Enantioselective Ti(IV) sulfoxidation catalysts bearing C3-symmetric trialkanolamine ligands: Solution speciation by 1H NMR and ESI-MS analysis

Article abstract of DOI:10.1021/ja983569x

Sulfoxidation catalysts generated from titanium(IV) isopropoxide and enantiopure trialkanolamine ligands promote the enantioselective oxidation of aryl alkyl sulfides to the corresponding sulfoxides even at low (1-2%) catalyst loadings. Electrospray ionization mass spectrometry (ESI-MS), in combination with conventional low-temperature NMR techniques, provides a powerful tool for understanding the unique nature of these catalysts. Tetradentate ligation of the titanium atom by the trialkanolamine ligand provides a highly robust titanatrane core which is retained even in hydroxylic solvents and/or under acidic conditions. In contrast, the remaining apical coordination site is shown to be substitutionally labile. Previously ill-defined species formed when the catalyst is generated in situ with a slight excess of trialkanolamine are shown to consist of discrete 2:1, 3:2, and 4:3 oligomers in which the excess trialkanolamine bridges multiple titanatrane units. In the presence of excess tert-butyl hydroperoxide, all of the precatalyst species are cleanly converted to a mononuclear titanium(IV) peroxo complex which serves as the active sulfoxidation catalyst. Ab initio molecular orbital calculations were used to probe the structure and position of protonation of the catalytic species. Other ionization techniques (fast ion bombardment or electron impact) proved less useful than ESI-MS due to high levels of fragmentation during the ionization process.

Full text of DOI:10.1021/ja983569x

6258
J. Am. Chem. Soc. 1999, 121, 6258-6268
Enantioselective Ti(IV) Sulfoxidation Catalysts Bearing C3-Symmetric
1
Trialkanolamine Ligands: Solution Speciation by H NMR and
ESI-MS Analysis
Marcella Bonchio,† Giulia Licini,*,† Giorgio Modena,† Olga Bortolini,*,‡ Stefano Moro,§ and
William A. Nugent
Contribution from the Dipartimento di Chimica Organica, CMRO del CNR, Via Marzolo 1, and
Dipartimento di Chimica Farmaceutica, Via Marzolo 9, UniVersita` di PadoVa, I-35131 PadoVa, Italy,
Dipartimento di Chimica, Via Borsari 46, UniVersita` di Ferrara, 44100 Ferrara, Italy, and The Du Pont
Company, Central Research & DeVelopment, P.O. Box 80328, Wilmington, Delaware 19880-0328
ReceiVed October 9, 1998
Abstract: Sulfoxidation catalysts generated from titanium(IV) isopropoxide and enantiopure trialkanolamine
ligands promote the enantioselective oxidation of aryl alkyl sulfides to the corresponding sulfoxides even at
low (1-2%) catalyst loadings. Electrospray ionization mass spectrometry (ESI-MS), in combination with
conventional low-temperature NMR techniques, provides a powerful tool for understanding the unique nature
of these catalysts. Tetradentate ligation of the titanium atom by the trialkanolamine ligand provides a highly
robust titanatrane core which is retained even in hydroxylic solvents and/or under acidic conditions. In contrast,
the remaining apical coordination site is shown to be substitutionally labile. Previously ill-defined species
formed when the catalyst is generated in situ with a slight excess of trialkanolamine are shown to consist of
discrete 2:1, 3:2, and 4:3 oligomers in which the excess trialkanolamine bridges multiple titanatrane units. In
the presence of excess tert-butyl hydroperoxide, all of the precatalyst species are cleanly converted to a
mononuclear titanium(IV) peroxo complex which serves as the active sulfoxidation catalyst. Ab initio molecular
orbital calculations were used to probe the structure and position of protonation of the catalytic species. Other
ionization techniques (fast ion bombardment or electron impact) proved less useful than ESI-MS due to high
levels of fragmentation during the ionization process.
Introduction
As far as oxygen-transfer catalysis is concerned, the Sharp-
less-Katsuki epoxidation system7,8 launched the rich field of
titanium-mediated asymmetric oxidations.7-18 Since that break-
through, chiral Ti(IV) alkoxides have been used to catalyze a
variety of oxidative transformations, affording highly stereo-
selective processes in the allylic alcohol epoxidation,7,8 â-hy-
droxyamine N-oxidation,9 and sulfoxidation,10-17 as well as the
Baeyer-Villiger oxidation of cyclobutanones.18 The general
reaction protocol employs the achiral precursor titanium tetra-
Central to the field of homogeneous asymmetric catalysis is
the discovery of new transition-metal-based systems that are
able to mediate organic reactions with high efficiency and
selectivity.1-3 Toward this end, an intensive research effort has
been focused on the engineering of suitable chiral ligands4,5 and
on the study of their interactions with selected metal nuclei.
The ultimate goal is the solution-phase production of a stable,
kinetically competent, and highly stereoselective catalyst.1-5
A
(6) Consider, for example, the detailed study on osmium(VIII)-mediated
stereoselective dihydroxylation: Kolb, H. C.; VanNieuwenhze, M. S.;
Sharpless, K. B. Chem. ReV. 1994, 94, 2483.
(7) Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 5974.
(8) Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune, H.;
Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765.
(9) Miyano, S.; Lu, L. D.-L.; Viti, S. M.; Sharpless, K. B. J. Org. Chem.
1983, 48, 3608.
rational approach toward this objective requires delineation of
the key elements that are involved in such catalyst assembly,
namely, the chiral ligand or its components, the achiral metal
precursor, and any other additive eventually present in the
reaction mixture.1-6
* To whom correspondence should be addressed.
† Dipartimento di Chimica Organica, Universita` di Padova.
‡ Universita` di Ferrara.
(10) Modena, G.; Furia, F. D.; Seraglia, R. Synthesis 1984, 325.
(11) Pitchen, P.; Dunach, E.; Deshmuck, M. N.; Kagan, H. B. J. Am.
Chem. Soc. 1984, 106, 8188.
(12) Brunel, J.-M.; Kagan, H. B. Synlett 1996, 404.
(13) Komatsu, N.; Hashizume, M.; Sugita, T.; Uemura, S. J. Org. Chem.
1993, 58, 4529.
(14) Reetz, M. T.; Merk, C.; Naberfeld, G.; Rudolph, J.; Griebenow,
N.; Goddard, R. Tetrahedron Lett. 1997, 38, 5273.
(15) Watson, K. G.; Fung, Y. M.; Gredley, M.; Bird, G. J.; Jackson, W.
R.; Gountzos, H.; Matthews, B. R. J. Chem. Soc., Chem. Commun. 1990,
1018.
§ Dipartimento di Chimica Farmaceutica, Universita` di Padova.
The DuPont Company.
(1) Ojima, I. Catalytic Asymmetric Synthesis; VCH Publisher: New York,
1993.
(2) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley-
Interscience: New York, 1994.
(3) Nugent, W. A.; RajanBabu, T. V.; Burk, M. Science 1993, 259, 479-
483.
(4) Seyden-Penne, J. Chiral Auxiliaries and Ligands in Asymmetric
Synthesis; John Wiley & Sons: New York, 1995.
(5) Brunner, H.; Zettlmeier, W. Handbook of EnantioselectiVe Catalysts
with Transition Metal Compounds; VCH: Weinheim, 1993.
(16) Superchi, S.; Rosini, C. Tetrahedron: Asymmetry 1997, 8, 349.
(17) Yamanoi, Y.; Imamoto, T. J. Org. Chem. 1997, 62, 8560.
(18) Lopp, M.; Payn, A.; Kanger, T.; Pehk, T. Tetrahedron Lett. 1996,
37, 7583.
10.1021/ja983569x CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/16/1999

Solution Speciation of Ti(IV) Sulfoxidation Catalysts
J. Am. Chem. Soc., Vol. 121, No. 26, 1999 6259
isopropoxide [Ti(O-i-Pr)4] in the presence of an enantiopure C2-
symmetric diol19 and an alkyl hydroperoxide as the primary
oxidant. Under such conditions, ligand exchange occurs in the
coordination sphere of the titanium nucleus,20,21 producing,
among all the possible complexes, a chiral, highly reactive
species (ligand-accelerating effect) which is responsible for the
enantioselective outcome of the process.22
Scheme 1a
Due to the tendency of the Ti(IV) alkoxides to form mixtures
of oligomeric complexes equilibrating in solution and because
of the rapid ligand exchange equilibria,22-24 for the majority of
the processes cited above, the actual catalyst structure remains
elusive. In all cases, a chiral, metal-centered η2-peroxo species
has been postulated as the reactive intermediate.21 This latter
hypothesis has been recently substantiated by the solid-state
structure of the achiral η2 tert-butylperoxotitanatrane dimer,
which is, so far, the only reported X-ray structure of Ti(IV)
alkylperoxo complex.25 Despite the general lack of structural
information, the majority of evidence collected in the literature
for the different Ti(IV)/C2-symmetric diol systems seems to
agree about the aggregate nature of the enantioselective oxidant
formed in solution.11,13,20 As a matter of fact, in some protocols
a supplementary dose of hydroxyl functionalities in the form
of water or aliphatic alcohols is added to the reaction mixture
to promote the assembly of oligomeric complexes, which are
then much better catalysts than the precursors.11-14
a Reagents and conditions: (a) stoichiometric Ti(i-PrO)4, CHCl3, 20
°C; (b) Ti(i-PrO)4 (0.75 equiv), CH2Cl2, 20 °C, followed by solvent
removal; (c) alkyl hydroperoxide; (d) excess of 1.
Our approach employs electrospray ionization mass spec-
trometry (ESI-MS),27 combined with standard 1H NMR analysis,
to assess the nature of the various titanium species equilibrating
in solution under different reaction conditions and their evolution
to the active species originating by interaction with tert-butyl
hydroperoxide (TBHP). From the results obtained, structures
of the key Ti(IV) complexes are proposed and discussed by
comparison with experiments performed using different ioniza-
tion techniques: fast ion bombardment (FIB+ ) and electron
impact (EI) and ab initio calculations.
Following the above discussion, it is possible to highlight
the weak points affecting the Ti(IV)-based systems described
thus far:
(1) The solution integrity of the active catalyst, likely
oligomeric, is seriously jeopardized under turnover conditions
by the excess of the hydroperoxide that, acting as a bidentate
ligand toward the titanium center, can alter the nuclearity and/
or substitute the chiral ligand in the coordination sphere of the
metal.24 Consequently, only ca. 20 catalytic cycles can be
achieved without depletion of the system enantioselectivity.8,12
(2) The lack of information about the actual structure of the
reactive titanium species hampers both rational catalyst engi-
neering and detailed mechanistic studies.20,21
We have reported that a novel series of tetradentate alkoxide
ligands, namely C3-symmetric enantiopure trialkanolmines 1,
provide very stable titanium(IV) complexes.24,26 In the presence
of alkyl hydroperoxides, such species are able to mediate the
asymmetric sulfoxidation of alkyl aryl sulfides with ee’s up to
84% and with unprecedented catalytic efficiency, reaching 50-
100 turnover numbers (Scheme 1).26
Results and Discussion
Ligands 1 bind tightly to the titanium center in a tetradentate
fashion, producing complexes whose nature depends on the
stoichiometry of the reaction with Ti(O-i-Pr)4 (Scheme 1).24,26,28
When a precise 1:1 ligand/Ti(IV) ratio is employed, catalyst
2 is obtained (Scheme 1, path a), whose spectroscopic behavior
is consistent with a monomeric structure. In contrast, a complex
mixture of polynuclear Ti(IV)-based species 3 forms from the
interaction of Ti(IV) with a slight excess of ligand 1 (Scheme
1, path b). Upon removal of the solvent containing the
2-propanol released in the reaction, catalyst 3 is isolated as a
white powder and is routinely used to mediate asymmetric
sulfoxidation.26 In fact, it provides faster reaction rates than the
in situ-formed system 2, where the liberated 2-propanol behaves
as a competing ligand. Furthermore, it offers a practical
advantage in handling.
In this work we present a detailed study focused on (i) the
structural evaluation of the catalytic Ti(IV)/1 complexes, (ii)
the establishment of their solution integrity both in the absence
of the hydroperoxide and under catalytic conditions, and (iii)
the direct observation of the monomeric reactive peroxometal
complex involved in the enantioselective sulfoxidation.
The room-temperature 1H NMR spectrum of 3 shows broad
resonances, indicating the presence of a mixture of slowly
equilibrating aggregates. In a recent communication, we have
shown that the characterization of the diverse species present
in 3 can been achieved through ESI-MS and 1H NMR analysis.28
Here we report the complete study that has been performed both
on the in situ-formed species 2 and on the isolated catalyst 3.
Our results also include the investigation of such systems in
the presence of TBHP, which reacts with 2 and 3, yielding the
same monomeric Ti(IV) peroxo species 4 (Scheme 1, path c).
(19) Derivatives of tartaric acid were employed in refs 7-12 and 18,
1,1′-binaphthol and its derivatives in refs 13 and 14, and 1,2-diaryl and
di-tert-butylethane-1,2-diols in refs 15-17.
(20) Woodward, S. S.; Finn, M. G.; Sharpless, K. B. J. Am. Chem. Soc.
1991, 113, 106.
(21) Johnson, R. A.; Sharpless, K. B. In ref 1a, Chapter 4.1, pp 103-
158.
(22) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1059-1070.
(23) Bradley, D. C.; Mehrota, R. C.; Gaur, D. P. Metal Alkoxides;
Academic Press: London, 1978.
(24) Nugent, W. A.; Harlow, R. L. J. Am. Chem. Soc. 1994, 116, 6142
and references therein.
(25) Boche, G.; Mo¨bus, K.; Harms, K.; Marsh, M. J. Am. Chem. Soc.
1996, 118, 2770.
(26) Di Furia, F.; Licini, G.; Modena, G.; Motterle, R.; Nugent, W. A.
J. Org. Chem. 1996, 61, 5175.
(27) Colton, R.; D’Agostino, A.; Traeger, J. C. Mass Spectrom. ReV.
1995, 14, 79 and references therein.
(28) Bonchio, M.; Licini, G.; Modena, G.; Moro, S.; Bortolini, O.; Traldi,
P.; Nugent, W. A. Chem. Commun. 1997, 869.

6260 J. Am. Chem. Soc., Vol. 121, No. 26, 1999
Bonchio et al.
latter resonance has the same chemical shift found for the
methine of the isopropoxide ligand.
Considering the similar structure of the two competing
alkoxides (Me2CHO- and R2NCH2CHMeO-), this result
suggests that these new signals can be assigned to the methine
and methylene protons of one arm of ligand 1a that has displaced
the apical isopropoxy ligand in 2a. Such a modification of the
coordination sphere of the titanium nucleus does not influence
the local C3 symmetry of the chelated titanatrane moiety, whose
resonances remain basically unchanged. Therefore, the new
Ti(IV) species displays two molecules of 1a in its coordination
sphere with a diverse spatial arrangement. However, the
integration of the signals assigned to the different protons of
the coordinated ligands, namely the one involved in the chelated
atrane structure and the other in the apical position, does not
provide significative evidence for the straightforward assignment
of the structure of the new complex. Signals corresponding to
the nonligated trialkanolamine 1a are at δ ) 3.8-4.0 (CH),
2.37 and 2.19 (CH2), and 1.20 ppm (CH3).
On the basis of the NMR evidence reported above, eq 1 can
be put forward to describe the equilibria occurring in solution.
Likely, the polyalcoholic nature of the ligand can promote the
formation of Ti(IV)-based oligomers. As a matter of fact, a total
of 5 equiv of ligand 1a versus Ti(O-i-Pr)4 (Figure 1, spectrum
c) causes the total displacement of the apical isopropoxy ligand.
Figure 1. 1H NMR titration of in situ-formed complex 2a (2.5 × 10-2
M) by 1a in deuteriochloroform at -10 °C for Ti(IV): 1a ratios (a)
1:1, (b) 1:3, and (c) 1:5. Resonances are relative to (9) complex 2a,
(2) displaced 2-propanol, ([) ligand 1a, and (g) newly formed titanium
species (see discussion in the text and eq 1).
The collected ESI-MS spectra, together with tandem MSn
analysis and the simulation of the observed isotopic distributions,
have provided clear insight in the structural elucidation of the
Ti(IV) complexes under study. There are precedents in the
literature where ESI-MS has been applied successfully to the
identification of different polynuclear metal complexes and has
proved to be an effective tool for the investigation of rapid
exchanging systems.27,29 In this study, ESI-MS has emerged as
a diagnostic technique which provides a powerful alternative
to low-temperature NMR spectroscopy.
Ti(IV)/1a in Situ-Formed System: Catalyst 2a. System 2a
is prepared in situ by reacting stoichiometric amounts of (S,S,S)-
tri-2-propanolamine 1a with Ti(i-PrO)4. Its 1H NMR spectrum
at -10 °C30 (Figure 1, spectrum a) in deuteriochloroform shows
a single set of signals at δ ) 1.10, 2.80-2.95, and 4.80-4.98
ppm respectively for the methyl, methylene, and methine groups
of coordinated 1a. Signals corresponding to the two diastereo-
meric methyl groups and to the methine of the bonded
isopropoxy ligand can be observed at δ ) 1.31 (CH3) and 4.62
ppm (CH). According to the stoichiometry of the equilibrium,
favoring the formation of the chelated complex 2a, 3 equiv of
i-PrOH are displaced in solution [δ ) 1.20 (CH3) and 4.01 ppm
(CH)].
The 1H NMR data are consistent with a solution-phase
monomeric structure for complex 2a. Addition of increasing
amounts of ligand 1a to the system (Figure 1, spectra b and c)
causes the disappearance of the apical isopropoxy ligand
resonances and the synchronous appearance of two broader
signals at δ ) 2.7 and 4.6-4.8 ppm. It is noteworthy that the
The presence of polynuclear aggregates 3a′, 3a′′, and 3a′′′
has been unequivocally confirmed by positive ion ESI mass
spectrometry (Chart 1 and Figure 2).
The ESI-MS experiments were conducted to investigate the
nature of the titanium complexes originating when an acidic
chloroform solution31 of Ti(O-i-Pr)4 is “titrated” with progressive
additions of the chiral ligand 1a (Figure 2). The positive ion
spectrum collected for the 1:1 system shows a single intense
peak at m/z ) 724 (I) (Figure 2, spectrum a). As the molecular
ion peak corresponding to the protonated monomeric complex
2a would be expected at m/z ) 295, this observation stands in
1
contrast with the previously discussed H NMR analysis (see
Figure 1, spectrum a). Indeed, the experimental cluster ion peaks
of ion I (see Experimental Section) correspond to a trinuclear
titanium species, namely [TiN(CH2CHMeO)3]3O+. Although the
molecular composition of ion I retains the correct 1:1 stoichi-
ometry, as established between Ti(IV) and ligand 1a in the initial
experimental conditions, the lack of the isopropoxy residues
and its aggregate nature indicate that the transfer of complex
(29) Løver, T.; Henderson, W.; Bowmaker, G. A.; Seakins, J. M.;
Cooney, R. P. Inorg. Chem. 1997, 36, 3711 and references therein.
(30) 1H NMR spectra have been recorded at low temperature in order to
shift the resonance of the free hydroxy group in a region far away from the
signals of interest. At -10 °C, it was observed at 1.65 ppm (see Figure 1,
spectrum a).
(31) Stock solutions of chloroform (stabilized with silver foils) were
routinely used with no acidic additives.

Solution Speciation of Ti(IV) Sulfoxidation Catalysts
J. Am. Chem. Soc., Vol. 121, No. 26, 1999 6261
Chart 1
Figure 2. ESI-MS titration of in situ-formed complex 2a (2.5 × 10-3
M) by 1a in acidic chloroform as mobile phase for Ti(IV): 1a ratios
(a) 1:1, (b) 1:3, (c) 1:5, and (d) 1:7.
Table 1. MS2 Fragment Ions after Collision-Induced Dissociation
of Ions I (m/z ) 724), II (m/z ) 192), III (m/z ) 427), IV (m/z )
662), and V (m/z ) 897)33
principal ions,a
ion m/z (relative intensity)
identification
I
724 (10)
680 (100)
636 (6)
M+
M+ - MeCHO
M+ - 2MeCHO
II
192 (2)
[M + H]+
174 (100)
156 (13)
138 (2)
[M + H - H2O]+
[M + H - 2H2O]+
[M + H - 3H2O]+
[M + H]+
III
IV
V
427 (3)
409 (16)
294 (16)
174 (100)
662 (29)
644 (5)
489 (14)
409 (100)
897 (8)
[M + H - H2O]+
2a from solution into the gas phase occurs with decomposition
and produces a major reorganization of the metal species.32
Tandem MS2 analysis (Table 1) shows that ion I fragments
by losing one and two acetaldehyde molecules. Such a
fragmentation pattern is strong evidence for the coordination
of ligand 1a to the titanium nucleus. In fact, the MS2 spectrum
of ion II (m/z ) 192) collected for a chloroform solution of 1a
under identical experimental conditions and corresponding to
the protonated trialkanolamine 1a produces only fragments
resulting from the loss of one, two, or three molecule of water
(see Table 1). This observation suggests that the chelated
titanatrane structure retains its integrity during the electrospray
process and that oxonium ion I likely originates from the loss
of the more labile apical isopropoxy ligand. Following this
decomposition process, yielding the highly reactive cationic
species [TiN(CH2CHMeO)3]+ (vide infra), reaction with traces
of water present in the mobile phase could eventually lead to
the formation of ion I. Indeed, the isotopic cluster of ion I shows
[M + H - C6H15NO2]+
[1a + H - H2O]+
[M + H]+
[M + H - H2O]+
[M + H - 1a + H2O]+b
[M - TiN(CH2CHMeO)3 - H2O]+
[M + H]+
644 (18)
409 (100)
[M + H - TiN(CH2CHMeO)3 - HO]+
[M + H - 2TiN(CH2CHMeO)3 - O]+
a Ions are identified by the peak of great intensity in the isotope
distribution pattern. b Ion likely originating from the reactionof fragment
[M + H - 1a]+ with water molecules present in the ion trap (ref 33).
18O incorporation (65%) when H218O is intentionally added to
the mobile phase (see Experimenal Section).
The other experiments reported in Figure 2, spectra b-d,
provide an unequivocal picture of the modifications occurring
in the system under examimation when an excess of 1a with
respect to the Ti(IV) precursor is added. In fact, notwithstanding
the presence of ESI-induced reactions (in all the spectra, the
base peak is still ion I), the progressive appearance of three
(32) Variation of the spray voltage (2-7 kV), of the capillary voltage
(10-50 V), and of the temperature (80-180 °C) does not affect the
speciation pattern in the in situ 2a system.

6262 J. Am. Chem. Soc., Vol. 121, No. 26, 1999
Bonchio et al.
Figure 4. MS2 fragment ions after collision-induced dissociation (CID)
of the isobaric m/z ) 409 peaks originating from the fragmentation of
ions III, IV, and V.
Figure 3. Observed (s) and calculated (|||) isotope distribution patterns
of ions (a) III (m/z ) 427), (b) IV (m/z ) 662), and (c) V (m/z )
897).
new titanium-based ions is monitored at m/z ) 427 (III), 662
(IV), and 897 (V). Their composition corresponds respectively
to the 2:1, 3:2, and 4:3 protonated adducts of ligand 1a with
the titanium nucleus, as revealed by the inspection of the
experimental isotopic clusters (Figure 3). This observation
confirms the solution-phase formation of titanium-based oligo-
mers in line with the hypothesis of eq 134 and their capability
to survive, at least in part, to the ESI ionization conditions.
Tandem MS2 analysis provides a further characterization of
species III, IV, and V (Table 1). The comparative analysis of
their fragmentation patterns allows the recognition of a common
cascade motif, thus indicating that ions III, IV, and V are
homologues (Figure 4).
In particular, the higher aggregate, ion V (m/z ) 897),
containing three chiral titanatrane units, decays to species at
m/z ) 644 and 409 with respectively two and one titanium
center. In a similar fashion, ion IV leads to the mononuclear
titanium species at m/z ) 409 that is also generated in the MS2
spectrum of ion III (Table 1). Moreover, the isobaric ions at
m/z ) 409, originating from the collision-induced dissociation
(CID) of the three precursors III, IV, and V, afford identical
MS3 spectra (Figure 4). On the basis of this evidence, we
attribute the molecular ions under examination to the protonated
species 3a′, 3a′′, and 3a′′′, whose existence has been postulated
according to eq 1.
Figure 5. ESI-MS spectra of complex 3a (2.5 × 10-3 M) in (a) acidic
chloroform31 and (b) methanol as mobile phases with peak assignments
indicated.
No signals corresponding to the coordinated and free 2-propanol
can be detected, and broader peaks relative to the chiral
titanatrane unit can be recognized at δ ) 4.93 (CH), 2.80-
2.98 (CH2), and 1.08-1.21 ppm (CH3). While 1H NMR analysis
confirms that the trialkanolamine 1a is the only ligand present
in the coordination sphere of the titanium(IV) nucleus, it does
not provide direct evidence of the existence of species 3a′, 3a′′,
and 3a′′′.
Again, ESI-MS has proved to be a more diagnostic tool. The
positive ion ESI-MS spectrum of 3a has been collected in two
different mobile phases, namely acidic chloroform31 and metha-
nol (Figure 5).
In both spectra, ions III, IV, and V are present, together with
the oxonium complex I and the protonated free ligand II. When
methanol is the mobile phase (Figure 4, spectrum b), additional
peaks corresponding to other titanium complexes deriving from
the interaction of 3a with the solvent itself and water can be
detected. Low-abundance ions at m/z ) 254 (VI) and 268 (VII)
Ti(IV)/1a Preformed System: Catalyst 3a. Compared with
the monomeric system 2a, the 1H NMR of the preformed
catalyst 3a (Scheme 1, path b) shows a more complex system.
(33) A similar process has been reported for Pt(II) allylic compounds:
Favaro, S.; Pandolfo, L.; Traldi, P. Rapid Commun Mass Spectrom. 1997,
11, 1859-1866.
(34) The location of the protonation site in ions III, IV, V, XI, and XII
cannot be precisely assigned via ESI-MS. We will assume that the
protonation occurs at the nitrogen atom of the apical trialkanolamine ligand.

Solution Speciation of Ti(IV) Sulfoxidation Catalysts
J. Am. Chem. Soc., Vol. 121, No. 26, 1999 6263
Table 2. FIB+-MS and EI-MS Data for Complex 3a
principal
FIB+-MSb EI-MS
ions,a m/z
indentification
(%)
(%)
192
236
280
294
381
II
X
94
17
48
35
52
22
100
ndc
nd
[Ti(N(CH2CHMeO)3(MeCHO))]+
[Ti(N(CH2CHMeO)3(CH2CHMeO))]+
[Ti(N(CH2CHMeO)3N(CH2CH-
MeOH)2(CH2CHMeO))]+
X + NO2C6H4CH2OH
III
nd
389
427
624
19
20
22
nd
nd
nd
[(Ti(N(CH2CHMeO)3))2OCH-
C6H4NO2]+
724
I
30
nd
Figure 6. Ab initio [RHF/3-21G(*)] optimized structures and energies
(Hartrees) of (S,S,S)-Ti[N(CH2CHMeO)3MeO) 5 and cationic species
X.
a Ions are identified by the peak of great intensity in the isotope
distribution pattern. b 3-nitrobenzyl alcohol as matrix. c Not determined.
Table 3. Selected Bond Distances (Å) and Atomic Charges for
(S,S,S)-Ti(NCH2CH2O)3(MeO) (5) and Cationic Species (S,S,S)-X
Calculated at the RHF/3-21G(*) Theoretical Level
can be assigned to the mononuclear chiral titanatrane moiety
bearing an axial molecule respectively of water or methanol.
Ions at m/z ) 489 (VIII) and 503 (IX) show an isotopic
distribution and tandem MS2 spectra (see Experimental Section),
consistent with binuclear structures containing a hydroxy and
methoxy residue, likely bridging between two titanium nuclei.
Isotopic labeling experiments performed with perdeuteriometha-
nol further confirm the assignment of ion IX28 (see Experimental
Section). It is important to notice that, despite the strongly
competing solvent, the chelated titanatrane complexes are
reasonably stable and can be detected in spectrum b (Figure 5).
compound
distances (Å)
atomic charges
5
Ti-O(1) ) 1.815
Ti-O(2) ) 1.814
Ti-O(3) ) 1.815
Ti-O(4) ) 1.747
Ti-N ) 2.397
Ti ) 1.53
O(1) ) -0.81
O(2) ) -0.82
O(3) ) -0.81
O(4) ) -0.53
N ) 0.56
ion X
Ti-O(1) ) 1.753
Ti-O(2) ) 1.753
Ti-O(3) ) 1.753
Ti-N ) 2.142
Ti ) 2.10
O(1) ) -0.91
O(2) ) -0.91
O(3) ) -0.91
N ) 0.37
MS spectra of 3a have also been obtained using different
ionization techniques. Data reported in Table 2 allow the
comparison of the results obtained under fast ion bombardment
(FIB+) and electron impact (EI) ionization conditions.
condensed phase.35 Significantly, ion X is generated only under
FIB+ and EI conditions, while in the ESI-MS experiments only
its formal adducts with oxygenated ligands (water, methanol,
and the trialkanolamine itself) are detected.
Although the FIB+-MS technique has been used in solution-
phase complexation studies to assess the correspondence
between solution- and gas-phase phenomena, in the case of the
titanium catalyst under examination it does not provide a
straightforward answer. The problem affecting the FIB+-MS
experiment is twofold. First, there are ions produced from the
reaction of the titanium catalyst with the 3-nitrobenzyl alcohol
matrix or with the water present in it, i.e., species at m/z ) 624
and 389, and the already discussed ion I (m/z ) 724). Second,
it shows a higher degree of fragmentation as compared with
the ESI-MS results. This last conclusion originates from the
observation that the base peak of the FIB+-MS spectrum
corresponds to ion X (m/z ) 236), whose composition, supported
by isotopic cluster and MS2 analysis (see Experimental Section),
can be ascribed to the cationic metal species [TiN(CH2-
CHMeO)3]+. Moreover, at higher mass-to-charge ratios, there
is no trace of the oligomeric titanium-based ions IV and V,
and the monomeric homologue III is detected only with low
intensity (∼20%). Major fragmentation is expected in the mass
spectrum collected for 3a in the EI ionization mode. Indeed,
the only detected peaks occur in the monomeric region at m/z
) 381 (Table 2), deriving from the breakdown of the apical
trialkanolamine ligand with consequent destruction of the Ti(IV)
oligomers. Moreover, also in the EI-MS spectrum, ion X is one
of the dominant peaks.
While the gas-phase reactivity of ion X with neutral oxygen
donor ligands will be the object of a separate communication,
in this paper we report the results of ab initio theoretical
calculations, performed at the RHF/3-21G(*) level, to determine
the geometries, energies, and atomic charge distributions of the
neutral complex (S,S,S)-Ti[(MeO)N(CH2CHMeO)3] (5) and of
the cationic species X. As shown in Figure 6 and Table 3, the
high-energy cation X displays a compressed structure with
shorter titanium-oxygen and titanium-nitrogen bonds. The
positive charge is distributed on the entire molecule, as indicated
by the variation of all atomic charges, with respect of the neutral
species 5. It is possible to conclude that the polydentate ligand
1a has a stabilizing effect on the metal ion, which retains the
tetracoordination around the Ti(IV) center and survives the
conditions of the FIB+ and EI experiments.
Ti(IV)/1b or 1c Preformed Systems: Catalysts 3b and 3c
and Crossed -Type Complexes. The preformed Ti(IV) com-
plexes with ligands 1b and 1c were prepared according to
Scheme 1, path b, and examined by positive ESI-MS. Methanol
solutions of species 3b and 3c show ESI-MS patterns in perfect
analogy with the one described above for system 3a (see Table
4).28
Therefore, a common feature of all isolated catalysts 3a-c
is that they consist of a mixture of oligomeric Ti(IV) complexes
with similar structure. It should be pointed out that, among the
catalysts examined, 3b affords the most enantioselective sulf-
oxidation system, whose merits are mentioned in the Introduc-
tion. In this light, its solution-phase composition has been further
The gas-phase production of a titanium-centered cation X
bearing the polyfunctional ligand 1a is of particular interest. In
general, the study of transition-metal-containing ions offers the
unique opportunity to investigate the “intrinsic” properties and
reactivity of such species, in the absence of complicating effects
such as solvation and counterion effect, and at the same time,
it can provide important clues about their behavior in the
(35) Frieser, B. S. Acc. Chem. Res. 1994, 27, 353-360.

6264 J. Am. Chem. Soc., Vol. 121, No. 26, 1999
Bonchio et al.
Table 4. ESI-MS Data for Complexes 3b and 3c (2.5 × 10-3 M,
Methanol)a
system
iona
I′
m/z
system
iona
I′′
m/z
4b
nda,b
378
799
1220
ndb
4c
598
150
343
536
729
212
226
nd
II′
II′′
III′
IV′
V′
VI′
VII′
VIII′
IX′
III′′
IV′′
V′′
VI′′
VII′′
VIII′′
IX′′
nd
454
861
875
Figure 8. ESI-MS spectrum in methanol of 3a (2.5 × 10-3 M) upon
addition of a large excess of 1b, with peak assignments indicated. In
ions XI (m/z ) 613) and XII (m/z ) 1034), the relative position of
ligands 1a and 1b in the coordination sphere of Ti(IV) is arbitrary.
419
a Ions are identified by the peak of great intensity in the isotope
distribution pattern. b “nd” indicates m/z values over instrument detec-
tion limit.
Figure 9. 1H NMR spectra in deuteriochloroform at -10 °C of (a)
complex 2a (2.5 × 10-2 M), (b) complex 2a (2.5 × 10-2 M) + 1b
(2.5 × 10-2 M), and (c) complex 2b (2.5 × 10-2 M). Resonances are
relative to (9) complex 2a, ([) complex 2b, and (2) displaced
2-propanol. The other resonances are relative to the free ligands 1a
and 1b.
1
Figure 7. Temperature-dependent H NMR spectra of complex 3b
(2.5 × 10-2 M) in deuteriochloroform at (a) - 17, (b) 25, and (c) 55
°C.
investigated through 1H NMR at variable temperatures (Figure
7).28
The spectral changes, observed within the range from -17
to +55 °C for a deuteriochloroform solution of 3b, definitely
indicate the presence of Ti(IV)-based polynuclear complexes
equilibrating in solution. Specifically, while at high temperature
only three broad signals are detected (Figure 7, spectrum c),
each one corresponding to the different type of protons present
in ligand 1b (CH, CH2, and C6H5), at lower temperature the
system freezes out (Figure 7, spectrum a), yielding a major
number of resonances with different intensities and spanning a
wide range of frequencies. The latter observation is consistent
with the existence in solution of stable nonsymmetric aggregates,
thus supporting the ESI-MS results.
613 (XI) and 1034 (XII) are, indeed, observed whose composi-
tion reveals the simultaneous coordination of both ligands 1a
and 1b to the Ti(IV) center (Figure 8).
The proposed assignment of their structures is indicated
directly on the spectrum (Figure 7) and follows the straight-
forward analogy with species 3′ and 3′′. Moreover, ions at m/z
) 799 (III′) and 1220 (IV′) (see also Table 4) indicate the
presence of complexes where the original ligand 1a has been
completely substituted by 1b in the coordination to the tita-
nium(IV) nucleus.
Direct proof of ligand scrambling occurring in solution is
also provided by the modification of the 1H NMR spectrum of
the in situ-formed catalyst 2a upon addition of 1 equiv of ligand
1b. In Figure 9, spectra a and c allow the identification of the
resonances pertaining to the single monomeric catalysts 2a and
2b and their comparison with the mixed system 2a + 1b (Figure
9, spectrum b).36 In the latter spectrum, the simultaneous
Further evidence both on the nature of oligomers 3 and on
the occurrence of ligand-exchange equilibria affecting their
1
solution integrity has been achieved by ESI-MS and H NMR
analysis of system 3a in the presence of ligand 1b as additive.
The experiments have been designed to monitor the presence
of crossed-type Ti(IV)/1a,b complexes eventually formed
through ligand scrambling. Following this idea, the ESI-MS
spectrum of system 3a has been collected under the same
conditions of the Figure 5b experiment, but adding an excess
of ligand 1b to the mobile phase. New peaks arising at m/z )
(36) An analogous result can be obtained if a different addition order of
the reagents is used, namely by adding Ti(i-PrO)4 to equimolar amounts of
1
premixed ligands 1a and 1b (see H NMR spectra reported in Figure 1S,
Supporting Information).

Solution Speciation of Ti(IV) Sulfoxidation Catalysts
J. Am. Chem. Soc., Vol. 121, No. 26, 1999 6265
Table 6. 1H NMR Spectral Data for tert-Butyl Peroxo Complexes
4a-c in Deuterochloroform at 21 °C
compd
1H NMR (δ, ppm)a
4a
1.09 (d, J ) 6.0 Hz, 9H), 1.47 (s, 9H), 2.95-3.03
(AB system, 6H), 4.91-5.08 (m, 3H)
1.58 (s, 9H), 3.32 (dd, J ) 11.9, 4.4 Hz, 3H);, 3.43 (dd,
J ) 11.9, 10.6 Hz, 3H), 5.98 (dd, J ) 10.6, 4.4 Hz, 3H),
7.20-7.48 (m, 15H)
4b
4c
1.48 (s, 9H), 3.26 (t, J ) 6.0 Hz, 6H), 4.50 (t, J ) 6.0 Hz,
6H)
a Relative to TMS.
the nature of the active peroxometal complex generated under
the conditions of catalytic sulfoxidation, which will be addressed
in the next paragraph.
Ti(IV)/1 Alkyl Hydroperoxides. As already mentioned in
the Introduction, titanium η2-alkyl peroxo complexes are
considered the active oxidants in the catalytic cycle employing
Ti(IV) alkoxide precursors and alkyl hydroperoxides as oxygen
donors20,21 (Scheme 1). Direct evidence of the existence of such
reactive intermediates is provided by the solid-state structure
of η2-tert-butyl peroxotitanatrane dimer which is, so far, the
only characterized Ti(IV) peroxo complex.25 However, X-ray
data may be of limited utility or even misleading for the
assignment of the solution-phase structure of complexes that
exchange ligands as readily as titanium(IV) alkoxides do. In
the case of the Ti(IV)/1 systems, 1H NMR allows the monitoring
of the formation of the peroxidic species 4 when an excess of
tert-butyl hydroperoxide (TBHP) is added to catalysts 2 or 3
(Scheme 1, path c). In particular, under catalytic conditions,
e.g., in the presence of an excess of the primary oxidant TBHP,
both the monomeric system 2 and the oligomeric complexes 3
Figure 10. Plot of benzyl p-tolyl sulfoxide formation vs time (min)
in the oxidation of benzyl p-tolyl sulfide (0.27 M) by CHP (0.027 M)
in DCE at -20 °C, catalyzed by (b) 2a, (2) 2b, and (9) 2a + 2b (for
experimental conditions, see footnote to Table 5).
Table 5. Stereoselective Sulfoxidation of Benzyl p-Tolyl Sulfide
(0.27 M) by CHP (0.027 M) Catalyzed by in Situ-Formed Systems
2a, 2b, and 2a + 2b in DCE at -20 °Ca
ee (%)
(absolute
configuration)
ligand
kobs
entry
system
× 103 (M)
× 104 (s-1
)
1
2
3
(S,S,S)-2a
1a (3.6)
1b (3.6)
1a (1.8) + 1.57
1b (1.8)
0.12 ( 0.014
2.51 ( 0.13
12 (R)
53 (S)
52 (S)
(R,R,R)-2b
(S,S,S)-2a +
(R,R,R)-2b
1
quantitatiVely evolve to species 4. The H NMR resonances
pertaining to 4 are those of a highly symmetric Ti(IV) species.
In fact, a single set of signals is observed for each type of
protons of the chiral titanatrane unit, and a singlet is observed
corresponding to the coordinated tert-butylperoxo group (Table
6).
a In all the reactions, [Ti(i-PrO)4] ) 3.3 × 10-3 M, and ligand 1
according to the table.
appearance of signals corresponding to the chiral titanatrane
moiety with chelated 1b and to the displaced ligand 1a is
observed, thus indicating the coexistence in solution of both
catalytically active complexes 2a and 2b.
The spectroscopic evidence points to a monomeric structure
for the peroxotitanium species formed in solution under catalysis
conditions.37 Additional support for a monomeric structure
comes from the characterization via ESI-MS of the protonated
Ti(IV) peroxo complexes 4a (m/z ) 326, ion XIII) and 4c (m/z
) 284, ion XIII′′), whose formation is monitored upon addition
of TBHP to the mobile phase (methanol or acidic chloroform)
containing 3a or 3c. Analysis of the isotopic clusters28 and MS2
studies (see Experimental Section) of both ions XIII and XIII′′
provides direct proof of the monomeric structure of peroxo
complexes 4.
In this respect, a key observation is provided by the
examination of the asymmetric oxidation of benzyl p-tolyl
sulfide with cumyl hydroperoxide (CHP) in the presence of the
different catalytic systems. The aim of such investigation is to
test the single catalysts 2a and 2b and the mixed system (Ti(i-
PrO)4 + 1a + 1b) under the twofold profile of reaction kinetics
and enantioselectivity (Figure 10 and Table 5).
Inspection of the pseudo-first-order rate constant (kobs) values
reported in Table 5 reveals that (i) catalysis by 2a is negligible
with respect to 2b, whose reactivity is superior by 1 order of
magnitude (cf. entries 1 and 2 in Table 5) and (ii) in the mixed
system, where 2a:2b ≈ 50:50, the two catalysts behave
independently, displaying a “solo” reactivity, and their contribu-
tion to the overall reaction rate is additive, so that a predictable
reduction of the rate constant to one-half is, indeed, observed
(cf. entries 2 and 3 in Table 5). As a consequence of the ligand-
accelerating effect,22 induced by the chiral trialkanolamine 1b,
the enantiomeric excess of the sulfoxide recovered either in the
reaction catalyzed by 2b alone or by the mixed system (2a +
2b) is basically the same (cf. entries 2 and 3 in Table 5). The
above observations are unanimous in support of the independent
reactivity of the two enantiopure catalysts produced in solution
through ligand exchange equilibria involving 1a, 1b, and the
achiral precursor Ti(O-i-Pr)4; moreover, they raise the issue of
Upon collision-induced dissociation (Scheme 2), XIII and
XIII′′ give analogous daughter ion spectra. Their common
fragmentation pattern involves mainly the peroxidic ligand, so
that ions at m/z ) 270 and 228 (ions XIV and XIV′′,
respectively) are produced from the cleavage of the t-Bu-O
bond and peaks at m/z ) 254 (VI) and 212 (VI′′) derive from
the peroxidic O-O bond scission. Moreover, the monomeric
nature of the peroxometal complex accounts for the previously
discussed results obtained in the asymmetric oxidation of the
model substrate benzyl p-tolyl sulfide, which show no contribu-
tion from crossed-type Ti(IV)/1 active oxidants, displaying more
than one molecule of the trialkanolamine ligand in the coordina-
tion sphere of the metal.
(37) The achiral peroxo complex 4c is unstable at 21 °C and decomposes
readily in the presence of an excess of TBHP.

6266 J. Am. Chem. Soc., Vol. 121, No. 26, 1999
Bonchio et al.
Table 7. Energies (Hartrees) and Selected Bond Distances (Å) of Peroxidic Complex 4a and of Isomers A-D of Ion XIII Calculated at the
RHF/3-21G(*) Theoretical Level
4a
A
B
C
D
energies
in Hartrees
in kcal/mol
-1776.103 077
-1776.445 272
+24.9
-1776.471 126
+13.3
-1776.483 918
+5.3
-1776.492 499
0.00
Bond Distances
Ti-O(1)
1.82
1.81
1.79
1.91
2.10
2.48
1.50
1.48
1.78
1.77
1.82
1.90
2.00
3.43
1.50
1.50
1.77
1.79
1.78
2.01
2.57
2.27
1.51
1.58
1.77
1.77
1.78
2.05
2.72
2.25
1.48
1.53
1.81
1.79
1.82
1.91
2.10
2.48
1.50
1.48
Ti-O(2)
Ti-O(3)
Ti-O(4)
Ti-O(5)
Ti-N
O(4)-O(5)
O(5)-C(1)
Scheme 2. MS2 Fragment Ions after CID of Peroxidic Ions
XIII (m/z ) 326) and XIII′′ (m/z ) 284)
of the small energy difference, the protonated isomers A, B,
and C are all likely to be present.40
Conclusions
In this work, the solution structure of chiral Ti(IV)/trialkanol-
amine complexes has been elucidated throughout the combined
1
use of ESI-MS and H NMR techniques.
The main results, which have provided a definite picture of
the catalytic system under examination, can be summarized as
follows:
(a) The catalyst structure depends on the Ti(IV)/ligand 1
stoichiometric ratio. When a 1:1 ratio is used, monomeric, highly
symmetric complexes are obtained. An excess of ligand 1
induces the production of polynuclear Ti(IV) species, whose
architecture is organized by a central molecule of trialkanol-
amine 1 with one, two, or three chiral titanatrane units appended.
In such complexes, the apical ligand is too labile and only
partially survives the electrospray ionization conditions. There-
fore, ESI-MS analysis does not allow the determination of the
thermodynamic distribution of the different Ti(IV) species
equilibrating in solution.
(b) Fast ligand exchange equilibria occur in solution among
1
the Ti(IV) complexes. In fact, both via ESI-MS and H NMR,
Concerning the structural assignment of ion XIII, ab initio
calculations (RHF/3-21G(*)) have been performed to gain
insight about the favored protonation site in peroxo complex
4a. In Table 7, energies obtained for the neutral complex (S,S,S)-
4a and for the different isomers of ion XIII are reported.38,39
These latter derive from protonation at the four possible basic
sites of the molecule namely: the two peroxygen atoms
(structures A and B), the equatorial alkoxy functionality, and
the tertiary nitrogen atom of the coordinated trialkanolamine
ligand (structures C and D, respectively).
crossed-type Ti(IV)/1a,b adducts were detected in the presence
of both ligands. Nevertheless, kinetic experiments dealing with
the stereoselective sulfoxidation catalyzed by the scrambled
complexes point out the independent reactivity of the two
enantiopure catalysts.
(c) The key observation is that the active Ti(IV) peroxo
complexes 4, obtained both from the monomeric 2 and from
the polynuclear aggregates 3 in the presence of an excess of
TBHP, are themselves monomeric.
According to the theoretical calculations, structure D is highly
destabilized, as the N-protonated ligand loses its tetracoordi-
nation to the titanium center (cf. Ti-N bond distance in 4a and
in D, Table 7). To the contrary, protonation at the equatorial
trialkoxyamine oxygen affords the most stable isomer C and a
minor reorganization of the geometry of the complex. Because
Taken as a whole, the present study has evidenced the
important role played by the tetradentate ligands 1 in stabilizing
Ti(IV) catalytic precursors and active species. It should be
pointed out that the Ti(IV)/1 sulfoxidation system is remarkably
robust, reaching up to 1000 turnovers.41 The fact that the highly
reactive peroxo complexes 4 survive the conditions of ESI-MS
bodes well for the extension of this approach to the study of
other metal-mediated processes42
(38) The calculated geometrical parameters for peroxide 4a are compa-
rable with previously reported data for Ti(IV) trialkanolamine complexes
from difrattometric analysis (see refs 24 and 25) and calculations performed
on other monomeric transition metal peroxides; see: (a) Bonchio, M.; Di
Furia, F.; Licini, G.; Modena, G.; Moro, S.; Nugent, W. A. J. Am. Chem.
Soc. 1997, 119, 6935. (b) Wu, Y.-D.; Lai, D. K. W. J. Am. Chem. Soc.
1995, 117, 11327. (c) Wu, Y.-D.; Lai, D. K. W. J. Org. Chem. 1995, 60,
673. (d) Jørgensen, K. A. J. Chem. Soc., Perkin Trans. 2 1994, 117. (e)
Boche, G.; Bosold, F.; Lohrenz, J. C. W. Angew. Chem., Int. Ed. Engl.
1994, 33, 1161.
(40) Worthy of notice is that protonation at the peroxidic ligand causes
the loosening of its η2 coordination geometry, thus indicating that both forms
A and B could afford the fragmentation pattern observed in the MS2
experiments. Indeed, elongation of the O-t-Bu bond occurring in B could
explain the loss of the isobutene residue hypothesized in Figure 11.
(41) Bonchio, M.; Licini, G. Unpublished results.
(42) For a direct proof of reactive manganese(V) oxo species, see:
Feichtinger, D.; Plattner, D. A. Angew. Chem., Int. Ed. Engl. 1997, 36,
1718.
(39) The optimized geometries of complex (S,S,S)-4a and of the isomers
A-D of ion XIII are reported in the Supporting Information, Figure S2.

Solution Speciation of Ti(IV) Sulfoxidation Catalysts
J. Am. Chem. Soc., Vol. 121, No. 26, 1999 6267
Ti(IV)/1 in Situ-Formed System (Catalyst 2). A 0.025 M solution
of catalyst 2 in CHCl3 or CDCl3 was prepared in a 2-mL volumetric
flask by mixing 0.500 mL of a mother solution of ligand 1 (0.100 M)
in the same solvent with a stoichiometric amounts of Ti(i-PrO)4 (0.015
mL, 0.05 mmol). For performing the ESI-MS experiments, the mother
solution of catalyst 2 was diluted by a factor of 10. Solutions of catalyst
2 containing increasing amount of ligand 1 (Figures 1 and 2) were
prepared by appropriate dilutions of the starting ligand 1 mother
solution.
Experimental Section
General Methods. 1H NMR spectra were recorded on a Bruker AC-
200 SY (200 MHz) or an AC-250 (250 MHz) instrument. Gas
chromatographic analyses were performed using a Hewlett-Packard
5890 series II GC equipped with an SE-30 15-m × 0.25-mm-i.d.
capillary column, using 4-methyl benzophenone as internal standard.
Benzyl p-tolyl sulfoxide enantiomeric excesses and absolute configu-
ration were determined directly on reaction mixtures by HPLC analysis
performed on a Water Associates HPLC/GPC (FDP) 201 pump and a
Water Associates 440 UV detector (λ ) 254 nm) with a Lichrosorb
S100 (S,S)-CSP-DACH-DNB [(250 × 4.0 mm i.d.] chiral column43,44
(n-hexane/2-propanol (8:2) as eluent, flow rate of 2.0 mL/min, P )
1000 psi) according to the elution order [the (S) enantiomer is eluted
before the (R) one]. Yields and product distributions were determined
via quantitative GC analysis. Benzyl p-tolyl sulfide and sulfoxide
spectral data match those already reported.45 ESI-MS experiments were
performed with a Finnigan LCQ instrument, with an upper mass limit
of m/z ≈ 1850, through direct infusion via a syringe pump. Standard
experimental conditions were as follow: sample concentration, 10-3M;
flow rate, 8 µL min-1; nebulizing gas, N2 (40 units flow rate); spray
voltage, 4 kV; capillary voltage, 25 V; capillary temperature, 120 °C;
tube lenses offset, 30 V. The parameters related to octapoles and
detector were those achieved by the automatic setup procedure.
Collision-induced decompositions of selected ions were obtained by
applying a supplementary radio frequency voltage (tickling voltage)
to the end-cap electrodes of the ion trap (resonance activation). In the
experiments aimed at the detection of peroxometal species, the capillary
temperature was set at 80 °C. Cs+ fast ion bombardment (FIB+) mass
spectra have been collected on a multiple quadrupole instrument
(VGQuattro, VGBiotech, Altrincham, UK). The cesium ion gun was
operated at an accelerating voltage of 15 kV and a heating current of
2.4 A. Test samples were prepared by dissolving the investigated
precursor (0.1 mg) in 3-nitrobenzyl alcohol (NBA) used as the liquid
matrix. Electron impact (EI) mass experiments have been performed
on a VG AUTOSPEC instrument operating at 70 eV.
Computational Study. Ab initio calculations were carried out with
the program systems Spartan v.446 and Gaussian 94,47 running on IBM
RS/6000 workstations. The molecular geometries were optimized using
the 3-21G(*) basis set at the Hartree-Fock (HF) level of theory.
Chemicals. HPLC grade solvents were generally used. Methanol-
d4 (Fluka) and water-18O, 20% (C.I.L.) were used in the labeling
experiments. Dichloromethane was distilled over CaH2 and stored over
molecular sieves. 1,2-Dichloroethane (DCE) was washed 3 times with
10% concentrated H2SO4 and with water several times to a pH of 7,
dried over CaCl2 overnight, distilled over P2O5, and stored over
molecular sieves. tert-Butyl hydroperoxide (Fluka, 80%, 20% di-tert-
butylperoxide) was purified by distillation under vacuum (bp 31-32
°C/16 Torr) and stored at 0 °C. Cumyl hydroperoxide (80% in cumene,
Fluka) was stored over molecular sieves at 0 °C. Titanium(IV)
tetraisopropoxide (Aldrich) was distilled under vacuum (bp 60-63 °C/
0.1 Torr). Benzyl p-tolyl sulfide was prepared by alkylation of sodium
p-tolylthiolate.48 Enantiopure trialkanolamines 1a,b were prepared
following the literature procedure.24 Triethanolamine (Aldrich) was
distilled under reduced pressure (bp 190-193 °C/5 Torr).
Ti(IV)/1 Preformed System (Catalyst 3). Ti(O-i-Pr)4 (0.1 mL, 0.33
mmol) was added to a solution of ligand 1 (0.44 mmol) in anhydrous
CH2Cl2 (5 mL) under a nitrogen atmosphere. After the solution was
stirred for 5 min at room temperature, the solvent was removed under
reduced pressure. The recovered material was twice dissolved in
dichloromethane (2 × 5 mL), and the solvent was removed again under
vacuum. After the material was washed with hexane (6 mL), the solvent
was removed under vacuum, yielding a white solid that was dried under
high vacuum (0.1 mmHg) for 1 h and stored under nitrogen.
Procedure for the Kinetic Study of the Stereoselective Oxidation
of Benzyl p-Tolyl Sulfide. A typical procedure is the following: in a
10-mL volumetric flask, Ti(i-PrO)4 (0.015 mL, 0.05 mmol), ligand 1
(0.05 mmol), and benzyl p-tolyl sulfide (1.164 g, 5.440 mmol) were
dissolved in dry DCE. After the mixture was cooled to -20 °C, cumyl
hydroperoxide (0.100 mL, 0.544 mmol) was added under magnetic
stirring. At the reaction times stated in Figure 10, a sample of the
mixture (0.025 mL) was taken out and immediately quenched with an
excess of di-n-butyl sulfide for the determination of conversion and
product distribution (GC analysis), and for the ee and absolute
configuration determination of benzyl p-tolyl sulfoxide (HPLC analysis).
The asymmetric oxidation performed in the presence of the mixed
catalytic system (2a + 2b) was performed by introducing both ligand
1a (0.025 mmol) and ligand 1b (0.025 mmol) in the reaction mixture.
Labeling Experiments Performed by ESI-MS. The experiment
aimed at the detection of ion I (18O) was run by injecting into the mass
spectrometer a methanol solution of catalyst 3a (2 mL), in the presence
of H218O (20%, 20 µL added to the mobile phase). 18O incorporation
can be calculated from the two isotopic clusters collected for ion I
before and after the labeling of the mobile phase according to the
following formula: %[18Oinc] ) [%(M + 2) O - %(M + 2) O]/%M
18
16
18
O
18
16
+ [%(M + 2) O - %(M + 2) O]. Based on the experimental isotopic
distributions, the following isotopic distributions (ID) are found: for
I(16O), ID (exp %, calc %) ) 722 (19, 31); 723 (25, 39); 724 (100,
100); 725 (50, 53); 726 (28, 33); 727 (7, 11); for I(18O), ID (exp %) )
722 (19); 723 (25); 724 (100); 725 (48); 726 (43); 727 (14), 728 (8).
Considering the actual 18O content of the labeled water used (20%),
we calculated that [18Oinc] g 65% (the diluition of the H2O18 with
unlabeled water present in the mobile phase was not considered in the
former calculation).
Detection of Peroxometal Complexes 4a and 4c by ESI-MS. For
these experiments, a solution of catalyst 3 in methanol or chloroform
(2 mL) containing TBHP (200 µL) was injected into the ESI-MS
instrument. The resulting spectra showed a significant reduction of the
signals relative to the polynuclear Ti(IV) species and the appearance
of a new peak in the monomeric region as follows: for ion XIII, ID
(exp %, calc %) ) 324 (12, 11); 325 (19, 12); 326 (100, 100); 327
(22, 23); 328 (9, 10); MS2 ) 270 (M+ - C4H8), 254 (M+ - C4H8O),
268 (ion VII);49 for ion XIII′′, ID (exp %, calc %) ) 282 (10, 11);
283 (10, 11); 284 (100, 100); 285 (19, 19); 286 (10, 10); MS2 ) 228
(M+ - C4H8), 212 (M+ - C4H8O), 226 (ion VII′′).49
ESI-MS Isotopic Distributions and MS2 Data for Selected Ions
Not Reported in the Body of the Paper: for VIII, ID (exp %, calc
%) ) 487 (20, 21); 488 (25, 24); 489 (100, 100); 490 (37, 36); 491
(20, 20); MS2 (%) ) 445 (M+ - CH3CHO); for IX, ID (exp %, calc
%) ) 501 (19, 21); 502 (22, 25); 503 (100, 100); 504 (33, 37); 505
(19, 20); MS2 ) 459 (M+ - CH3CHO). When perdeuterated methanol
was used as the mobile phase, the isotopic cluster of ion IX showed
an increment of 3 mass units consistent with the formula C19H36D3O7N2-
(43) Gargano, G.; Gasparrini, F.; Misiti, D.; Palmieri, G.; Pierini, M.;
Villani, C. Chromatographia 1987, 24, 505.
(44) Altomare, C.; Carotti, A.; Cellamare, S.; Fanelli, F.; Gasparrini, F.;
Villani, C.; Carrupt, P.-A.; Testa, B. Chirality 1993, 5, 527.
(45) Lonoshita, M.; Sato, Y.; Kunieda, N. Chem. Lett. 1974, 377.
(46) Spartan v.4.0, Wavefunction, Inc., 18401 von Karman Ave. #370,
Irvine, CA 92715, 1995.
(47) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Oritz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Reploge, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D.; Baker, J. J.; Steward, J. P.; Head-
Gordon, M.; Gonzales, C.; Pople, J. A. Gaussian 94, Revision C.2; Gaussian
Inc.: Pittsburgh, PA, 1995.
(48) Peach, M. E. In The Chemistry of Thiol Group, Patai, S., Ed.; John
Wiley and Sons, Ltd.: Bristol, 1974; Part 2, pp 721-784.
(49) Ion likely originating from the reaction of ion X with neutral solvent
molecules (methanol) (see also ref 33).

6268 J. Am. Chem. Soc., Vol. 121, No. 26, 1999
Bonchio et al.
Ti2 as follows: IX(D3), ID (exp %, calc %) ) 504 (27, 21); 505 (23,
assistance. Financial support by CNR, MURST, and the DuPont
Aid to Education (ATE) program are gratefully acknowledged.
25); 506 (100, 100); 507 (29, 37); 508 (16, 21); MS2 ) 462 (M+
-
CH3CHO).
Supporting Information Available: 1H NMR spectra in
deuteriochloroform of ligands 1a and 1b, ligands 1a (2.5 × 10-2
M) and 1b (2.5 × 10-2 M) + Ti(i-PrO)4, complex 2a + 1b,
complex 2a, and complex 2b; ab initio [RHF/3-21G(*)]
optimized structures and energies (Hartrees) of peroxo complex
4a and of isomers of ion XIII; and 1H NMR of peroxo
complexes 4 (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
FIB+-MS Isotopic Distributions and MS2 Data for Selected Ions
Not Given in the Body of the Paper: for ion X, ID (exp %, calc %)
) 234 (16, 11); 235 (15, 11); 236 (100, 100); 237 (21, 18); 238 (12,
9); MS2 ) 205 (M+ - CH3O), 192 (M+ - CH3CHO).
Acknowledgment. We thank Miss G. Santoni for performing
the kinetic measuraments and the “Servizio di Spettrometria di
Massa del CNR, Area di Ricerca” Padova for technical
JA983569X