Activation of carbon dioxide by divalent tin alkoxides complexes

Article abstract of DOI:10.1021/ic102273n

A series of terminal tin(II) alkoxides have been synthesized utilizing the bulky β-diketiminate ligand [{N(2,6-_iPr₂C ₆H₃)-C(Me)}₂CH] (BDI). The nucleophilicities of these alkoxides have been examined,

Full text of DOI:10.1021/ic102273n

Inorg. Chem. 2011, 50, 1879–1888 1879
DOI: 10.1021/ic102273n
Activation of Carbon Dioxide by Divalent Tin Alkoxides Complexes
Lorenzo Ferro, Peter B. Hitchcock, Martyn P. Coles, Hazel Cox, and J. Robin Fulton*
Department of Chemistry, University of Sussex, Falmer, Brighton BN1 9QJ, U.K.
Received November 12, 2010
A series of terminal tin(II) alkoxides have been synthesized utilizing the bulky β-diketiminate ligand [{N(2,6-iPr2C6H3)-
C(Me)}2CH] (BDI). The nucleophilicities of these alkoxides have been examined, and unexpected trends were
observed. For instance, (BDI)SnOR only reacts with highly activated aliphatic electrophiles such as methyl triflate, but
reacts reversibly with carbon dioxide. Both the rate of reaction and the degree of reversibility is dependent upon minor
changes in the alkoxide ligand, with the bulkier tert-butoxide ligand displaying slower reactivity than the corresponding
isopropyl ligand, although the latter system is a more exergonic reaction. Density Function Theory (DFT) calculations
show that the differences in the reversibility of carbon dioxide insertion can be attributed to the ground-state energy
differences of tin alkoxides while the rate of reaction is attributed to relative bond strengths of the Sn-O bonds. The
mechanism of carbon dioxide insertion is discussed.
Introduction
ligand acts as a nucleophile onto the electrophilic carbon atom
concurrent with coordination of one of the carbon dioxide
oxygen atoms to the metal center and a weakening of the
M-O and the CdO bonds (Scheme 1). An open coordination
site is unnecessary for this nucleophilic insertion reaction13
and the driving force for this reaction is generally thought
to be the strong nucleophilicity of the alkoxide or amide
ligand.1,9,11
Although tin(IV) alkoxides complexes are known to react
with carbon dioxide,7,14 the reactivity of the less Lewis acidic
tin(II) alkoxides has not been explored outside of lactide
polymerization studies.15,16 Divalent tin has one of the lowest
aqueous acidities of any metal species (pKa = 2),17,18 sig-
nificantly lower than what is predicted based upon electro-
static parameters. As such, the resultant tin hydroxide species
is very weakly basic and should not exhibit nucleophilic be-
havior. As the reactivity of alkoxides is similar to that of
hydroxides, divalent tin alkoxides should also be very weakly
basic and non-nucleophilic.
The activation of carbon dioxide by metal complexes has
recently received fresh attention because of the obvious
implication for carbon dioxide sequestering or converting
carbon dioxide into a useful carbon feedstock. Metal com-
plexes used in carbon dioxide activation generally possess
nucleophilic ligands such as alkoxides and amides; in these
systems, carbon dioxide is inserted into the M-O or M-N
bond to form a metallo-alkylcarbonate (or -alkylcarbamate)
complex.1-11 In general, most metal alkoxide complexes
react reversibly with carbon dioxide,1,3-9 presumably be-
cause of similar M-O bond strengths of the metal alkoxide
and metallo-alkylcarbonate. Metal amide complexes exhibit
both reversible12 and irreversible reactivity with carbon
dioxide.1,10 The mechanism for carbon dioxide insertion into
M-O or M-N bonds has been postulated to occur via a
four-membered transition state in which the alkoxide or amide
*To whom correspondence should be addressed. E-mail: j.r.fulton@
sussex.ac.uk.
(1) Simpson, R. D.; Bergman, R. G. Organometallics 1992, 11, 4306.
(2) Darensbourg, D. J.; Sanchez, K. M.; Rheingold, A. L. J. Am. Chem.
Soc. 1987, 109, 290.
(3) Darensbourg, D. J.; Sanchez, K. M.; Reibenspies, J. H.; Rheingold,
A. L. J. Am. Chem. Soc. 1989, 111, 7094.
We have recently synthesized a series of monomeric divalent
lead alkoxides complexes utilizing the bulky β-diketiminate
monoanionic ligand [{N(2,6-iPr2C6H3)C(Me)}2CH]- (BDI)
to stabilize the low-coordinate complexes.4,19 As with divalent
(4) Tam, E. C. Y.; Johnstone, N. C.; Ferro, L.; Hitchcock, P. B.; Fulton,
J. R. Inorg. Chem. 2009, 48, 8971.
(5) Tsuda, T.; Saegusa, T. Inorg. Chem. 1972, 11, 2561.
(6) Chisholm, M. H.; Cotton, F. A.; Extine, M. W.; Reichert, W. W.
J. Am. Chem. Soc. 1978, 100, 1727.
(7) Choi, J.-C.; Sakakura, T.; Sako, T. J. Am. Chem. Soc. 1999, 121, 3793.
(8) Mandal, S. K.; Ho, D. M.; Orchin, M. Organometallics 1993, 12, 1714.
(9) Darensbourg, D. J.; Lee, W. Z.; Phelps, A. L.; Guidry, E. Organome-
tallics 2003, 22, 5585.
(10) Boyd, C. L.; Clot, E.; Guiducci, A. E.; Mountford, P. Organome-
tallics 2005, 24, 2347.
(13) Darensbourg, D. J.; Mueller, B. L.; Bischoff, C. J.; Chojnacki, S. S.;
Reibenspies, J. H. Inorg. Chem. 1991, 30, 2418.
(14) Kizlink, J. Collect. Czech. Chem. Commun. 1993, 58, 1399.
(15) Dove, A. P.; Gibson, V. C.; Marshall, E. L.; White, A. J. P.; Williams,
D. J. Chem. Commun. 2001, 283.
(16) 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.
(17) Burgess, J. Metal Ions in Solution; Ellis Horwood Ltd: Chichester, U.K.,
1978.
(18) Cox, H.; Stace, A. J. J. Am. Chem. Soc. 2004, 126, 3939.
(19) Fulton, J. R.; Hitchcock, P. B.; Johnstone, N. C.; Tam, E. C. Y.
Dalton Trans. 2007, 3360.
(11) Brombacher, H.; Vahrenkamp, H. Inorg. Chem. 2004, 43, 6042.
(12) McCowan, C. S.; Groy, T. L.; Caudle, M. T. Inorg. Chem. 2002, 41, 1120.
r
2011 American Chemical Society
Published on Web 01/13/2011
pubs.acs.org/IC

1880 Inorganic Chemistry, Vol. 50, No. 5, 2011
Ferro et al.
Scheme 1. Insertion of Carbon Dioxide into a M-O Bond via a Four-
Membered Transition State. R is Either an Aliphatic or Aromatic
Group
increasing bulk of the alkoxy ligand, and a very slight
increase of pyramidalization around the tin metal center
with increasing bulk of the alkoxy ligand.
tin, divalent lead has a low aqueous acidity of 7.2.17 As such,
lead hydroxides, and by analogy, lead alkoxides, should only
be weakly nucleophilic. Our (BDI)PbOR (R=iPr, sBu, and
tBu) complexes displayed only sluggish reactivity with methyl
iodide, suggesting that the lead alkoxides are non-nucleophilic.
However, these lead alkoxide complexes do react readily and
reversibly with carbon dioxide to form lead-alkylcarbonate
complexes. Both the rate of reaction as well as the degree
of reversibility was found to be dependent upon the alkoxide
substituent. For instance, when (BDI)PbOiPr was treated with
carbon dioxide, quantitative formation of the corresponding
carbonate was observed. In contrast, when the alkoxide sub-
stituent was tert-butoxide, (BDI)PbOtBu, the carbonate was
only observed in solution and application of reduced pressure
resulted in the reformation of the alkoxide. Although our in-
vestigations of the lead system allowed for a qualitative study
on these observations, a more quantitative examination was
thwarted by decomposition of the tert-butyl carbonate, pre-
venting accurate equilibrium measurements. Fortunately, the
isostructural tin system has similar, but slower, reactivity pat-
terns to that of the lead alkoxides; as such, we report the results
of the thermodynamic and computational studies for the inser-
tion of carbon dioxide into a Sn(II)-O bond, as well as factors
behind the apparent non-nucleophilicity of these group 14
alkoxide complexes.
The reactivity of tin alkoxides 2a, 2b, and 2c with
aliphatic electrophiles was examined. No reactivity was
observed between the tin alkoxides and methyl iodide,
even at elevated temperatures. This is in contrast to the
apparent nucleophilic reactivity observed with the iso-
structural lead alkoxides complexes; formation of (BDI)-
PbI and the corresponding methyl ether was observed upon
heating a solution of (BDI)PbOR with methyl iodide.4 In
addition, this lack of reactivity is also in contrast to trans-
ition metal alkoxide complexes in which facile reactivity
with aliphatic electrophiles is observed.11,22 Addition of a
more reactive electrophile, methyl triflate, to 2a, 2b, and
2c does result in apparent nucleophilic behavior of the al-
koxides as the known (BDI)SnOTf complex (3) is formed
along with the corresponding methyl ether (eq 2).23 Inter-
estingly, the half-life for this reaction is dependent upon
the alkoxide group, with the tin isopropoxide 2a proving
to be the most reactive (t1/2=8 min), tin sec-butoxide 2b
to be slightly slower (t1/2=15 min) and the tin tert-but-
oxide 2c to be the most sluggish (t1/2 = 150 min).
Results
Synthesis and Reactivity. The synthesis of tin isoprop-
oxide (BDI)SnOiPr (2a) has been reported elsewhere.15
Treatment of a THF solution of (BDI)SnCl(1) with KOsBu
and KOtBu affords tin alkoxides (BDI)SnOsBu (2b) and
(BDI)SnOtBu (2c), respectively, after three days at room
temperature (eq 1). The X-ray crystal structures were de-
termined for 2b and 2c; Figure 1 shows the ORTEP dia-
grams of sec-butoxide 2b and tert-butoxide 2c. Selected
bond lengths and angles for 2a, 2b, and 2c are reported in
Table 1 and data collection parameters for 2b and 2c are
given in Table 2. Analogous to the solid-state structures
reported for 2a, as well as the isostructural lead compounds,
a pyramidal ligand arrangement is observed around the tin
metal center. This geometry is presumably a result of the
relativistic contraction of the 5s tin orbital, resulting in
minimal hybridization of the s and p orbitals, thus produ-
cing a stereochemically active lone pair.20,21 The alkoxy
group in the solid state structure of 2a, 2b, and 2c points
away from the BDI-Sn core and, in contrast to the (BDI)-
SnCl precursor, the Sn and alkoxide ligand lie on opposite
sides of the BDI backbone N2C3 plane. Only minor varia-
tions are observed between the tin alkoxide complexes,
such as an increase in the Sn-O bond length with the
In contrast to aliphatic electrophiles, compounds 2a,
2b, and 2c readily react with unsaturated electrophiles.
Although addition of CS2 and phenylisocyanate to the tin
alkoxide complexes gave intractable product mixtures,
treatment of isopropoxide 2a withmaleic anhydride results
i
in ring opened product (BDI)SnO2CCHCHCO2 Pr 4 in
87% yield (eq 3). The X-ray crystal structure was deter-
mined (Figure 2); the bond lengths and bond angles are
listed in Table 3 and data collection parameters are listed in
Table 2. Although this reactivity is known for transition
metal alkoxides this is a rare example a structurally char-
acterized terminal metal maleate complex.24,25 The unit
cell consists of two independent molecules that differ in the
conformation of the maleate ligand. In both molecules, the
metal center and the maleate ligand lie on the same side of
the N2C3 backbone plane and the ligand bond angles around
the metal center are more acute than in the alkoxide exa-
mples. The Sn-O bond length is longer than the alkoxide
analogues, potentiallydue to a stabilizing Sn O2 interaction
3 3 3
(22) Fulton, J. R.; Holland, A. W.; Fox, D. J.; Bergman, R. G. Acc. Chem.
Res. 2002, 35, 44.
(23) Ding, Y.; Roesky, H. W.; Noltemeyer, M.; Schmidt, H.-G.; Power,
P. P. Organometallics 2001, 20, 1190.
(24) Cuesta, L.; Hevia, E.; Morales, D.; Perez, J.; Riera, L.; Miguel, D.
Organometallics 2006, 25, 1717.
(25) Fulton, J. R.; Sklenak, S.; Bouwkamp, M. W.; Bergman, R. G.
J. Am. Chem. Soc. 2002, 124, 4722.
(20) Chen, M.; Fulton, J. R.; Hitchcock, P. B.; Johnstone, N. C.; Lappert,
M. F.; Protchenko, A. V. Dalton Trans. 2007, 2770.
(21) For a good description of this hybridization, see reference 16.

Article
Inorganic Chemistry, Vol. 50, No. 5, 2011 1881
Figure 1. ORTEP diagrams of (BDI)SnOtBu 2b (left) and (BDI)SnOtBu 2c (right) with H atoms omitted and BDI aryl groups C atoms minimized for
clarity; ellipsoid probability shown at 30%.
˚
Table 1. Selected Bond Lengths (A) and Angles (deg) for Compounds 2a, 2b, 2c, and 2d
(BDI)SnOiPr (2a)a
(BDI)SnOsBu (2b)
(BDI)SnOtBu (2c)
(BDI)SnOtBuF (2d)
Sn-O
2.000(5)
2.206(4)
2.208(4)
1.418
83.6(2)
94.1(2)
95.2(2)
118.58
0.938
2.013(3)
2.202(3)
2.202(3)
1.442(7)
82.79(11)
94.08(12)
94.29(12)
118.1(3)
0.974
2.0179(16)
2.2015(19)
2.2100(19)
1.419(3)
83.02(7)
92.66(7)
93.89(7)
122.40(16)
0.990
2.110(2)
2.1837(16)
2.1837(16)
1.360(3)
83.53(8)
93.65(6)
93.65(6)
128.97(18)
1.056
Sn-N(1)
Sn-N(2)
O-C(30)
N(1)-Sn-N(2)
N(1)-Sn-O(1)
N(2)-Sn-O(1)
Sn-O-C(30)c
Sn-NCCCN plane
sum of angles around Sn
DP (%)b
272.9
97
271.16
99
269.57
100
270.83
99
a See ref 14. b Degree of pyramidalization, DP (%) = (360 - [sum of angles])/0.9. c C(16) for 2d
results in an increase of the 13C NMR carbonate resonance.
Further conformation of the nature of the product was
given by treatment of tin chloride 1 with the potassium salt
of monoalkyl carbonate (KO2COR, R=sBu, tBu), which
produced a mixture of tin alkoxide 2 and tin carbonate 5.
Although KO2COR does revert to KOR and CO2, this
latter reaction only happens at elevated temperatures.
Interestingly, and similar to the lead alkoxide system,
the rate of reaction between 2 and CO2, as well as the ob-
served equilibria between alkoxide 2, CO2 and metallo-
alkylcarbonate 5 is dependent upon the nature of the alkyl
group on the alkoxide ligand. For instance, the t1/2 for the
reaction between isopropoxide 2a or sec-butoxide 2b with
1 atm of CO2 is <10 min, and complete conversion to the
metallo-alkylcarbonates was observed, but the reaction
between tert-butoxide 2c with 1 atm of CO2 is signifi-
cantly slower and only goes to 74% completion after one
week at room temperature.
˚
of 2.954 A (molecule A) and 2.898 A (molecule B), which
˚
are inside the combined Sn-O van der Waals radii of
3.69 A.
˚
All three tin alkoxides react with carbon dioxide to give
the carbonate insertion product (BDI)SnOCO2R (5a-5c)
(Scheme 2). Although the facile reversibility of this reac-
tion has prevented any solid-state characterization of the
tin alkylcarbonates 5, IR spectroscopy (CCl4) showed the
characteristic solution phase carbonyl stretching frequen-
cies at 1621 cm-1. In addition, the 13C NMR spectrum of
all three products reveals a resonance at δ 158.7 (5a),
158.8 (5b), and 158.9 (5c) ppm, indicative of a carbonate
carbon.4,8,26 Addition of 13CO2 to all of the alkyl carbonates
Equilibrium Studies. To further understand the influ-
ence of the alkyl group on the reaction, we investigated
both the equilibria formed between the reactants (2 and
CO2) and products (5), as well as the rate of reaction
between the alkoxides and carbon dioxide. These studies
(26) Campora, J.; Matas, I.; Palma, P.; Alvarez, E.; Graiff, C.; Tiripicchio, A.
Organometallics 2007, 26, 3840.

1882 Inorganic Chemistry, Vol. 50, No. 5, 2011
Table 2. Crystallographic Data for Compounds 2b, 2c, 2d, and 4
(BDI)SnOsBu (2b)
Ferro et al.
(BDI)SnOtBu (2c)a
(BDI)SnOtBuF (2d)b
(BDI)Sn(Ma)OiPr (4)c
chemical formula
formula weight
temp (K)
C33H50N2OSn
609.44
173(2)
C33H50N2OSn
609.44
173(2)
C33H41F9N2OSn
771.37
173(2)
C36H50N2O4Sn
693.47
173(2)
˚
wavelength (A)
0.71073
0.4 ꢀ 0.3 ꢀ 0.3
0.71073
0.71073
0.71073
0.23 ꢀ 0.20 ꢀ 0.10
cryst size (mm3)
cryst syst
0.15 ꢀ 0.10 ꢀ 0.10
monoclinic
P21/n (No. 14)
13.3880(2)
16.7487(3)
15.2140(2)
90
107.654(1)
90
3250.80(9)
4
1.25
0.16 ꢀ 0.08 ꢀ 0.06
monoclinic
P21/m (No. 11)
8.9128(2)
19.8833(5)
10.4482(2)
90
113.394(1)
90
1699.96(7)
2
1.51
triclinic
triclinic
space group
˚
P1 (No.2)
8.7788(2)
9.8810(2)
20.5060(4)
92.884(2)
90.071(1)
113.948(1)
1623.12(6)
2
1.25
0.81
P1 (No.2)
14.7295(3)
16.2188(2)
17.6171(2)
110.624(1)
99.628(1)
108.961(1)
3531.49(9)
4
a (A)
˚
b (A)
˚
c (A)
R (deg)
β (deg)
γ (deg)
3
˚
V (A )
Z
pc (Mg m-3
)
1.30
0.76
abs coeff (mm-1
)
0.81
0.83
θ range for data collection (deg)
measured/indep rflns/R(int)
rflns with I >2σ(I)
3.64-26.06
23390/6364/0.043
5816
3.42-26.01
45959/6387/0.051
5479
3.74-27.11
27173/3861/0.063
3446
3.40-27.10
56849/15530/0.058
12240
data/restraints/params
GOF on F2
6364/0/331
1.110
6387/0/347
1.057
3861/0/263
1.023
15530/0/779
1.011
final R indices [I > 2σ(I)]
R indices (all data)
largst diff. peak and hole (e A )
R1 = 0.045, R2 = 0.117
R1 = 0.050, R2 = 0.121
2.63 and -1.13
R1 = 0.029, R2 = 0.067
R1 = 0.037, R2 = 0.071
0.56 and -0.50
R1 = 0.028, R2 = 00.059
R1 = 0.035, R2 = 0.062
0.36 and -0.35
R1 = 0.033, R2 = 0.070
R1 = 0.051, R2 = 0.076
0.44 and -1.10
3
˚
a The isobutyl group attached to O is disordered over two partially overlapping orientations with the C30 and C31 alternative positions resolved, and
was included with isotropic C atoms. b The molecule lies on a mirror plane with the alkoxide group disordered over two positions. c There are two
independent molecules in the unit cell that differ in the conformation of the maleate ligand.
Figure 2. ORTEP diagrams of 4 showing molecule A (left) and molecule B (right) with H atoms omitted for clarity; ellipsoid probability shown at 30%.
Table 3. Selected Bond Lengths and Angles for Compound 4
calculate the mole fraction (χCO2) solubility corrected to a
partial pressure of 1.013 bar.27
At 298 K and 1 atm of CO2, the H NMR spectrum
molecule A
molecule B
1
Sn-O(1)
2.1489(16) Sn(1b)-O(1b)
2.1812(17) Sn(1b)-N(1b)
2.1784(18) Sn(1b)-N(2b)
2.1534(16)
2.1866(17)
2.1851(18)
2.954
86.09(7)
88.78(6)
86.09(7)
260.96
shows that the tin isopropoxide 2a is completely con-
verted to the corresponding carbonate 5a. However, at
313 K, the conversion is only 92%. As such, we used this
temperature to perform our comparative equilibrium mea-
surements. Table 4 shows the ratio of 2 and 5 at 313 K, and
the calculated Keq and ΔG° for each system. Although the
apparent equilibrium favors the reactants at elevated tem-
peratures, we are unfortunately limited as to a “usable”
temperature range due to lack of carbon dioxide solubility
data outside of the 283 - 313 K temperature range at baro-
metric pressures.27 The equilibria between tert-butoxide 2c
and metallo-alkylcarbonate 5c was measured at four dif-
ferent temperatures and the ΔH° (-13.8 kcal mol-1), ΔS°
(-39.6 cal mol-1 K-1), and ΔG°298 (-2.0 kcal mol-1) were
calculated for the insertion of CO2 into the Sn-OtBu
bond. Asthere was no measurable amountof isopropoxide
Sn-N(1)
Sn-N(2)
Sn-O(2)
N(1)-Sn-N(2)
O(1)-Sn-N(1)
O(1)-Sn-N(2)
sum of angles
DP (%)a
2.898
84.88(7)
87.32(7)
87.90(7)
260.1
Sn(1b)-O(2b)
N(1b)-Sn(1b)-N(2b)
O(1b)-Sn(1b)-N(1b)
O(1b)-Sn(1b)-N(2b)
sum of angles
111
DP (%)a
110
a Degree of pyramidalization, DP (%) = (360 - [sum of angles]/0.9).
are complicated by an additional equilibria between gas
phase and solution phase CO2, which was partially over-
come by flooding the reaction with carbon dioxide (1 atm),
allowing us to maintain a steady CO2 concentration and
(27) Fogg, P. G. T.; Gerrard, W. In Solubility of Gases in Liquids; John
Wiley & Sons: Chichester, England, 1991, p 249.

Article
Inorganic Chemistry, Vol. 50, No. 5, 2011 1883
Scheme 2. Reaction of Alkoxides 2a-2c with CO2 to Form Carbo-
nates 5a-5c and the Various Equilibria Involved
Figure 3. ORTEP diagram of (BDI)SnOtBuF 2d (fluorine atoms in
green) with H atom omitted and BDI aryl groups C atoms minimized
for clarity; ellipsoid probability shown at 30%.
Table 4. Ratio of Alkoxide 2 to Carbonate 5, Keq ΔG°313 (kcal mol-1), and
Relative Time to Equilibrium (tequil) for the Equilibrium between 2/CO2 and 5a
Treatment of (BDI)SnOtBuF 2d with methyl triflate pro-
duces an intractable mixture of products; no evidence for
either (BDI)Sn-OTf or MeOC(CF3)3 was found. Addition
of maleic anhydride to 2d leads to the protonated ligand LH
and no isolable tin containing species. In sharp con-
trast to the perprotonated alkoxides 2a-2c, perfluoro
2d does not react with carbon dioxide.
2: 5
Keq
ΔG°313
(tequil)
iPr (2a/5a)
sBu (2b/5b)
tBu (2c/5c)
7.6:92.4
11.1:88.9
45.4:54.6
191
178
20
-3.27
-3.22
-1.88
3 h
1 d
3 d
a All measurements were performed in C6D6 at 313 K using 36.4 mM
alkoxide in a sealed NMR tube with a 50 cm3 head-space (to maintain a
steady CO2 concentration). The calculated CO2 concentration at this
temperature is 0.091 M.
Computational Studies. Density Functional Theory
(DFT) studies were performed to further understand the
influence of the alkyl group on both the thermodynamics
and kinetics for the formation of tin alkylcarbonates from
the reaction of tin alkoxides with carbon dioxide. Geometry
optimization and single point energy calculations were per-
formed on tin alkoxides 2a, 2b, 2c, and 2d and the cor-
responding alkylcarbonates 5a, 5b, 5c, and 5d using a
B3LYP/Lanl2dz/3-21 g level of theory for the geometry
optimization of the entire molecule followed by a single
point energy calculation using a larger basis set (see
Experimental Section).30 The geometry optimization for
the alkoxide complexes (2a, 2b, 2c, and 2d) are in good
agreement with the experimental results, although the
N-Sn-O bond angles are overestimated by 5 - 7% for
the nonfluorinated alkoxide complexes 2a-2c (see Sup-
porting Information). The ΔG°298 for the insertion of car-
bon dioxide intothe tinalkoxidewascalculatedtobecloser
to neutrality than our experimental results, with a positive
ΔG2°98 for the sec-butyl and tert-butyl systems and a nega-
tive ΔG2°98 for the isopropyl system. However, the trends in
reactivity of the different alkyl groups correspond to what
we observe; that is, the ΔG°298 was more positive for the
tert-butoxide system and the least positive for the isoprop-
oxide system.
An abbreviated molecule in which the N-aryl groups, as
well as the backbone methyl groups, are replaced with
protons, [CH{CHCNH}2]- (L*) was used to give further
mechanistic insight into the insertion of carbon dioxide
into the tin-oxygen bond.31 This smaller ligand was uti-
lized in looking for transition states in the reaction path-
way and, because of the minimal steric influence of this
ligand, providing information about the different electro-
nic contributions from each of the different alkoxide
substituents. The reaction profile for the conversion of
L*SnOiPr to the corresponding metallo-alkylcarbonate
or sec-butoxide at room temperature, the ΔH° and ΔS°
was not determined for these latter systems.
Although we were able to flood with carbon dioxide to
obtain accurate equilibrium measurements, determination
of the reaction rate using NMR spectroscopy was impos-
sible becauseof limited ability to mix the solution, resulting
in diffusion of carbon dioxide to be rate influencing. We
were able to qualitatively gauge the reaction rate; for the
isopropoxide system, equilibrium was established after 3 h
at 313K, for the sec-butoxide system, equilibrium was esta-
blished after 1 d at 313 K, and for the tert-butoxide system,
equilibria wasestablishedafter3 dat313K; however, all re-
actions proceeded faster when the NMR tube was shaken.
Unfortunately, UV-vis spectroscopic measurements gave
unreliable results, thus thwarting our attempts to examine
the relative rates in greater detail.
Nonafluoro-tert-butoxide Tin Complex. The nona-
fluoro-tert-butoxide tin complex 2d was synthesized in
order to gauge the role of steric and electronic factors in
the reactivity trends of the tin alkoxides 2a-2c. The X-ray
crystal structure was determined for 2d, and revealed a
significantly shorter O-C31 bond length when compared
to the per-protio tert-butoxide analogue 2c, presumably
due to the electronic withdrawing nature of the fluoroalkyl
˚
groups (Figure 3). The Sn-O bond length of 2.110(2) A is
significantly longer than 2c (2.0179(16) A). This can be
˚
attributed to the increased ionic nature of the Sn-O bond,
which is also reflected in the shortened Sn-N bond lengths
(average 2.1837 A for 2d versus average 2.2058 A for 2c).
˚
˚
Potential long-rangeSn F interaction is observed between
3 3 3
˚
˚
˚
Sn and F1 (3.204 A), F2 (3.353 A) and F9 (3.433 A) as all of
these distances are shorter than the sum of the Sn-F radii of
3.64 A, consistent with other fluoroalkoxide complexes.
28,29
˚
(28) Reisinger, A.; Trapp, N.; Krossing, I.; et al. Organometallics 2007, 26,
2096.
(29) Samuels, J. A.; Lobkovsky, E. B.; Streib, W. E.; Folting, K.; Huffman,
J. C.; Zwanziger, J. W.; Caulton, K. G. J. Am. Chem. Soc. 1993, 115, 5093.
(30) Frisch, M. J. et al. et. al. Gaussian 03, revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(31) Liu, C.; Munjanja, L.; Cundari, T. R.; Wilson, A. K. J. Phys. Chem.
A 2010, 114, 6207.

1884 Inorganic Chemistry, Vol. 50, No. 5, 2011
Ferro et al.
Table 5. Calculated Thermodynamic (kcal mol-1) and Kinetic (s-1) Parameters
at the B3LYP/[4333111/433111/43]/6-31g(d,p) Level of Theory, for the Trans-
formation of (BDI)SnOR to (BDI)SnO(CO2)R and L*SnOR to L*SnO(CO2)R
L = BDI
L* = [CH {CHCNH}2]
ΔG° ΔH° ΔE° ΔG°
-0.0 -9.7 -9.5 -4.0
ΔH°
ΔE° ΔGq ΔHq ΔEq
iPr
þ7.0 -13.7 þ10.4 þ0.8 þ0.8
sBu þ 0.5 -10.8 -10.2 -2.9 -12.4 -12.4 þ12.4 2.5 þ2.5
tBu þ 1.1 -10.0 -9.4 -0.3 -12.0 -11.3 þ15.6 3.6 þ4.2
tBuF þ 17.5 þ7.0 þ7.6 þ20.4
9.4 þ10.1 NA NA NA
i
L*SnOCO2 Pr is shown in Figure 4, and includes the
relative energies of the reactants (Alk-a), transition state
(TS-a), intermediate (INT-a) and carbonate (Carb-a).
Because of the low reactivity of the tin alkoxides
with aliphatic electrophiles, we initially examined a non-
nucleophilic pathway in which carbon dioxide coordi-
nates directly to the metal center, and insertion of carbon
dioxide into the Sn-O bond would follow. Although
rare, coordination of carbon dioxide to coordinatively
unsaturated metal centers has been observed, most com-
monly in an η2-fashion,32-35 but end-on coordination has
also been found.36 A high energy transition state (TSNP
,
ΔG‡=56.6 kcal mol-1) was located in which carbon di-
oxide coordinates in an η2-fashion to the tin center of
isoproxide Alk-a. This interaction involves the tin 5s lone
pair binding to an antibonding orbital located on the car-
bon dioxide molecule. As a result, there is a predictable
decrease in electron density at the metal center and an in-
crease in electron density at the carbon atom (see Figure 5
for Mulliken charge distribution). The C-O2 bond length
is significantly longer than experimentally derived η2-CO2
complexes,32-35 opening up the possibility that this transi-
tion state structure is best described as a metallacycle in
which there has been an oxidative addition of CdO to the
metal center, generating a tin(IV) complex.
Figure 4. Gibbs free energy diagram (kcal mol-1) and drawings for the
reaction of L*SnOiPr with CO2 at the B3LYP level. Electronic energy
values are given in parentheses.
˚
The Sn O2 interaction of 2.733 A is well within their
3 3 3
˚
combined Van der Waal radii of 3.69 A and is shorter
than the long-range interaction observed for maleate 4.
Although the WBI of 0.09 shows very little direct electron-
ic interaction, the Mulliken population analysis shows
there is a build-up of negative charge on O2 (-0.45 for the
O2 atom in the TS-a compared to -0.33 fro the free car-
bon dioxide). Taken with the increase of positive charge
from þ0.71 to þ0.83 of the tin atom, there could be an
electrostatic Sn-O2 interaction that has not been ac-
counted for in our calculations. Once TS-a is formed, a
smooth transition to the metallo-alkylcarbonate inter-
mediate, INT-a, is observed and no intermediates be-
tween TS-a and INT-a were calculated during an Intrinsic
Reaction Coordinate (IRC) calculations.
Because of both the high-energy nature of this structure
and our inability to find a pathway from TSNP to the cor-
responding carbonate (Carb-a), this transition state was
deemed nonproductive. An alternative and more energe-
tically accessible transition state was found arising from
an interaction between the oxygen lone pair (HOMO -1)
on the tin alkoxide and the LUMO of CO2 (TS-a, ΔGq =
16.6 kcal mol-1). This four-membered transition state
involves formation of a new C-O1 bond, weakening of
the CdO2 double bond, formation of a new Sn-O2 bond
and a weakening of the Sn-O1 bond. The C O1 inter-
3 3 3
INT-a possesses a formal Sn-O2 bond, the Sn O1
distance is 2.913 A, which is shorter than the sum of the
˚
action of 1.869 A results in a slight CdO bond elongation
3 3 3
˚
˚
˚
of 0.017 A (C-O2) and 0.039 A (C-O3), a decrease in the
O2-C-O3 bond angle to 151° and an elongation of the
combined van der Waals radii (WBI = 0.05) and no inter-
action exists between Sn and O3. A similar intermediate
has also been observed by Wakamatsu and Otera.37 After
formation of INT-a, the Sn-O2 bond rotates such that the
bulky alkyl group projects away from the metal center and
˚
Sn-O bond by 0.121 A. The Wiberg bond index (WBI) of
the formation C-O1 bond is 0.33, whereas the WBI of the
Sn-O1 bond has decreased by 0.19- 0.37, indicating that
although there has been minimal geometric change within
the Sn-O1 bond, there has been significant weakening.
˚
anewSn O3 interaction is formed (2.728 A, WBI = 0.11).
The final product, Carb-a, is analogous to that reported for
3 3 3
the lead isopropoxide system, in which there is a long distance
Pb O3 interaction.
The corresponding calculations were performed on the
other alkoxide complexes, sec-butoxide Alk-b, tert-butoxide
Alk-c and nonafluoro-tert-butoxide Alk-d, and their corre-
sponding metallo-alkylcarbonates Carb-b-d. The ΔG°298
(32) Hirano, M.; Akita, M.; Tani, K.; Kumagai, K.; Kasuga, N. C.;
Fukuoka, A.; Komiya, S. Organometallics 1997, 16, 4206.
(33) Fu, P. F.; Khan, M. A.; Nicholas, K. M. J. Organomet. Chem. 1996,
506, 49.
(34) Alvarez, R.; Carmona, E.; Marin, J. M.; Poveda, M. L.; Gutierrezpuebla,
E.; Monge, A. J. Am. Chem. Soc. 1986, 108, 2286.
3 3 3
(35) Gambarotta, S.; Strologo, S.; Floriani, C.; Chiesivilla, A.; Guastini,
C. J. Am. Chem. Soc. 1985, 107, 6278.
(36) Castro-Rodriguez, I.; Nakai, H.; Zakharov, L. N.; Rheingold, A. L.;
Meyer, K. Science 2004, 305, 1757.
(37) Wakamatsu, K.; Orita, A.; Otera, J. Organometallics 2010, 29, 1290.

Article
Inorganic Chemistry, Vol. 50, No. 5, 2011 1885
Figure 5. Optimized geometries and Mulliken charges for (a) Alk-a, (b) CO2, (c) TSNP, and (d) TS-a. Only Sn, N, O, and carbon dioxide atoms of the
model complex are shown for clarity.
and ΔE° for the reactions follow the same trends as the
larger systems (Table 4); that is, the isopropyl system
as the most negative ΔG°298 and ΔE°, followed by the
sec-butyl and tert-butyl system, respectively. The fluorinated
system has a significantly positive ΔG°298 and ΔE°, which is
in agreement with our experimental results in with the nona-
fluoro-tert-butoxide 2d doe not react with carbon dioxide. A
transition state was located for the sec-butyl system (TS-b)
and the tert-butyl system (TS-c); however, no transition state
was found for the fluorinate derivative (TS-d).
transition state, similar to what we observed in our calcula-
tions and have been proposed by others.3,9,13,37
A few transition metal systems have been reported to react
with both methyl iodide and carbon dioxide, unfortunately
most reports have been vague on the reaction conditions.40,41
Vahrenkamp’s pyrazolylborate-zinc methoxide complex,
(TpPh,Me)ZnOMe, reacts readily with methyl iodide but very
slowly with carbon dioxide, in sharp contrast to our results.11
The former reaction is attributed to the nucleophilicity of the
zinc-alkoxide. Thus, it can not only be the nucleophilicity of
the alkoxide ligand that is governing the reactivity of alk-
oxides with carbon dioxide (vide infra). Interestingly, trispyr-
azolborate zinc hydroxides do react readily and reversibly
with carbon dioxide,42,43 although this could be because of
hydrogen-bonding effects.
Another intriguing aspect about our results is the strong
dependence on both the reaction rate as well as final equilib-
rium on minor variations of the alkyl group on the alkoxide
ligand. This it true both with the tin and lead systems. The
insertion of carbon dioxide into the Sn-O bond to generate
metallo-alkylcarbonates is measurably reversible; as such, by
using the equilibrium studies as well as the computational
studies on the transition state, the factor governing the
influence of the alkyl group can be examined. In order to
normalize the different alkyl systems onto one scale, a “zero-
point” was set by assuming that the alkyl group has a greater
inductive influence on the ground state energy of the alkoxide
Alk than the metallo-alkylcarbonate Carb. Electronic influ-
ences of the different alkyl groups are apparent upon con-
sidering the pKa’s of aliphatic alcohols;39,44 however, these
electronic influences are not apparent upon considering either
the bond dissociation energies (BDE’s) or pKa’s of the
corresponding carboxylic acids,45 and only minor variations
within experimental error of each other are observed in the
BDE’s of the corresponding alcohols. As such, the relative
energies of our metallo-alkylcarbonates Carb-a-c can be set to
“zero” and all other energies can be compared to these
metallo-alkylcarbonates (Figure 6).
Discussion
The tin(II) alkoxide complexes 2a-2d do not exhibit the
standard reactivity as their transition metal counterparts;
their lack of reactivity with methyl iodide indicates that the
oxygen atom is essentially non-nucleophilic toward anything
but very strong electrophiles such as methyl triflate. The
notable difference in reactivity between the alkoxides 2a-2c
might be attributable to steric effects; the combination of the
bulky β-diketiminate ligand as well as the tert-butoxide li-
gand of 2c results in a more hindered approach to the oxygen
atom than the corresponding isopropoxide system 2a. How-
ever, there is also a compelling electronic argument as the
reactivity trend mirrors the relative basicity of the alkoxide
moieties; that is, the greater the basicity, or the greater the
pKa of the conjugate acid, the more reactive the alkoxide.38,39
Because the reaction between the tin alkoxides and methyl
triflate is irreversible, it is impossible to differentiate between
steric and electronic arguments based solely on this reaction.
The activation of carbon dioxide by a seemingly non-
nucleophilic metal alkoxide complex is counterintuitive. It
has previously been show that the nucleophilicity of the alk-
oxide ligand drives insertion reactions.1 Studies on rhenium
alkoxide and aryloxide complexes, (fac(CO)3(PMe3)2ReOR,
revealed that only the alkoxide (R=Me) reacts with carbon
dioxide, yet both alkoxide and aryloxide (R = 4-MeC6H4)
react with carbon disulfide. Darensbourg observed that (fac)-
(CO)3(dppe)MnOMe reacts more readily with carbon diox-
ide than (CO)3(dppe)MnOCH2CF3, a less nucleophilic alk-
oxide ligand.9 The mechanism of carbon dioxide insertion in
this case was postulated to proceed via a four-membered
(40) Park, S.; Rheingold, A. L.; Roundhill, D. M. Organometallics 1991,
10, 615.
(41) Bryndza, H. E.; Tam, W. Chem. Rev. 1988, 88, 1163.
(42) Ruf, M.; Vahrenkamp, H. Inorg. Chem. 1996, 35, 6571.
(43) Looney, A.; Han, R.; McNeill, K.; Parkin, G. J. Am. Chem. Soc.
1993, 115, 4690.
(38) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.; Bercaw, J. E.
J. Am. Chem. Soc. 1987, 109, 1444.
(39) Brauman, J. I.; Blair, L. K. J. Am. Chem. Soc. 1970, 92, 5986.
(44) Olmstead, W. M.; Margolin, Z.; Bordwell, F. G. J. Org. Chem. 1980,
45, 3295.
(45) Blanksby, S. J.; Ellison, G. B. Acc. Chem. Res. 2003, 36, 255.

1886 Inorganic Chemistry, Vol. 50, No. 5, 2011
Ferro et al.
tin and lead systems can be compared. The higher reactivity
of the lead system can be attributed to a combination of a more
polarized and weaker Pb-O bond, resulting in a more nucleo-
philic alkoxide ligand and a more Lewis acidic lead metal
center, which forms a more stable four-membered transition
state due to a greater interaction with the CO2 oxygen atom.
The tin system suffers both from a stronger Sn-O bond, low
nucleophilicity of the alkoxide ligand, as well as a smaller Lewis
acidity of the metal center, thus requiring a higher activation
barrier in its reaction with carbon dioxide.
Conclusions
Tin alkoxide complexes 2 are weakly nucleophilic species
that only react with highly reactive aliphatic electrophiles;
however, these complexes do react reversibly with carbon
dioxide, due to a combination of a weakly nucleophilic alk-
oxide ligand and a Lewis acidic metal center that can help
support the four-membered transition state. The rate of reac-
tion as well as the relative equilibrium between reactants and
products is dependent upon the alkoxide ligand. The iso-
propoxide complex 2a reacts with carbon dioxide faster than
the bulkier tert-butoxide complex 2c, and the former reaction
has a greater equilibrium constant than the latter. Although
the sterics of the alkoxide ligand might play a role in these
results, computational studies have shown that there is a sig-
nificant electronic contribution to both the activation barrier
as well as the established equilibrium. If it is assumed that the
Sn-O BDE of the metallo-alkylcarbonate 5 complexes are
less influenced by the alkyl substituent than the Sn-O BDEof
the alkoxide complexes 2, then a relative ordering of the
Sn-O BDE can be made, with the tert-butoxide complex 2c
possessing the strongest Sn-O BDE, and the isopropoxide
complex 2a possessing the weakest Sn-O BDE. Interestingly,
the tert-butoxide complex 2c has the longest Sn-O bond and
the isopropoxide complex 2a has the shortest Sn-O bond.
Figure 6. Energy diagram for the interconversion of Alk þ CO2 and
Carb via transition state TS. Energies are reported in kcal mol-1. ΔGf° is
the change in Gibbs free energy in the forward sense, ΔGf‡ is the activation
energy in the forward reaction and ΔGr‡ is the activation energy for the
reverse reaction.
In the isopropoxide system, the conversion from Alk-a to
Carb-a has a relatively low activation barrier and the overall
reaction is thermodynamically favored. In contrast, the con-
version from Alk-c- to Carb-c in the tert-butoxide system is
almost thermodynamically neutral and a significantly higher
activation barrier is found. Steric arguments cannot fully ex-
plain the thermodynamic differences, these must be a con-
sequence of the relative bond dissociation energies (BDE’s) of
the Sn-OR bond. If it is assumed that the Sn-O2 BDE for the
metallo-alkylcarbonates is relatively similar due to reasons
previously described, then the BDE of the Sn-OtBu bond is
greater than that of the Sn-OsBu bond and the Sn-OiPr
bond is the weakest of the three. Thus, in the forward sense of
the reaction, in order to achieve the transition state, the energy
required is greater for the tert-butoxide system than the
isopropoxide system, thus the Sn-O bond of isopropoxide
Alk-a must be weaker than the Sn-O bond of tert-butoxide
Alk-c. Although the activation barriers to the reverse reaction
follow a similar pattern in which the tert-butoxide system has
a greater activation barrier than the isopropoxide system, the
difference is not as great and can be attributed to a combina-
tion of steric arguments and the differences in the O1-C
bond strengths. For the real molecule, (BDI)SnOR, the re-
verse reaction should be more affected by the sterics of the
alkoxide group due to the interactions of both the BDI N-aryl
groups and the alkoxide.
Although nonafluoro-tert-butoxide tin complex 2d is non-
reactive with carbon dioxide, the calculations show that the
postulated carbon dioxide insertion product is significantly
higher in energy than just 2d and carbon dioxide. However, a
direct comparison with the other alkoxide complexes 2a-2c
cannot be made because, in contrast to the nonfluorinated
derivatives, the pKa of the fluorinated alcohol and carboxylic
acid are heavily influenced by the nearby fluorine atoms.46
The contrast between the reactivity of group 14 alkoxides
and Vahrenkamp’s zinc alkoxides is striking and reveals that
it is not just the nucleophilicity of the alkoxide ligand that
determines its reactivity with carbon dioxide. Although there
are too many variables between group 14 alkoxides and
(TpPh,Me)ZnOMe to make a meaningful comparison, the
Experimental Section
General. All manipulations were carried out under an atmo-
sphere of dry nitrogen using standard Schlenk techniques or in
an inert-atmosphere glovebox. Solvents were dried from the ap-
˚
propriate drying agent, distilled, degassed, and stored over 4 A
sieves. (BDI)SnCl and (BDI)SnOiPr were prepared according to
the literature.15,23 Potassium alkoxide salts were prepared by the
slow addition of the relevant alcohol (dried and distilled) to a
suspension of potassium hydride, except NaOtBuF that was pre-
pared by the treatment of a water solution of (CF3)3COH with
1eq of NaOH and dried overnight under vacuum over P2O5.
Methyl trifluoromethanesulfonate was freshly dried and dis-
tilled before use. Carbon dioxide was used as received (Union
Carbide, 99.999%), and 13CO2 was 99 atom %. 1H, 13C and 119Sn
NMR spectra were recorded on a Varian 400 MHz or Varian
500 MHz spectrometer. 1H and 13C NMR spectroscopy chemi-
cal shifts are given relative to residual solvent peaks and 119Sn
was externally referenced to SnMe4. The UV spectra were re-
corded in Varian Cary 50 UV-vis spectrophotometer. The data for
X-ray structures were collected at 173 K on a Nonius KappaCCD
˚
diffractometer, k(Mo-KR) = 0.71073 A and refined using the
SHELXL-97 software package.47
[CH{(CH3)CN-2,6-iPr2C6H3}2SnOiPr] (2a). (BDI)SnOiPr
was prepared following literature procedures. The H and 13C
1
NMR spectra were identical to the ones reported earlier.15
(46) Arnett, E. M.; Venkatasubramaniam, K. G. J. Org. Chem. 1983, 48,
(47) Sheldrick, G. M. SHELXL-97, Program for the Refinement of Crystal
€ €
Structures; University of Gottingen: Gottingen, Germany, 1997.
1569.

Article
Inorganic Chemistry, Vol. 50, No. 5, 2011 1887
UV-vis (benzene): λmax (ε, L mol-1 cm-1) 366 nm (11000). IR
(CCl4, ν/cm-1): 2964 (s), 2925 (s), 1464 (s), 1430, 1359, 1315 (s),
1170, 1101 (s).
(benzene): λmax (ε, L mol-1 cm-1) 320 nm (22000). IR (Nujol,
ν/cm-1): 2024, 1945, 1548, 1518, 1319, 1262 (s), 1233 (s), 1167 (s),
1017, 966 (s), 795. Anal. Calcd for C33H41F9N2OSn: C, 51.38; H,
5.36; N, 3.63. Found: C, 51.41 ; H, 5.30; N, 3.64.
[CH{(CH3)CN-2,6-iPr2C6H3}2SnOsBu] (2b). A suspension of
KOsBu (98 mg, 0.87 mmol) in THF (5 mL) was added to a
solution of (BDI)SnCl (0.50 g, 0.87 mmol) in THF (5 mL) at
room temperature, and the reaction mixture was stirred for 3 d.
The solvent was removed under vacuum, the yellow crude prod-
uct was extracted with toluene and the solution was filtered through
Celite. Removal of the volatiles and recrystallization from pentane
overnight afforded yellow crystals of (BDI)SnOsBu (0.42 g, 80%).
1H NMR (400 MHz, C6D6, 303 K): δ 7.22 (d, J = 7.6, 2H, ArH),
7.14 (t, J = 7.6, 2H, ArH), 7.07 (d, J = 7.7, 2H, ArH), 4.71 (s, 1H,
γσ-CH), 3.81 (m, 3H, CHMe2 þ OCH(Me)Et), 3.22 (m, 2H,
CHMe2), 1.55 (s, 3H, NCMe), 1.54 (s, 3H, NCMe), 1.50 (d, J =
6.1, 3H, CHMe),1.48(d, J=6.1,3H,CHMe),1.25(d, J=6.8,6H,
CHMe), 1.19 (d, J = 6.9, 6H, CHMe), 1.11 (d, J = 6.8, 6H,
CHMe), 0.75 (d, J = 6.0, 3H, OCH(Me)Et, 0.46 (t, J = 7.4,
OCH(Me)CH2Me). 13C{1H} NMR (500 MHz, C6D6, 303 K): δ
164.6 (NCMe), 144.9 (ipso-C), 142.6 (o-C), 141.1 (o-C), 126.2 (p-C),
124.3 (m-C), 123.8 (m-C), 96.3 (γ-CH), 70.7 (OCH(Me)Et), 34.9
(OCH(Me)CH2Me), 28.2 (NCMe), 28.1 (NCMe), 26.6 (CHMe),
26.0 (CHMe), 24.6 (CHMe), 24.4 (CHMe), 24.3 (CHMe), 24.2
(CHMe), 23.1 (OCH(Me)Et), 10.3 (OCH(Me)CH2Me). 119Sn
NMR (400 MHz, C6D6, 303 K): δ -181.5. UV-vis (benzene):
i
[CH{(CH3)CN-2,6-iPr2C6H3}2SnO(CO)C2H2CO2 Pr] (4). A
solid mixture of (BDI)SnOiPr (250 mg, 0.41 mmol) and maleic
anhydride (40 mg, 0.41 mmol) was dissolved in toluene (10 mL)
and stirred for 2 h, affording an orange solution. After evaporation
of the solvent, the residue was dissolved in the minimum amount of
hexane. Storage at -32 °C for 3 days afforded deep red crystals
of (BDI)Sn(Ma)iPr (248 mg, 87%). 1H NMR (400 MHz, C6D6,
303 K): δ 7.21 - 6.93 (m, 6H, ArH), 5.96 (d, J = 11.9 Hz, 1H,
COCHCHCO2), 5.83 (d, J = 12.0 Hz, 1H, COCHCHCO2), 5.09
(sept, 6.0 Hz, 1H, CO2CHMe2), 4.91 (s, 1H, γ-CH), 3.62 (sept,
J = 6.7 Hz, 2H, CHMe2), 3.04 (sept, J = 6.7 Hz, 2H, CHMe2),
1.59 (s, 6H, NCMe), 1.37 (d, J = 6.6 Hz, 6H, CHMe2), 1.27 (d,
J = 6.7 Hz, 6H, CHMe2), 1.12 (d, J = 6.8 Hz, 6H, CHMe2), 1.07
(d, J = 6.2 Hz, 6H, CHMe2), 1.04 (d, J = 6.8 Hz, 6H, CHMe2).
13C{1H} NMR (400 MHz, C6D6, 303 K): δ 170.0 (SnOCO),
i
169.3 (CO2 Pr), 165.40 (NCMe), 145.31 (ipso-C), 142.43 (o-C),
141.27 (o-C), 131.0 (CH=CH), 130.0 (CHdCH), 127.1 (p-C),
124.8 (m-C), 123.7 (m-C), 99.8 (γ-CH), 67.5 (CO2CHMe2), 28.7
(NCMe), 27.7 (NCMe), 26.4 (CHMe), 24.6 (CHMe), 24.1
(CHMe), 23.60 (CHMe), 21.4 (CHMe). 119Sn NMR (400 MHz,
C6D6, 303 K): δ -374.2. IR (Nujol, v/cm-1): 1735 (s), 1600, 1552,
1524, 1261 (s), 1211, 1172, 1102 (s), 1019 (s), 797(s). Anal. Calcd for
C36H50N2O4Sn: C, 62.35; H, 7.27; N, 4.04. Found: C, 62.39; H,
7.15; N, 3.94.
λ
max (ε, L mol-1 cm-1) 366 nm (12000). IR (Nujol, ν/cm-1): 1554
(s), 1518 (s), 1260 (s), 1098, 1018 (s), 795 (s). IR (CCl4, ν/cm-1): 2966
(s), 2925 (s), 1463 (s), 1423, 1314 (s), 1172, 1104. Anal. Calcd for
C33H50N2OSn: C, 65.03; H, 8.27; N, 4.60. Found: C, 64.96; H, 8.23;
N, 4.57.
Generation of Metallocarbonates 5a, 5b, and 5c: General
Procedure. (BDI)SnOiPr (8.5 mg, 0.014 mmol) was dissolved
in C6D6 (400 μL) in an NMR tube sealed with a Young’s tap.
The NMR tube was submerged in a dry ice/ethanol bath, the gas
inside the NMR tube was evacuated and CO2 was introduced at a
pressure of 1 atm. A pale yellow solution mixture was observed,
the reaction mixture was kept at room temperature for 48 h and
was monitored by 1H NMR spectroscopy. To generate samples
for IR spectroscopy, a similar procedure was followed using CCl4
as solvent.
[CH{(CH3)CN-2,6-iPr2C6H3}2SnOtBu] (2c). Yellow crystals
of (BDI)SnOtBu can be obtained in good yield (76%) using a
similar methodology as used in the preparation of (BDI)SnOsBu
2b. 1H NMR (400 MHz, C6D6, 303 K): δ 7.18 (d, J = 7.6, 2H,
ArH),7.14-7.06(m,J=7.6,2H,ArH),7.03(d,J=7.6,2H,ArH),
4.61 (s, 1H, γ-CH), 3.86-3.71 (m, 2H, CHMe2), 3.29-3.14 (m, 2H,
CHMe2),1.54(s,6H,NCMe),1.53(d,J=6.9,6H,CHMe),1.28(d,
J=6.8, 6H,CHMe),1.17(d,J=6.9, 6H, CHMe), 1.12(d, J=6.8,
6H, CHMe), 0.86 (s, 9H, OC(Me)3). 13C{1H} NMR (400 MHz,
C6D6, 303 K): δ 169.4 (NCMe), 165.0 (NCMe), 144.9 (ipso-C),
142.6 (o-C), 140.9 (o-C), 126.1 (p-C), 123.9 (m-C), 123.8 (m-C), 95.5
(γ-CH), 69.4 (OC(Me)3), 34.9 (OC(Me)3), 28.2 (NCMe), 28.2
(NCMe), 26.1 (CHMe), 24.4 (CHMe), 24.3 (CHMe), 24.2
(CHMe), 23.0 (CHMe). 119Sn NMR (400 MHz, C6D6, 303 K): δ
-149.1. UV-vis (benzene): λmax (ε, Lmol-1 cm-1) 365 nm (13000).
IR(Nujol, ν/cm-1): 1555 (s), 1261 (s), 1093, 1018 (s), 939 (s), 796 (s).
IR (CCl4, ν/cm-1): 2965 (s), 2927 (s), 1463 (s), 1424, 1361, 1171 (w).
Anal. Calcd for C33H50N2OSn: C, 65.03; H, 8.27; N, 4.60. Found: C,
64.97; H, 8.32; N, 4.55.
[CH{(CH3)CN-2,6-iPr2C6H3}2SnO(CO2)iPr] (5a). 1H NMR
(400 MHz, C6D6, 303 K): δ 7.27-6.95 (m, 6H, ArH), 4.95 (s, 1H,
γ-CH), 4.89 (dq, J = 6.2, 1H, CO2CHMe2), 3.71 (hept, J = 6.7,
2H, CHMe2),3.07(hept,2H,J=6.9,CHMe2), 1.59 (s, 6H, NCMe),
1.44 (d, J = 6.7, 6H, CHMe), 1.22 (d, J = 6.8, 6H, CHMe), 1.18 (d,
J =6.3, 6H, CHMe), 1.16(d, J=7.0, 6H, CHMe), 1.06(d, J=6.8,
6H, CO2CHMe). 13C{1H} NMR (400 MHz, C6D6, 303 K): δ 165.2
(NCMe), 158.7 (OCO2), 145.4 (ipso-C), 142.5 (o-C), 141.3 (o-C),
127.1 (p-C), 124.7 (m-C), 123.6 (m-C), 99.7 (γ-CH), 68.5 (OCO2-
CHMe2), 28.7 (OCO2CHMe2), 27.7 (NCMe), 26.0 (NCMe),
24.5 (CHMe), 24.1 (CHMe), 23.9 (CHMe), 23.5 (CHMe), 22.0
(CHMe). 119Sn NMR (400 MHz, C6D6, 303 K): δ -379.7. UV-
vis (benzene): λmax 321 nm. IR (CCl4, ν/cm-1): 3963, 2963 (s),
2927, 2871, 2339 (s), 1718 (w), 1621, 1463, 1437, 1385, 1318.
[CH{(CH3)CN-2,6-iPr2C6H3}2SnO(CO2)sBu] (5b). 1H NMR
(400 MHz, C6D6, 303 K): δ 7.27-7.01 (m, 6H, ArH), 4.93 (s, 1H,
γ-CH), 4.72 (m, 1H, CO2CH(Me)Et), 3.72 (hept, J = 6.8, 2H,
CHMe2), 3.06 (m, 2H, CHMe2), 1.59 (s, 3H, NCMe), 1.58 (s,
3H, NCMe), 1.47 (d, J = 6.7, 3H, CHMe), 1.45 (d, J = 6.7, 3H,
CHMe), 1.27-1.20 (m, 11H, CO2CH(Me)CH2Me þ CO2CH-
(Me)CH2Meþ CHMe), 1.16(d, J = 6.9 6H, CHMe), 1.06 (d,
J = 6.8, 6H, CHMe), 0.85 (t, J = 7.5, 3H, CO2CH(Me)CH2Me).
13C{1H} NMR (400 MHz, C6D6, 303 K): δ 167.0 (NCMe), 165.2
(NCMe), 158.8 (OCO2), 145.4 (ipso-C), 142.5 (o-C), 141.2 (o-C),
127.0 (p-C), 124.7 (m-C), 123.6 (m-C), 99.8 (γ-CH), 73.3
(OCO2C(Me)Et), 29.0 (OCH(Me)CH2Me), 28.7 (NCMe), 27.6
(NCMe), 26.1 (CHMe), 24.5 (CHMe), 24.0 (CHMe), 23.8
(CHMe), 23.4 (CHMe), 22.9 (CHMe), 19.4 (OCH(Me)Et), 9.6
(OCH(Me)CH2Me). 119Sn NMR (400 MHz, C6D6, 303 K): δ -
379.1. UV-vis (benzene): λmax 321 nm. IR (CCl4, ν/cm-1): 2965 (s),
2928, 2869, 2336 (s), 1720 (w), 1621, 1463, 1437, 1385, 1319, 1101.
[CH{(CH3)CN-2,6-iPr2C6H3}2SnOC(CF3)3] (2d). A solution
of NaOtBuF (0.14 g, 0.55 mmol) in THF (5 mL) was added to a
solution of (BDI)SnCl (0.31 g, 0.55 mmol) in THF (5 mL) at
room temperature, and the reaction mixture was stirred for 3 d.
The solvent was removed under vacuum, the yellow crude prod-
uct was extracted with toluene and the solution was filtered through
Celite. The volatiles were removed and the yellow/green solid was
washed with pentane. Recrystallization from THF overnight af-
1
forded green crystals of (BDI)SnOtBuF (0.29 g, 69%). H NMR
(400 MHz, C6D6, 303 K): δ 7.16 (d, J = 7.6, 2H, ArH), 7.09 (d, J =
7.7, 2H, ArH), 7.03 (d, J = 7.7, 2H, ArH), 4.55 (s, 1H, γ-CH), 3.80
- 3.34 (m, 2H, CHMe2), 3.16 - 2.76 (m, 2H, CHMe2), 1.45 (d, J =
5.7, 6H, CHMe), 1.44 (s, 6H, NCMe), 1.23 (d, J = 6.8, 6H, CHMe),
1.10 (d, J = 6.7, 6H, CHMe), 1.09 (d, J = 6.5, 6H, CHMe).
13C{1H} NMR (500 MHz, C6D6, 303 K): δ 166.0 (NCMe), 143.8
(ipso-C), 142.2 (o-C), 139.7 (o-C), 127.5 (p-C), 126.8 (m-C), 124.3
(m-C), 96.4 (γ-CH), 67.4 (OC(CF3)3), 28.6 (OC(CF3)3), 28.3
(NCMe), 25.4 (CHMe), 24.6 (CHMe), 24.3 (CHMe), 23.0
(CHMe), 23.0 (CHMe). 19F NMR (400 MHz, C6D6, 303 K): δ
-73.0. 119Sn NMR (400 MHz, C6D6, 303 K): δ -256.3. UV-vis

1888 Inorganic Chemistry, Vol. 50, No. 5, 2011
Ferro et al.
[CH{(CH3)CN-2,6-iPr2C6H3}2SnO(CO2)tBu] (5c). 1H NMR
(400 MHz, C6D6, 303 K): δ 7.23-7.06 (m, 6H, ArH), 4.89 (s, 1H,
γ-CH), 3.68 (hept, J = 6.8, 2H, CHMe2), 3.05 (hept, 2H, J =
6.8, CHMe2), 1.44 (d, J = 6.7, 6H, CHMe), 1.42 (s, 6H, NCMe),
1.18 (d, J = 6.8, 6H, CHMe), 1.13 (d, J = 6.8, 6H, CHMe), 1.02
(d, J = 7.0, 6H, CHMe), 0.87 (s, 9H, OCMe3). 13C{1H} NMR
(400 MHz, C6D6, 303 K): δ 167.0 (NCMe), 165.3 (NCMe), 157.9
(OCO2), 145.3 (ipso-C), 142.4 (o-C), 141.3 (o-C), 127.0 (p-C),
124.6 (m-C), 123.6 (m-C), 99.7 (γ-CH), 77.1 (OC(Me)3), 34.8
(OC(Me)3), 28.7 (NCMe), 28.0 (NCMe), 27.6 (CHMe), 26.3
(CHMe), 24.4 (CHMe), 24.1 (CHMe), 23.8 (CHMe), 23.4
(CHMe). 119Sn NMR (400 MHz, C6D6, 303 K): δ -377.4. UV-
vis (benzene): λmax 325 nm. IR (CCl4, ν/cm-1): 2966, 2928, 2870,
2337 (s), 1717 (w), 1621, 1463, 1438, 1386, 1318.
γ-CH 1H NMR signal to the internal standard ferrocene. In order
to ensure accurate integration, a 10s delay between 30° pulses was
utilized (number of scans = 4, acquisition time = 4).
Computational Details. All the calculations were carried out
using density functional theory (DFT) in the Gaussian 03
program.30 Geometry optimization for the full complexes
((BDI)SnOR) was performed at the b3lyp level by using a
double-ζ basis set (Lanl2dz) plus a d type polarization function
(d exponent 0.183) along with the effective core potential (Lanl2
ECP)48 for Sn atoms and 3-21 g for all other atoms. Geometry
optimization for the model complexes (L#SnOR) was per-
formed at the b3lyp level by using a double-ζ basis set
(Lanl2dz) plus a d type polarization function (d exponent
0.183) along with the effective core potential (Lanl2 ECP)48
for Sn atoms and 6-31þg* for all other atom. Single point
calculations were carried on all the found saddle points with the
larger basis set [4333111/433111/43] plus two d polarization
function (d exponents 0.253 and 0.078) for Sn49,50 and 6-31 g(d,
p) for all the other atoms. Zero-point vibrational energy correc-
tions were also included. All the transition structures (TS) were
located on the potential energy surfaces and verified by one
imaginary frequency. Intrinsic reaction coordinates (IRC) cal-
culations confirmed the connectivity of TS with reactants and
products.
Determination of t1/2 for the Reaction of 2a-2c with MeOTf:
General Procedure. A sealable NMR tube was charged with
1.7 ꢀ 10-5 mol of (BDI)SnOR in 0.25 mL in C6D6, 1.1 μL of di-
chloromethane (1.7 ꢀ 10-5 mol) and 1.9 μL of MeOTf (1.7 ꢀ 10-5
mol) to give a total volume of 0.25 mL. The NMR tube was then
placed in the thermostatted probe (298.1 K) of a Varian 400
MHz spectrometer. The conversion was monitored by 1H NMR
spectroscopy. NMR spectra were acquired in 5 min intervals. In
order to ensure accurate integration, a 10s delay between 30°
pulses was utilized (number of scans =4, acquisition time = 4).
The t1/2 was determined by comparing the γ-CH signal integra-
tion of (BDI)SnOR and (BDI)SnOTf, using the DCM peak as
standard.
Acknowledgment. The authors are grateful for the financial
support from the EPSRC (LF, grant no EP/E032575/1).
Equilibrium Studies of the Reaction of 2a-2c with CO2:
General Procedure. A 0.45 mL C6D6 solution consisting of
1.64 ꢀ 10-5 mol of (BDI)SnOR and 2.68 ꢀ 10-6 mol of ferrocene
was added to a sealable NMR tube with a 50 cm3 headspace. The
solution was submerged in a dry ice/ethanol bath, the gas inside the
NMR tube was evacuated and CO2 was introduced at a pressure of
1 atm. The reaction mixture was kept at 313.1 K in a thermostatted
bath for 72 h and the 2/5 ratio was monitored by comparing the
Supporting Information Available: Cartesian coordinates and
total energies of all the structures, basis set information for the
large basis set single-point calculations, crystallographic data for
2b, 2c, 4, and 6 (CIF), and full citation for ref 30. This material is
available free of charge via the Internet at http://pubs.acs.org.
(49) Takagi, N.; Nagase, S. Organometallics 2007, 26, 469.
(50) Gaussian Basis Sets for Molecular Calculations; Elsevier: Amsterdam,
1984.
(48) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.