.18g
0.28
0.17
0.22g
tBu
4969.2
0.68
H
(CO)Ph
NO2
4968.6
4968.9
0.78
0.75
SiO2
a Labeled as per Scheme 1. b Determined by Ti ICP/MS or by TGA. c ( 0.02 eV. d Energy of pre-edge maximum ( 0.2 eV, component peak area ( 0.07.
e Spectra acquired in chloroform solution. All materials spectra were collected in diffuse reflectance mode. f LMCT characteristic of Ti-O-Si. g Assumes
additional combustion of four C3H7O surface groups per calixarene fragment; other materials assume additional combustion of 1 C3H7O surface group per
calixarene fragment.
concentration in the reactor as for the heterogeneous reactions.
Homogeneous catalysts were removed from the reaction solution before
injecting liquid aliquots into a gas chromatograph by passing over a
dimethylaminopropyl-modified SiO2 column (∼0.5 g, Aldrich), which
adsorbed alkyl hydroperoxides and calixarene-Ti species without
altering epoxide concentrations. As discussed elsewhere,32 epoxidation
rates on calixarene-Ti materials depend on reactant and catalyst
concentrations as described by
∂[epoxide]
rate )
) k[Ti][alkene][TBHP]
(1)
∂t
in which [Ti] is the total Ti concentration in the reactor, and all
concentrations are in moles L-1, resulting in a rate constant k with
units of M-2 s-1. This rate expression was previously shown to
accurately describe reaction rates in the range of reactant concentrations
used here. The rate equation and the rate constants were independent
of the surface density of Ti on SiO2, consistent with uniform single-
site catalytic structures.32
Figure 1. Solid-state 13C CP/MAS NMR spectra of 1-SiO2, 1-TiO2, and
1-Al2O3 collected at a spinning rate of 14 kHz. Inset shows 10×
magnification of region around ipso carbon resonance at ∼161 ppm. Curves
shown result from 20-point finite Fourier transform smoothing of the
experimental data to remove high frequency noise.
Results and Discussion
Spectroscopic Consequences of Support. Calixarene surface
densities (from TGA) and Ti surface densities (from ICP/MS)
are given in Table 1 for all materials. A ∼1:1 Ti:calixarene
molar ratio was measured for all grafted calixarenes and oxide
supports, except 1-TiO2, for which Ti content cannot confirm
the concentration of grafted complexes. These surface densities
were achieved upon contacting supports with excess calixarene-
Ti complexes; similar surface densities on SiO2, TiO2, and Al2O3
indicate that surface concentrations are predominantly limited
by the calixarene footprint and not by the number or type of
surface defect sites. Surface densities were measured after
extensive extraction in hot toluene and therefore represent
covalent grafting on SiO2, TiO2, and Al2O3.
Solid-state 13C CP/MAS NMR spectra of grafted materials
1-SiO2, 1-Al2O3, and 1-TiO2 (Figure 1) exhibit sharp resonances
at 23 ppm and 63-70 ppm, consistent with surface Si,51 Ti,51
and Al alkoxides52 formed from the titanium isopropoxide
precursor. Resonances at 30-35 ppm and 120-145 ppm are
consistent with the spectrum of calixarene 1a in solution.
Homogeneous complex 1c, 1-SiO2, 1-TiO2, and 1-Al2O3 all
possess a resonance at ∼161 ppm due to the calixarene ipso
carbon (adjacent to the phenol oxygen), which is shifted
significantly from the corresponding resonance in the free
calixarene ligand 1a (152 ppm), indicative of calixarene-Ti
coordination in all samples. In complexes with additional
electron-withdrawing ligands, this resonance is shifted further
downfield, as in tert-butylphenol-TiCl3 (169 ppm36) and calix-
arene-TiCl 1d (163 ppm49). The calixarene-Ti LMCT edge
energies of 1-SiO2, 1-Al2O3, and 1-TiO2, obtained from diffuse-
reflectance UV-vis spectroscopy (Figure 2) are all at ∼2.20 eV,
irrespective of the identity of the support. Calixarene-TiOPr
1b and homogeneous calixarene-TiOSiPh3 1c also give similar
LMCT energies (2.34 and 2.37 eV, respectively, Figure S1).
These latter complexes have slightly higher energies because
charge separation in the excited state is less favorable in low
dielectric solvents (e.g., chloroform) than in high dielectric
environments typical of hydroxylated oxide surfaces.53 The edge
energy of calixarene-TiCl 1d is lower at 1.9 eV.32 The position
of the pre-edge maximum and relative areas of the pre-peaks
in the pre-edge region of Ti K-edge (Figure 3, Table 1) for 1c
and 1-SiO2 are the same within the accuracy of the spectra and
their analysis. The A2 + A3 area (see Figure 3) in 1-Al2O3 is
(51) Dire, S.; Babonneau, F. J. Non-Cryst. Solids 1994, 167, 29-36.
(52) Abraham, A.; Prins, R.; van Bokhoven, J. A.; van Eck, E. R. H.; Kentgens,
A. P. M. J. Phys. Chem. B 2006, 110, 6553-6560.
(53) Lever, A. B. P. Inorganic Electronic Spectroscopy; Elsevier: New York,
1984. Shepherd, R. E.; Hoq, M. F.; Hoblack, N.; Johnson, C. R. Inorg.
Chem. 1984, 23, 3249-3252.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15589

A R T I C L E S
Notestein et al.
Scheme 3. Steps of Ti-Catalyzed Epoxidation of Alkenes with
Alkyl Hydroperoxide, as Described in Text
Figure 2. Diffuse-reflectance UV-visible spectra of supported complexes
1-SiO2, 1-Al2O3, and 1-TiO2. All spectra use the Kubelka-Munk formalism
and assume the LMCT transition to be indirect. Edge energies are defined
as the intercept of the indicated tangent lines, calculated from the first
derivative of the transformed spectra.
with our assertion that 1-Al2O3, 1-TiO2, and 1-SiO2 all consist
of single-site, isolated Ti centers with inner-sphere coordination
environments similar to those in soluble compound 1c. In the
absence of outer-sphere effects, which are largely undetectable
by these techniques, these four Ti species should give similar
epoxidation turnover rates on the basis of previously established
quantitative correlations between catalytic reactivity and Ti
electronic structure.38
Catalytic Consequences of Altering the Support. Alkene
epoxidation with hydroperoxides involves equilibrated coordina-
tion of hydroperoxides to Ti centers, electrophilic attack by
alkenes at O1 oxygens in bound hydroperoxides, and desorption
of alcohol coproducts (steps 2, 3, and 4 in Scheme 3).6,7 The
higher rates measured for electron-rich alkenes4,7,32 and the
theoretical estimates of activation barriers for cluster models14,15
indicate that epoxidation events occur either as the rate-
determining step or before it. Thus, the rate constant k, (eq 1)
used here to report epoxidation reactivity, is given by k ) k3K2.
These steps have been proposed as a common mechanism on
Ti sites in titanosilicates (e.g., TS-1), surface grafted Ti species,
and homogeneous silsesquioxanes, which share in common the
required isolated Ti centers with 4-fold coordination (discussed
above).
The [Ti] nomenclature used here is intended to represent the
Ti atom and all ligands that remain connected to Ti during
kinetically relevant steps; these “static” ligands can be the oxide
framework, as in TS-1, an organic ligand, or both. The
intermediates are drawn to reflect the expansion of the Ti
coordination sphere required for epoxidation.21 The structure
for A used in Scheme 3 was proposed by Sinclair and Catlow12
on the basis of a comparison of density functional theory
estimates and measured epoxidation rates for different solvents
and additives. The identity of ligand X participating in k3, and
therefore the source of the acidic protons involved in stabilizing
intermediate A, remains unclear.11-16 Moreover, the group to
which the proton is transferred from the hydroperoxide in K2 is
often written as distinct from the species providing the proton-
stabilizing intermediate A. Scheme 3 allows, but does not
require, that these groups be indistinguishable. In this section,
Figure 3. Pre-edge region of Ti K-edge XANES for 1c and supported
complexes 1-SiO2 and 1-Al2O3. The spectrum of anatase TiO2, a support
material and a representative six-coordinate Ti material, is also shown for
comparison. Intensities are normalized with respect to that at 5050-5200
eV. The A2 and A3 feature intensities are a measure of the availability of
unoccupied electronic states at the Ti 3d level. For a detailed description
about the origin of the four different features, see experimental section and
previous reference.38
smaller than in 1c and 1-SiO2 by ∼0.1 units and appears at
higher energies (∼0.4 eV), but this difference is significantly
smaller than would indicate an increase in average coordination
number from 4 to 5.38,40 Ti K-edge XANES data were not
specific to grafted species in 1-TiO2 because of the overlapping
features and strong contributions from ubiquitous Ti in the
support.
The 13C CP/MAS NMR, UV-visible, and XANES Ti K-edge
spectra for all materials discussed thus far are similar and quite
distinct from those for complexes with deliberate modifications
to the Ti inner-sphere. These spectra are therefore consistent
9
15590 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Alkene Epoxidation Catalyzed by Calizarene-Ti(IV)
A R T I C L E S
Figure 4. Catalytic turnovers as a function of time for (A) grafted 1-SiO2 pretreated at 393 K (k ) 6.9 M-2 s-1, O) as compared to homogeneous complexes
1b (k ) 0.4 M-2 s-1, +) and 1c (k ) 0.4 M-2 s-1, X), and for (B) grafted 1-SiO2 (k ) 6.0 M-2 s-1, O), 1-TiO2 (k ) 0.2 M-2 s-1, X), and 1-Al2O3 (k )
0.1 M-2 s-1, x) pretreated at 473 K. Catalytic turnovers are expressed as mol cyclohexene epoxide per mol Ti per s. Rate constants for the homogeneous
complexes were calculated for initial four turnovers.
epoxidation rates in soluble calixarene-TiOPr 1b and calix-
arene-TiOSiPh3 1c complexes are compared to those on grafted
1-SiO2, 1-Al2O3, and 1-TiO2 to probe the catalytic consequences
of hydroxylated supports acting as the source of outer-sphere
Brønsted acidity for Ti centers that are otherwise electronically
similar.
Figure 4a shows that epoxide turnover rates on homogeneous
species 1b and 1c are similar to each other but ∼20 times lower
than on 1-SiO2 treated at 393 K.54 Complexes 1b and 1c also
deactivate with time, while 1-SiO2 is stable (Figure S3). As
previously reported for incompletely condensed Ti-silsesqui-
oxanes,29 these data indicate that Ti-OSi connectivity is, by
itself, insufficient for epoxidation catalysis, even though triph-
enylsilanol ligands can provide steric bulk and electron-
withdrawing properties comparable to those for monodentate
55
silsesquioxane ligands or isolated silanols on SiO2 and have
been previously used as ligands in Ti epoxidation catalysts.34,56
Figure 4b shows that initial epoxide turnover rates on 1-TiO2
and 1-Al2O3 are ∼50 times lower than on 1-SiO2 samples treated
Figure 5. The epoxidation rate constant (pretreated at 393 K, epoxidation
of cyclohexene with TBHP, assuming rate eq 1) is plotted against XANES
similarly. The integrated hydroperoxide selectivity (mol epoxide
produced per mol hydroperoxide consumed) is ∼35% at ∼25%
hydroperoxide conversion on 1-TiO2, which is similar to
selectivities measured on 1-SiO2 for nonactivated alkenes (e.g.,
n-octene).32 At high conversions, selectivities on 1-SiO2 are near
100%, indicating that the low apparent selectivities on other
samples reflect competing noncatalytic pathways that become
apparent for catalysts with low epoxidation reactivity. This
noncatalytic hydroperoxide consumption is not significant during
the initial stages of reaction, where the reported epoxidation
rate constants are measured. Products of other cyclohexane
pre-edge component peaks relative area (AA2 + AA3)/AT (treated before
analysis at 383 K at 10-6 Torr, 1 h), for materials 1c (X), 1-SiO2 (O),
1-SiO2-c (b), 1-Al2O3 (x), 3-SiO2 (4), and 4-SiO2 ()). The dashed line
marks a correlation developed previously38 that spans all tested 4- and
5-coordinate Ti-SiO2 materials between the two extremes shown (9). This
correlation fails to account for the low reactivity of soluble 1c and 1-Al2O3,
consistent with outer-sphere effects.
oxidation pathways (cyclohexenol and cyclohexenone) were not
detected in any of the experiments.
Figure 5 extends our previously reported rate constants for
alkene epoxidation as a function of the area of XANES pre-
edge features38 to include those measured on 1-Al2O3 and 1c.
Turnover rates on these catalysts are ∼10 times lower than
predicted from Ti centers with the electronic configuration
represented by their near-edge X-ray absorption spectra, on the
basis of our previous correlations (Figure 5).38 The similar solid-
state NMR and UV spectra for 1-TiO2 and 1-Al2O3 indicate
that 1-TiO2 also deviates from the trends shown by the other
(54) Calixarene-Ti(IV) complexes have been used elsewhere for epoxidation
of alkenes (see reference 48): Massa, A.; D’ Ambrosi, A.; Pronto, A.;
Scettri, A. Tetrahedron Lett. 2001, 42, 1995-1998.
(55) Duchateau, R.; Cremer, U.; Harmsen, R. J.; Mohamud, S. I.; Abbenhuis,
H. C. L.; van Santen, R. A.; Meetsma, A.; Thiele, S. K. H.; van Tol, M. F.
H.; Kranenburg, M. Organometallics 1999, 18, 5447-5459.
(56) Attfield, M. P.; Sankar, G.; Thomas, J. M. Catal. Lett. 2000, 70, 155-
158. Oldroyd, R. D.; Thomas, J. M.; Sankar, G. Chem. Commun. 1997,
2025-2026.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15591

A R T I C L E S
Notestein et al.
catalysts in Figure 5. Thus, we conclude that the low turnover
rates measured on 1-TiO2 and 1-Al2O3 do not reflect inner-
sphere effects on Ti centers caused by these supports acting as
monodentate ligands, but instead arise from outer-sphere effects
that do not cause detectable changes in the electronic state of
Ti centers.
and by fluorous alcohols.63 Higher rates of peroxide heterolytic
cleavage on Fe heme-type catalysts correlate with decreasing
pKa of alcohol co-solvents.64,65 Outer-sphere effects caused by
protons also lead to an increase in epoxidation rates when using
peroxyacid oxidants,8 heme enzymes, such as the cytochrome
P450 family,66 and bioinspired non-heme iron catalysts.67 The
“pull” effect of vicinal acid centers in heme enzymes has also
been demonstrated in “hangman” metalloporphyrins.68 Similar
effects have been implicated in the accelerating effects of
increasingly acidic alcohols in the epoxidation of alkenes with
aqueous H2O2 on TS-1.26,69 On TS-1, reactant solubility also
correlates with the pKa of alcohol cosolvents,28 but these effects
are too weak to account for the strong effects of alcohols on
epoxidation rates.26
In view of these precedents, the higher epoxidation rates on
1-SiO2, relative to Ti sites with similar inner-sphere coordination
environments in solution or grafted on TiO2 or Al2O3, are
attributed to cooperative interactions between Ti Lewis acid sites
and weakly acidic protons provided by the SiO2 support. SiO2
surfaces, with a high density of weakly acidic protons, provide
adsorbed species with an environment resembling that in high-
dielectric solvents with a pKa of ∼7.70 Cooperative H-bonding
intermediates between organic and inorganic active sites and
surface silanols are easily formed.71 In contrast, the outer-sphere
environment in homogeneous 1c is imposed by octane, a low
dielectric aprotic solvent, as demonstrated by the higher energy
of the LMCT band (discussed above). Exposure of 1c to TBHP
at concentrations used in our catalytic experiments led to the
quantitative formation of free triphenylsilanol ligands (by
solution 1H NMR), indicating that large values of K2 are favored
by the weak acidity of the triphenylsilanol ligand (∼2 pKa units
higher than for isolated silanols on SiO255). However, the
resulting silanol concentration is many orders of magnitude
lower than the concentrations for which alcohols were found
previously to alter the rate of heterolytic cleavage of metal-
alkylhydroperoxides.64 The overall epoxidation rate for homo-
geneous 1c is therefore determined by a low value of k3. We
have previously observed a correlation between decreasing
cyclohexene epoxidation rates for 1-SiO2 and increasing tem-
perature of catalyst pretreatment,38 which would decrease silanol
densities and the ability to form intermediate A.
Except in the patent literature,10 there have been few
systematic studies of isolated Ti sites supported on oxides other
than SiO2,57 such as Al2O3 or TiO2, which would provide
58
evidence for outer-sphere effects caused by support surfaces.
Sol-gel mixed oxides containing TiO259 are not well-suited to
explore support effects because of their complex and poorly
defined Ti environments. Strong Brønsted acids34,60 open
epoxide rings rapidly in the presence of traces of water and
lead to low epoxide yields and to deactivation via chelation of
Lewis acid active sites by the alkane diols.61 Cluster density
functional theory studies of titanosilicate structures suggest that
Ti Lewis centers with vicinal weakly acidic protons form
hydrogen-bonded structures such as intermediate A (Scheme
3), which decrease the activation barrier for the kinetically
relevant step 3 in the catalytic cycle (Scheme 3).11-15 This
decrease in activation barriers is accompanied by a concomitant
increase in [Ti]O-OR bond lengths and by a decrease in the
energy level of the σ* orbital involved in the O-O bond.13,15
These outer-sphere effects caused by protons have been
difficult to detect unambiguously in solid oxidation catalysts,
but have been more clearly established in homogeneous systems.
H-bonding has been implicated in hydroperoxide activation by
homogeneous Ti(OSiMe3)4 catalysts in high alkyl hydroperoxide
concentrations,17 by protonated analogues of Ti-salen salts,62
(57) Gianotti, E.; Bisio, C.; Marchese, L.; Guidotti, M.; Ravasio, N.; Psaro, R.;
Coluccia, S. J. Phys. Chem. C 2007, 111, 5083-5089. Guidotti, M.;
Ravasio, N.; Psaro, R.; Ferraris, G.; Moretti, G. J. Catal. 2003, 214, 242-
250. Fraile, J. M.; Garc´ıa, J. I.; Mayoral, J. A.; Vispe, E.; Brown, D. R.;
Naderi, M. Chem. Commun. 2001, 1510-1511. Fraile, J. M.; Garc´ıa, J. I.;
Mayoral, J. A.; Vispe, E. Appl. Catal. A: 2004, 276, 113-122.
(58) van Vliet, M. C. A.; Mandelli, D.; Arends, I.; Schuchardt, U.; Sheldon, R.
A. Green Chem. 2001, 3, 243-246. Rinaldi, R.; Schuchardt, U. J. Catal.
2004, 227, 109-116.
(59) Beck, C.; Mallat, T.; Burgi, T.; Baiker, A. J. Catal. 2001, 204, 428-439.
Hutter, R.; Mallat, T.; Baiker, A. J. Catal. 1995, 153, 177-189. de Farias,
R. F.; Arnold, U.; Martinez, L.; Schuchardt, U.; Jannini, M. J. D. M.;
Airoldi, C. J. Phys. Chem. Solids 2003, 64, 2385-2389.
(60) Iengo, P.; Aprile, G.; Di Serio, M.; Gazzoli, D.; Santacesaria, E. Appl.
Catal. A 1999, 178, 97-109. Gianotti, E.; Oliveira, E. C.; Coluccia, S.;
Pastore, H. O.; Marchese, L. Inorg. Chim. Acta 2003, 349, 259-264.
(61) Davies, L.; McMorn, P.; Bethell, D.; Page, P. C. B.; King, F.; Hancock, F.
E.; Hutchings, G. J. Chem. Commun. 2000, 1807-1808. Corma, A.;
Domine, M.; Gaona, J. A.; Jorda´, J. L.; Navarro, M. T.; Rey, F.; Pe´rez-
Pariente, J.; Tsuji, J.; McCulloch, B.; Nemeth, L. T. Chem. Commun. 1998,
2211-2212.
(62) Sawada, Y.; Matsumoto, K.; Kondo, S.; Watanabe, H.; Ozawa, T.; Suzuki,
K.; Saito, B.; Katsuki, T. Angew. Chem., Int. Ed. 2006, 45, 3478-3480.
(63) van Vliet, M. C. A.; Arends, I.; Sheldon, R. A. Synlett 2001, 1305-1307.
van Vliet, M. C. A.; Arends, I.; Sheldon, R. A. Synlett 2001, 248-250.
Berkessel, A.; Adrio, J. A.; Huttenhain, D.; Neudorfl, J. M. J. Am. Chem.
Soc. 2006, 128, 8421-8426. Berkessel, A.; Adrio, J. A. J. Am. Chem. Soc.
2006, 128, 13412-13420.
(64) Traylor, T. G.; Xu, F. J. Am. Chem. Soc. 1990, 112, 178-186.
(65) Stephenson, N. A.; Bell, A. T. Inorg. Chem. 2006, 45, 2758-2766. Lee,
W. A.; Yuan, L. C.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110, 4277-
4283.
(66) Sono, M.; Roach, M. P.; Coulter, E. O.; Dawson, J. H. Chem. ReV. 1996,
96, 2841-2887. Ozaki, S. I.; Roach, M. P.; Matsui, T.; Watanabe, Y. Acc.
Chem. Res. 2001, 34, 818-825.
(67) Chen, K.; Costas, M.; Kim, J.; Tipton, A. K.; Que, L. J. Am. Chem. Soc.
2002, 124, 3026-3035. Taktak, S.; Ye, W.; Herrera, A. M.; Rybak-
Akimova, E. V. Inorg. Chem. 2007, 46, 2929-2942.
Fully hydroxylated anatase TiO2 and Al2O3 (e.g., boehmite
or steam-treated γ-alumina) surfaces exhibit higher densities
of OH groups than SiO2,72 and, unlike SiO2, protonate ammonia
and piperidine, indicating higher acid strength than SiO2.73 On
Ti centers grafted on Al2O3 and TiO2 surfaces, we conclude
that intermediate A (Scheme 3) is very reactive (large k3) but
present at low concentrations (low K2) because its formation
requires the transfer of a proton from the hydroperoxide to
(70) Rouxhet, P. G.; Sempels, R. E. J. Chem. Soc., Faraday Trans. 1 1974, 70,
2021-2032. Anderson, J. H.; Lombardi, J.; Hair, M. L. J. Colloid Interface
Sci. 1975, 50, 519-524. Hair, M. L.; Hertl, W. J. Phys. Chem. 1970, 74,
91-94.
(71) Bass, J. D.; Anderson, S. L.; Katz, A. Angew. Chem., Int. Ed. 2003, 42,
5219-5222; Bass, J. D.; Solovyov, A.; Pascall, A. J.; Katz, A. J. Am. Chem.
Soc. 2006, 128, 3737-3747.
(72) Knozinger, H. In The Hydrogen Bond: Recent DeVelopments in Theory
and Experiments; Schuster, P., Zundel, G., Sandorfy, C., Eds.; American
Elsevier Pub. Co.: New York, 1976; pp 1267-1353.
(68) Chang, C.-H.; Chng, L. L.; Nocera, D. G. J. Am. Chem. Soc. 2003, 125,
1966-1976. Dempsey, J. L.; Esswein, A. J.; Manke, D. R.; Rosenthal, J.;
Soper, J. D.; Nocera, D. G. Inorg. Chem. 2005, 44, 6879-6892.
(69) Sato, T.; Dakka, J.; Sheldon, R. A. In Zeolites and Related Microporous
Materials: State of the Art 1994, Proceedings of the 10th International
Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17-22, 1994;
Weitkamp, J., Ed.; Elsevier: Amsterdam, Netherlands, 1994; Vol. 84, p
1853-1860.
(73) Moss, J. H.; Parfitt, G. D.; Wright, A. Colloid Polym. Sci. 1978, 256, 1121-
1130. Busca, G. Catal. Today 1998, 41, 191-206. Majors, P. D.; Raidy,
T. E.; Ellis, P. D. J. Am. Chem. Soc. 1986, 108, 8123-8129. Tsyganenko,
A. A.; Pozdnyakov, D. V.; Filimonov, V. N. J. Mol. Struct. 1975, 29, 299-
318. Tsyganenko, A. A.; Storozheva, E. N.; Manoilova, O. V.; Lesage, T.;
Daturi, M.; Lavalley, J. C. Catal. Lett. 2000, 70, 159-163.
9
15592 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Alkene Epoxidation Catalyzed by Calizarene-Ti(IV)
A R T I C L E S
Table 2. Tabulated Values of the Upper Rim (R2) Group Hammett
78 a
Inductive Parameter (σ*),77 and Monomeric Phenol pKa
calixareneb
ipso 13C NMR shift (ppm)c
R2
σ*
phenol pKa
(H2O)
R2
ligand (Xa)
SiO2 complex
1
2
3
4
tBu
H
(CO)Ph
NO2
151.6
150.6
155.5
161.2
161
164
167
169
-0.3
+0.5
+2.2
+4.0
10.31
9.99
∼9.0
7.18
a Data for the monomeric phenols are provided in the absence of data
on the calixarenes themselves. Data for calixarene 3 are estimated from
correlations of pKa to Hammett constants presented in the cited references.
b Labeled as per Scheme 1. c Considered to be the resonance at lowest field.
Figure 7. Correlation between observed ipso carbon 13C CP/MAS NMR
shifts and tabulated values of the Hammett constant and individual phenol
pKa.
Figure 8. Catalytic turnovers as a function of time for materials 1-SiO2 (k
) 5.8 M-2 s-1, O), 2-SiO2 (k ) 4.3 M-2 s-1, 0), 3-SiO2 (k ) 3.5 M-2s-1
,
4), and 4-SiO2 (k ) 3.7 M-2s-1, )), after pretreatment at 523 K and
Figure 6. Solid-state 13C CP/MAS NMR spectra of upper-rim modified
grafted species collected at a spinning rate of 14 kHz and a contact time of
1.0 ms. A resonance at ∼196 ppm corresponding to carbonyl species in
3-SiO2 was observed using a contact time of 2.5 ms. Signals are normalized
per g of material. Inset shows 10× (20× for 3-SiO2) magnification of region
near ipso carbon resonance and fit with a 20-point finite Fourier transform
to remove high frequency noise.
expressed as mol cyclohexene epoxide per mol Ti per s.
Spectroscopic and Catalytic Consequences of Altering
Calixarene Ligands. We next test whether the phenolic groups
on the calixarene ligand can create an outer-sphere Brønsted
acid effect on epoxidation rates similar to those we have
described here for oxide supports. The upper rim groups in
calixarene-Ti complexes 1b-4b were chosen to cover a broad
range of Hammett constants, from electron-donating (tBu) to
electron-withdrawing (NO2); these systematic modifications are
expected to lead in turn to large changes in the pKa of the
calixarene ligand. The tabulated pKa values for monomeric
phenol (Table 2) can be used to provide an upper bound for
the respective free calixarenes, because macrocycles stabilize
the conjugate base more effectively than phenol monomers.76
In lieu of a direct measure of the acidity of the protonated
form of the grafted calixarene-Ti complexes, 13C NMR spectra
(Figure 6) were used to probe the degree of ligand electron
delocalization. In particular, the position of the ipso carbon
resonance is sensitive to reactions of free phenols to form Ti-
phenolate complexes.36 In all SiO2-grafted materials, this
resonance is shifted downfield by ∼10 ppm (Table 2) from the
corresponding free calixarene in solution,32,49 consistent with
calixarene-Ti connectivity in these materials. The resonances
for the free ligand and the SiO2-grafted complexes shift further
downfield as the macrocycle aromatic π system becomes more
electron-deficient (Table 2), as expected from the changes in
surfaces more acidic than SiO2. These compensating thermo-
dynamic and kinetic factors lead to optimal rates for outer
spheres with protons of intermediate acid strengths. These effects
are consistent with intermediates resembling A in Scheme 3,
for which stronger acids would decrease K2 but concurrently
increase k3. Therefore, rates reflect the value of k3 for weak
acids, for which Ti sites tend to be saturated with bound
peroxides; as acid strength increases, peroxide binding weakens
and rates become proportional to (k3K2) values, which decrease
with increasing acid strength. Such “volcano curves” are
ubiquitous in catalysis and have been reported previously for
mechanistically distinct Pt-based74 and MoOx-based75 epoxi-
dation catalysts. However, this behavior has not, to our
knowledge, been previously considered in Ti-based epoxidation
catalysis.
(74) Pizzo, E.; Sgarbossa, P.; Scarso, A.; Michelin, R. A.; Strukul, G.
Organometallics 2006, 25, 3056-3062.
(75) Shimura, K.; Fujita, K.; Kanai, H.; Utani, K.; Imamura, S. Appl. Catal. A
2004, 274, 253-257.
(76) Araki, K.; Iwamoto, K.; Shinkai, S.; Matsuda, T. Bull. Chem. Soc. Jpn.
1990, 63, 3480-3485. Shinkai, S.; Araki, K.; Koreishi, H.; Tsubaki, T.;
Manabe, O. Chem. Lett. 1986, 1351-1354.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15593

A R T I C L E S
Notestein et al.
Scheme 4. Epoxidation Intermediate A for Grafted Calixarene-Ti(IV) Structures Differing in the Destination of the Proton Transferred from
the Hydroperoxide
the Hammett constants of the upper rim groups and from the
stronger acidity of the respective conjugate phenol (Figure 7).
These systematic changes in NMR resonances for the calixarene
aromatic π system were not accompanied by significant changes
in the energy of ligand-to-metal charge-transfer bands in their
UV-visible spectra (Table 1, Figure S4) or in the energy
position of the maximum intensity of their pre-edge features in
X-ray absorption spectra (Table 1). These findings are consistent
with previous theoretical studies,37 reported NMR spectra,79 and
electrochemical data,80 which report electronic effects on Ti as
a result of changes in phenolate geometry, but not phenolate
substituent. This property enables outer-sphere effects of calix-
arene substituents to be determined independently of concomi-
tant changes in the electronic structure of Ti centers.
Figure 8 shows that epoxide formation turnover rates on 1-,
2-, 3-, and 4-SiO2 are only weakly influenced by the calixarene
upper rim. 1-SiO2 gave slightly higher rates than 2-, 3-, and
4-SiO2, consistent with weak inner-sphere effects expected from
XANES spectra and their correlation with reactivity (Figure 5).
These small changes in epoxidation rates with >3 pKa changes
in calixarene phenol acid strength between 1-SiO2 and 4-SiO2
suggest that Ti-OCalix bonds are not cleaved in kinetically
relevant steps required for epoxidation turnovers. The prefer-
ential cleavage of Ti-OCalix bonds over Ti-OSi bonds was
previously proposed on the basis of thermodynamic arguments,
such as the strained nature and greater number of Ti-OCalix
bonds,32 and of the relative acidity of isolated phenols and
silanols. The stoichiometric replacement of alkyl hydroperoxides
for alkoxides has been reported for Ti(OiPr)4 grafted onto SiO2.81
The data reported here indicate that phenols are either not
generated during hydroperoxide binding (k5 , k2, Scheme 4)
or are generated only momentarily as minority species leading
to a kinetic dead-end (K5[TBHP] , 1 and k6 , k3, Scheme 4).
A large kinetic barrier to Ti-OC cleavage is consistent with
previously observed larger losses of Ti-OSi species compared
with Ti-OC species during reaction of molecules containing
both groups with SiO2.82 Scheme 4 summarizes our conclusions
that the “static” [Ti] core in Scheme 3 represents an intact
calixarene-Ti complex, in which Ti-OSi bonds are broken in
kinetically relevant intermediates in Ti-catalyzed epoxidation
reactions.
Mechanistic Conclusions
The synthesis and grafting of calixarene-Ti complexes onto
oxide surfaces can be used to control independently the inner-
sphere ligands at active Ti centers and the surrounding outer
sphere. In this approach, macrocyclic multidentate calixarenes
lead to Ti centers with similar inner-sphere environments,
confirmed by UV-visible spectra and pre-edge features at the
Ti K-edge. Grafting onto supports, however, places Ti centers
in environments where vicinal OH groups differ markedly in
acid strength as a result of the Brønsted acid properties of SiO2,
TiO2, and Al2O3 supports. These latter outer-sphere effects lead
to ∼50-fold changes in epoxidation turnover rates. The absence
of comparable catalytic consequences when altering the pKa of
calixarene ligands via changes in their upper rim lead us to
conclude that the single Ti-OSi bond (instead of any of the
three Ti-OCalix bonds) is cleaved in kinetically relevant
epoxidation steps on Ti-SiO2.
Epoxidation rates increased as homogeneous 1c was grafted
onto SiO2 (1-SiO2), and rates were observed previously to
increase for mild pretreatment temperatures that would retain
the high local hydroxyl surface density on SiO2. We attribute
both effects to an increase in the rate of O transfer from Ti-
hydroperoxide complexes to alkenes (k3) via cooperative H-
bonding assemblies (intermediate A) between Ti centers and
vicinal weakly acidic protons. We note that because of the high
densities of OH groups pre-existing on these SiO2 surfaces, the
SiOH generated by Ti-OSi bond breaking is not required to
be the same SiOH involved in increasing k3, but the two groups
are thermodynamically indistinguishable. These H-bonded
structures also lead to rapid and thermodynamically favored
regrafting of Ti intermediates onto the SiO2 surface following
O transfer to the alkene, consistent with the observed lack of
(77) Dean, J. A. Lange’s Handbook of Chemistry, 15th ed.; McGraw-Hill, Inc.:
New York, 1999.
(78) Rived, F.; Roses, M.; Bosch, E. Anal. Chim. Acta 1998, 374, 309-324.
(79) Harrod, J. F.; Taylor, K. R. Inorg. Chem. 1974, 14, 1541-1545.
(80) Fussing, I. M. M.; Pletcher, D.; Whitby, R. J. J. Organomet. Chem. 1994,
470, 109-117.Chen, L. X.; Rajh, T.; Wang, Z. Y.; Thurnauer, M. C. J.
Phys. Chem. B 1997, 101, 10688-10697.
(83) Corma, A.; Camblor, M. A.; Esteve, P.; Martinez, A.; Perez-Pariente, J. J.
Catal. 1994, 145, 151-158. Camblor, M. A.; Costantini, M.; Corma, A.;
Gilbert, L.; Esteve, P.; Martinez, A.; Valencia, S. Chem. Commun. 1996,
1339-1340.
(84) Imamura, S.; Nakai, T.; Utani, K.; Kanai, H. J. Catal. 1996, 161, 495-
497.
(81) Bouh, A. O.; Rice, G. L.; Scott, S. L. J. Am. Chem. Soc. 1999, 121, 7201-
7210.
(82) Jarupatrakorn, J.; Tilley, T. D. J. Am. Chem. Soc. 2002, 124, 8380-8388.
9
15594 J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007

Alkene Epoxidation Catalyzed by Calizarene-Ti(IV)
A R T I C L E S
leaching of active structures during epoxidation catalysis on
grafted calixarene-Ti complexes.32 In TS-1 and related materi-
als, silanols formed during synthesis or by hydroperoxide
binding can form the required H-bonded intermediate A (akin
to structures shown in Scheme 4); the low concentrations of
pre-existing OH defect sites and the restricted mobility of Ti
centers in these materials, however, decrease the likelihood that
these structures form. Then, protic solvents with increasing pKa
can bridge silanols and Ti active sites and create the requisite
H-bonded intermediate A and, in this manner, increase epoxi-
dation turnover rates.15,26
Epoxidation rates decrease when stronger acid sites inhibit
the formation of H-bonded intermediates (K2). The lower
epoxidation turnover rates on Al-containing Ti-silicalites may
reflect, in part, the presence of strong Brønsted acids,83 which
destabilize intermediate A (decrease K2). Similarly, the unre-
active nature of 1-TiO2 catalysts indicates that the unreactive
nature of Ti centers in anatase TiO2 and Ti-SiO2 with high Ti
contents19,84 reflects the unfavorable equilibrium dictated by low
K2 values, even though coordinatively unsaturated Ti centers
are present on surfaces of small anatase crystallites.85 We
conclude that a “volcano”-type relationship leads to optimal
epoxidation turnover rates at intermediate acid strengths of
vicinal protons, which maximize the values of the relevant
(k3‚K2) product that determines epoxidation rates.
The grafting of catalytic entities with uniform structure onto
oxide supports extends the range of outer-sphere environments
accessible with protic cosolvents and represents a novel route
to tune and to probe the function of active Ti centers. The range
of acid strengths and their high density of surface species make
oxide surfaces well-suited to promote the cooperative behavior
between active sites and acidic protons2 observed for condensa-
tion71 and hydrogenation86 reactions. The data and conclusions
reported here illustrate how outer-sphere protons positioned by
supports also influence epoxidation rates on Ti centers with
otherwise identical electronic properties. As a result, an accurate
description of the structure and function of inorganic catalysts
such as TS-1 and Ti-SiO2 requires that we include outer-sphere
effects that complement inner-sphere effects intrinsic to Ti
centers and their catalytic properties.
Acknowledgment. The authors thank Sonjong Hwang of the
Caltech Solid-State NMR facility for technical assistance and
useful discussion and Dr. Herman van Halbeek of the UC-
Berkeley facility for technical assistance. The authors acknowl-
edge the U.S. DOE Office of Basic Energy Sciences (Grant
DE-FG02-05ER15696) and ANPCYT (PICT 06-17492), Ar-
gentina; CONICET (PIP 6075), Argentina; CONICET-CNPq-
NSF collaborative research agreement (CIAM program) and
LNLS, Campinas, Brazil (project D04B-XAFS1-3492) for
financial support. J.M.N. acknowledges the National Science
Foundation for a graduate fellowship.
Supporting Information Available: 1H, 13C, and 29Si NMR
resonances, FAB-MS, and elemental composition for 1c and
2c; UV-visible spectra of homogeneous calixarene-Ti com-
plexes 1b and 1c (Figure S1); a correlation between Ti surface
densities determined by Ti ICP/MS, calixarene surface densities
determined by TGA, and 13C CP/MAS NMR intensities for
resonance corresponding to calixarene methylene bridges (Figure
S2); linearized form of epoxide production vs time for 1b, 1c,
1-SiO2, and 1-Al2O3 showing stable rate constants for grafted
catalysts (Figure S3); UV-visible spectra of upper-rim modified
calixarene-Ti materials 1-4 on SiO2 (Figure S4); extended
comparison of grafted calixarene-Ti catalyst 1-SiO2 with
sequentially assembled calixarene and Ti materials and calcined
Ti-SiO2 catalysts including synthesis (Scheme S1), UV-visible
spectroscopic comparison (Figure S5), and comparison of
catalytic turnovers to epoxide (Figure S6). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA074614G
(86) Pugin, B.; Landert, H.; Spindler, F.; Blaser, H. U. AdV. Synth. Catal. 2002,
344, 974-979. Gonzalez-Arellano, C.; Corma, A.; Iglesias, M.; Sanchez,
F. AdV. Synth. Catal. 2004, 346, 1316-1328. Goettmann, F.; Boissiere,
C.; Grosso, D.; Mereier, F.; Le Floch, P.; Sanchez, C. Chem. Eur. J. 2005,
11, 7416.
(85) Chen, L. X.; Rajh, T.; Jager, W.; Nedeljkovic, J.; Thurnauer, M. C. J.
Synchrotron Rad. 1999, 6, 445-447.
9
J. AM. CHEM. SOC. VOL. 129, NO. 50, 2007 15595