The reversible amination of tin(II)-ligated imines: Latent initiators for the polymerization of rac-lactide

Article abstract of DOI:10.1021/ic701671s

The 1:1 reactions of nine potentially tridentate salicylaldimines with tin(II) diamides, Sn(NR₂)₂ (R = Me, Et, ⁱPr, SiMe₃) have been investigated. With Sn(NⁱPr ₂)₂ and Sn(NTMS

Full text of DOI:10.1021/ic701671s

Inorg. Chem. 2008, 47, 5417-5424
The Reversible Amination of Tin(II)-Ligated Imines: Latent Initiators for
the Polymerization of rac-Lactide
Nonsee Nimitsiriwat,† Vernon C. Gibson,*,† Edward L. Marshall,*,† and Mark R. J. Elsegood‡
Department of Chemistry, Imperial College London, South Kensington Campus,
London SW7 2AZ, U.K., and Chemistry Department, Loughborough UniVersity,
Leicestershire LE11 3TU, U.K.
Received August 24, 2007
i
The 1:1 reactions of nine potentially tridentate salicylaldimines with tin(II) diamides, Sn(NR2)2 (R ) Me, Et, Pr,
SiMe3) have been investigated. With Sn(NiPr2)2 and Sn(NTMS2)2, the anticipated products of amine elimination,
iminophenoxy tin(II) mono(amide)s, are formed. However, for R ) Me and R ) Et, nucleophilic attack of the
amide at the imino carbon occurs to generate tin(II) complexes of tetradentate, dianionic aminoamidophenoxide
ligands. The transfer of the amide is shown to be reversible, with both alcoholysis and the initiation of rac-lactide
polymerization apparently mediated by the terminal amide tautomer.
Introduction
purposes, a constraint which almost entirely eliminates the
influence of the ancillary ligand over the stereochemistry of
the ring-opening event.2b
The industrial preparation of poly(lactic acid), PLA, from
plant starch epitomizes the current interest in using renewable
rather than depleting chemical feedstocks.1 PLA is manu-
factured via the ring-opening polymerization of lactide, LA,
using a poorly defined initiating species comprising tin(II)
bis(2-ethylhexanoate) and alcohol. In order to produce single-
site analogues of this system, tin(II) alkoxide and amide
complexes of both ꢀ-diketiminate2 and amidinate3 ligand
families have been reported by ourselves and Tolman et al.
Despite deploying a wide range of ancillary ligand substit-
uents, however, all of these initiators polymerize rac-LA with
a similar level of heteroselectivity (Pr ) 0.62–0.65), an
observation attributed to the nonbonding nature of the tin
5s2 electrons: poor spatial overlap with the 4d and 5p orbitals
(due to a relativistic contraction) means that the tin center
uses orthogonally aligned p and d orbitals for coordination
Stereochemical invariance at single-site catalytic centers
is unusual, so we decided to explore the divalent tin
chemistry of different monoanionic ancillary ligand systems.
Salicylaldiminates (R-iminophenoxides) are one of the most
commonly studied ligand classes in contemporary coordina-
tion chemistry, and their ease of synthesis makes them
particularly attractive subjects for catalytic studies. The
opportunity to realize enhanced polymerization activity and
microstructural control by varying their steric and electronic
properties has already been exploited in early4 and late5
transition (and even main group6) metal initiators for the
polymerization of ethylene and higher R-olefins.
Although their use as ligands for other monomer systems
is somewhat less advanced, some salicylaldiminate-based
initiators have been reported for the ring-opening polymer-
* E-mail: v.gibson@imperial.ac.uk (V.C.G.); e.marshall@imperial.ac.uk
(E.L.M.).
(4) (a) Mitani, M.; Mohri, J.; Yoshida, Y.; Saito, J.; Ishii, S.; Tsuru, K.;
Matsui, S.; Furuyama, R.; Nakano, T.; Tanaka, H.; Kojoh, S.; Matsugi,
T.; Kashiwa, N.; Fujita, T. J. Am. Chem. Soc 2002, 124, 3327–3336.
and references therein. (b) Tian, J.; Coates, G. W. Angew. Chem., Int.
Ed. 2000, 39, 3626–3629. (c) Tian, J.; Hustad, P. D.; Coates, G. W.
J. Am. Chem. Soc. 2001, 123, 5134–5135. (d) Hustad, P. D.; Tian, J.;
Coates, G. W. J. Am. Chem. Soc. 2002, 124, 3614–3621. (e) Coates,
G. W.; Hustad, P. D.; Reinartz, S. Angew. Chem., Int. Ed. 2002, 41,
2236–2257. (f) Gibson, V. C.; Mastroianni, S.; Newton, C.; Redshaw,
C.; Solan, G. A.; White, A. J. P.; Williams, D. J. J. Chem. Soc., Dalton
Trans. 2000, 1969–1971. (g) Jones, D. J.; Gibson, V. C.; Green, S. M.;
Maddox, P. J. Chem. Commun. 2002, 1038–1039.
†
Imperial College London.
Loughborough University.
‡
(1) Drumright, R. E.; Gruber, P. R.; Henton, D. E. AdV. Mater. 2000, 12,
1841–1846.
(2) (a) Dove, A. P.; Gibson, V. C.; Marshall, E. L.; White, A. J. P.;
Williams, D. J. Chem. Commun. 2001, 283–284. (b) Dove, A. P.;
Gibson, V. C.; Marshall, E. L.; Rzepa, H. S.; White, A. J. P.; Williams,
D. J. J. Am. Chem. Soc. 2006, 128, 9834–9843.
(3) (a) Aubrecht, K. B.; Hillmyer, M. A.; Tolman, W. B. Macromolecules
2002, 35, 644–650. (b) Nimitsiriwat, N.; Gibson, V. C.; Marshall, E. L.;
White, A. J. P.; Dale, S. H.; Elsegood, M. R. J. Dalton Trans., DOI
10.1039/b706663e.
(5) Younkin, T. R.; Connor, E. F.; Henderson, J. I.; Friedrich, S. K.;
Grubbs, R. H.; Bansleben, D. A. Science 2000, 287, 460–462.
10.1021/ic701671s CCC: $40.75  2008 American Chemical Society
Inorganic Chemistry, Vol. 47, No. 12, 2008 5417
Published on Web 05/14/2008

Nimitsiriwat et al.
Scheme 1
Scheme 2. Syntheses of Potentially Tridentate Salicylaldimines
ization of lactide, most notably with Zn,7 Mg,7c and Ca.8 A
clear message to arise from these studies is the difficulty in
obtaining single-site behavior if the salicylaldiminate binds
merely as a bidentate chelate. For example, Chisholm and
co-workers have successfully employed bidentate phenoxy-
imines to support three-coordinate zinc-based initiators, but
single-site behavior requires both the Schiff base and the
initiating nucleophile (either an aryloxide or amide) to be
bulky (e.g., I, Scheme 1): slow consumption of L-LA ensues,
with 20 equiv polymerized over 3 h at 25 °C.7a Lin et al.
have shown that dimeric complexes of tridentate variants
are more readily accessible, and less bulky counterparts of
II (lacking the phenoxy tert-butyl substituents) afford much
more active initiators than I, capable of consuming 200 equiv
of LA in <30 min at room temperature.7b
In light of these results, we chose to investigate the tin(II)
complexation chemistry of potentially tridentate salicylaldi-
mines, with a view to producing single-site LA polymeri-
zation initiators. We were particularly interested to discover
whether the presence of a third donor site might help
overcome the influence of the 5s2 electrons, leading to
different PLA tacticities. As described below, this study
instead led to the discovery of a rather unusual ligand
amination reaction, aspects of which have been com-
municated.9
halogenated substituents on the phenoxy ring have also been
examined within the O-donor family.
Single-site initiators for the ring-opening polymerization
of cyclic esters are typically alkoxide (-OR) complexes,
although other nucleophilic ligands such as amides (-NR2)
may also be used.11 In order to prepare [RONL]Sn(OR)
alkoxide complexes, a two-step route successfully used to
synthesize ꢀ-diketiminate tin(II) analogues2 was initially
envisaged, that is, formation of [RONL]SnCl and subsequent
metathesis with an alkali metal alkoxide. Thus, lithiation of
Synthetic Studies
[
tBuONNMe ]H, [tBuONPPh ]H, [tBuONOMe]H, and [tBuONSPh]H
2
2
n
with BuLi, followed by addition to a suspension of SnCl2
in toluene at room temperature, afforded the salicylaldiminate
tin(II) chloride complexes 1-4 in recrystallized yields of
74, 76, 49, and 64%, respectively (Scheme 3).10
Salicylaldimines with a neutral donor atom bound to the
imine functionality are readily obtained from the condensa-
tion of a 2,4-disubstituted salicylaldehyde with an ap-
propriately functionalized amine or aniline (Scheme 2);10 for
the sake of brevity, these are abbreviated to [RONR′]H, where
R represents the ortho- and para-phenol substituents, and
R′ indicates the N-substituent. These were chosen to allow
a probe of both steric and electronic factors, particularly with
respect to the imino donor arm, which features a range of
N-, P-, O-, and S-based donor groups. The effects of
However, all reactions of the chloride complexes with
LiOiPr and NaOtBu led to intractable mixtures, with the
exception of the reaction of 4 with NaOtBu, for which
recrystallization of the crude product mixture resulted in the
isolation of the bischelate [tBuONSPh]2Sn, 5. This synthetic
route was therefore abandoned in favor of direct metalation
with Sn(NMe2)2.12 Due to high solubilities, the products
(6) (a) Cameron, P. A.; Gibson, V. C.; Redshaw, C.; Segal, J. A.; Bruce,
M. D.; White, A. J. P.; Williams, D. J. Chem. Commun. 1999, 1883–
1884. (b) Cameron, P. A.; Gibson, V. C.; Redshaw, C.; Segal, J. A.;
White, A. J. P.; Williams, D. J. Dalton. Trans. 2002, 415–422. (c)
Annunziata, L.; Pappalardo, D.; Tedesco, C.; Pellecchia, C. Organo-
metallics 2005, 24, 1947–1952.
arising from the reactions with [tBuONNMe ]H, [tBuONOPh]H,
2
and [tBuONSMe]H could not be readily purified by either
washing or recrystallization. Nonetheless, complexes 6–11,
formed from the treatment of Sn(NMe2)2 with [tBuONquin]H,
(7) (a) Chisholm, M. H.; Gallucci, J.; Zhen, H.; Huffman, J. C. Inorg.
Chem. 2001, 40, 5051–5054. (b) Chen, H.-Y.; Tang, H.-Y.; Lin, C.-
C. Macromolecules 2006, 39, 3745–3752. (c) Wu, J.-C.; Huang, B.-
H.; Hsueh, M.-L.; Lai, S.-L.; Lin, C-C. Polymer 2005, 46, 9784–9792.
(8) Darensbourg, D. J.; Choi, W.; Richers, C. P. Macromolecules 2007,
40, 3521–3523.
(11) (a) O’Keefe, B. J.; Hillmyer, M. A.; Tolman, W. B. Dalton Trans.
2001, 2215–2224. (b) Gibson, V. C.; Marshall, E. L. In ComprehensiVe
Coordination Chemistry II; McCleverty, J. A., Meyer, T. J., Eds.;
Elsevier: Oxford, 2003; Vol. 9, pp 1–74; (c) Nakano, K.; Kosaka, N.;
Hiyama, T.; Nozaki, K. Dalton Trans. 2003, 4039–4050. (d) Dechy-
Cabaret, O.; Martin-Vaca, B.; Bourissou, D. Chem. ReV. 2004, 104,
6147–6176.
(9) Nimitsiriwat, N.; Marshall, E. L.; Gibson, V. C.; Elsegood, M. R. J.;
Dale, S. H. J. Am. Chem. Soc. 2004, 126, 13598–13599.
(10) See Supporting Information.
(12) Foley, P.; Zeldin, M. Inorg. Chem. 1975, 14, 2264–2267.
5418 Inorganic Chemistry, Vol. 47, No. 12, 2008

The ReWersible Amination of Tin(II)-Ligated Imines
Scheme 3. The Syntheses of the Tin(II) Chloride Complexes 1-4
Scheme 4. The Syntheses of the Bridged Amide Complexes 6-11
[
tBuONPPh ]H, [tBuONOMe]H, [tBuONSPh]H, [ClONOMe]H, and
2
[IONOMe,]H, respectively, were isolated in analytically pure
forms by recrystallization.
The 1H and 13C solution-state NMR spectra of the
recrystallized samples of 6–11 share similar gross features
but are all inconsistent with [RONL]Sn(NMe2) formulations.
Most notably, no HCdN imino resonances are observed in
1
in the diphenylphosphino-based 7, the H and 13C NMR
1
the expected window of the H NMR spectra (δ 7.44–7.61
spectra of which contain two sharp singlets for the inequiva-
lent NMe2 groups (δH 2.08 and 1.93; δC 43.80 and 40.87).
A similar effect is observed in the phenylthioether-containing
9 (δH 1.94 and 1.80; δC 43.65 and 40.62). The reason why
the Sn-N interaction appears stronger in these two com-
pounds is not clear: one might anticipate that it would
actually be weakest in these structures, with sterically bulky
and relatively basic -PPh2 and -SPh substituents likely to
result in congested metal centers of reduced electron
deficiency. One explanation is that despite the presence of
119Sn-31P coupling in the NMR spectra of 7 (see discussion
of complex 15 below) the Sn-P and Sn-S interactions might
be hemilabile in 7 and 9, allowing stronger coordination of
the amine (whereas the smaller OMe donors of 8, 10, and
11 are tightly bound to the tin center). Alternatively, the ring
strain within the SnN2C fragment may promote dissociation
of the neutral amine donor: once this has occurred, the
remainder of the ligand becomes free to maneuver into a
less strained conformation, a process possibly disfavored for
the heavier PPh2 and SPh units.
The molecular structure of 11 (Figure 1) shares many of
the features previously described by us for complex 7.9 The
phenoxy-based ligand is now tetradentate, with anionic amide
[Sn(1)-N(1) ) 2.111(3) Å] and coordinative dative amine
[Sn(1)-N(2) ) 2.404(3) Å] ligation [analogous bond lengths
in 7 are 2.1649(12) and 2.4490(13) Å]. The removal of
unsaturation from the former salicylaldimine CdN func-
tionality is confirmed by the similar C(8)-N(1) and
C(8)-N(2) bond distances [1.448(5) and 1.503(4) Å respec-
tively]. The ethereal oxygen atom in 11 is displaced
significantly further from the tin than its phenoxy counterpart
(Sn-O(1) ) 2.809(3) Å; Sn-O(2) ) 2.123(2) Å), but this
for 1–4; δ 7.86 for 5), although in every case a singlet,
integrating as a solitary proton, is present between 5.35 and
5.70 ppm. The 13C NMR spectra all include a corresponding
upfield-shifted resonance for the associated carbon atom (e.g.,
1
87 ppm for 7 versus 169 ppm for 2). H/13C HMBC NMR
experiments further indicated that in each of the six
complexes the same carbon atom is separated from the NMe2
protons by just three bonds: a five-bond separation would
be anticipated for [RONL]Sn(NMe2).
X-ray diffraction studies on 7 and 11 revealed that both
complexes contain tetradentate, dianionic ligands arising from
amination of the imino sp2-carbon atom. The NMe2 ligand
becomes a neutral dative σ-donor to the metal, with the
former imino N converted into an anionic amido ligand. As
the NMR spectra of these two compounds are entirely
representative of the family of 6–11, the structures of 6 and
8–10 are therefore assigned by analogy, as summarized in
Scheme 4.
In solution, the Sn-NMe2 interaction appears to be
hemilabile, at least for complexes 6, 8, 10, and 11, all of
which give one broad resonance for what should be in-
equivalent N-methyl protons; one broad signal is also
observed for the N(CH3)2 resonance in the 13C spectrum.
Although one could interpret these data in terms of the
reversible migration of the NMe2 unit from the tin to the
imino carbon, sharp resonances for the HCN proton indicate
that the broadening instead arises from the breaking and re-
forming of the dative N-Sn interaction. This is supported
by low-temperature 1H NMR spectra in CD2Cl2 which reveal
that the NMe2 signals resolve into two sharp singlets.10
stronger interaction is apparently present at room temperature
A
Inorganic Chemistry, Vol. 47, No. 12, 2008 5419

Nimitsiriwat et al.
Scheme 5. Alternative Synthesis of Complex 7
Scheme 6. The syntheses of complexes 8 and 12–14
Figure 1. The molecular structure of complex 11. Ellipsoids are at the
50% probability level. Selected bond lengths (Å) and bond angles (°):
Sn(1)-N(1) ) 2.111(3), Sn(1)-N(2) ) 2.404(3), Sn(1)-O(2) ) 2.123(2),
Sn(1)···O(1) ) 2.809(3), N(1)-C(8) ) 1.448(5), N(2)-C(8) ) 1.503(4);
N(1)-Sn(1)-N(2) ) 59.64(11), N(1)-Sn(1)-O(2) ) 88.50(11), N(2)-
Sn(1)-O(2) ) 80.91(10); N(1)-C(8)-N(2) ) 99.9(3), Sn(1)-N(1)-C(8)
) 99.7(2), Sn(1)-N(2)-C(8) ) 86.41(19), O(1)-Sn(1)-N(1) ) 61.84(10),
O(1)-Sn(1)-N(2) ) 119.45(9), O(1)-Sn(1)-O(2) ) 111.89(9).
NMR spectra (Figure 2) confirmed that 12 possesses a similar
ligand skeleton to that seen in 8 (e.g., δ CH(NEt2) ) 5.79
cf. δ CH(NMe2) ) 5.44); by contrast, the spectra of 13 and
14 clearly demonstrate that ligand amination does not occur
so readily with the bulkier tin(II) amides, with singlets
characteristic of HCdN imine environments observed at 7.57
and 7.48 ppm, respectively, and no CH(NR2) resonances
separation is still shorter than the sum of the respective van
der Waals radii (2.17 Å (Sn) + 1.52 Å (O) ) 3.69 Å13) and
is comparable in length to another weak Sn-O interaction
recently reported by us.2b Intermolecular approaches are
shown in the Supporting Information, but none of these is
of sufficient proximity to be considered a bonding interaction.10
The coordination sphere of the metal therefore contains a
large region of seemingly unoccupied space. Previous
computational studies2b suggest that for ꢀ-diketiminate
counterparts the tin 5s orbital undergoes a relativistic
contraction: assuming a similar situation exists here, the
metal-based orbitals used to bind the tetradentate ligands will
be orthogonally orientated 4d and 5p lobes: the presence in
the crystal structure of 11 (and 7) of several interligand angles
at tin of ca. 90° (e.g., N(1)-Sn(1)-O(2) ) 88.50(11);
N(2)-Sn(1)-O(2) ) 80.91(10); O(1)-Sn(1)-O(2) )
111.89(9)°) lends support to this hypothesis.
between δ 5 and 6. Similarly, the reaction of [tBuONPPh ]H
2
with Sn[N(SiMe3)2]2 cleanly afforded the terminal amide
[
tBuONPPh ]SnN(SiMe3)2, 15 (δ HCdN ) 7.15 ppm).
2
An X-ray crystallographic analysis of 15 confirms the
terminal, monoanionic nature of the NTMS2 amide ligand
(Figure 3) and also reveals that in the solid state the
salicylaldiminate ligand adopts a bidentate chelation mode.
Unlike the Me2N-bridged complexes, the anilino ring is
aligned in a near-orthogonal manner to the plane of the N(1)
Two pathways are advanced to account for the formation
of complexes 6–11: one of the amide ligands of Sn(NMe2)2
could attack the imino carbon atom prior to salicylaldimine
deprotonation, or the desired [RONL]SnNMe2 species might
form first and migration of the amide then occurs to complete
the ligand transformation. The latter of these mechanisms
appears more likely since the addition of LiNMe2 to the
chloride complex 2 also results in the clean formation of 7
(Scheme 5). 1H NMR confirms that complexes 3 and 4 react
similarly (though the products of these reactions were again
too soluble to allow ready isolation of analytically pure
products).
In order to further probe the versatility of the amination
reaction, [tBuONOMe]H was treated with Sn(NEt2)2, Sn(NiPr2)2,
and Sn(NTMS2)2 to yield complexes 12–14, respectively
(Scheme 6), though 13 gradually decomposes in solution and
1
has therefore only been characterized spectroscopically. H
1
Figure 2. H NMR spectra of complexes 8, 12, and 14 (C6D6, 250 MHz,
(13) Bondi, A. J. Phys. Chem. 1964, 68, 441–451.
5420 Inorganic Chemistry, Vol. 47, No. 12, 2008
298 K; * ) C6D5H).

The ReWersible Amination of Tin(II)-Ligated Imines
trigonal pyramid, with the three-coordinate tin atom displaced
by 1.109 Å from the O(1), N(1), N(2) plane. The Sn(1)-N(2)
bond is approximately perpendicular to the O(1), Sn(1), N(1)
plane [N(2)-Sn(1)-N(1) ) 92.6(3), N(2)-Sn(1)-O(1) )
96.4(3)°]. The retention of the imine functionality is con-
firmed by the short C(15)-N(1) bond [1.323(11) Å] and the
dative nature of the Sn(1)-N(1) interaction [2.328(7) Å, cf.
Sn(1)-N(2) ) 2.115(8) Å].
Despite the crystallographic data obtained on 15, we
believe that in solution the salicylaldiminate ligands in
complexes 14 and 15 probably coordinate in a hemilabile
tridentate manner. Their 119Sn chemical shifts (δ -150 and
-151, respectively) are downfield to some of the related four-
coordinate bridged amine species (e.g., δ -166, 8; -172,
12; -271, 7) but are not shifted sufficiently to indicate purely
tricoordinate tin centers.15 Tellingly, the 31P NMR spectrum
of 15 consists of a singlet at -15.60 ppm with satellites
showing coupling to 117Sn and 119Sn (450 and 470 Hz,
respectively); its 119Sn spectrum is a doublet at δ -151 with
JSnP ) 460 Hz. When compared to the analogous spectra of
complex 7 (31P NMR: singlet resonance δ -16.38, unre-
solved satellites with JSnP ≈ 540 Hz; 119Sn NMR: doublet
resonance δ –271, JSnP ) 538 Hz), it would appear that
phosphorus coordination in solution is only slightly weaker
in 15.
Figure 3. The molecular structure of complex 15. Ellipsoids are at the
50% probability level. Selected bond lengths (Å) and bond angles (°):
Sn(1)-N(1) ) 2.328(7), Sn(1)-N(2) ) 2.115(8), Sn(1)-O(1) ) 2.112(6),
N(1)-C(15) ) 1.323(11), N(1)-C(16) ) 1.425(11); N(1)-Sn(1)-N(2) )
92.6(3), N(1)-Sn(1)-O(1) ) 79.9(2), N(2)-Sn(1)-O(1) ) 96.4(3),
C(1)-O(1)-Sn(1) ) 124.7(5), C(15)-N(1)-Sn(1) ) 119.6(6), Si(1)-
N(2)-Si(2) ) 123.5(5), Si(1)-N(2)-Sn(1) ) 120.0(4), Si(2)-N(2)-Sn(1)
) 114.2(4).
In order to probe whether the amide migration is a
reversible process, 10 equiv of Me2NH was added to a
solution of 12 in benzene-d6 at ambient temperature and
1
monitored by H NMR spectroscopy (Figure 5a). Within a
few hours, the broad resonances at δ 2.57 and 0.72
corresponding to the bridging diethylamine group were seen
to disappear and Et2NH was generated, along with clean
formation of the Me2N-bridged complex 8. Conversely,
addition of excess Et2NH (12 equiv) to a solution of 8 led
to an incomplete conversion to 12 (Figure 5b; [8]/[12] ≈
1.6:1; Keq ) 0.021); the tin(II) complex containing the
smaller amide bridge therefore appears to be thermodynami-
cally more stable. Although direct observation of the terminal
amide isomer of either 8 or 12 has not proved possible, we
believe that transamination most likely proceeds via these
intermediates (Scheme 7), particularly as precedence exists
for the exchange of organic amines with metal-bound
monoanionic amide ligands.16
Figure 4. The asymmetric unit of complex 15. Selected nonbonding
distances (Å): Sn(1)···P(1) ) 3.767, Sn(1)···P(2) ) 4.126, Sn(2)···P(2) )
3.748, Sn(2)···P(1) ) 4.089, Sn(1)···Sn(2) ) 6.152.
substituents, placing the PPh2 moiety on the opposite side
of the molecule to the N(SiMe3)2 unit probably in order to
minimize steric interaction between these two bulky groups.
The Sn(1)···P(1) distance is 3.767 Å, and though slightly
shorter than the sum of the van der Waals radii (2.17 (Sn)
+ 1.80 (P) ) 3.97 Å13) such a distance far exceeds the
longest reported Sn-P bond [3.036(2) Å].14
Complex 15 crystallizes with two molecules in the
asymmetric unit related by a noncrystallographic pseudo-2-
fold rotation axis (Figure 4). There are no intermolecular
interactions between the phosphorus and tin atoms, with both
the Sn(1)-P(2) and Sn(2)—P(1) separations >4 Å. The
geometry of the tin center is therefore best described as a
Typically, metal amides also undergo facile alcoholysis,17
so we next examined whether the bridged amine complexes
might function as alkoxide precursors via their terminal
amide tautomers. As anticipated, complex 8 reacts cleanly
with Ph2CMeCOH to afford the terminal alkoxide [tBuO-
N
OMe]Sn(OCMePh2), 16 (Scheme 8). However, smaller
alcohols such as iPrOH and tBuOH led to intractable product
mixtures, although the absence of HC(OR)NAr methine
(15) (a) Hani, R.; Geanangel, R. A. Coord. Chem. ReV. 1982, 44, 229–
246. (b) Hahn, F. E.; Wittenbecher, L.; Le Van, D.; Zabula, A. V.
Inorg. Chem. 2007, 46, 7662–7667.
(16) Duan, Z.; Naiini, A. A.; Lee, J.-H.; Verkade, J. G. Inorg. Chem. 1995,
34, 5477–5482.
(17) Chisholm, M. H.; Rothwell, I. P. In ComprehensiVe Coordination
Chemistry; Wilkinson, G., Gillard, R. D., McCleverty, J. A., Eds.;
Pergamon: Oxford, 1987; Vol. 2, pp 161–188.
(14) Barney, A. A.; Heyduk, A. F.; Nocera, D. G. Chem. Commun. 1999,
2379–2380.
Inorganic Chemistry, Vol. 47, No. 12, 2008 5421

Nimitsiriwat et al.
Table 1. Polymerization of rac-LA Using Initiators 6–11a
time conversionb induction
c
d
d
(min)
(%)
period (min) kapp (s-1
)
Mn
Mw/Mn Pre
6
7
8
9
360
110
120
120
60
91
94
91
95
92
95
74
8
7
9
<1
0
1.5 × 10-4 19800 1.17 0.61
4.7 × 10-4 20500 1.25 0.60
4.0 × 10-4 18200 1.29 0.59
4.4 × 10-4 19100 1.27 0.59
7.4 × 10-4 16000 1.57 0.65
5.2 × 10-4 23400 1.40 0.63
10
11 100
a All polymerizations were carried out at 60 °C in toluene; [LA]/[Sn] )
100:1; [LA] ) 0.28 M. b Determined by 1H NMR spectroscopy, integration
of methine resonances of LA and PLA (CDCl3, 250 MHz). c -d[LA]/dt )
kapp[LA], where kapp ) kp[Sn]x (x ) rate law order dependence on the
initiator) - kapp values determined only after the induction period has
elapsed. d Determined by gel permeation chromatography, calibrated with
PS standards in CHCl3. e Determined from the methine region of the
1
homonuclear decoupled H NMR spectrum.
Polymerization Studies
The primary goal at the outset of this study was the
synthesis of tin(II) alkoxide or amide complexes as single-
site initiators for the ring-opening polymerization (ROP) of
cyclic esters: the alkoxide 16 and terminal amide complexes
14 and 15 were therefore screened for rac-LA polymerization
activity. Of these three, 16 is the most active, consuming
100 equiv of monomer in 240 min in toluene at 60 °C (92%
conversion; Mn ) 21 000; Mn
) 13 300), but the
(calcd)
molecular weight distribution is broader than expected for a
well-behaved polymerization (Mw/Mn ) 1.46). The amide
complexes are much less active, with 48 h at 60 °C required
to attain conversions of 64 and 93% for 14 and 15,
respectively. Furthermore, both initiators give very broad
molecular weight distributions (14: Mn ) 27 400; MwMn )
6.1; 15: Mn ) 20 200; Mw/Mn ) 7.9), consistent with several
previous studies which describe inefficient initiation by bulky
amide ligands.2b,18
1
Figure 5. H NMR spectra (C6D6, 250 MHz, 298 K) of the reactions of
(a) 12 + Me2NH (10 equiv); (b) 8 + Et2NH (10 equiv).
Scheme 7. Proposed Mechanism for the Amine Exchange Reactions
Given that even the OCMePh2 ligand serves as a relatively
poor initiating group, the use of smaller nucleophilic units
was deemed preferable. Therefore, we next investigated
whether the bridged Me2N complexes 6-11 might function
as polymerization initiators: it was envisaged that insertion
of a lactide monomer into the Sn-NMe2 bond of the terminal
amide [RONL]Sn(NMe2) tautomers could lead to a terminal
propagating species. As summarized in Table 1, all six
examples are active with high conversions typically observed
within just 2 h at 60 °C and molecular weights close to
predicted values (Mn (calcd) (PLA)100 ) 14 400).
Scheme 8. The Synthesis of Complex 16 via Alcoholysis of 8
The mode of action of 8 has been examined in detail by
1H NMR spectroscopy. The addition of 10 equiv of rac-LA
to 8 in C6D6 leads to the loss of the bridging amide group
[δ HC(NMe2) 5.55] and the regeneration of imino signals
attributed to the propagating species:9 all resonances are
consistent with the presence at the metal center of a terminal
propagating polyester group and a tridentate salicylaldiminato
NMR resonances (in both 1H and 13C spectra) indicates that
a bridged ether ligand analogous to the bridged amine
skeleton found in 6-11 is probably not present in the product
mixture. Although the structure of 16 has not been estab-
lished crystallographically, solution-state NMR data (C6D6)
are entirely consistent with the regeneration of the imino
bond (δ HCdN 7.75; δ HCdN 166.44), and a 119Sn chemical
shift of -379 ppm points toward a four-coordinate metal
center.
1
ligand. The H NMR spectra (CDCl3) of the PLA samples
obtained from the polymerizations initiated by 6–11 also
feature two singlet resonances (δ 2.94 and 3.03) attributable
to the diastereotopic methyl groups of the C(O)NMe2 chain
(18) Chamberlain, B. M.; Cheng, M.; Moore, D. R.; Ovitt, T. M.;
Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc. 2001, 123, 3229–
3238.
5422 Inorganic Chemistry, Vol. 47, No. 12, 2008

The ReWersible Amination of Tin(II)-Ligated Imines
Figure 6. MALDI-TOF mass spectrum of PLA generated from initiator 8
([LA]0/[8] ) 20, toluene, 60 °C, terminated with MeOH; matrix ) 2,5-
dihydroxybenzoic acid).
Figure 8. Plots of ln{[LA]0/[LA]t} versus time (s) for: (a) complexes 6, 7,
and 8; (b) complexes 9, 10, and 11 ([LA]0/[Sn] ) 100, toluene, 60 °C).
The polymerizations initiated by complexes 6–11 all
exhibit a first-order dependence upon the concentration of
rac-LA (Figure 8) and, with the exception of the quinolyl-
functionalized 6, all attain >90% monomer conversion within
2 h, despite the variation in the ligand pendant arms.
Nonetheless, some electronic effects may be delineated: for
example, the diiodophenoxy-based 11 affords faster polym-
erization than its tert-butyl analogue 8, in accord with its
more Lewis acidic metal center.2b The chloro-substituted 10
is even more active, achieving high conversion in just 60
min. The particularly slow nature of 6 most likely results
from the rigid nature of the quinolyl donor group obstructing
the interconversion of at least one pair of stationary points
along the insertion reaction coordinate.21
A more detailed kinetic study has been performed on
initiator 7. The relationship between kapp and [7] (Figure S3)
is nonlinear,10 and this indicates a non-first-order dependence
on the initiator (-d[LA]/dt ) kapp[LA], where kapp ) kp[Sn]x
and kp ) rate of propagation).22 As shown in Figure 9,
however, the plot of lnkapp versus ln[7] is linear, and its
gradient indicates a dependence upon the initiator of 0.57,
that is, -d[LA]/dt ) kp[LA][7]0.57. The activity of the tin(II)
centers is therefore apparently tempered by aggregation, an
effect also observed in amidinate tin(II) systems.3,23 We have
attempted to investigate the nature and degree of aggregation
Figure 7. A plot of monomer conversion (%) versus Mn and correlation
with theoretical Mn for the polymerization of rac-LA using 8 as an initiator
([LA]0/[8] ) 100, toluene, 60 °C; polydispersities shown in parentheses).
end group.10 Furthermore, a MALDI-TOF mass spectrum
of a sample of PLA prepared using 8 and 20 equiv of rac-
LA confirms the presence of NMe2-terminated chains (Figure
6, series A). Two other distributions, H-[O-C(H)CH3-
CO]n-OMe, Na+ (B) and cyclic PLA, Na+ (C) are observed:
the latter is consistent with intramolecular transesterifica-
tion,19 whereas B might originate either from intermolecular
transesterification of propagating chains by [Sn]-OMe
species generated when the polymerization was terminated
with excess MeOH or by partial acylation of the amide
termini of A with methanol.20
These six initiators all give rise to well-controlled chain
growth, with linear correlations between monomer conver-
sion and Mn (exemplified in Figure 7 for 8). The gradual
rise in polydispersity values as the conversion increases
reflects the increased prevalence for transesterification of the
propagating chain as the concentration of monomer de-
creases. The notably broader distribution of chain lengths
observed at high conversions with initiators 10 and 11 are
also consistent with heightened levels of chain backbiting
with more Lewis acidic active sites.2b
(19) Montaudo, G.; Montaudo, M. S.; Puglisi, C.; Samperi, F.; Spassky,
N.; LeBorgne, A.; Wisniewski, M. Macromolecules 1996, 29, 6461–
6465.
(21) Marshall, E. L.; Gibson, V. C.; Rzepa, H. S. J. Am. Chem. Soc. 2005,
127, 6048–6049.
(22) Williams, C. K.; Breyfogle, L. E.; Choi, S. K.; Nam, W., Jr.; Hillmyer,
M. A.; Tolman, W. B. J. Am. Chem. Soc. 2003, 125, 11350–11359.
(23) Aubrecht, K. B.; Hillmyer, M. A.; Tolman, W. B. Macromolecules
2002, 35, 644–650.
(20) Nishiura, M.; Hou, Z.; Koizumi, T.; Imamoto, T.; Wakatsuki, Y.
Macromolecules 1999, 32, 8245–8251. We wish to thank one of the
reviewers for highlighting this possibility.
Inorganic Chemistry, Vol. 47, No. 12, 2008 5423

Nimitsiriwat et al.
bias (Pr ) 0.62 ( 0.03); this observation mirrors our findings
with analogous ꢀ-diketiminate2 and amidinate3,23 tin(II)
initiator families.
Discussion
Although the metal complexation chemistry of salicyla-
ldiminato ligands has been extensively studied, this is, to
the best of our knowledge, the first observation of amide
migration to the imine carbon. Although we have not been
able to unambiguously explain why it should happen in this
system, we believe that the divalent nature of the tin center
is key. The solid-state structures obtained during this and
related2,3 studies indicate that the 5s orbital is reluctant to
hybridize with the 4d or 5p orbitals, and the ligand donor
atoms therefore occupy approximately orthogonally displaced
coordination sites. This results in the proximal arrangement
of the NMe2 ligand and the electrophilic imino carbon, and
this presumably lowers the barrier to the migratory process.
The amination reaction is reversible as shown by the
exchange reactions of 8 with Et2NH and of 12 with Me2NH,
and there is a clear preference for the bridging amide
structure to incorporate smaller amide substituents. This
observation most likely reflects steric crowding in the
transition state associated with migration and in the resultant
tetradentate ligand. Indeed, with the bulkier amide ligands
NiPr2 and N(SiMe3)2 amide transfer is hindered to such an
extent that the terminal amide isomers 12–15 can be isolated.
Furthermore, their isolation may be facilitated by potential
hemilability of the salicylaldiminate ligands, as suggested
by the bidentate coordination observed in the molecular
structure of 15.
Figure 9. (a) Plots of ln{[LA]o/[LA]t} versus time for the polymerization
of rac-LA using 7 as the initiator ([LA] ) 0.28 mol L-1, [7] ) 0.0007 (I),
0.0014 (II), 0.0028 (III), 0.0056 (IV) mol L-1, 60 °C, toluene); (b) a plot
of ln kapp versus ln[7] for the polymerization of rac-LA using 7 as the
initiator (60 °C, toluene).
Although the precise mechanism for the amination reaction
can not yet be stated with complete certainty, the interme-
diacy of terminal amide isomers is consistent with all the
results obtained. For example, the reaction of alcohol with
8 can be likened to the enol-mediated reactivity of carbonyls;
though present in low concentration, the terminal amide
tautomer undergoes rapid alcoholysis and the tin(II)-contain-
ing equilibrium shifts accordingly. Similarly, the molecular
weight control observed when these complexes are used to
polymerize rac-lactide indicates that initiation via the
terminal amide is rapid and quantitative.
further using cryoscopic measurements; unfortunately, to
date, these have proved inconclusive.
All but the most active (i.e., 10 and 11) of these initiators
display induction periods, though these are relatively short
(<10 min) with the exception of 6 for which a delay of ca.
74 min exists before significant levels of propagation occur.
Although the delay could arise from the formation of the
active, terminal NMe2 initiators, we believe a more likely
explanation is that recently proposed for ꢀ-diketiminate tin(II)
complexes2b and salen-supported titanium(IV) initiators;24
that is, rapid formation of a particularly stable first insertion
product is followed by much slower insertion of a second
repeat unit with subsequent insertions gradually becoming
more exothermic as dispersion forces associated with the
propagating polymer increase with chain length.
Acknowledgment. The Royal Thai Government is thanked
for a studentship (to N.N.).
Supporting Information Available: Full experimental details
for the synthesis of complexes 1–16; .cif formatted files containing
crystallographic data for complexes 11 and 15; crystallographic
refinement details. This material is available free of charge via the
Internet at http://pubs.acs.org.
Despite the variation in the ancillary ligand substituents,
the microstructures of the PLA materials generated using
6–11 show little divergence from a moderate heterotactic
(24) Gregson, C. K. A.; Blackmore, I. J.; Gibson, V. C.; Long, N. J.;
Marshall, E. L.; White, A. J. P. Dalton Trans. 2006, 3134–3140.
IC701671S
5424 Inorganic Chemistry, Vol. 47, No. 12, 2008