Ethylene-linked bis(phenol) ligands: Efficient synthesis, Ti(IV) coordination chemistry, and Lewis acid catalysis

Article abstract of DOI:10.1016/j.jorganchem.2003.08.046

We report herein an efficient synthesis of a series of chelating bis(aryloxide) ligands that can be diverged at a late stage to generate a variety of structures. Based on structural differences between linked and unlinked analogs of six-coordinate acid-ba

Full text of DOI:10.1016/j.jorganchem.2003.08.046

Journal of Organometallic Chemistry 688 (2003) 121ꢀ124
/
www.elsevier.com/locate/jorganchem
Ethylene-linked bis(phenol) ligands: efficient synthesis, Ti(IV)
coordination chemistry, and Lewis acid catalysis
Jeffrey W. Anthis, Andrew O. Larsen, Peter S. White 1, Michel R. Gagne´ *
Department of Chemistry CB # 3290, University of North Carolina, Chapel Hill, NC 27599-3290, USA
Received 9 May 2003; received in revised form 28 August 2003; accepted 29 August 2003
Abstract
We report herein an efficient synthesis of a series of chelating bis(aryloxide) ligands that can be diverged at a late stage to generate
a variety of structures. Based on structural differences between linked and unlinked analogs of six-coordinate acidꢀbase adducts of
(ArO)2TiCl2(dmpe), a hypothesis predicted that a difference between the two structure types would be apparent when the
compounds shuttled between four- and six-coordinate structures. Comparing their efficiencies in the Lewis acid accelerated Dielsꢀ
/
/
Alder reaction, however, did not support this notion.
# 2003 Published by Elsevier B.V.
Keywords: Ethylene-linked bis(phenol) ligands; Ti(IV) coordination chemistry; Lewis acid catalysis
1. Introduction
tion chemistry of the 6,6?-dimethyl and diisopropyl
ligands with Ti, and a comparative analysis of their
activity as Lewis acid catalysts.
Extensive experimental evidence has reinforced the
paradigm that chelating ligands are most efficient at
transferring their stereochemical information to transi-
tion metals and subsequently to their catalyzed reactions
(e.g. BINAP, TADDOL, tartrate esters and ansa-
metallocenes) [1]. Moreover, the chelate effect can be a
significant stabilizing influence on complexes prone to
ligand redistribution. For these and other reasons we
recently had the need to synthesize a series of chelating
aryloxide ligands for use in Ti chemistry [2]. While
methylene, and oxo linkages were relatively well known
we found that simple ethylene linked structures were
rare, and that the reported method for their synthesis
was not amenable to delivering the quantity or struc-
tural variability that we desired [3]. This note reports a
new synthetic approach to ethylene-bridged bis(phenol)
ligands that introduce the ortho-substituents at a late
stage for maximum structural variability, the coordina-
2. Ligand synthesis
Our ligand synthesis was designed to be high yielding,
inexpensive and amenable to late stage diversification
(Scheme 1). McMurry coupling [4] of salicylaldehyde
under Mukaiyama conditions [5] (TiCl4, Zn) affords the
crystalline 2,2?-ethylenebis(phenol) (80:20 trans:cis ra-
tio, 48% yield), which was hydrogenated over Pd/C to
give the diphenol 1 in 94% yield. The diphenol was
protected without purification as the methoxymethyl
ether with NaH and chloromethyl methyl ether, and
purified by distillation (78%). Orthometallation of 2 (2.2
equivalents tert-BuLi, 00
afforded multi gram quantities of 3 (Rꢁ
/
25 8C) followed by MeI
/Me, 95%
yield), after acidic workup and chromatography. Ortho-
metallated 2 could also be quenched with other electro-
philes like 1,2-Br2Cl4C2 or DMF to provide the 6,6?-Br2,
3b, and 6,6?-diformyl, 3c, diphenols, respectively. Ortho-
metallation with an electrophilic quench offers viable
routes to a wide variety of additional ortho functional
groups (e.g. acetates, acids, and boronic acids), making
this synthetic scheme highly applicable to the synthesis
* Corresponding author. Tel.: ꢂ
6342.
E-mail address: mgagne@unc.edu (M.R. Gagne´).
/
1-919-962-6341; fax: ꢂ1-919-062-
/
1
P l e a s e d i r e c t c r y s t a l l o g r a p h i c i n q u i r i e s t o
pwhite@pyrite.chem.unc.edu.
0022-328X/03/$ - see front matter # 2003 Published by Elsevier B.V.
doi:10.1016/j.jorganchem.2003.08.046

122
J.W. Anthis et al. / Journal of Organometallic Chemistry 688 (2003) 121ꢀ124
/
(ArO)2TiCl2 complexes like 7, but that unknown 1:2
Ti:dmpe adducts (up to 25%) always accompanied
formation of the desired product [9]. Surprisingly, the
linked analog 5 reacted smoothly with dmpe to provide
the desired adduct 9 free of byproducts, and in good
yield (Eq. (1)). The 4-vinyl analog of 9 [10] provided X-
ray quality crystals for analysis.
ð1Þ
Scheme 1.
of additional ligand variants [6]. Attempts to quench the
dianion with iso-propyl iodide were unsuccessful, and so
the 6,6?-iPr2-derivative 4 was obtained by a different
procedure [7]. A 4-vinyl version of 3 was also readily
As expected, the six-coordinate compound (vinyl-
3)TiCl2(dmpe) had a distorted octahedral geometry
[11] (Fig. 1) that was similar to its unlinked analog in
most regards [10], but had a ligand conformation that
differed because of the ethylene link. Like (2,6-dimethy-
laryloxy)2TiCl2(dmpe), the aryloxide oxygen atoms are
trans to the phosphorus atoms with most bond distances
and angles being nearly identical. The orientations of
the phenyl rings, however, are quite different as they
obtained by selective bromination (Br2, ꢀ
/98%) and
Heck coupling [8] to give vinyl-3 (52% yield).
The (ArO)2TiCl2 Lewis acid complexes containing the
new diaryloxy ligands were synthesized using the con-
ditions of Okuda (ligandꢂTiCl4 in pentane) [4]. This
/
methodology proved amenable to both monodentate
and bidentate phenols, and provided good yields of
linked and unlinked variants of 2,6-Me2 and 2,6-Me,iPr
orient perpendicular to the Oꢀ
/
Oꢀ
/
Tiꢀ
/
PꢀP plane in the
/
unlinked case and are skewed almost parallel to the
plane in the linked case, a natural consequence of the
linkage (Fig. 2). This structural difference results in
significantly different projections of the ortho methyl
groups relative to the neutral-ligand binding plane.
Contrasting these orientational differences are the
nearly superimposable structures of the linked and
complexes (Scheme 2). The dichlorides 5ꢀ7 were iso-
/
lated by crystallization/precipitation in moderate to
good yields as red to brown solids. Catalyst 8 was
isolated as a viscous brick red liquid, the physical
characteristics of which made purification difficult and
necessitated accurate reagent stoichiometry. Each com-
plex was utilized as a catalyst for the DielsꢀAlder
/
reaction (vide infra).
3. Results and discussion
Previous experiments have shown that stable, isolable
dmpe adducts could be obtained from tetrahedral
Fig. 1. ORTEP diagram of (vinyl-3)TiCl2(dmpe). Selected bond lengths
˚
(A) and bond angles (8): Tiꢀ
/
P(2)ꢁ
/
2.699(3), Tiꢀ
/
O(1)ꢁ
/
1.802(7), Tiꢀ
/
Cl(1)ꢁ
/
2.346(3), Pꢀ
/
Tiꢀ
/
Pꢁ
/
74.53(8), Clꢀ
/
TiꢀClꢁ
/
/
158.72(11), Oꢀ
/
Tiꢀ
/
Scheme 2.
Oꢁ105.5(3).
/

J.W. Anthis et al. / Journal of Organometallic Chemistry 688 (2003) 121ꢀ
/
124
123
accelerate the DielsꢀAlder reaction between the two-
/
point binding substrate 3-acryoyloxazolidinone and 1,3-
cyclohexadiene.
The first catalyst set included (2,6-Me2-C6H3O)2TiCl2
(7) [14] and its linked analog 5. The second, more
sterically bulky catalyst set, included (2-iPrꢀ
/
6-Meꢀ
/
C6H3O)2TiCl2 (8) and its linked analog (6). The reac-
tions were carried out in CH2Cl2 at ambient temperature
using 5 mol% catalyst and 15 equivalents of diene to
achieve pseudo-first order conditions. Reaction progress
was monitored by quenching an aliquot on silica gel and
analyzed by GC. As Fig. 3 indicates, the catalysts have
similar initial rates with the linked methyl complex 5
being slightly faster (1.1ꢄ
(kobs 3.4ꢄ
10ꢃ4 and 3.0ꢄ
Since the linked Pr complex 6 was 1.2 times slower
/
) than its unlinked analog 7
ꢁ
/
/
/
10ꢃ4 sꢃ1, respectively).
i
than its unlinked analog 8 (kobs
10ꢃ4 sꢃ1, respectively), these data indicate that the
conformational rigidity observed in the solid state does
ꢁ
/
2.4ꢄ
/
10ꢃ4 and 2.9ꢄ
/
not translate into significant Dielsꢀ/Alder rate differ-
ences.
Based on the simple mechanism in Scheme 2 and the
rate of reaction, one possible explanation for the
observed insensitivity to ligand structure is that cycload-
dition is not turnover limiting (k2), and that the effect
that putatively diminishes the binding constant also
compensates by enhancing the rate of product release
(reforming the tetrahedral complex) [15]. Additional
mechanistic studies will be necessary to distinguish
between the reasonable mechanisms.
Fig. 2. CHEM3D representations of the X-ray structures of (2,6-Me2 ꢀ
/
C6H3O)2 TiCl2×/dmpe [10] (top) and 9 (bottom) highlighting the
different aryloxide rotomers. The dmpe portion is removed in both
cases for clarity.
unlinked derivatives of tetrahedral (3)TiBr2 complexes
[4,12,13], Thus, the linkage joining the two aryloxy
ligands is structurally significant only in the octahedral
form of the Lewis acid; the base-free tetrahedral forms
are isostructural.
4. Conclusion
These data led us to hypothesize that the structural
differences between linked and unlinked octahedral
forms of otherwise similar complexes would be mani-
fested in their Lewis acid properties. That is, although
the ground state structure of linked and unlinked
tetrahedral Lewis acids are nearly identical, ligand
torsional strain generated upon substrate coordination
would attenuate the performance of the linked ligands in
In summary, this paper reports a new, efficient
methodology for the synthesis of ethylene-linked diphe-
a reaction like the DielsꢀAlder (k1/kꢃ1, Scheme 3),
/
which necessitates the intermediacy of an octahedral
activated complex. This hypothesis was evaluated by
comparing the ability of two sets of catalysts to
Fig. 3. Plot of ln[oxazolidinone] vs. time. mꢁ
/
7; kꢁ
/
5; Iꢁ/8; and
Scheme 3.
jꢁ6.
/

124
J.W. Anthis et al. / Journal of Organometallic Chemistry 688 (2003) 121ꢀ124
/
[5] T. Mukaiyama, T. Sato, J. Hanna, Chem. Lett. (1973) 1041.
[6] (a) For examples of quenching with formyl, acetate, amide,
carbinol and carboxylate precursors, see: C.A. Townsend, L.M.
Bloom, Tetrahedron Lett. 22 (1981) 3923;
nol ligands that introduces the ortho-substituent in the
next to last step. This approach provides a highly
convergent route to many ligand structures in addition
to the ones reported herein. The coordination chemistry
of these ligands parallels that of the unlinked cases in
many ways, though octahedral coordination environ-
ments bring out their differences.
(b) For quenching with a boronic acid precursor, see: B.I. Alo, A.
Kandil, P.A. Patil, M.J. Sharp, M.A. Siddiqui, V. Snieckus, P.D.
Josephy, J. Org. Chem. 56 (1991) 3763.
[7] (a) J.W. Anthis, I. Filipov, D.E. Wigley, Inorg. Chem., in press;
(b) J.W. Anthis, PhD dissertation, University of Arizona, April
2001.
[8] (a) For several reviews of the Heck coupling, see: N.J. Whit-
combe, K.K.M. Hii, S.E. Gibson, Tetrahedron 57 (2001)
7449;
Acknowledgements
We warmly thank the National Science Foundation
for support of this research (CHE-0075717). M.R.G. is a
(b) A. Biffis, M. Zecca, M. Basato, J. Mol. Cat. A: Chem. 173
(2001) 249;
(c) A. de Meijere, F.E. Meyer, Angew. Chem. Int. Ed. Engl. 33
(1994) 2379.
CamilleꢀDreyfus Teacher Scholar.
/
[9] A.O. Larsen, P.S. White, M.R. Gagne´, Inorg. Chem. 38 (1999)
4824.
References
[10] The 4-vinyl analog of 5 was obtained by a different route since the
vinyl group is hydrohalogenated under the conditions of Scheme
2. It is obtained in pure form by treating the ligand first with
Ti(NEt2)2(OiPr)2, and then SiCl4, see footnote 13.
[11] (a) For several other examples of this octahedral distortion see:
L.M. Kobriger, A.K. McMullen, P.E. Fanwick, I.P. Rothwell,
Polyhedron 8 (1989) 77;
[1] J. Seyden-Penne, Chiral Auxiliaries and Ligands in Asymmetric
Synthesis, Wiley, New York, 1995.
[2] (a) See, for example: F.G. Sernetz, R. Mulhaupt, S. Fokken, J.
¨
Okuda, Macromolecules 30 (1997) 1562;
(b) S. Fokken, T.P. Spaniol, H.-C. Kang, W. Massa, J. Okuda,
Organometallics 15 (1996) 5069;
(b) J.R. Clark, A.L. Pulvirenti, P.E. Fanwick, M. Sigalas, O.
Eisenstein, I.P. Rothwell, Inorg. Chem. 36 (1997) 3623.
[12] B.P. Santora, P.S. White, M.R. Gagne´, Organometallics 18 (1999)
2557.
(c) D. Takeuchi, T. Nakamura, T. Aida, Macromolecules 33
(2000) 725;
(d) J. Okuda, S. Fokken, T. Kleinhenn, T.P. Spaniol, Eur. J.
Inorg. Chem. (2000) 1321.;
[13] S.A. Waratuke, M.G. Thorn, P.L. Fanwick, A.P. Rothwell, I.P.
Rothwell, J. Am. Chem. Soc. 121 (1999) 9111.
(e) T.J. Davis, J. Balsells, P.J. Carroll, P.J. Walsh, Org. Lett. 3
(2001) 2161.
[3] S. Fokken, T.P. Spaniol, J. Okuda, F.G. Sernetz, R. Mulhaupt,
¨
[14] A.W. Duff, R.A. Kamarudin, M.F. Lappert, R.J.Norton, J.
Chem. Soc. Dalton Trans. (1986) 489.
Organometallics 16 (1997) 4240.
[4] (a) J.E. McMurry, T. Lectka, J.G. Rico, J. Org. Chem. 54 (1989)
3748;
[15] Attempts to measure comparative rates and thermodynamics of
binding dieneophile to linked and unlinked analogs (5/7 and 6/8)
were frustrated by unresolvable broad peaks.
(b) J.E. McMurry, Chem. Rev. 89 (1989) 1513.