A step too far? Assessment of the boroxide ligand in ring-opening polymerization

Article abstract of DOI:10.1021/om049520j

Dimesityl borinic acid, (mes)₂BOH [mes = 2,4,6-Me ₃C₆H₂], has been employed as a source of boroxide ligand in lithium, zinc, and magnesium chemistry; for structural comparisons, the alkoxide ligand [OCH(mes)₂]^- has also been used. The lithium compounds [Li{OB(mes)₂}-(L)_n] ₂ [1a, L = Et₂O, n = 1; 1b, L = Py, n = 1; 1c, L = MeCN, n = 2] have been structurally characterized, showing a common bimetallic core with bridging boroxide ligands that differ in the relative rotation of the BC₂ unit about the Li₂O₂ metallacycle. The zinc compounds [Zn{OR}Me]₂ [2, R = B(mes) ₂; 3, R = CH(mes)₂] have been synthesized from the reaction of ZnMe₂ with the corresponding ROH species. As for the Li salts, 2 and 3 display a dimeric solid-state structure with μ-OR ligands. Reaction of the in situ generated lithium boroxide 1a with MgBr ₂·Et₂O afforded an unprecedented tetrametallic compound, [Mg{OB(mes)2}-Br·LiBr(OEt₂)₂]₂ (4), in which the lithium bromide side product is retained in a tricyclic structure containing a Li(μ-Br)₂Mg{μ-OB(mes) ₂}Mg(μ-Br)₂Li core. The magnesium alkyl compounds [Mg{OB(mes)₂}R(THF)]₂ [5, R = Me; 6, R = Bu] were obtained from the reaction of the Li salt with the Grignard reagent, MgMeBr, and protonolysis of MgBu₂ with (mes)₂-BOH, respectively. X-ray crystallography showed formation of a dimer in each case, where in 6 only the n-butyl group was observed in the solid state. Attempts at combining the boroxide ligand with the widely used β-diketiminate ligand, [HC{C(Me)NAr}₂]^- (BDI), at a zinc center are discussed, and a brief study into the potential of selected zinc complexes to initiate the ring-opening polymerization of ε-caprolactone and rac-lactide is described.

Full text of DOI:10.1021/om049520j

Organometallics 2004, 23, 5159-5168
5159
A Step Too F a r ? Assessm en t of th e Bor oxid e Liga n d in
Rin g-Op en in g P olym er iza tion
Sarah C. Cole, Martyn P. Coles,* and Peter B. Hitchcock
Department of Chemistry, University of Sussex, Falmer, Brighton BN1 9QJ , U.K.
Received J uly 1, 2004
Dimesityl borinic acid, (mes)2BOH [mes ) 2,4,6-Me3C6H2], has been employed as a source
of boroxide ligand in lithium, zinc, and magnesium chemistry; for structural comparisons,
the alkoxide ligand [OCH(mes)2]- has also been used. The lithium compounds [Li{OB(mes)2}-
(L)n]2 [1a , L ) Et2O, n ) 1; 1b, L ) Py, n ) 1; 1c, L ) MeCN, n ) 2] have been structurally
characterized, showing a common bimetallic core with bridging boroxide ligands that differ
in the relative rotation of the “BC2” unit about the Li2O2 metallacycle. The zinc compounds
[Zn{OR}Me]2 [2, R ) B(mes)2; 3, R ) CH(mes)2] have been synthesized from the reaction of
ZnMe2 with the corresponding ROH species. As for the Li salts, 2 and 3 display a dimeric
solid-state structure with µ-OR ligands. Reaction of the in situ generated lithium boroxide
1a with MgBr2‚Et2O afforded an unprecedented tetrametallic compound, [Mg{OB(mes)2}-
Br‚LiBr(OEt2)2]2 (4), in which the lithium bromide side product is retained in a tricyclic
structure containing a “Li(µ-Br)2Mg{µ-OB(mes)2}Mg(µ-Br)2Li” core. The magnesium alkyl
compounds [Mg{OB(mes)2}R(THF)]2 [5, R ) Me; 6, R ) Bu] were obtained from the reaction
of the Li salt with the Grignard reagent, MgMeBr, and protonolysis of MgBu2 with (mes)2-
BOH, respectively. X-ray crystallography showed formation of a dimer in each case, where
in 6 only the n-butyl group was observed in the solid state. Attempts at combining the
boroxide ligand with the widely used â-diketiminate ligand, [HC{C(Me)NAr}2]- (BDI), at a
zinc center are discussed, and a brief study into the potential of selected zinc complexes to
initiate the ring-opening polymerization of ꢀ-caprolactone and rac-lactide is described.
1. In tr od u ction
metathesis polymerization (ROMP), it was shown that
the tacticity of polymeric materials was also strongly
influenced by which alkoxide group was used,4 demon-
strating that not only the reactivity but also the mech-
anism by which metals interact with substrates can be
influenced by the supposedly innocent “ancillary”
ligands.5
Alkoxide anions remain one of the most widely
applied ligands, and the chemistry surrounding their
metal complexes has grown steadily in the last fifty
years.1 The properties of these ligands are routinely
altered through derivitization of the O-substituent,
allowing the generation of different steric and electronic
environments at metal centers. For instance, sterically
demanding groups have been extensively used to restrict
the aggregation of metal alkoxides to molecular species
in contrast to cluster systems by disfavoring bridging
through the oxygen. Substituted phenyl frameworks,
affording aryloxide compounds, are often used in this
context,2 where the availability of a large number of
substituted phenols as precursors ensures access to a
wide number of variants.
In addition to varying the steric properties of alkoxide
ligands, it is possible to exercise control over the
electronic properties, which may have a profound effect
on the reactivity of the metal center. For example, it
was demonstrated at an early stage in the development
of the “Schrock” alkylidene metathesis catalysts, Mo-
(NAr)(CHR)(OR′)2, that employing electron-poor fluo-
roalkoxide ligands markedly increased the reactivity
toward certain substrates.3 In the area of ring-opening
An alternative approach to exercising control of the
donor properties of alkoxide-type ligands that circum-
vents the use of (often expensive) fluorinated alcohols
is to incorporate an element that has π-acceptor proper-
ties adjacent to oxygen atom. Siloxide [OSiR3]- and the
boroxide [OBR2]- ligands fit this criteria, where in the
latter class, the empty p-orbital on the boron is able to
accept electron density from the oxygen lone pairs,
affording an “electron-poor” alkoxide. Work with this
system was pioneered by Power and co-workers, in
which it was demonstrated that the boroxides [OBR2]-
(R ) mes ) 2,4,6-Me3C6H2; 2,4,6-iPr3C6H2) coordinated
to group 8 and 9 metals in either a bridging or terminal
coordination mode.6 More recently this ligand has been
used in main group systems,7 and a recent structural
report of the dimeric zinc compound [Zn{OB(mes)2}Et]2
has been presented.8
(3) Freudenberger, J .; Schrock, R. R.; Churchill, M.; Rheingold, A.;
Ziller, J . Organometallics 1984, 3, 1563. Schrock, R. R. Polyhedron
1995, 14, 3177.
* To whom correspondence should be addressed. Fax: 01273 677196.
Tel: 01273 877339. E-mail: m.p.coles@sussex.ac.uk.
(1) Bradley, D. C.; Mehrotra, R. C.; Gaur, D. P. Metal Alkoxides;
Academic: New York, 1978. Mehrotra, R. C. Adv. Inorg. Chem. 1983,
26, 269.
(4) Feast, W. J .; Gibson, V. C.; Marshall, E. L. J . Chem. Soc., Chem.
Commun. 1992, 1157.
(5) Schrock, R. R. Tetrahedron 1999, 55, 8141.
(6) Olmstead, M. M.; Power, P. P.; Sigel, G. Inorg. Chem. 1986, 25,
1027. Chen, H.; Power, P. P.; Shoner, S. C. Inorg. Chem. 1991, 30,
2884.
(2) Rothwell, I. P. Acc. Chem. Res. 1988, 21, 153.
10.1021/om049520j CCC: $27.50 © 2004 American Chemical Society
Publication on Web 09/25/2004

5160 Organometallics, Vol. 23, No. 22, 2004
Cole et al.
advantage of removing the steric bulk from the vicinity
of the metal, hopefully generating a site whereby
polymerization will initiate. We report here a series of
lithium, zinc, and magnesium complexes incorporating
the dimesitylboroxide ligand, [OB(mes)2]-. Structural
analyses have been performed to further our under-
standing of the way in which this anion interacts with
metals, and a preliminary assessment of the potential
for this ligand to initiate ROP of lactide is described.
2. Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. All manipulations
were carried out under dry nitrogen using standard Schlenk
and cannula techniques, or in a conventional nitrogen-filled
glovebox. Solvents were dried over appropriate drying agent
and degassed prior to use. The compounds (mes)2BOH,14
(mes)2CHOH,15 MgMe2,16 H2C{C(Me)NAr}2H(BDI)H,17 Zn-
(BDI)Me,18 and Zn(BDI)Et19 were synthesized according to
literature procedures. nBuLi (2.5 M solution in hexanes, Acros),
ZnMe2 (∼2 M solution in toluene, Fluka), MgBr2‚Et2O (Ald-
rich), and MgMeCl (3.0 M solution in THF, Aldrich) were
purchased from the indicated sources and used as received.
MgBu2 was purchased as a 1.0 M solution in heptane from
Aldrich. The volatiles were removed under vacuum, and the
residue was redissolved in THF prior to use.
F igu r e 1.
Work in our group has attempted to show that the
effect of the boron atom is real in both the solid state
and solution, using a series of d0-early transition metal
imido compounds (Figure 1).9 Crystallographic analysis
showed a trend for the lengthening of the metal-oxygen
bond in boroxide compounds, compared to the sterically
similar [OCH(mes)2]- alkoxide, and solution-state mea-
surements of the ∆δ value10 for Mo(NtBu)2{OB(mes)2}2
and Ti(NtBu){OB(mes)2}2(Py)2 were in agreement, sup-
porting the idea that boroxide ligands behave as electron-
deficient alkoxides.
NMR spectra were recorded using a Bruker Avance DPX
300 MHz spectrometer at 300 (1H) and 75 (13C{1H}) MHz.
Proton and carbon chemical shifts are internally referenced
to deuterated solvent and reported relative to TMS; coupling
constants, J , are quoted in Hz. Elemental analyses were
performed by S. Boyer at London Metropolitan University.
Gen er a l P r oced u r e for Syn th esis of th e Li Sa lts, 1a ,
1b, a n d 1c. A solution of (mes)2BOH (0.50 g, 1.88 mmol) in
ca. 35 mL of THF was cooled to 0 °C, and nBuLi (0.75 mL of a
2.5 M solution in hexanes, 1.88 mmol) was added. The mixture
was allowed to warm to room temperature and stirred for 1
h, after which time clear dark yellow solutions were observed.
The volatile components were removed in vacuo, and the
residue was dissolved in the relevant solvent and stored at
room temperature (1a and 1c) or 4 °C (1b), affording colorless
crystals of the solvated lithium boroxide. Accurate elemental
analyses proved difficult to obtain in the case of 1b and 1c.
[Li{OB(m es)2}(Et2O)]2 (1a ). Anal. Calc for C44H64B2O4Li2
(692.49): C, 76.32; H, 9.32. Found: C, 76.00; H, 9.51. 1H NMR
(C6D6, 298 K): δ 6.64 (s, 4H, C6H2), 3.13 (q, 4H, CH2), 2.15 (s,
6H, 4-Me), 2.09 (s, 12H, 2,6-Me2), 1.02 (t, 6H, CH3). 13C NMR
(C6D6, 298 K): δ 143.0 (br, C), 138.9 (CH), 136.9 (C), 128.9
(C), 65.8 (CH2, Et2O), 22.2 (CH3), 21.2 (CH3), 15.3 (CH3, Et2O).
There has been a recent surge in metal-alkoxide
chemistry fueled by the application of certain com-
pounds as well-defined initiators for ring-opening po-
lymerization (ROP) of cyclic monomers.11 In these
instances, the M-OR linkage provides the active site
for initiation via a coordination-insertion pathway, and
much of the work has focused therefore on the develop-
ment of suitable ancillary ligands that support this
moiety during the course of the polymerization. To date,
relatively nontoxic and cheap metals including alumi-
num, zinc, and magnesium have received the most
attention in this area, although more “exotic” metals
including lanthanum,12 neodymium, and gadolinium13
have also been investigated, demonstrating that this
type of chemical reactivity may be promoted by many
elements. As new systems are developed, the reactivity
of the metal-alkoxide unit will need to be addressed,
where, for example, a balance between adequate bulk
that deters cluster formation while maintaining an
active site that initiates ROP is required. We were
therefore interested in studying the potential for the
boroxide ligand to be applied in this context, given that
evidence from our earlier studies indicates the electron-
donating ability is similar to the aryloxide, which often
suffers from poor initiation due to the bulk. However,
the incorporation of the boron atom in boroxides has the
1
[Li{OB(m es)2}(P y)]2 (1b). H NMR (C6D6, 298 K): δ 8.11
(m, 2H, Py), 6.89 (m, 1H, Py), 6.85 (s, 4H, C6H2), 6.55 (m, 2H,
Py), 2.58 (s, 12H, 2,6-Me2), 2.24 (s, 6H, 4-Me). 13C NMR (C6D6,
298 K): δ 150.1 (CH, Py), 145.8 (br, C), 140.4 (C), 135.9 (CH,
Py), 135.6 (C), 128.4 (CH), 123.4 (CH, Py), 22.9 (CH3), 21.2
(CH3).
1
[Li{OB(m es)2}(MeCN)2]2 (1c). H NMR (C6D6, 298 K): δ
(7) Gibson, V. C.; Mastroianni, S.; White, A. J . P.; Williams, D. J .
Inorg. Chem. 2001, 40, 826. Anulewicz-Ostrowska, R.; Lulinski, S.;
Serwatowski, J .; Suwinska, K. Inorg. Chem. 2000, 39, 5763.
(8) Anulewicz-Ostrowska, R.; Lulinski, S.; Pindelska, E.; Serwa-
towski, J . Inorg. Chem. 2002, 41, 2525.
(9) Cole, S. C.; Coles, M. P.; Hitchcock, P. B. J . Chem. Soc., Dalton
Trans. 2002, 4168.
6.70 (s, 4H, C6H2), 2.32 (s, 12H, 2,6-Me2) 2.17 (s, 6H, 4-Me),
-0.54 (s, 6H, MeCN). 13C NMR (C6D6, 298 K): δ 143.7 (br, C),
(14) Brown, H. C.; Dodson, V. H. J . Am. Chem. Soc. 1957, 79, 2302.
(15) Grilli, S.; Lunazzi, L.; Mazzanti, A.; Casarini, D.; Femoni, C.
J . Org. Chem. 2001, 66, 488.
(10) Nugent, W. A.; McKinney, R. J .; Kasowski, R. V.; van-Catledge,
F. A. Inorg. Chim. Acta 1982, 65, L91.
(11) Darensbourg, D. J .; HoltcampM. W. Coord. Chem. Rev. 1996,
153, 155. O’Keefe, B. J .; Hillmyer, M. A.; Tolman, W. B. J . Chem. Soc.,
Dalton Trans. 2001, 2215. Coates, G. W. J . Chem. Soc., Dalton Trans.
2002, 467.
(12) Giesbrecht, G. R.; Whitener, G. D.; Arnold, J . Dalton Trans.
2001, 923.
(13) Luo, Y.; Yao, Y.; Shen, Q.; Sun, J .; Weng, L. J . Organomet.
Chem. 2002, 662, 144.
(16) Bailey, P. J .; Dick, C. M. E.; Fabre, S.; Parsons, S. J . Chem.
Soc., Dalton Trans. 2000, 1655.
(17) Stender, M.; Wright, R. J .; Eichler, B. E.; Prust, J .; Olmstead,
M. M.; Roesky, H. W.; Power, P. P. J . Chem. Soc., Dalton Trans. 2001,
3465. Feldman, J .; McLain, S. J .; Parthasarathy, A.; Marshall, W. J .;
Calabrese, J . C.; Arthur, S. D. Organometallics 1997, 16, 1514.
(18) Prust, J .; Stasch, A.; Zheng, W.; Roesky, H. W.; Alexpoulos, E.;
Uson, I.; Bohler, D.; Schuchardt, T. Organometallics 2001, 20, 3825.
(19) Cheng, M.; Moore, D. R.; Reczek, J . J .; Chamberlain, B. M.;
Lobkovsky, E. B.; Coates, G. W. J . Am. Chem. Soc. 2001, 123, 8738.

Boroxide Compounds of Li, Zn, and Mg
Organometallics, Vol. 23, No. 22, 2004 5161
140.1 (CH), 136.1 (C), 127.9 (C), 116.0 (CH3CN), 22.8 (CH3),
21.2 (CH3), 1.4 (CH3CN).
tion and storage at -30 °C afforded 6 as colorless crystals.
Yield: 0.320 g, (25%). Anal. Calc for C52H78B2Mg2O4 (837.41):
C, 74.59; H, 9.39. Found: C, 69.05; H, 9.06. 1H NMR (C6D6,
298 K): δ 6.76 (s, 4H, C6H2), 3.24 (br m, 4H, THF), 2.58 (s,
12H, 2,6-Me2), 2.14 (s, 6H, 4-Me), 1.60 (m, 2H, CH2), 1.44 (m,
2H, CH2), 1.17 (t, J ) 7.2, 3H, CH3), 1.04 (br m, 4H, THF),
-0.20 (m, 2H, MgCH2). 13C NMR (C6D6, 298 K): δ 141.9 (br,
C), 140.6 (C), 137.3 (C), 128.6 (CH), 69.2 (CH2, THF), 33.0
(CH2), 32.8 (CH2), 25.0 (CH3), 23.4 (CH3), 21.2 (CH2, THF),
14.6 (CH3), 9.8 (MgCH2).
Cr ysta llogr a p h y. Details of the crystal data, intensity
collection, and refinement for complexes 1a and 1c are listed
in Table 1, for compounds 2 and 3 in Table 3, and for
compounds 4, 5, and 6 in Table 5. Crystals were covered in
oil, and suitable single crystals were selected under a micro-
scope and mounted on a Kappa CCD diffractometer. The
structures were refined with SHELXL-97.20 Additional fea-
tures are described below.
[Li{OB(m es)2}(Et2O)]2 (1a ). The crystals diffracted only
very weakly. The molecule has a center of inversion.
[Li{OB(m es)2}(MeCN)2]2 (1c). The molecule lies on a
2-fold rotation axis.
[Zn {µ-OB(m es)2}Me]2 (2). The molecule lies on a crystal-
lographic mirror plane. The toluene solvate molecule was
unresolved and was modeled only by including four peaks as
carbon atoms at half occupancy, close to a mirror plane. The
large values for the largest diffraction peak and hole (0.95 and
-0.42 e Å3, respectively) were located near the disordered
solvate.
[Zn {OB(m es)2}Me]2 (2). A solution of (mes)2BOH (0.750
g, 2.82 mmol) in toluene (20 mL) was added dropwise at room
temperature to a stirred solution of ZnMe2 (1.41 mL of 2 M
solution in toluene, 2.82 mmol, further diluted with an
additional 20 mL of toluene), to afford a colorless solution and
a white precipitate. After stirring for 14 h under ambient
conditions the volatiles were removed under reduced pressure,
and the resultant white solid was extracted by filtration from
a small amount of insoluble material using toluene. Concen-
tration and storage at 4 °C yielded 2 as colorless crystals.
Yield: 0.914 g (94%). Anal. Calc for C45H58B2O2Zn2 (783.35)†:
C, 69.00; H, 7.46. Found: C, 68.30; H, 7.62. 1H NMR (C6D6,
298 K): δ 6.73 (s, 4H, C6H2), 2.35 (s, 12H, 2,6-Me2), 2.10 (s,
6H, 4-Me), -0.87 (s, 3H, ZnMe). 13C NMR (C6D6, 298 K): δ
140.1 (br, C), 138.4 (CH), 128.8 (C), 127.7 (C), 23.0 (CH3), 21.4
(CH3), -15.1 (ZnCH3) (†calculated for the mono-toluene sol-
vate).
[Zn {OCH(m es)2}Me]2 (3). Compound 3 was prepared using
the general procedure outlined for 2, using 0.350 g of
(mes)2CHOH (1.30 mmol) and 0.65 mL of a 2 M solution of
ZnMe2 in toluene (1.30 mmol). Concentration and storage at
room temperature yielded 3 as opaque white crystals. Yield:
0.378 g (42%). Anal. Calc for C40H52O2Zn2 (695.61): C, 69.07;
1
H, 7.53. Found: C, 69.10; H, 7.61. H NMR (C6D6, 298 K): δ
6.70 (s, 4H, C6H2), 6.65 (s, 1H, CH), 2.28 (s, 12H, 2,6-Me2) 2.08
(s, 6H, 4-Me), -0.75 (s, 3H, ZnMe). 13C NMR (C6D6, 298 K): δ
139.2 (br, C), 136.3 (C), 135.9 (C), 131.3 (CH), 78.0 (CH), 21.9
(CH3), 20.7 (CH3), -16.9 (ZnCH3).
[Zn {µ-OCH(m es)2}Me]2 (3). The molecule has a center of
inversion.
[Mg{OB(m es)2}Br ‚LiBr (Et2O)2]2 (4). nBuLi (0.61 mL of a
2.5 M solution in hexanes, 1.58 mmol) was added dropwise to
a cooled (0 °C) solution of (mes)2BOH (0.400 g, 1.50 mmol) in
Et2O (25 mL). The mixture was allowed to warm to room
temperature and stirred for 45 min, affording a fine white
precipitate of 1a . The slurried lithium salt was added to a
solution of MgBr2‚Et2O (0.387 g, 1.50 mmol) in Et2O (25 mL)
and the resultant mixture stirred for 14 h at room tempera-
ture. Removal of the volatiles afforded a white solid that was
extracted from a small amount of insoluble material with Et2O.
Concentration and storage at -30 °C yielded pure 4 as
colorless crystals. Yield: 0.812 g (60%). Anal. Calc for C36H44B2-
Br4Li2Mg2O2 (912.47)‡: C, 47.39; H, 4.86. Found: C, 47.49; H,
4.81 (‡formula corresponds to the desolvated complex, Mg{OB-
(mes)2}Br‚LiBr).
[Mg{OB(m es)2}Br ‚LiBr (OEt2)2]2 (4). Two independent
molecules were present in the unit cell. Two of the Et2O ligands
showed partial disorder and were included with isotropic O
and C atoms and SADI bond length constraints.
[Mg{OB(m es)2}Me(THF )]2 (5). The molecule has a center
of inversion.
[Mg{OB(m es)2}n Bu (THF )]2 (6). The molecule has a center
of inversion.
3. Resu lts a n d Discu ssion
3.1. Lith iu m Bor oxid e Com p ou n d s. Synthesis of
metal boroxide compounds employing a borinic acid as
the ligand precursor has been achieved using two
general synthetic protocols. Protonolysis of a metal-
carbon bond with the parent acids, R2BOH, provides a
clean route to the target species due to the volatile
hydrocarbon side products, while transmetalation em-
ploying the lithiated reagent, [Li{OBR2}(solvent)x]n, is
a valuable approach due to the large number of metal-
halide starting materials. Work in our group toward
establishing the influence that boron exerts on the
bonding parameters in titanium and molybdenum borox-
ide complexes9 successfully utilized the latter approach,
employing the imido complexes Mo(NR)2Cl2(DME)21 and
[Mg{OB(m es)2}Me(THF )]2 (5). nBuLi (0.61 mL of a 2.5 M
solution in hexanes, 1.58 mmol) was added dropwise to a cooled
(0 °C) solution of (mes)2BOH (0.400 g, 1.50 mmol) in Et2O (25
mL). The mixture was allowed to warm to room temperature
and stirred for 45 min, affording a fine white precipitate of
1a . The slurried lithium salt was added to a cooled (-78 °C)
solution of MgMeCl (0.50 mL of a 3 M solution in THF, 1.50
mmol, further diluted by addition of 20 mL of Et2O). After
stirring for 14 h under ambient conditions the volatiles were
removed under reduced pressure, affording a white solid.
Compound 5 was extracted from the lithium chloride side
product with hot toluene (100 °C). Concentration and storage
at 4 °C yielded 5 as colorless crystals. Yield: 0.167 g (30%).
1H NMR (CD2Cl2, 298 K): δ 6.70 (s, 4H, C6H2), 3.32 (br m,
4H, THF), 2.29 (s, 12H, 2,6-Me2), 2.20 (s, 6H, 4-Me), 1.62 (br
m, 4H, THF), -1.86 (s, 3H, CH3). 13C NMR (CD2Cl2, 298 K):
141.6 (br, C), 140.4 (C), 137.4 (C), 128.2 (CH), 69.3 (CH2, THF),
25.4 (CH2, THF), 23.0 (CH3), 21.1 (CH3), -14.4 (MgCH3).
[Mg{OB(m es)2}n Bu (THF )]2 (6). A solution of (mes)2BOH
(0.400 g, 1.50 mmol) in THF (20 mL) was added dropwise to a
stirred solution of MgBu2‚THF (1.50 mL of a 1.0 M solution
in THF, 1.50 mmol, further diluted with an additional 20 mL
of THF) at -78 °C. The mixture was allowed to warm to room
temperature and stirred for 14 h. Removal of the volatiles
afforded a sticky white solid that was extracted from a small
amount of insoluble white material with hexane. Concentra-
22
Ti(NR)Cl2(Py)3 as starting reagents. In these in-
stances, synthesis of the compounds Mo(NR)2{OB-
(mes)2}2 and Ti(NR){OB(mes)2}2(Py)2 proved most con-
venient using lithiated reagents that were generated in
situ, although in previous studies in other groups,
[LiOB{CH(SiMe3)2}2]2,23 [Li{OB(mes)2}(THF)]2,24 and
(20) Sheldrick, G. M. SHELXL-97, Program for the Refinement of
Crystal Structures; Go¨ttingen, 1997.
(21) Bell, A.; Clegg, W.; Dyer, P. W.; Elsegood, M. R. J .; Gibson, V.
C.; Marshall, E. L. J . Chem. Soc., Chem. Commun. 1994, 2547.
(22) Blake, A. J .; Collier, P. E.; Dunn, S. C.; Li, W.-S.; Mountford,
P.; Shishkin, O. V. J . Chem. Soc., Dalton Trans. 1997, 1549.

5162 Organometallics, Vol. 23, No. 22, 2004
Cole et al.
Sch em e 1
[Li{OB(fmes)2}(THF)]225 (fmes ) 2,4,6-(CF3)3C6H2) have
been isolated and structurally characterized.
selected bond lengths and angles are given in Table 2.
A poor quality data set for the pyridine adduct, 1b, was
of insufficient accuracy to allow meaningful discussion
of bond lengths and angles. It did however confirm the
molecular structure as consisting of the dimeric monoad-
duct, [Li{OB(mes)2}(Py)]2.
During the course of our work a further three ex-
amples of lithium boroxide complexes, [Li{OB(mes)2}-
(Et2O)]2, 1a, [Li{OB(mes)2}(Py)]2, 1b, and [Li{OB(mes)2}-
(MeCN)2]2, 1c, were isolated and structurally charac-
terized (Scheme 1). Significant differences are evident
in the molecular structures, dependent on the coordi-
nated solvent, warranting a brief discussion. The mo-
F igu r e 2. (a) Molecular structure of 1a (thermal ellipsoids
30%; ′:-x, -y, -z). (b) Core of 1a viewed along the
transannular O‚‚‚‚O vector of the Li2O2 metallacycle.
F igu r e 3. (a) Molecular structure of 1c (thermal ellipsoids
30%; ′:-x, y, -z+1/2). (b) Core of 1c viewed along the
transannular O‚‚‚‚O vector of the Li2O2 metallacycle.
lecular structures of 1a and 1c are illustrated in Figures
As in previously reported structures of this general
2 and 3, crystal data are summarized in Table 1, and
type, lithium salts 1a -c consist of dimeric molecules

Boroxide Compounds of Li, Zn, and Mg
Organometallics, Vol. 23, No. 22, 2004 5163
Ta ble 1. Cr ysta l Str u ctu r e a n d Refin em en t Da ta
for [Li{OB(m es)2}(Et2O)]2 (1a ) a n d
[Li{OB(m es)2}(MeCN)2]2 (1c)
1a a
1c
formula
fw
C
44H64B2Li2O4
C
44H56B2Li2N4O2
692.45
173(2)
0.71073
708.43
173(2)
0.71073
temperature (K)
wavelength (Å)
cryst size (mm)
cryst syst
space group
a (Å)
0.10 × 0.05 × 0.01 0.30 × 0.30 × 0.20
monoclinic
P21/n (No.14)
9.0303(4)
9.1269(4)
25.4398(11)
91.214(2)
2096.2(2)
2
monoclinic
C2/c (No.15)
27.7092(13)
9.4616(6)
16.3004(7)
91.119(3)
4272.7(4)
4
b (Å)
c (Å)
â (deg)
V (Å3)
Z
D
calc (Mg/m3)
1.10
0.07
3.91 to 22.98
1.10
0.07
3.82 to 25.04
F igu r e 4. Molecular structure of 2 (thermal ellipsoids
30%; ′: x, -y+3/2, z).
absorp coeff (mm-1
θ range for data
collection (deg)
)
solid state, a consequence of the different symmetry
elements (i.e., 1a has a center of inversion; 1c lies on a
2-fold rotation axis). Therefore in 1a , each -B(mes)2
group is effectively rotated in an opposite direction along
the B-O bond, with respect to the Li2O2 plane, while
in 1c, both groups are rotated in the same direction.
Another consequence of locating the solvate in the Li2O2
plane within 1a is a large distortion from ideal trigonal
planar geometry at the oxygen atom [Li-O(1)-B, 148.6-
(4)°; Li′-O(1)-B, 126.3(4)°], whereas equivalent angles
are observed in 1c. The corresponding Li-O bond
distances are essentially identical for 1a and 1c (within
esd’s), although notable asymmetry is present within
the boroxide bridge in 1a [Li-O(1) ) 1.827(8) Å; Li-
O(2) ) 1.889(9) Å].
no. of reflns collected
no. of indep reflns
no. of reflns with
I>2σ(I)
11 272
10 157
2766 [Rint ) 0.129] 3764 [Rint ) 0.058]
1871
2616
no. of data/restraints/ 2766/0/239
params
3764/0/275
goodness-of-fit on F2
final R indices
[I>2σ(I)]
1.157
R1 ) 0.090
1.049
R1 ) 0.056
wR2 ) 0.191
R1 ) 0.135
wR2 ) 0.212
0.28 and -0.23
wR2 ) 0.139
R1 ) 0.086
wR2 ) 0.157
0.22 and -0.18
R indices (all data)
largest diff peak and
hole (e Å-3
)
a
Very weak diffraction.
Ta ble 2. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for [Li{OB(m es)2}(Et2O)]2 (1a ) a n d
[Li{OB(m es)2}(MeCN)2]2 (1b)
3.2. Zin c Bor oxid e Com p ou n d s. Initial studies
toward the introduction of a boroxide ligand at a zinc
center focused on the reaction of the in situ generated
lithium salts, [Li{OB(mes)2}(solvent)x]n, with zinc di-
bromide. Analysis of the crude product by 1H NMR
spectroscopy indicated three different environments for
the mesityl groups, and no clean product could be
isolated from the reaction. The protonolysis reaction
between 1 equiv of (mes)2BOH and ZnMe2, however,
proceeded cleanly at room temperature to afford color-
less crystals of 2 in excellent yield (Scheme 1). NMR
data and combustion analysis were consistent with
formation of the desired mixed alkylzinc boroxide
complex, [Zn{OB(mes)2}Me]n,26 although no data con-
cerning the extent of aggregation could be obtained by
mass spectral analysis (EI+), the highest molecular
weight fragment observed corresponding to the ligand
[OB(mes)2]+.
1a
1b
B-O
1.331(6)
1.827(8)
1.889(9)
96.5(4)
1.315(3)
1.860(4)
1.892(4)
97.4(3)c
95.2(3)d
83.66(19)
Li-Oa
Li-Ob
O-Li-O
Li-O-Li
83.5(4)
a
b
1a Li-O(1), 1b Li(1)-O. 1a Li-O(1′), 1b Li(2)-O. c O-
d
Li(1)-O′. O-Li(2)-O′.
containing a bimetallic Li2O2 core, with bridging borox-
ide ligands. The coordination sphere at lithium is
completed by solvent in each case where, due to the
smaller size and “rodlike” shape of acetonitrile, two
molecules are accommodated in 1c, affording distorted
tetrahedral centers. In contrast, a single solvate is
present in 1a and 1b, generating a distorted trigonal
planar lithium with slight pyramidalization [∑angles
)
358.5°] for the Et2O adduct. 1c therefore extends this
series of compounds of general formula [Li{µ-OBR2}-
(solvent)x]2, from x ) 0 [R ) CH(SiMe3)2]23 and x ) 1 [R
) mes; solvent ) THF,24 Et2O (1a ), Py (1b)] to x ) 2 [R
) mes; solvent ) MeCN (1c)].
The Li2O2 metallacycles of 1a and 1c are both planar,
with the major structural difference being the way in
which the mesityl substituents are positioned in the
The molecular structure of 2 is illustrated in Figure
4, crystal data are summarized in Table 3, and selected
bond lengths and angles are given in Table 4. Analogous
to the zinc ethyl analogue,8 compound 2 crystallizes as
the dimeric complex [Zn{µ-OB(mes)2}Me]2. The zinc
atom is distorted trigonal planar [∑angles ) 359.96°],
coordinated to two bridging boroxides and a terminal
methyl group. The bridging coordination mode for the
boroxide ligand is unremarkable6,7,24 and may be favored
in these systems by the presence of the boron, which
(23) Beck, G.; Hitchcock, P. B.; Lappert, M. F.; Mackinnon, I. A. J .
Chem. Soc., Chem. Commun. 1989, 1312.
(24) Weese, K. J .; Bartlett, R. A.; Murray, B. D.; Olmstead, M. M.;
Power, P. P. Inorg. Chem. 1987, 26, 2409.
(25) Gibson, V. C.; Redshaw, C.; Clegg, W.; Elsegood, M. R. J .
Polyhedron 1997, 16, 2637.
(26) Cole, S. C.; Coles, M. P.; Hitchcock, P. B. Dalton Trans. 2003,
3663.

5164 Organometallics, Vol. 23, No. 22, 2004
Cole et al.
Ta ble 3. Cr ysta l Str u ctu r e a n d Refin em en t Da ta
for [Zn {OB(m es)2}Me]2 (2) a n d
[Zn {OCH(m es)2}Me]2 (3)
2
3
formula
fw
C38H50B2O2Zn2.(C7H8) C40H52O2Zn2
783.27
695.56
temperature (K)
wavelength (Å)
cryst size (mm)
cryst syst
space group
a (Å)
173(2)
173(2)
0.71073
0.71073
0.2 × 0.2 × 0.2
orthorhombic
Pnma (No. 62)
16.3684(2)
26.5187(5)
9.7433(2)
90
0.2 × 0.2 × 0.1
triclinic
P1h (No. 2)
8.0099(1)
10.5839(2)
11.6288(2)
105.530(1)
108.855(1)
93.507(1)
887.01(3)
1
b (Å)
c (Å)
R (deg)
â (deg)
90
90
γ (deg)
V (Å3)
4229.3(1)
4
1.23
1.17
3.92 to 25.22
Z
D
F igu r e 5. Molecular structure of 3 (thermal ellipsoids
30%; ′: -x+1, -y+1, -z+1).
calc (Mg/m3)
1.30
1.38
3.80 to 25.05
absorp coeff (mm-1
θ range for data
collection (deg)
)
Me units. In contrast, the related aryloxide [Zn(µ-
Omes*)(CH2SiMe3)]2 contains one long [2.021(4) Å] and
one short [1.958(4) Å] Zn-O bond and may therefore
be considered as consisting of a more loosely held dimer,
presumably on account of the steric bulk at zinc.
To evaluate the electronic influence of the boron atom
on the metal-oxygen bond in 2, we investigated the
protonolysis reaction between the alcohol, (mes)2CHOH,15
and ZnMe2 (eq 1). We have previously described the use
no. of reflns collected
no. of indep reflns
22824
3686 [Rint
0.074]
9558
3094 [Rint
0.034]
)
)
no. of reflns with
I>2σ(I)
2859
2862
no. of data/restraints/ 3686/0/227
params
3094/0/205
goodness-of-fit on F2
final R indices
[I>2σ(I)]
1.055
0.575
R1 ) 0.056,
wR2 ) 0.171
R1 ) 0.075,
wR2 ) 0.185
0.95 and -0.42
R1 ) 0.026,
wR2 ) 0.070
R1 ) 0.029,
wR2 ) 0.074
0.25 and -0.37
R indices (all data)
largest diff peak and
hole (e Å-3
)
Ta ble 4. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for [Zn {OB(m es)2}Me]2 (2) a n d
[Zn {OCH(m es)2}Me]2 (3)
2
3
Zn-Oa
1.965(3)
1.970(3)
1.922(7)
1.921(6)
1.353(5)
1.954(1)
1.946(1)
1.938(2)
of this group as a system sterically comparable to
dimesityl boroxide,9 enabling a more representative
investigation of the electronic influences of the B atom
to be conducted. The reaction proceeded cleanly to afford
colorless crystals of [Zn{OCH(mes)2}Me]n (3). The key
features of the 1H NMR spectrum are the low-field shift
of δ 6.65 observed for the CH of the alkoxide and the
high-field ZnMe resonance, δ -0.75. The molecular
structure is illustrated in Figure 5, crystal data are
summarized in Table 3, and selected bond lengths and
angles are given in Table 4.
Compound 3 is also dimeric in the solid state, with
the molecular formula [Zn{µ-OCH(mes)2}Me]2. The
Zn2O2 metallacycle is planar, consisting of distorted
trigonal zinc and oxygen atoms, with the methyl carbon
atom displaced from this plane by 0.16 Å. The uncer-
tainty in the metal-oxygen bond distances for 2 and 3
precludes any definitive conclusions from being made
regarding a possible reduction in bond order for the
boroxide, although a more pronounced tendency toward
inequivalent zinc-oxygen bonds [1.946(1) and 1.954(1)
Å] is noted in 3. The zinc-carbon distance [1.938(2) Å]
is experimentally the same as the corresponding values
in 2 [1.922(7) and 1.921(6) Å], again preventing any
conclusions to be made concerning the effect of the boron
atom.
Zn-Ob
Zn-Cc
Zn(2)-C(20)
B-O
C(1)-O
1.423(2)
79.46(6)
O-Zn-Od
O-Zn-Oe
O-Zn-Cf
O-Zn-Cg
Zn-O-Znh
81.26(15)
81.00(15)
139.35(8)
139.48(8)
98.85(12)
138.28(8)
142.22(8)
100.54(6)
2 Zn(1)-O(1), 3 Zn-O. 2 Zn(2)-O(1), 3 Zn-O′. c 2 Zn(1)-
a
b
C(19), 3 Zn-C(20). 2 O(1)-Zn(1)-O(1′). e 2 O(1)-Zn(2)-O(1′).f 2
d
g
O(1)-Zn(1)-C(19), 3 O-Zn-C(20). 2 O(1)-Zn(2)-C(20), 3 O′-
h
Zn-C(20). 2 Zn(1)-O(1)-Zn(2), 3 Zn-O-Zn′.
displaces the bulk of the mesityl groups from the anionic
oxygen atom, allowing the interaction with two metal
fragments.
The planarity of the metallacycle extends to include
the boron atom and terminal methyl substituent (max.
deviation from the plane ) 0.09 Å) with a twist angle
between the metallacycle and the plane defined by the
-BC2 fragment of 13.95(11)°. The internal angles at zinc
and oxygen [81.26(15)°/81.00(15)° and 98.85(12)°, re-
spectively] are similar to the corresponding angles in
the zinc ethyl complex8 and in the related dimeric
alkoxides [Zn(µ-OAr)(CH2SiMe3)]2 (Ar ) 2,6-iPr2C6H3
and 2,4,6-tBu3C6H2).27 The Zn-O bond lengths are equal
within experimental error [1.965(3) and 1.970(3) Å],
suggesting a strong association of the Zn{OB(mes)2}-
(27) Olmstead, M. M.; Power, P. P.; Shoner, S. C. J . Am. Chem. Soc.
1991, 113, 3379.

Boroxide Compounds of Li, Zn, and Mg
Organometallics, Vol. 23, No. 22, 2004 5165
the lithium-magnesium mixed amide species [{LiMg-
(TMP)[CH2SiMe2N(SiMe3)]}2].29 This structure is how-
ever fundamentally different, as the central component
consists of a 5-fold system of fused four-membered rings
rather than the three distinct metallacycles present in
4. The geometry at magnesium is distorted tetrahedral,
with bond angles in the range 83.83(16)-128.37(14)°
and 84.08(16)-125.65(14)° for Mg(1) and Mg(2), respec-
tively. The most acute angle at magnesium is within
the central Mg2O2 metallacycle, which forms an es-
sentially planar component (max. deviation from the
plane ) 0.015 Å). Large twist angles to the two MgBr2-
Li rings are noted [74.33° and 77.25° for Mg(1)-Br(1/
2)-Li(1) and Mg(2)-Br(3/4)-Li(2) respectively], where
these two units are rotated in opposite direction relative
to each other (Figure 6b), presumably to minimize steric
interactions. The Mg-O bond lengths within the core
[1.96 Å av] are slightly contracted compared with the
corresponding lengths in the related dimers [Mg{OB-
n
(mes)2}R(THF)]2 (R ) Me, Bu, vide infra).
Making use of readily available Grignard reagents,
formation of the Mg-O linkage can be envisaged in two
ways: transmetalation of the Mg-X bond with lithiated
borinic acid to afford [Mg{OB(mes)2}R]n and protonoly-
sis of the Mg-C bond to give [Mg{OB(mes)2}X]n (X )
halide). The Grignard reagent, MgMeCl, was investi-
gated as an alternative source of magnesium employing
the transmetalation protocol where, if successful, the
reaction would afford a potential catalyst, directly
building on Chen’s observation that methyl ligands
behave as ancillary groups in Al complexes active for
ROP.30 The reaction afforded a colorless crystalline
material, which from 1H and 13C NMR spectroscopic
data was consistent with the monosolvated alkylmag-
nesium boroxide [MgMe{OB(mes)2}(THF)]n (5). How-
ever, repeated attempts at obtaining accurate elemental
analysis failed, due we believe to the high sensitivity of
this class of compound and incomplete separation from
the lithium chloride side product during workup. Struc-
tural analysis of a representative crystal, however,
confirmed the product as the mono-THF adduct (vide
infra).
F igu r e 6. (a) Molecular structure of 4 (thermal ellipsoids
30%). (b) View of the tricyclic core of 4.
3.3. Ma gn esiu m Bor oxid e Com p ou n d s. Magne-
sium is another metal of interest in the context of
initiators for the ring-opening polymerization of cyclic
esters.28 Initial attempts at the synthesis of magnesium
boroxide catalysts focused on a salt metathesis route
employing MgBr2‚Et2O as the source of metal. Addition
of a slurry of 1 equiv of the in situ generated salt [Li-
{OB(mes)2}(Et2O)x]n (1a ) to a cooled solution of MgBr2‚
Et2O afforded the lithium magnesate [Mg{OB(mes)2}Br‚
LiBr(Et2O)2]n (4). The low solubility and apparent
decomposition of 4 in a range of solvents precluded the
collection of NMR data. However, combustion analysis
on a sample that had been exposed to vacuum for 6 h
was consistent with the desolvated species Mg{OB-
(mes)2}Br‚LiBr, and crystals suitable for X-ray analysis
were obtained from cooling a saturated Et2O solution
to -35 °C. The molecular structure is illustrated in
Figure 6, crystal data are summarized in Table 5, and
selected bond lengths and angles are given in Table 6.
Two molecules are present in the unit cell with small
differences in the bond lengths and angles; for brevity,
values referring to only one molecule will be referred
to in the discussion.
Further difficulties associated with this methodology
became apparent, as, despite numerous attempts, the
synthesis of 5 could not be repeated using this protocol.31
Despite being able to identify the product in the crude
1H NMR spectrum on occasion, isolation of the pure
compound was not successful due, we believe, to the
high sensitivity of this synthetic approach to reaction
conditions and the possible instability of the products.
In addition, difficulties in purification via selective
crystallization (due to the similar solubility properties
of the magnesium compound and the lithium halide side
products) resulted in low yields and inaccurate elemen-
tal analyses. Alternative methods for the introduction
of the boroxide ligand at magnesium were therefore
investigated.
Compound 4 exists as the dimeric magnesate [Mg-
{OB(mes)2}Br‚LiBr(Et2O)2]2, containing the previously
unreported tetrametallic Li(µ-X)2Mg{µ-OR}2Mg(µ-X)2Li
core [X ) Br and R ) B(mes)2] formed of four tetrahe-
dral metal centers bridged by pairs of bromine or oxygen
atoms. The only other structurally characterized ex-
ample containing this arrangement of metal atoms is
(29) Barr, L.; Kennedy, A. R.; MacLellan, J . G.; Moir, J . H.; Mulvey,
R. E.; Rodger, P. J . A. J . Chem. Soc., Chem. Commun. 2000, 1757.
(30) Chakraborty, D.; Chen, E. Y. X. Organometallics 2002, 21, 1438.
(31) Variations in conditions and reagents that were investigated
in an attempt to overcome these problems included the use of fresh
batches of recrystallized ligand, isolation of the lithium salt rather than
reacting the reagents in situ, changing the Grignard from the chloride
to the bromide, and performing the reaction in THF or Et2O.
(28) Chisholm, M. H.; Huffman, J . C.; Phomphrai, K. J . Chem. Soc.,
Dalton Trans. 2001, 222.

5166 Organometallics, Vol. 23, No. 22, 2004
Cole et al.
Ta ble 5. Cr ysta l Str u ctu r e a n d Refin em en t Da ta for [Mg{OB(m es)2}Br ‚LiBr (OEt2)2]2 (4)
[Mg{OB(m es)2}Me(THF )]2 (5), a n d [Mg{OB(m es)2}n Bu (THF )]2 (6)
4
5
6
formula
fw
C
52H84B2Br4Li2Mg2O6
C
46H66B2Mg2O4
C
52H78B2Mg2O4
1208.95
173(2)
0.71073
0.40 × 0.40 × 0.40
monoclinic
P2/c (No. 13)
23.7839(3)
16.7263(2)
32.1496(3)
90
98.462(1)
90
12650.4(2)
8
1.27
753.23
173(2)
0.71073
837.38
173(2)
0.71073
temperature (K)
wavelength (Å)
cryst size (mm)
cryst syst
space group
a (Å)
0.30 × 0.30 × 0.20
monoclinic
P21/n (No. 14)
14.9529(2)
10.4052(2)
15.4761(2)
90
114.518(1)
90
2190.78(6)
2
0.35 × 0.35 × 0.15
triclinic
P1h (No. 2)
10.2085(5)
11.3243(6)
12.9560(7)
111.292(2)
94.887(2)
111.725(2)
1253.99(11)
1
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
D
calc (Mg/m3)
1.14
0.10
1.11
0.09
absorp coeff (mm-1
)
2.61
θ range for data collection (deg)
no. of reflns collected
3.71 to 25.70
169 906
3.74 to 25.03
22 701
3.70 to 25.64
20 077
no. of indep reflns
23 430 [Rint ) 0.097]
3839 [Rint ) 0.048]
4585 [Rint ) 0.063]
no. of reflns with I>2σ(I)
no. of data/restraints/params
goodness-of-fit on F2
16 672
3283
3447
23 430/56/1206
0.950
3839/0/249
1.032
4585/0/277
1.044
final R indices [I>2σ(I)]
R1 ) 0.061, wR2 ) 0.163
R1 ) 0.095, wR2 ) 0.189
0.69 and -0.68
R1 ) 0.050, wR2 ) 0.129
R1 ) 0.060, wR2 ) 0.136
0.40 and -0.51
R1 ) 0.058, wR2 ) 0.125
R1 ) 0.087, wR2 ) 0.138
0.22 and -0.23
R indices (all data)
largest diff peak and hole (e Å-3
)
Ta ble 6. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for [Mg{OB(m es)2}Br ‚LiBr (OEt2)2]2 (4)
The final magnesium reagent used in this study was
dibutylmagnesium, commercially available as the 1:1
mixture of n-butyl and sec-butyl species in heptane.32
Initial protonolysis reactions using this solution were
unsuccessful, and so in an attempt to overcome this, a
1.0 M THF solution was made.33 One equivalent of
(mes)2BOH was added at room temperature and the
crystallized product analyzed using spectroscopic and
Mg(1)-O(1)
Mg(1)-O(2)
Mg(1)-Br(1)
Mg(1)-Br(2)
Li(1)-Br(1)
Li(1)-Br(2)
B(1)-O(1)
1.966(4)
1.969(4)
2.4776(18)
2.474(2)
2.552(13)
2.576(11)
1.357(7)
Mg(2)-O(1)
Mg(2)-O(2)
Mg(2)-Br(3)
Mg(2)-Br(4)
Li(2)-Br(3)
Li(2)-Br(4)
B(2)-O(2)
1.966(4)
1.960(4)
2.4674(19)
2.472(4)
2.566(12)
2.588(12)
1.346(7)
1
X-ray diffraction techniques. H and 13C NMR spectra
O(1)-Mg(1)-O(2)
83.83(16) O(1)-Mg(2)-O(2)
84.08(16)
were consistent with the formula [Mg{OB(mes)2}Bu-
(THF)]n, and the resonance at δ -0.20 in the proton
NMR spectrum is attributable to the R-CH2 of the
n-butyl ligand, with no corresponding resonance in the
region of ∼+0.2 ppm predicted for the sec-butyl isomer.34
These data suggest that the nBu isomer crystallizes
preferentially over the sBu one, subsequently confirmed
by X-ray crystallography. This is in contrast to the
related compound [Mg{OAr′}Bu]2 (Ar′ ) 2,6-tBu2C6H3),
where the molecular structure consists of the sec-butyl
isomer.35
Early studies of the solution-state structures of alkyl-
magnesium alkoxides concluded that bulky alkoxides
favor formation of solvated dimers, while smaller alkoxy
groups form unsolvated aggregates containing between
four and seven monomeric units.36 Only a few examples
of structurally characterized [Mg{OR}R′]n complexes
have been reported, showing a number of different
structural types. For example, dimeric aryloxide com-
pounds [Mg{OAr′′}(n-hex)]2 (Ar′′ ) 2,6-tBu2-4-Me-
C6H2)37 and the aforementioned sec-butyl complex [Mg-
O(1)-Mg(1)-Br(1) 108.80(13) O(1)-Mg(2)-Br(3) 110.76(13)
O(1)-Mg(1)-Br(2) 128.37(14) O(1)-Mg(2)-Br(4) 125.65(14)
O(2)-Mg(1)-Br(1) 127.40(14) O(2)-Mg(2)-Br(3) 125.25(14)
O(2)-Mg(1)-Br(2) 110.95(14) O(2)-Mg(2)-Br(4) 110.24(13)
Br(1)-Mg(1)-Br(2) 100.24(7) Br(3)-Mg(2)-Br(4) 102.45(7)
O(1)-B(1)-C(1)
O(1)-B(1)-C(10)
C(1)-B(1)-C(10)
Mg(1)-O(1)-B(1)
Mg(2)-O(1)-B(1)
119.1(5)
120.1(5)
120.7(5)
132.2(4)
131.8(4)
O(2)-B(2)-C(19)
O(2)-B(2)-C(28)
119.6(5)
119.3(5)
C(19)-B(2)-C(28) 121.5(3)
Mg(1)-O(2)-B(2)
Mg(2)-O(2)-B(2)
132.3(4)
131.6(4)
Mg(1)-O(1)-Mg(2) 95.97(17) Mg(1)-O(2)-Mg(2) 96.06(17)
Mg(1)-Br(1)-Li(1) 81.8(3)
Mg(1)-Br(2)-Li(1) 81.4(3)
Br(1)-Li(1)-Br(2)
Mg(2)-Br(3)-Li(2) 80.7(3)
Mg(2)-Br(4)-Li(2) 80.1(3)
95.6(4)
Br(3)-Li(2)-Br(4)
96.7(4)
The first attempt at developing an alternative method
for the synthesis of Mg boroxide compounds was the
reaction of (mes)2BOH with MgMeCl, predicted to
proceed via protonolysis to afford [Mg{OB(mes)2}Cl]n.
The reaction resulted in formation of a highly insoluble
white powder, which resisted all attempts at purifica-
tion. The next approach employed MgMe2 as the start-
ing reagent, also predicted to react via protonolysis with
(mes)2BOH, in an alternative route to 5. Initial reactions
in toluene were hampered by poor solubility of both
reagents and products, and in an attempt to resolve this,
the THF adduct of MgMe2 was employed.16 In this case,
pure 5 could be isolated in very low (∼5%) yield,
demonstrating that this route is successful for the
synthesis of magnesium boroxides, although the small
amount of isolated material makes it impractical as a
viable entry point to the study of this chemistry.
(32) Duff, A. W.; Hitchcock, P. B.; Lappert, M. F.; Taylor, R. G.;
Segal, J . A. J . Organomet. Chem. 1985, 293, 271.
(33) Tenorio, M. J .; Puerta, M. C.; Salcedo, I.; Valerga, P. J . Chem.
Soc., Dalton Trans. 2001, 653.
(34) Dove, A. P.; Gibson, V. C.; Hormnirun, P.; Marshall, E. L.; Segal,
J . A.; White, A. J . P.; Williams, D. J . Dalton Trans. 2003, 3088.
(35) Henderson, K. W.; Honeyman, G. W.; Kennedy, A. R.; Mulvey,
R. E.; Parkinson, J . A.; Sherrington, D. C. Dalton Trans. 2003, 1365.
(36) Coates, G. E.; Ridley, D. Chem. Commun. 1966, 560. Ashby, E.
C.; Nackashi, J .; Parris, G. E. J . Am. Chem. Soc. 1975, 97, 3162.
(37) Gromada, J .; Mortreux, A.; Chenal, T.; Ziller, J . W.; Leising,
F.; Carpentier, J . F. Chem.-Eur. J . 2002, 8, 3773.

Boroxide Compounds of Li, Zn, and Mg
Organometallics, Vol. 23, No. 22, 2004 5167
Ta ble 7. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for [Mg{OB(m es)2}Me(THF )]2 (5) a n d
[Mg{OB(m es)2}n Bu (THF )]2 (6)
5
6
Mg-O(1)
Mg-O(1′)
Mg-C(23)
Mg-O(2)
B-O(1)
2.0077(15)
1.9910(14)
2.117(2)
2.0342(15)
1.347(2)
2.006(2)
2.005(2)
2.135(3)
2.040(2)
1.356(3)
O(1)-Mg-O(1′)
O(2)-Mg-O(1′)
O(1)-Mg-O(2)
O(1)-Mg-C(23)
O(1′)-Mg-C(23)
O(2)-Mg-C(23)
O(1)-B-C(1)
O(1)-B-C(10)
C(1)-B-C(10)
Mg-O(1)-B
Mg′-O(1)-B
Mg-O(1)-Mg′
84.31(6)
84.77(7)
113.78(6)
114.05(6)
116.89(9)
122.25(9)
105.07(9)
118.66(7)
118.93(17)
122.40(16)
126.45(13)
137.11(13)
95.69(6)
112.79(7)
109.04(7)
121.42(9)
118.40(9)
108.68(9)
118.11(19)
119.2(2)
122.63(19)
129.66(14)
134.81(14)
95.23(7)
F igu r e 7. Molecular structure of 5 (thermal ellipsoids
30%; ′: -x, -y, -z).
no additional metrical data to substantiate this idea
(e.g., significant shortening of the Mg-C distance), lack
of directly analogous compounds with which to compare
bond lengths, and no method to examine the bonding
in the solution state, we are cautious in claiming that
the observed bonding patterns are solely due to the
boron atom, and not at least partly due to differences
in the steric environment at the metal center and/or
crystal packing forces.
3.4. Attem pted Syn th esis of [Zn (BDI){OB(m es)2}]n
(BDI ) [HC{C(Me)NAr }2]-). A large family of zinc-
based ROP catalysts have been developed employing
monoanionic â-diketiminate ligands [HC{C(Me)NAr}2]-
(BDI)40 as a support for the metal alkoxide or amide
bonds, which are the active component in this process.11
To allow a more direct comparison to be made between
existing catalysts and postulated systems based on the
boroxide ligand, we targeted compounds of general
formula [Zn(BDI){OB(mes)2}]n using two different syn-
thetic approaches (eqs 2a-c).
F igu r e 8. Molecular structure of 6 (thermal ellipsoids
30%; ′: -x+2, -y+1, -z).
{OAr′}Bu]2 (Ar′ ) 2,6-tBu2C6H3) have been characterized,
in which the steric bulk around the metal is sufficient
to stabilize a three-coordinate magnesium, while reduc-
ing the steric volume in the tert-butoxy complex [Mg-
{OtBu}Me]4 allows formation of a tetrameric cubane
structure.38 The molecular structures of 5 and 6 are
illustrated in Figures 7 and 8, crystal data are sum-
marized in Table 5, and selected bond lengths and
angles are collected in Table 7.
Compounds 5 and 6 also exist as dimers in the solid
state with bridging boroxide ligands and terminal alkyl
groups. Displacement of the steric bulk of the mesityl
substituents due to the inclusion of boron allows the
coordination sphere of each magnesium to be completed
by a molecule of THF, which form an anti arrangement
with respect to the central Mg2O2 plane. The geometry
at the metal center is therefore distorted tetrahedral,
with bond angles in the range 84.31(6)-122.25(9)° for
5 and 84.77(7)-121.42(9)° for 6, where, as noted in 4,
the smallest value angle corresponds to the internal
angle within the metallacycle. The Mg-O bond lengths
[5, 2.0077(15) and 1.9910(14) Å; 6 2.005(2) and 2.006-
(2) Å] are on average longer than in the closely related
bis-alkoxide and -siloxide complexes [Mg(OR)2(THF)]2
(R ) CMePh2; SiPh3).39 This is consistent with the
proposed reduction of the metal-oxygen bond order
arising from reduced π-component. However, in view of
The reaction between [Zn{OB(mes)2}Me]2 (2) and
neutral (BDI)H did not proceed at room temperature
or in refluxing toluene, attributed to a steric conflict
between the mesityl substituents of the boroxide ligand
and the aryl group of the (BDI)H ligand disfavoring the
(38) Sung, M. M.; Kim, C. G.; Kim, J .; Kim, Y. Chem. Mater. 2002,
14, 826.
(39) Zechmann, C. A.; Boyle, T. J .; Rodriguez, M. A.; Kemp, R. A.
Inorg. Chim. Acta 2001, 319, 137.
(40) Bourget-Merle, L.; Lappert, M. F.; Severn, J . R. Chem. Rev.
2002, 102, 3031.

5168 Organometallics, Vol. 23, No. 22, 2004
Cole et al.
erization reaction. Unfortunately, the numerous dif-
ficulties encountered during the synthesis of magnesium
boroxide complexes prevented a similar assessment of
the activity of the Mg-OBR2 bond.
3.6. Con clu sion s. Two main protocols for the gen-
eration of metal boroxide complexes have been investi-
gated in the context of zinc and magnesium chemistry:
(i) protonolysis using the borinic acid and (ii) trans-
metalation from the lithium boroxide species. Starting
from dialkyl zinc complexes, method (i) proved highly
successful, with isolation of the mixed boroxide/alkyl
complex in good yield. For magnesium, however, many
difficulties were encountered using either (i) or (ii) as
an attempted entry point. While generation of magne-
sium boroxide complexes has frequently been demon-
strated, a high-yielding (reproducible) route failed to be
developed during the course of the study. The low
solubility and facile decomposition of many of these
complexes contribute to the difficulties in isolating pure
compounds in sufficient quantities to fully explore the
chemistry.
X-ray crystallography has been extensively used in
the characterization of the boroxide complexes de-
scribed, indicating exclusive formation of the µ2-
coordination mode. An important contributory factor in
the predilection for this ligand to bridge metal centers
is believed to be the displacement of the bulky mesityl
substituents away from the oxygen atom arising from
the incorporation of the boron. This allows the approach
of a second metal center to the oxygen and the formation
of a bridging coordination mode. Investigation of the
metal-oxygen bond distances indicates essentially equal
distances throughout the M2O2 metallacycle, suggesting
strongly held dimers in the solid state that are likely
to maintain their structure in solution.
A brief assessment of the ROP reactivity of zinc
boroxide 2 with rac-lactide demonstrated a total lack
of catalytic behavior under the conditions investigated.
Considering a coordination-insertion mechanism for
the propagation step during ROP and that many of the
most active systems for ROP consist of a three-
coordinate metal (or a loosely held dimer that is likely
to dissociate in solution), a ligand framework which
allows the approach of a monomer to the metal center
is essential in generating an active system. It is believed
therefore that despite a predicted weakening of the
metal-oxygen bond for boroxide species, the combined
electron deficiency and displaced bulk associated with
these ligands favors the formation of strongly held
dimers that are inactive for ring-opening polymeriza-
tion, and that perhaps we have gone a step too far in
modifying the properties of the metal-oxygen linkage.
F igu r e 9. Unit cell contents, crystallized from MeCN, of
the reaction between Zn(DBI)Et and (mes)2BOH that had
been refluxed in toluene.
approach of the diketimine to the metal center. We
therefore attempted the protonolysis reaction between
Zn(BDI)R (R ) Me, Et) and dimesitylborinic acid, in
which the BDI ligand is already bonded to the metal
center. As before no reaction was observed at room
temperature, suggesting that steric demands of the
mesityl substituents restrict the reactivity, particularly
since a similar approach using smaller alcohol reagents
i
(e.g., H2O, MeOH, PrOH) has previously been shown
to be a viable route to [Zn(BDI)OR]n compounds.19
Attempts at driving the reaction by using more forcing
conditions (refluxing toluene) had no effect for the zinc
methyl compound; however, a reaction was detected for
the zinc ethyl analogue, as indicated by the loss of the
1
(mes)2BOH resonance in the H NMR spectrum of the
crude product. An X-ray analysis of crystals obtained
by crystallization from acetonitrile showed that the
molecular structure consisted of the Zn(BDI)Et starting
reagent that had cocrystallized with neutral (BDI)H and
MeCN solvate molecule (Figure 9). This suggests that
the Zn-N bond is susceptible to protonolysis by the
borinic acid reagent, a result similar to that with the
related zinc guanidinate system, in which the reaction
between [Zn(Me2NC{NiPr}2)Me]2 and (mes)2BOH af-
fords the mixed guanidinate-guanidine boroxide, Zn-
[Me2NC{NiPr}2][OB(mes2)]‚Me2NC{NiPr}{NHiPr}.41 The
Zn(BDI)Et/(BDI)H/MeCN structure also serves to il-
lustrate the extremely effective protection that [BDI]-
offers the metal, preventing the sterically unobtrusive
acetonitrile from interacting with the zinc center.
3.5. At t em p t ed R in g-Op en in g P olym er iza t ion
Ca ta lysis. The initial aim of this study was to assess
the merits of the Zn-BOR2 bond in the context of lactide
polymerization. Attempted preparative scale polymer-
izations of ꢀ-caprolactone were unsuccessful using both
the dimeric boroxide, [Zn{OB(mes)2}Me]2 (2), and the
conventional alkoxide analogue, [Zn{OCH(mes)2}Me]2
(3). A more detailed study using NMR scale reactions
with rac-lactide also gave a negative result for the
initiation of polymerization using 2. In contrast 3 slowly
polymerized 100 equiv of rac-lactide under identical
conditions during a 16 h period. These results suggest
that, despite the predicted weakening of the zinc-
oxygen bond and the removal of the steric bulk from
the coordination sphere of the metal in the boroxide
complexes, additional factors must be in effect that
prevent the initiation and/or propagation of the polym-
Ack n ow led gm en t. We thank the EPSRC for a
research studentship (S.C.C.) and University of Sussex
for additional financial support. We also wish to ac-
knowledge the use of the EPSRC’s Chemical Database
Service at Daresbury.
Su p p or tin g In for m a tion Ava ila ble: 1H and 13C NMR
spectra of [Mg{OB(mes)2}Me(THF)]2 (5) and [Mg{OB(mes)2}nBu-
(THF)]2 (6). This material is available free of charge via the
Internet at http://pubs.ac.org.
(41) Birch, S. J .; Boss, S. R.; Cole, S. C.; Coles, M. P.; Haigh, R.;
Hitchcock, P. B.; Wheatley, A. E. H. Dalton Trans. 2004, in press.
OM049520J