Fig. 1 POV-ray depiction of 3, C: black, B: light-pink, Cl: green;
P: orange, O: red. H-atoms are omitted for clarity.
binding to boron is in slow exchange between the two
boron sites.
Fig. 2 POV-ray depiction of 4, C: black, B: light-pink, Cl: green;
P: orange, O: red. H-atoms are omitted for clarity.
Despite the formation of an adduct, the ability of 2 to act as
an FLP was queried. We have previously shown that borane/
donor FLP systems exhibit the ability to effect THF ring opening.
Thus, in an initial test, the species 2 was exposed to THF. This
resulted in the formation of a new species which was ultimately
obtained as colorless needles of 3 in 97% yield. The 31P {1H} and
11B NMR spectrum showed singlets at 48.80 and 12.00 ppm,
respectively. The corresponding 1H NMR spectrum revealed
resonances consistent with the ring-opening of THF and
thus the formulation of 3 as 1,2-C6H4(BCl2)2(O(CH2)4PtBu3).
consistent with the thermal stability of 4. The O-C-O angle
was found to be 127.0(4)1. It is interesting to note that despite
the asymmetry of the CO2 binding, both boron centers are four
coordinate. This is achieved via bridging of one chlorine atom
between the B centers. The bridging B-Cl distances are found to
be 1.999(5) and 2.038(5) A, while the terminal B-Cl bond
distances averaged 1.838(5) A.
Compound 4 proved to be remarkably stable with respect to
loss of CO2. No decomposition was observed on heating to
80 1C for 24 h. This stands in marked contrast to the CO2
products derived from the FLPs tBu3P/B(C6F5)3 and
(C6H2Me3)2PCH2CH2B(C6F5)2. These species were seen to
lose CO2 at 80 1C and ꢁ20 1C, respectively. Similarly, 4
is significantly more thermally robust than the symmetric
FLP-CO2 complexes derived from the bis-boranes Me2C =
C(BX2)2 (X = Cl, C6F5) and PtBu3 as these latter species were
observed to lose CO2 at 15 1C. The enhanced thermal stability
is thought to arise from the bridging chlorine atom between
the boron centers of 4 as this serves to enhance the Lewis
acidity of the boron bound to the oxygen of CO2. This
enhanced acidity results in a strong B–O bond and explains
the improved stability of the CO2 activation product.
This formulation of
3 was subsequently confirmed by
X-ray crystallography (Fig. 1).z The structural data affirm
the zwitterionic nature resulting from the ring-opening of
THF. The oxygen atom bridges the two boron centers giving
rise to B-O bond distances of 1.532(4) and 1.533(4) A with a
B-O-B angle of 112.2(2)1. It is noteworthy that the constraint
of the five-membered ring imposes a B-O-B angle that is
markedly less than that seen for unrestricted cases such as
O(B(C6F5)2)2 (B-O-B 139.5(2)1). The remaining metric para-
meters are unexceptional. The ability of 2 to activate THF is
consistent with the notion that this species provides access to a
reactive FLP.22–24
Having affirmed the ability of 2 to act as an FLP, a solution
of 2 in toluene was exposed to an atmosphere of CO2 and
allowed to stir overnight. Subsequent work-up afforded a white,
microcrystalline solid 4 in 77% yield. The 31P{1H} NMR
spectrum of 4 exhibits a resonance at 48.4 ppm, while the
1H NMR data are consistent with inclusion of phosphine and
bis-borane in a 1 : 1 ratio. A 13C{1H} NMR signal is observed at
156.3 ppm. This signal also exhibits phosphorus-carbon cou-
pling of 94.6 Hz, suggesting the inclusion of CO2. This was
further supported by the observation of an IR absorption at
1719 cmꢁ1. The 11B NMR spectrum reveals a single peak at
18.0 ppm. Although this latter observation infers a molecular
symmetry that would arise from the CO2 spanning the two
boron centers, X-ray diffraction data reveal the molecular
structure is in fact dissymmetric (Fig. 2). In the solid state,
the CO2 fragment is captured between the phosphine and one of
the boron centers. This gives rise to P-C, O-B, C-O and CQO
bond lengths of 1.894(4), 1.477(5), 1.320(4) and 1.177(5) A,
respectively. While the P-C bond in 4 is similar to that seen in
tBu3P(CO2)B(C6F5)3 (1.8931(12) A), Mes2PCH2CH2B(C6F5)2-
(CO2) (1.900(3) A)15 and Mes3PCO2(AlCl3)2 (1.886(3) A), the
O-B distance in 4 is significantly shorter tBu3P(CO2)B(C6F5)3:
1.5474(15) A, Mes2PCH2CH2B(C6F5)2(CO2): 1.550(4) A and
Mes3PCO2(AlCl3)2: 1.550(3) A). This latter observation is
Given the stability of 4, efforts were undertaken to reduce
the captured CO2. Treatment with the amine-borane
Me2NHBH3 for 15 min followed by quenching with D2O led
to the generation of methanol-d1 in 34% yield (Scheme 1).
The FLP derived from 2 failed to activate hydrogen. Treatment
of compound 4 with three equivalents of [C5H6Me4NH2]-
[HB(C6F5)3], a known salt derived from the activation of H2
by the FLP C5H6Me4NH/B(C6F5)3, reacted slowly. After
stirring for 24 h, subsequent treatment with D2O afforded
methanol in 15% yield. However, altering the salt to
[C5H6Me4NH2][HB(C6F5)2(C7H11)] (5), allowed the reaction
time to be reduced to 1 h, and boosted the yield of methanol
to 57%.
In conclusion, we have demonstrated that the bis-borane
1,2-C6H4(BCl2)2 in combination with PtBu3, reacts as an FLP
despite formation of a phosphorus-boron adduct. Moreover
this FLP is shown to bind CO2 strongly in an asymmetric
fashion, allowing further stoichiometric reactions with reducing
agents. Of particular interest is the ability to reduce this species
employing a product of FLP activation of H2. Although in
this case, B-O bond cleavage requires the addition of water,
this finding further demonstrates the potential of FLPs
to concurrently activate CO2 and H2 for CO2 reduction.
c
Chem. Commun.
This journal is The Royal Society of Chemistry 2012