Single-site bismuth alkoxide catalysts for the ring-opening polymerization of lactide

Article abstract of DOI:10.1039/c3dt50629k

Salen bismuth alkoxides, where the salen ligand contains 2,4-di-tert-butylphenoxy groups and one of ethylene, cyclohexane or ortho-phenyl as a backbone have been prepared from reactions involving Bi[N(SiMe ₃)₂]₃ and the free salen ligand followed by alcoholysis (Bu^tOH, PrⁱOH and 2,6-Bu^t ₂C₆H₃OH). The molecular structures of the salen ligand with the cyclohexyl back-bone have been determined for the complexes salenBiCl and salenBiOC₆H₃-2,6-Bu^t₂. The chloro compound is a dimer with chloride bridges while the phenoxide is monomeric with an unusually distorted five-coordinate geometry. The phenoxide and tert-butoxide complexes have been employed in the ring-opening polymerization of lactides (l- and rac-) to give polylactides, PLAs. With rac-LA heterotactic PLA is formed preferentially, P_r = ～0.9, in dichloromethane or toluene at room temperature. The reaction is first order in [Bi] and is notably faster than most aluminum and zinc initiators as well as tin(ii) octanoate. These results are discussed in terms of a recent report on the polymerization of LA by Peptobismol and bismuth subsalicylate.

Full text of DOI:10.1039/c3dt50629k

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Single-site bismuth alkoxide catalysts for the
ring-opening polymerization of lactide†
Cite this: Dalton Trans., 2013, 42, 11234
Vagulejan Balasanthiran, Malcolm H. Chisholm,* Christopher B. Durr and
Judith C. Gallucci
Salen bismuth alkoxides, where the salen ligand contains 2,4-di-tert-butylphenoxy groups and one of
ethylene, cyclohexane or ortho-phenyl as a backbone have been prepared from reactions involving
Bi[N(SiMe3)2]3 and the free salen ligand followed by alcoholysis (ButOH, PriOH and 2,6-But2C6H3OH). The
molecular structures of the salen ligand with the cyclohexyl back-bone have been determined for the
complexes salenBiCl and salenBiOC6H3-2,6-But2. The chloro compound is a dimer with chloride bridges
while the phenoxide is monomeric with an unusually distorted five-coordinate geometry. The phenoxide
and tert-butoxide complexes have been employed in the ring-opening polymerization of lactides (L- and
rac-) to give polylactides, PLAs. With rac-LA heterotactic PLA is formed preferentially, Pr = ∼0.9, in dichloro-
methane or toluene at room temperature. The reaction is first order in [Bi] and is notably faster than
most aluminum and zinc initiators as well as tin(II) octanoate. These results are discussed in terms of a
Received 6th March 2013,
Accepted 15th June 2013
DOI: 10.1039/c3dt50629k
www.rsc.org/dalton
recent report on the polymerization of LA by Peptobismol and bismuth subsalicylate.
®
Introduction
The ring-opening polymerization (ROP) of lactides to produce
the biodegradable and biocompatible polymers polylactides,
PLAs, has been shown to be accomplished by both organic1
and coordinate catalysis2 with the latter involving metal com-
plexes having nucleophilic initiator ligands such as alkoxides,
amides, and hydroxides. Commercially, Sn(Oct)2, where Oct =
2-ethyl hexanoate, is employed in a melt polymerization
process. The alkanoate ligand is not the initiator but rather a
Sn–OH bond that is formed reversibly in the presence of water
and the chain propagation can be understood in terms of the
reactions shown in Scheme 1.
This living or immortal system has an essential require-
ment that the M–OH bond be chemically persistent during the
reaction conditions. It should not react to form an inert oxo
metal derivative nor react with other species present such as
CO2 to form an inert carbonate. This led us to question
whether alternative M–OH bonds (but not those of alkali
metals that effect epimerization) might be similarly active in
melt polymerizations. We subsequently investigated the
Scheme 1 Mechanism of ROP of LA by tin(II) octanoate [where POH = hydroxyl
terminated oligomer].
reactivity of a number of biocompatible metal containing oral
relief aids and dietary supplements that contain hydroxyl
groups, and found that the most active of the bismuth contain-
ing species was bismuth subsalicylate (BSS) which is the active
ingredient in Peptobismol®.3 Both Bi(III) and Sn(II) has can
support M–OH bonds that will reversibly react to form an oxo
species as shown in eqn (1).
2½MꢀOH Ð ½Mꢀ2O þ H2O
ð1Þ
As a melt polymerization catalyst bismuth subsalicylate was
just slightly less active than Sn(Oct)2 based on its empirical
formula. Since bismuth subsalicylate is a polymeric material
of unknown structure,4 we reasoned that the activity of
bismuth should be higher than that of tin based on the fact
that only a few active sites were present when the BSS powder
was employed. Interestingly, no well defined bismuth alkoxide
Department of Chemistry and Biochemistry, The Ohio State University, 100W.
18th Avenue, Columbus, Ohio 43210-1185, USA.
E-mail: Chisholm@chemistry.ohio-state.edu
†Electronic supplementary information (ESI) available: Mn values with
times and plots of Mn and PDI with % conversion. CCDC 928067–928069. For
ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c3dt50629k
11234 | Dalton Trans., 2013, 42, 11234–11241
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
has been employed in the ROP of LAs, though Kricheldorf5 has cases we observed sample decomposition in the X-ray beam or
reported that bismuth alkanoates Bi(O2CR)3 may be used as upon placing the sample within the oil employed for air-sensi-
initiators. Here presumably the reactions proceed similarly to tive work. Decomposition in the oil may be due to either
Sn(Oct)2 due to adventitious water (see Scheme 1). We report solvent of crystallization loss or fast sample decomposition.
now on the preparation of single-site Bi(III) alkoxide complexes We were, however, able to obtain crystals of a phenoxide
supported by the Schiff base salen class of ligands and show derivative and note that the molecular structure of the homo-
that these are indeed active in the ROP of LA and more active leptic phenoxides Bi(OC6H3-2,6-R2)3 where R = Ph, Pri 6 or Me7
than Sn(Oct)2 under comparable conditions.
are known to be monomeric while that of the pentafluoro-
phenoxide complex is dimeric.8 Other alkoxides of the form
Bi(OR)3 are polymeric and the ethoxide has been shown to be
have a cyclic octameric structure [Bi(OEt)3]8·(7 + x)EtOH.9 As a
structural model for a salenBi(OR) compound, where R = Me
or Et, we have examined the molecular structure of a sale-
nBiCl. For a bulky alkoxide or aryloxide we might anticipate
a monomeric salenBi(OR) structure but for a lesser steric
demanding group we anticipate a higher degree of aggregation.
Results and discussion
Synthesis
The salen ligands employed in this study are depicted in
Chart 1. Each of these Schiff bases employs the 2,4-di-tert-
butylphenoxy groups and they differ only in the back-bone of
the ligand which is employed in the condensation reaction for
their synthesis, namely ethylene diamine, 1,2-cyclohexyldiamine
or ortho-phenyl diamine. For abbreviation we refer hereafter to
each of these as en-salen, cy-salen and ph-salen, respectively.
Two synthetic procedures toward the synthesis of the
salenBi(OR) compounds have been employed. The first
involves the synthesis of Bi[N(SiMe3)2]3 and its reaction with
the free salen-H2 ligands shown in Chart 1 leading to the prepa-
ration of salenBiN(SiMe3)2 which was then allowed to react with
the appropriate alcohol or phenol. The second was the syn-
thesis of salenBiCl followed by a metathetic reaction involving
KOBut. The salenBiCl compound can be readily prepared by
the direct reaction between BiCl3 and Na2salen or via the reac-
tion between Bi[N(SiMe3)2]2Cl and the free salen-H2 ligand.
The specific syntheses are reported in the Experimental
section. In general the alkoxide and phenoxide complexes were
soluble in toluene, hexane, pentane, THF and dichloro-
methane. However, chloroform was often found to be reactive,
not surprisingly as it is well known to react with bases via elimi-
nation of HCl. The synthesis of several of these alkoxides in
common organic solvents often presented a problem in
obtaining crystals suitable for single crystal X-ray diffraction
because of their high solubility.
SalenBiCl structure
Two crystalline samples were examined en-salenBiCl (3) and
cy-salenBiCl (2). The former complex was determined from a
twinned data set and it also suffered from disorder of various
types of solvent molecules. As a result, it was necessary to use
many restraints in the refinement of the model, and the final
results are less than optimal. However, the structure clearly
indicates an arrangement of tetranuclear [BiCl]4 units, which
indicate the ability of the Bi(III) ions to establish six-coordi-
nation by formation of a tetranuclear aggregate as opposed to
a more simple dimeric structure.
The central [BiCl]4 core is shown in Fig. 1 where the κ4-en-
salen ligands have been omitted. This 8 membered ring con-
tains a non-crystallographic two-fold rotation axis through its
center and a boat conformation.
The molecular structure of the cy-salenBiCl was more suc-
cessfully determined and its dimeric chloride bridged struc-
ture is shown in Fig. 2. There is a planar [BiCl]2 unit
supported by κ4-cy-salen ligands. Also we see that the N2O2
unit of the cy-salen ligand is planar or near planar and as such
the central N2O2BiCl2 core can simply be described as being
derived from an octahedral geometry. The view in Fig. 2 which
is almost parallel to the Bi2Cl2 plane emphasizes the tilt of the
two cy-salen ligands which allows one to speculate that this
distortion arises from what can be termed a stereochemically
active lone-pair. The present structure is seemingly most
closely related to that of tpClppBiCl (tpClpp = 5,10,15,20-tetra-
p-chlorophenylporphyrin)10 except that the latter structure
does not clearly implicate a stereochemically active lone pair.
In the view shown in Fig. 2, the two “lone pairs” would be
mutually anti as indicated in the simple representation shown
below in Fig. 3.
Single crystal X-ray studies
Several attempts were made to obtain a full structural determi-
nation of various tert-butoxides but with no success. In all
Selected bond distances and angles for [cy-salenBiCl]2 are
given in Table 1.
SalenBi-phenoxide structure
The molecular structure of cy-salenBiOC6H3-2,6-But2 (1) is
shown in Fig. 4. This structure can easily be described as that
of a distorted square based pyramid. The view shown in Fig. 4
Chart 1 Free salen ligands with different backbones employed in this work.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 11234–11241 | 11235

View Article Online
Paper
Dalton Transactions
Scheme 2 Reaction scheme for the general synthesis of salenBiOR.
Fig. 2 ORTEP representations of [cy-salenBiCl]2 (orange = bismuth, green =
chlorine, scarlet = oxygen, blue = nitrogen, gray = carbon) drawn at 50% prob-
ability. Hydrogen atoms, solvent molecules and tert-butyl groups are excluded
for clarity.
Fig. 3 Sketch of the locations and influence of the stereochemically active
lone-pairs in [cy-salenBiCl]2.
Table 1 Selected bond distances (Å) and bond angles (°) for cy-salenBiCl
Compound 2
Fig. 1 The central core of the tetranuclear en-salenBiCl molecule (top) and its
Bi(1)–O(2)
Bi(1)–O(1)
Bi(1)–N(1)
2.154(7)
2.233(7)
2.325(8)
Bi(1)–Cl(1)
Bi(1)–N(2)
2.932(3)
2.404(8)
central [BiCl]4 core (bottom).
O(2)–Bi(1)–O(1)
O(1)–Bi(1)–N(1)
O(1)–Bi(1)–N(2)
O(2)–Bi(1)–Cl(1)
N(1)–Bi(1)–Cl(1)
C(22)–N(1)–Bi(1)
C(15)–N(2)–Bi(1)
74.6(3)
115.9(3)
74.7(3)
82.50(19)
80.9(2)
125.7(7)
121.7(7)
O(2)–Bi(1)–N(1)
O(2)–Bi(1)–N(2)
N(1)–Bi(1)–N(2)
O(1)–Bi(1)–Cl(1)
N(2)–Bi(1)–Cl(1)
C(21)–N(1)–Bi(1)
C(16)–N(2)–Bi(1)
79.3(3)
118.6(3)
68.7(3)
147.7(2)
137.3(2)
110.4(6)
116.3(6)
emphasizes again the presence of the “stereochemically active
lone pair”. Not only does this influence the N2O2BiO geometry
but it also influences the Bi–O–C angle which is 130.7°. Typi-
cally in terminal metal–phenoxide bonds we see linear M–O–C
angles. This is favored by the oxygen lone pair interaction with
the π-system of the aryl ring and often in transition metal
complexes by Mdπ–Opπ bonding. In the present case if we
assume a linear M–O–C moiety one of the oxygen pπ orbitals Bi lone-pair. By rehybridization toward sp2 this interaction is
would be forced into a filled–filled orbital interaction with the minimized as schematically represented below in Fig. 5.
11236 | Dalton Trans., 2013, 42, 11234–11241
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Table 2 Selected bond distances (Å) and bond angles (°) for cy-salenBiOC6H3-
2,6-But2
Compound 1
N(1)–Bi
O(1)–Bi
O(3)–Bi
N(1)–C(16A)
N(2)–C(21B)
2.335(3)
2.237(3)
2.324(2)
1.569(7)
1.554(9)
N(2)–Bi
O(2)–Bi
N(1)–C(16B)
N(2)–C(21A)
2.343(3)
2.260(2)
1.478(9)
1.506(6)
C(37)–O(3)–Bi
C(1)–O(1)–Bi
O(1)–Bi–O(3)
O(1)–Bi–N(1)
O(3)–Bi–N(1)
O(2)–Bi–N(2)
N(1)–Bi–N(2)
130.7(2)
132.7(2)
133.96(9)
77.66(10)
6.72(10)
74.73(11)
68.72(11)
C(24)–O(2)–Bi
O(1)–Bi–O(2)
O(2)–Bi–O(3)
O(2)–Bi–N(1)
O(1)–Bi–N(2)
O(3)–Bi–N(2)
C(22)–N(2)–Bi
117.5(2)
83.54(9)
142.05(9)
111.38(11)
128.87(12)
82.11(11)
123.8(3)
polymerizations of L- and rac-lactide in either toluene or
dichloromethane solutions at room temperature. Interestingly,
the polymerization of rac-LA proceeds to give predominantly
heterotactic PLA with isi and sis tetrads. The preference for the
alternating ring-opening of L and D lactides is given by Pr ∼ 0.9
when reactions are carried out to 80% completion. The spec-
trum of a sample of heterotactic PLA is given in Fig. 6.
We have also studied the kinetics of these polymerizations
involving the initiator cy-salenBiOC6H3-2,6-But2 at room temp-
erature in dichloromethane. Plots of ln{[LA]0/[LA]t} versus time
are shown in Fig. 7 and reveal the living polymerization by the
catalyst system and moreover, that the order of polymerization
is first order in [Bi]. The plot of −ln kapp versus −ln[cat] is
shown in Fig. 8. The kp value of 5 × 10−2 M−1 s−1 is notably
faster than any salenAl(OR) system and just a little slower than
the most active Zn based catalyst systems.12,13 The kp value for
the polymerization of rac-LA by cy-salenBiOBut is 9.2 × 10−2
M−1 s−1 in toluene at room temperature. For a direct compari-
son with cy-salenAlOPri which has kp = 9.02 × 10−3 M−1 s−1 at
70 °C in toluene,14 we note that the bismuth complex is ∼10
times faster at room temperature.
Fig. 4 (Top) ORTEP representation of cy-salenBiOC6H3-2,6-tBu (orange
=
bismuth, scarlet = oxygen, blue = nitrogen, gray = carbon) drawn at 50% prob-
ability. Hydrogen atoms excluded for clarity. (Bottom) ORTEP representation of
cy-salenBiOC6H3-2,6-But (with tert-butyl groups removed for clarity) and empha-
sizing the nature of the BiO3N2 core.
Concluding remarks
This work describes the first single-site alkoxide catalyst for the
ROP of lactide by a bismuth complex. The reactivity of the Bi–OR
bond is greater than that of aluminum alkoxides of related
formula and notably much more active than Sn(Oct)2 at room
temperature. The compounds of formula salenBiOR are,
however, regrettably not tolerant to water which is clearly a limit-
ing feature when compared to Sn(Oct)2 or bismuth subsalicylate.
The reaction with water is believed to lead to ligand scrambling
and crystals of a bismuth salen complex of formula (en-salen)Bi-
(en-salenH) have been obtained and will be reported elsewhere.
Fig. 5 Minimization of lone pair–lone pair interactions.
A summary of crystallographic data for the three molecules
described above is given in Table 3 and selected bond dis-
tances for cy-salenBiOC6H3-2,6-But2 and bond angles are given
in Table 2.
Experimental
Ring-opening polymerization of lactide
General considerations
Both the tert-butoxide and the phenoxide complexes with the All the experiments were carried out under a rigorously dried
cy-salen supporting ligands are active in the ring opening nitrogen atmosphere either using Schlenk techniques or a
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 11234–11241 | 11237

View Article Online
Paper
Dalton Transactions
Table 3 Selected crystallographic information for cy-salenBiOC6H3-2,6-But (1), [cy-salenBiCl]2 (2) and en-salenBiCl (3)
Compound
1
2
3
Chemical formula
C
50H73BiN2O3
C76H112Bi2Cl10N4O4
C128H184Bi4Cl4N8O8, 0.598 (C6H19NSi2),
C4H8O, 0.2 (C4H10O),
3124.01
150(2)
Orthorhombic, P212121
17.4471(3)
Formula weight
Temperature (K)
Space group
a (Å)
b (Å)
c (Å)
959.08
150(2)
1918.16
150(2)
Monoclinic, P21/n
15.2904(5)
10.4994(3)
27.3011(8)
ˉ
Triclinic, P1
12.4091(1)
14.7848(2)
14.9983(1)
110.801(1)
102.945(1)
100.046(1)
2407.55(4)
2
17.4510(3)
58.3860(11)
α (°)
β (°)
γ (°)
103.756(1)
V (Å3)
Z
4257.2(2)
2
1.496
17 776.7(5)
4
1.167
Dcalcd (Mg m−3
)
1.323
Crystal size (mm)
0.19 × 0.12 × 0.12
1.66 to 27.48°
3.701
0.12 × 0.10 × 0.08
1.72 to 25.03°
4.488
0.19 × 0.19 × 0.35
1.05 to 25.19°
4.06
Theta range for data collection
μ (Mo, Kα) (mm−1
F(000)
)
988
1928
6297
Reflections collected
Unique reflections
51 970
11 011 [R(int) = 0.040]
99.9%
11 011/0/542
2.40 (6.08)
3.31 (8.33)
1.198
63 237
7490 [R(int) = 0.112]
99.6%
7490/0/433
6.92 (9.95)
18.44 (20.20)
1.046
214 449
31 482 [R(int) = 0.083]
98.7%
31 482/206/1426
5.43 (7.86)
13.25 (14.39)
1.047
Completeness to θmax
Data/restraints/parameters
R1a (%) (all data)
wR2b (%) (all data)
Goodness-of-fit on F2
Largest diff. peak and hole (e Å−3
)
2.154 and −0.941
5.126 and −2.851
1.748 and −0.851
a R1 = Σ||Fo| − |Fc||/Σ|Fo| × 100. b wR2 = {Σ[w(Fo2 − Fc2)2]/Σ[w(Fo2)2]}1/2 × 100.
Fig. 7 Linear plots of ln{[LA]0/[LA]t} versus time (min) for the polymerization of
rac-LA initiated by cy-salenBiOC6H3-2,6-But2 in CH2Cl2 at room temperature.
sublimation, followed by recrystallization in dry toluene. Then
the lactides were dried under reduced pressure at room temp-
erature overnight. Chloroform-d and benzene-d6 were pur-
chased from Cambridge Isotopes and distilled under nitrogen
over CaH2. Bi[N(SiMe3)2]3 was prepared according to the litera-
ture procedure.15
Fig. 6 1H NMR spectra (CDCl3, 400 MHz) of the homodecoupled CH resonance
of poly(rac-lactide) obtained by cy-salenBiOBut as initiator. For details of assign-
ments see ref. 11.
glove box. Pentane, hexane, THF, toluene, and dichloro-
methane were distilled under nitrogen over CaH2. Methanol,
ethanol, BiCl3, LiN(SiMe3)2, 1,2-cyclohexanediamine, ethylene-
diamine, ortho-aminoaniline, anhydrous ButOH, anhydrous
Measurements
PriOH and 2,6-di-ButC6H3OH were purchased from Sigma 1H and 13C spectra were recorded in CDCl3 (δ: 7.24) or C6D6
Aldrich. 3,5-Di-tert-butylsalicylaldehyde was purchased from (δ: 7. 15) and 13C6D6 (δ: 128.06) on Bruker DPX-400 NMR or
1
Alfa Aesar. All of the above chemicals were used without DRX-500 NMR spectrometers and referenced against the H or
further purification. L-Lactide (L-LA), and rac-lactide (rac-LA) 13C signal quoted. Gel permeation chromatography (GPC)
were purchased from Sigma Aldrich, and purified by measurements were carried out using a Waters 1525 binary
11238 | Dalton Trans., 2013, 42, 11234–11241
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
procedures were followed for the kinetics studies of cy-salen-
BiOBut in toluene at room temperature.
General procedure for the synthesis of salen ligands
All the salen ligands were prepared by the condensation reac-
tion of 3,5-di-tert-butyl-2-hydroxybenzaldehyde with the appro-
priate diamine (2 : 1 ratio) in ethanol under reflux for 3 h.16
The desired product was precipitated in ice bath and collected
by filtration. Then the ligands were dried under vacuum at
60 °C overnight.
Synthesis of cy-salenBiN(SiMe3)2
The solution of cy-salenH2 (1.0 g, 1.8 mmol) in 15 mL of THF
and the solution of Bi[N(SiMe3)2]3 (1.2 g, 1.8 mmol) in 15 mL
of THF were prepared in a glove box. Then the cy-salenH2 solu-
tion was transferred to the Bi[N(SiMe3)2]3 solution via cannula
and the solution was stirred at room temperature for 12 h in a
Schlenk flask under a N2 atmosphere. All the volatile components
were removed under vacuum. The completion of the reaction was
monitored by 1H NMR. Orange microcrystalline product was
obtained with 95% yield. The resulting product was employed
in the synthesis below without further purification.
Fig. 8 Plot of −ln(kapp) versus −ln[Cat] for the rac-LA initiated by cy-salen-
BiOC6H3-2,6-But2 in CH2Cl2 at room temperature.
HPLC pump and Waters 1525 differential refractometer
equipped with styragel HR 2 and 4 columns (100 to 1000 Å).
THF was used as an eluent at 1 mL min−1 at 40 °C and the cali-
bration was done with polystyrene standards. MALDI-TOF
mass spectra were collected on a Bruker Microflex mass
spectrometer.
Synthesis of cy-salenBiOBut
1.0 g (1.1 mmol) of cy-salenBiN(SiMe3)2 was dissolved in THF
and 0.15 mL (1.6 mmol) of ButOH was added and stirred for
6 h at room temperature. The volatile components were
removed under vacuum and the crude product was dissolved
in pentane and placed in a freezer at −25 °C for 2 days. An
orange precipitate (product) was obtained in 70% yield by fil-
tration. 1H NMR (C6D6, δ, ppm, 500 MHz) 1.41 (s, 18 H,
C(CH3)3), 1.76 (s, 18 H, C(CH3)3), 0.60–0.90, 1.42, 1.44, 2.52, 3.59
(cyclohexyl), 1.46 (s, 9H, OC(CH3)3), 7.03 (s, 1H, ArH), 7.07 (s,
1H, ArH), 7.80 (s, 2H, ArH), 7.82 (s, 1H, NvCH), 7.88 (s, 1H,
NvCH). 13C NMR (C6D6, δ, ppm, 500 MHz) 24.84, 25.05,
30.15, 30.24, 30.63, 31.48, 31.77, 31.79 (tBu), 33.96, 34.25,
34.39, 35.73 (tBu), 66.27, 69.77 (HC–N), 121.93, 122.02, 129.19,
129.43, 130.57, 130.85, 135.16, 135.26, 142.01, 142.08, 164.93,
166.23 (phenyl), 166.83, 167.23 (HCvN). MS(MALDI-TOF): m/z
M+ calculated 826.45; found 826.14.
General procedure for lactide polymerization
0.500 g (3.47 mmol) of L-LA/rac-LA was loaded into a Schlenk
flask containing a magnetic bar and dissolved in 10 mL
CH2Cl2 in the glove box. An appropriate amount of initiator
(0.0347 mmol initiator for [LA]/[initiator] = 100 and 3.47 × 10−3
mmol initiator for [LA]/[initiator] = 1000) was loaded into
another flask and dissolved in 10 mL CH2Cl2 in the glove box.
Both flasks were taken out from the glove box and attached to
the Schlenk line. The initiator solution was quickly added to
the LA containing flask using a cannula and stirred at room
temperature. The polymerization was quenched in 5 N acidic
methanol. The polymer was precipitated in excess methanol
and dried under high vacuum. The conversion of LA was esti-
mated by 1H NMR spectroscopy and the molecular weights
and PDI were determined by GPC.
Synthesis of cy-salenBiOC6H3-2,6But2
General procedures for kinetics studies of rac-lactide
polymerization
1.0 g (1.1 mmol) of cy-salenBiN(SiMe3)2 was dissolved in THF
and 0.25 mL (1.1 mmol) of OH-C6H3-2,6But2 was added and
0.500 g (3.47 mmol) of rac-LA was loaded in the Schlenk flask stirred for 6 h at room temperature. The volatile components
containing a magnetic bar and dissolved in 10 mL CH2Cl2 in were removed under vacuum and the crude product was dis-
the glove box. 33.0 mg of cy-salenBiOC6H3-2,6-OBut2 initiator solved in pentane and placed in a freezer at 0 °C for 2 days. A
(0.0347 mmol initiator for [rac-LA]/[initiator] = 100) was loaded yellow precipitate (product) was obtained in 75% yield. Crys-
in another Schlenk flask and dissolved in 10 mL CH2Cl2 in the tals suitable for single-crystal X-ray crystallography were
glove box. Both flasks were taken out from the glove box and obtained by placing a concentrated hexane solution in freezer
1
attached to the Schlenk line. The initiator solution was quickly at −25 °C for a week. H NMR (C6D6, δ, ppm, 500 MHz) 1.32,
transferred by cannula into the lactide solution and stirred at 1.34 (s, s, 18 H, C(CH3)3), 1.43 (s, 18 H, C(CH3)3), 1.61, 1.64 (s,
room temperature. Then ∼0.5 mL aliquots were removed at s, 18 H, C(CH3)3), 0.87, 1.23, 1.47, 1.71, 2.39, 5.51 (cyclohexyl),
appropriate time intervals and quenched with 5 N acidic 6.69 (t, 1H, ArH), 6.98 (d, 1H, ArH), 7.24 (d, 1H, ArH), 7.39 (d,
MeOH. The aliquots were dried under vacuum and the % con- 2H, ArH), 7.83 (d, 1H, ArH), 7.87 (d, 1H, ArH), 8.21 (s, 1H,
versions were obtained by 1H NMR spectroscopy. Similar NvCH), 8.26 (s, 1H, NvCH). 13C NMR (C6D6, δ, ppm,
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 11234–11241 | 11239

View Article Online
Paper
Dalton Transactions
500 MHz) 24.52, 25.65, 25.83, 27.78, (CH2 in cyclohexyl) 30.10, was added and stirred for overnight. All the volatile com-
30.43, 31.59, 31.70 (C(CH3)3), 32.18, 32.25, 34.05, 34.08 ponents were removed under vacuum and the crude product
(C(CH3)3), 34.40, 34.98, 35.63, 35.76 (C(CH3)3), 65.28, 69.93 was dissolved in pentane and centrifuged. The resulting solu-
(CH–N), 115.70, 121.69, 122.61, 124.95, 125.41, 128.56, 128.86, tion was transferred to another flask and placed in a freezer at
130.89, 132.29, 137.93, 138.29, 140.63, 142.57, 142.91, 162.12, −25 °C for three months. Yellow colored crystals were obtained
165.29 (phenyl), 164.71, 169.94 (HCvN). MS(MALDI-TOF): m/z with a yield of 85%.
M+ calculated 958.54; found 958.42.
Method 2. Bi[N(SiMe3)2]2Cl was prepared by adding
2 equivalents of LiN(SiMe3)2 to 1 equivalent of BiCl3 in THF at
0 °C. 1.0 g (2.0 mmol); en-salenH2 dissolved in THF was then
Synthesis of cy-salenBiOPri
1.0 g (1.1 mmol) of cy-salenBiN(SiMe3)2 was dissolved in THF added to the solution of 1.1 g (2.0 mmol) Bi[N(SiMe3)2]2Cl in
and 0.15 mL, (2.0 mmol) of PriOH was added and stirred for THF. The resulting solution was stirred overnight and all vola-
6 h at room temperature. The volatile components were tile components were removed under vacuum. The crude
removed under vacuum and the crude product was dissolved product was dissolved in pentane and placed in a freezer at
in pentane and placed in a freezer at −25 °C for 2 days. An −25 °C for one month. Yellow colored crystals were obtained
orange precipitate (product) was obtained in 65% yield. 1H in 70% yield.
i
NMR (C6D6, δ, ppm, 500 MHz) 1.03 (bs(broad)6H, CH(CH3)2),
1H NMR (C6D6, δ, ppm, 500 MHz) 1.33 (bs, 18 H, C(CH3)3),
1.39 (s, 18 H, C(CH3)3), 1.77 (s, 18 H, C(CH3)3), 0.68, 0.74, 1.15, 1.71 (bs, 18 H, C(CH3)3), 3.00 (bs, 2H, CH2), 4.16 (bs, 2H, CH2),
i
1.45, 2.59, 3.17 (cyclohexyl), 4.0–4.4 (m, 1H, CH(CH3)2), 7.06 6.89 (bs, 2H, ArH), 7.28 (bs, 2H, NvCH), 7.82 (bs, 2H, ArH).
(s, 1H, ArH), 7.09 (s, 1H, ArH), 7.82 (2, 2H, ArH), 7.82 (s, 1H, MS(MALDI-TOF): m/z M+ calculated 734.30; found 734.94.
NvCH), 8.04 (s, 1H, NvCH). 13C NMR (C6D6, δ, ppm,
500 MHz) 24.90, 24.99, 25.97, 27.15, 29.66, 30.05, 30.13, 31.77,
31.90, 34.00, 35.76, 35.82, 68.11, 68.23, 122.21, 129.40, 129.54,
130.85, 130.88, 135.40, 135.64, 141.89, 142.02, 164.22, 166.89,
166.99. MS(MALDI-TOF): m/z M+ calculated 812.43; found
812.81.
Synthesis of cy-salenBiCl
1.0 g of cy-salenH2 (1.8 mmol) dissolved in THF and 0.12 g
(5.0 mmol) of NaH were charged separately in Schlenk flasks
in the glove box. The cy-salenH2 solution was cooled in an ice
bath and transferred to the NaH containing Schlenk flask via
cannula and stirred overnight. The conversion was monitored
by the disappearance of OH the peak by 1H NMR. To the resul-
tant mixture 0.58 g (1.8 mmol) of BiCl3 was added and stirred
overnight. All the volatile components were removed under
vacuum and the crude product was dissolved in pentane and
centrifuged. The resulting solution was transferred to another
flask and placed in a freezer at −25 °C for one month. Yellow
colored crystals were obtained in 65% yield.
Synthesis of ph-salenBiOBut
A solution of ph-salenH2 (1.0 g, 1.8 mmol) in 15 mL of THF
and a solution of Bi[N(SiMe3)2]3 (1.2 g, 1.8 mmol) in 15 mL of
THF were prepared in a glove box. Then the cy-salenH2 solu-
tion was transferred to the Bi[N(SiMe3)2]3 solution via cannula
and the solution was stirred at room temperature for 12 h. The
volatile components were removed under vacuum. The com-
1
pletion of the reaction was monitored by H NMR. The resul-
1H NMR (C6D6, δ, ppm, 400 MHz) 0.70, 1.13, 1.35, 1.45,
2.14, 5.15 (cyclohexyl), 1.30, 1.34, 1.68, 1.75 (C(CH3)3), 6.89 (s,
1H, ArH), 7.08 (bs, 1H, ArH), 7.46 (bs, 1H, ArH), 7.77 (s, 1H,
NvCH), 7.79 (s, 1H, NvCH), 7.88 (s, 1H, ArH). MS(MALDI--
TOF): m/z M+ calculated 788.35; found 788.24.
tant product was redissolved in THF and 0.20 mL (2.1 mmol)
ButOH was added and stirred for another 6 h. All the volatile
components were removed under vacuum. The residue was
redissolved in pentane and placed in a freezer at −25 °C for
two days. A deep orange precipitate was obtained in 80% yield.
1H NMR (C6D6, δ, ppm, 500 MHz) 1.05 (bs, 9 H, C(CH3)3), 1.34
(bs, 18 H, C(CH3)3), 1.75 (bs, 18 H, C(CH3)3), 6.51 (s, 2H, ArH),
Crystallographic studies
6.90 (s, 2H, ArH), 7.04 (s, 2H, ArH), 7.84 (s, 2H, ArH), 8.18 (s, Single crystals of 1–3 were isolated under a pool of fluorinated
2H, NvCH). 13C NMR (C6D6, δ, ppm, 500 MHz) 30.30, 31.44, oil and were found to be quite reactive. Examination of the
33.89, 35.69 (C(CH3)3), 120.17, 122.78, 127.34, 129.64, 131.91, diffraction pattern was done on a Nonius Kappa CCD diffracto-
136.19, 141.95, 144.0, 162.77 (phenyl), 170.50 (HCvN). MS meter with Mo Kα radiation. All work was done at 150 K
(MALDI-TOF): m/z M+ calculated 820.40; found 820.94.
using an Oxford Cryosystems Cryostream Cooler. Data inte-
gration was done with Denzo, and scaling and merging of the
data was done with SADABS17 for 3 and Scalepack18 for 1 and
Synthesis of en-salenBiCl
Method 1. 1.0 g of en-salenH2 (2.0 mmol) dissolved in THF 2. The structures were solved by the direct methods program
and 0.12 g (5.0 mmol) of NaH were charged in Schlenk flasks in SHELXS-97. Full-matrix least-squares refinements based on
in the glove box. The flasks were taken out from the glove box F2 were performed in SHELXL-97,19 as incorporated in the
and attached to the Schlenk line. The en-salenH2 solution was WinGX package.20 For each methyl group, the hydrogen atoms
cooled in an ice bath and transferred to the NaH containing were added at calculated positions using a riding model with
Schlenk flask via cannula and stirred overnight. The conver- U(H) = 1.5Ueq (bonded carbon atom). The rest of the hydrogen
sion was monitored by the disappearance of the OH peak by atoms were included in the model at calculated positions
1H NMR. To the resultant mixture 0.64 g (2.0 mmol) of BiCl3 using a riding model with U(H) = 1.2Ueq (bonded atom).
11240 | Dalton Trans., 2013, 42, 11234–11241
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Neutral atom scattering factors were used and include terms
for anomalous dispersion.21
Kricheldorf, H. Hachmann-Thiessen and G. Schwarz, Bio-
macromolecules, 2004, 5, 492.
Structure 1 contains a disordered cyclohexane backbone
which crystallized in two conformations. The two orientations
were found in the difference map and their occupancy was
allowed to refine to 0.55/0.45. The pivot carbon atoms C17A/
C17B and C20A/C20B were restrained with EXYZ and EADP
commands.
Structure 3 consists of an en-salen[Bi-Cl] tetramer plus
three different solvent molecules: fully occupied THF,
(CH3)3Si–NH–Si(CH3)3 with a refined occupancy factor of 0.598(9),
and diethyl ether with an occupancy factor set to 0.2. There
are other regions of lower electron density, which are most
likely very disordered solvent molecules, and these were
not modeled. Pseudo-merohedral twinning is present and the
6 X. Kou, X. Wang, D. Mendoza-Espinosa, L. N. Zakharov,
A. L. Rheingdd, W. H. Watson, K. A. Brien, L. K. Jayarathna
and T. A. Hanna, Inorg. Chem., 2009, 48, 11002.
7 W. J. Evans, J. H. Hain and J. W. Ziller, J. Chem. Soc., Chem.
Commun., 1989, 1628.
8 C. M. Jones, M. D. Burkart, R. E. Bachman, D. L. Serra,
S. Hwu and K. H. Whitmore, Inorg. Chem., 1993, 32, 5136;
K. H. Whitmore, S. Hoppe, O. Sydora, J. L. Jolas and
C. M. Jones, Inorg. Chem., 2000, 39, 85.
9 V. G. Kessler, N. Ya. Turova and E. P. Turevskaya, Inorg.
Chem. Commun., 2002, 5, 549; T. Hatanpaa, M. Vehkamaki,
M. Ritala and M. Leskela, Dalton Trans., 2010, 39,
3219.
following twin law was applied to the data (0–1 0 −1 0 0/0 0 1) 10 B. Boitrel, M. Breede, P. J. Brothers, M. Hodgson,
with a twin fraction of 0.3633(8).
L. Michaudet, C. E. F. Rickard and N. A. Salim, Dalton
It was necessary to use many restraints (SADI, DFIX, and
Trans., 2003, 1803.
FLAT) during the refinement for both the Bi tetramer and the 11 M. T. Zell, B. E. Padden, A. J. Patrick, K. A. M. Thakur,
solvent molecules.
R. T. Kean, M. A. Hillmyer and E. J. Munson, Macro-
molecules, 2002, 35, 7700.
12 L. E. Breyfogle, C. K. Williams, V. G. Young Jr,
M. A. Hillmyer and W. B. Tolman, Dalton Trans., 2006, 928.
13 C. A. Wheaton and P. G. Hayes, Chem. Commun., 2010, 46,
8404.
Acknowledgements
We thank the Department of Energy, Office of Basic Sciences,
Chemistry Division for support of this work.
14 Z. Zhong, P. J. Dijkstra and J. Feijen, J. Am. Chem. Soc.,
2003, 125, 11291.
15 C. J. Carmalt, N. A. Compton, R. J. Errington, G. A. Fisher,
I. Moenandar, N. C. Norman and K. H. Whitmire, Inorg.
Synth., 1997, 31, 98.
References
1 F. Nederberg, E. F. Connor, M. Moller, T. Glauser and 16 P. Hormnirun, E. L. Marshall, V. C. Gibson, R. I. Pugh and
J. L. Hedrick, Angew. Chem., Int. Ed., 2001, 40, 2712.
2 H. R. Kricheldorf and R. Dunsing, Makromol. Chem., 1986,
187, 1611.
3 V. Balasanthiran, T. L. Beilke and M. H. Chisholm, Dalton
Trans., 2013, 42, 9274.
4 P. C. Andrews, G. B. Deacon, C. M. Forsyth, P. C. Junk,
I. Kumar and M. Maguire, Angew. Chem., Int. Ed., 2006, 45,
5638.
A. J. P. White, Proc. Natl. Acad. Sci. U. S. A., 2006, 103,
15343.
17 G. M. Sheldrick, SADABS, University of Göttingen,
Germany.
18 Z. Otwinowski and W. Minor, Methods in Enzymology,
Vol. 276: Macromolecular Crystallography, Part A, Academic
Press, 1997.
19 G. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystallogr.,
2008, 64, 112.
5 (a) H. R. Kricheldorf, H. Hachmann-Thiessen and G. J.
Schwarz, J. Biomater. Sci., Polym. Ed., 2006, 17, 721; 20 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.
(b) H. R. Kricheldorf, K. Bornhorst and H. Hachmann- 21 International Tables for Crystallography, Kluwer Academic
Thiessen, Macromolecules, 2005, 38, 5017; (c) H. R.
Publishers, Dordrecht, 1992, vol. C.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 11234–11241 | 11241