Single oxygen-atom insertion into p-b bonds: On- and off-metal transformation of a borylphosphine into a borylphosphinite

Article abstract of DOI:10.1002/chem.201500378

An oxygen atom is selectively inserted into the P-B bond of a borylphosphine (L₁) by reaction with Me₃NO to afford the corresponding borylphosphinite (L₂). This transformation can also be effected when L₁ is coordinated to rhodium. The ν(CO) values for trans-[RhCl(CO)(L)₂] reveal very different electronic properties for coordinated L₁ and L₂ which translate into the strikingly different performances of the complexes [RhCl(L)(cod)] (L=L₁ or L₂, cod=1,5-cyclooctadiene) in hydrosilylation and hydroboration catalysis. What a difference an atom makes: A single O atom can be readily incorporated into a borylphosphine-rhodium catalyst to produce a borylphosphinite-rhodium catalyst with sharply contrasting activity and selectivity.

Full text of DOI:10.1002/chem.201500378

DOI: 10.1002/chem.201500378
Communication
&
Ligand Design
Single Oxygen-Atom Insertion into PÀB Bonds: On- and Off-Metal
Transformation of a Borylphosphine into a Borylphosphinite
Jonathan A. Bailey, Hazel A. Sparkes, and Paul G. Pringle*[a]
of 1-phospha-2-bora-acenaphthene has been previously re-
Abstract: An oxygen atom is selectively inserted into the
ported[15] but, in other cases, elemental chalcogens have been
PÀB bond of a borylphosphine (L1) by reaction with
shown to react with compounds containing PÀB bonds to give
Me3NO to afford the corresponding borylphosphinite (L2).
PV products.[16]
This transformation can also be effected when L1 is coor-
The borylphosphine L1 was prepared in quantitative yield
dinated to rhodium. The n(CO) values for trans-
from the chlorosilane elimination reaction[6,17] shown in
[RhCl(CO)(L)2] reveal very different electronic properties for
Scheme 1; L1 has been fully characterised (see the Supporting
coordinated L1 and L2 which translate into the strikingly
different performances of the complexes [RhCl(L)(cod)]
(L=L1 or L2, cod=1,5-cyclooctadiene) in hydrosilylation
and hydroboration catalysis.
The varied chemistry of phosphinoboranes can be rationalised
by reference to the PÀB bonding, which can be described by
the Lewis structures I (phosphinoborane) and II (borylphos-
phine) shown below. The PÀB bond order and its reactivity
depend on the stereoelectronic effects of the P- and B-sub-
stituents.[1] Phosphinoboranes have been shown to heterolyti-
cally cleave H2,[2] dehydrogenate amine-boranes[3] and react
with compounds containing CꢀN, C=O and C=C functionali-
ties.[4] Furthermore, they have the capacity to ligate as k1-P[1a]
or h2-P=B[5] and we have shown that they can be ligands for
efficient homogeneous alkene hydrogenation catalysis.[6]
Scheme 1. Synthesis of ligands L1 and L2 and complexes 1, 2, 4 and 5 via
Here we report the insertion of an oxygen atom (from O2 or
N-oxides) into a PÀB bond to convert a borylphosphine into
a borylphosphinite, which have contrasting s-donor/p-accept-
or ligand properties and show very different performance in
catalysis. The insertion of an oxygen atom into single bonds is
rare[7] outside the realm of monoxygenases.[8] Peroxides will
insert O atoms into the CÀC bond of ketones to give esters
(the Bayer–Villiger reaction[9]) and amine-oxides[10] have been
off-metal and on-metal routes. Reagents: i) [RhCl(CO)2]2; ii) [RhCl(cod)]2.
Information for the data), including its 31P NMR spectrum,
which shows a signal at À48.4 ppm that is broad (w1/2 =67 Hz)
due to unresolved coupling to quadrupolar 11B. Treatment of
L1 with one equivalent of Me3NO (TMAO) led to complete con-
version to a product assigned structure L2 on the basis of the
used for oxygen insertion into SiÀSi,[11] CÀB,[12] BÀB[13] and BÀ following NMR data. The 31P NMR signal for L2 is a sharp singlet
Ru[14] bonds. The insertion of S and Se atoms into PÀB bonds
(consistent with the loss of 1JPB coupling) at 134.8 ppm, a chem-
ical shift that is similar to that of the phosphinite iPr2POPh
(149 ppm).[18] The 11B NMR chemical shift of L2 is 22.2 ppm
(compared with 32.0 for L1).
[a] J. A. Bailey, Dr. H. A. Sparkes, Prof. Dr. P. G. Pringle
School of Chemistry, University of Bristol
Cantock’s Close, Bristol, BS8 1TS (UK)
Solid TMAO is a convenient reagent for the conversion of L1
to L2 because it is easy to control the stoichiometry and the
volatile NMe3 by-product is readily removed, although the
E-mail: paul.pringle@bristol.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500378.
Chem. Eur. J. 2015, 21, 5360 – 5363
5360
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
same transformation was observed in the reaction of L1 with
O2.
Thus, when a CH2Cl2 solution of L1 was stirred under 2 atm
of O2 for 2 days, unexpectedly the main species present (ca.
80% according to 31P NMR spectroscopy) was unreacted L1
and, of the several products detected, L2 was prominent (see
the Supporting Information for the 31P NMR spectrum of the
reaction mixture).
The complexes trans-[RhCl(CO)(L)2] (1, L=L1) and (2, L=L2)
were prepared (see Scheme 1) in order to compare the donor
characteristics of the two ligands (see Table 1). IR spectroscopy
Table 1. IR[a] and 31P NMR[b] data for the complexes trans-[RhCl(CO)(L)2].
Figure 1. Thermal ellipsoid (50% probability) plot of 1 (left) and 2 (right),
omitting all hydrogen atoms. Selected bond lengths [Š] and angles [8] for
complex 1: Rh1ÀC33 1.7967(15), Rh1ÀP1 2.3333(4), Rh1ÀP2 2.3574(4), Rh1À
Cl1 2.3888(3), P1ÀB1 1.9585(16), P2ÀB2 1.9533(16), P1ÀC11 1.8584(14), P1À
C14 1.8526(15), P2ÀC27 1.8634(15), P2ÀC30 1.8562(15), N1ÀB1 1.415(2), N2À
B1 1.406(2), N3ÀB2 1.410(2), N4ÀB2 1.4129(19), O1ÀC33 1.1548(18), B1-P1-
P2-B2 96.2. Complex 2: Rh1ÀC33 1.813(2), Rh1ÀP1 2.3176(6), Rh1ÀP2
2.3278(6), Rh1ÀCl1 2.3670(6), P1ÀO1 1.6254(17), P2ÀO2 1.6183(16), P1ÀC11
1.827(2), P1ÀC14 1.837(2), P2ÀC27 1.838(2), P2ÀC30 1.827(2), C33ÀO3
1.152(3), O1ÀB1 1.385(3), O2ÀB2 1.379(3), B1ÀN1 1.416(3), N2ÀB1 1.411(3),
N3ÀB2 1.415(3), N4ÀB2 1.417(3), O1-P1-P2-O2 143.8.
Complex
Ligand n(CO) [cmÀ1
]
dP [ppm] 1JPRh [Hz]
1
2
3
L1
L2
L3
1956
1977
1984
À4.9
145.1
161.8
42.6
ꢁ110 (br)
125
134
[RhCl(CO)(iPr2PPh)2][19] iPr2PPh 1965
124.4
[a] Measured in CH2Cl2. [b] Measured in CD2Cl2 (for 1 and 2) or CDCl3 (for
3 and trans-[RhCl(CO)(iPr2PPh)2]).
revealed a difference of 21 cmÀ1 between the n(CO) stretching
frequencies for 1 (1956 cmÀ1) and 2 (1977 cmÀ1), consistent which might be expected to react with TMAO to generate co-
with L1 being a much more electron-donating ligand than L2. ordinative unsaturation by release of CO2.[21] The facility with
Comparison of these values to those for the analogues trans- which 1 is converted to 2 shows that the P lone pair need not
[RhCl(CO)(L)2] for L=iPr2PPh (1965 cmÀ1 [19]
)
and L=iPr2POPh be involved in the oxygenation of the PÀB bonds of L1 and
(L3) (3, 1984 cmÀ1; see the Supporting Information for experi- leads to the mechanism suggested in Scheme 2, in which
mental details) shows that, as previously demonstrated,[17] the
naphthalene-derived boryl fragment is more electron donating
than phenyl. The 31P NMR spectra of 1 and 2 were markedly
different: the signal for 1 is a doublet at À4.9 ppm (1JPRh
1
ꢁ110 Hz), which is broad due to the unresolved JPB coupling,
whereas the signal for 2 is a sharp doublet at 145.1 ppm
(1JPRh =125 Hz).
Complexes 1 and 2 have been characterised crystallographi-
cally and as is clear from Figure 1, the conformations adopted
in the solid state are quite different (B1-P1-P2-B2 and O1-P1-
Scheme 2. Suggested mechanism for the on-metal oxygen insertion into
1 and 4.
P2-O2 torsion angles of 96.28 and 143.88 in 1 and 2 respective-
ly). The BÀO bond lengths in 2 (1.385(3) and 1.379(3) Š) are
typical of single bonds with little evidence of BÀO p-overlap
(cf. the B=O determined in an oxoborane was 1.304(2) Š[20]).
It is of interest to compare the product of oxidation of sec- a Lewis acid-base adduct is the intermediate; this mechanism
ondary phosphines R2PH with that of the boryl phosphine L1. is reminiscent of that suggested by Yamashita et al.[14] for the
Oxidation of secondary phosphines gives PV secondary phos- oxygenation of a BÀRu bond with morpholine-N-oxide (NMO).
phine oxides R2P(=O)H as the major species in prototopic
The consequences of the different ligating properties of bor-
equilibrium with their PIII hydroxyphosphine R2POH tautomers. ylphosphine L1 and borylphosphinite L2, in catalysis have been
In the oxidation of L1 by TMAO, no PV products were detected. investigated by comparing the performance of complexes of
The results of oxidation of coordinated L1 described below the type [RhCl(L)(cod)][22] (cod=1,5-cyclooctadiene) in hydrosi-
show that a mechanism in which the phosphorus lone pair is lylation and hydroboration catalysis (see below). The reaction
not involved in the oxygenation of L1 to give L2 is feasible.
of L1 or L2 with [RhCl(cod)]2 produced [RhCl(L)(cod)] (4, L=L1;
Addition of two equivalents of TMAO to a solution of com- 5, L=L2). Complex 5 was also prepared quantitatively by the
plex 1 in CH2Cl2 gave 2 cleanly and in less than 5 min accord- addition of TMAO to 4 (see Scheme 1 and the Supporting In-
ing to 31P NMR spectroscopy. The chemoselectivity of this oxi- formation for the characterising data). Crystals of 4 and 5 suita-
dation for the PÀB bond was a surprise considering the com- ble for X-ray diffraction were grown from CH2Cl2/pentane and
plex has a low-valent RhI centre as well as a coordinated CO, CH2Cl2/hexane respectively and their crystal structures are
Chem. Eur. J. 2015, 21, 5360 – 5363
www.chemeurj.org
5361
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Table 2. Hydrosilylation of 4-methoxystyrene with Ph3SiH.
Pre-catalyst[a]
Conversion [%]
A
B
C
4
5
6
8
96
6
4
53
5
1
20
<1
3
23
1
[a] Conditions: 608C, C6D6, 6 h, alkene/silane ratio: 1:1.05, 2 mol% pre-cat-
1
alyst. Conversion to products monitored by H NMR spectroscopy.
The conversion obtained with the L2 complex 5 was 96% com-
pared with only 8% with the L1 analogue 4. This large differ-
ence in activity prompted us to screen the analogous complex
[RhCl(iPr2POPh)(cod)] (6) featuring the ligand iPr2POPh (L3) (see
the Supporting Information for details). Complex 6 was tested
under the same catalysis conditions and just 6% conversion
was observed which shows that the boryloxy ligand L2 has
considerable advantages in terms of catalyst activity over both
L1 and L3 for this reaction.
Figure 2. Thermal ellipsoid (50% probability) plot of 4 (left) and 5 (right),
omitting all hydrogen atoms. For 5, only one orientation of the disordered
iPr groups attached to P are shown for clarity. Selected bond lengths [Š] for
complex 4: Rh1ÀP1 2.3498(9), Rh1ÀCl1 2.3775(9), P1ÀB1 1.962(4), P1ÀC11
1.874(3), P1ÀC14 1.873(3), N1ÀB1 1.404(5), N2ÀB1 1.407(5). Complex 5: Rh1À
P1 2.261(3), Rh1ÀCl1 2.397(2), P1ÀO1 1.624(7), P1ÀC11 1.841(11), P1ÀC14
1.865(11), O1ÀB1 1.373(14), N1ÀB1 1.413(10), N2ÀB1 1.417(10).
The Rh-catalysed hydrosilylation of olefins is often accompa-
nied by numerous by-products of isomerisation, oligomerisa-
tion, polymerisation, dehydrogenative silylation and hydroge-
nation of the alkene.[23] In our case, three major products were
obtained: the expected hydrosilylated product A, the hydro-
genated species B and the dehydrogenative silylation product
C. In each case, the hydrosilylated product A was the major
product with a selectivity of over 50% with the productive cat-
alyst 5.
shown in Figure 2. The 31P NMR signal for 4 is a singlet at
À16.2 ppm with the expected 103Rh coupling subsumed by the
broadness of the signal (w1/2 =285 Hz), whereas a sharp dou-
blet at 152.3 ppm (1JPRh =174 Hz) was observed for 5 (Figure 3).
The hydroboration of p-methoxystyrene with pinacol borane
(HBPin, see Scheme 4) is catalysed by complexes 4–6 (Table 3).
Complex 5 is very active and highly selective for the linear
product D in contrast to the less active and unselective cata-
lyst 4. The POC catalyst 6 showed similarly high activity to the
POB catalyst 5 but interestingly the selectivity is switched to
the branched product E. It is known that the hydroboration of
styrene derivatives can be steered to Markovnikov and anti-
Figure 3. 31P{1H} spectra of 4 (bottom) and 5 (top) in CD2Cl2.
The catalytic activities of complexes 4 and 5 were measured
for the hydrosilylation and hydroboration of 4-methoxystyrene.
For the hydrosilylation reactions (Scheme 3), solutions of 4-me-
thoxystyrene (1 eq), Ph3SiH (1.05 eq) and the catalyst (2 mol%)
in C6D6 were heated at 608C for 6 h and the conversions moni-
Scheme 4. Hydroboration of 4-methoxystyrene with pinacolborane (HBPin).
Table 3. Hydroboration of 4-methoxystyrene with pinacolborane.
1
tored by H NMR spectroscopy; the results are given in Table 2.
Pre-catalyst[a]
Conversion [%]
D
E
Other[b]
4
5
6
53
>99
>99
20
80
6
17
4
77
16
15
16
[a] Conditions: RT, C6D6, 2 h, alkene/borane ratio: 1:1.1, 2 mol% pre-cata-
lyst. Conversion to products monitored by 1H NMR spectroscopy. [b] Sev-
eral unidentified by-products (see 1H NMR spectra in the Supporting In-
formation).
Scheme 3. Hydrosilylation of 4-methoxystyrene with Ph3SiH.
Chem. Eur. J. 2015, 21, 5360 – 5363
www.chemeurj.org
5362
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Markovnikov selectivities by modifications of the catalyst;[24]
the very different results obtained with L1, L2 and L3 for this re-
action reflect their fundamentally different ligating properties.
It has been shown that a single oxygen atom insertion into
the PÀB bond of L1 occurs upon reaction with TMAO to give
L2, which is a new type of ligand that we have called a boryl-
phosphinite to stress the relationship to phosphinites, the
well-known hydrocarbyl analogues. The PÀB oxygenation can
be performed on free or bound ligand L1. This O insertion
transforms the ligand from an electron-rich to an electron-poor
donor in one step; a similar process with conventional PÀC li-
gands is inconceivable. The large difference detected in the
ligand binding properties of L1 and L2 is reflected in the cata-
lytic activity and selectivity of their complexes 4 and 5 in the
catalytic hydrosilylation and hydroboration of p-methoxystyr-
ene. The clean insertion of O into a BÀP bond is a new reaction
with the potential to be general.
[6] J. A. Bailey, M. F. Haddow, P. G. Pringle, Chem. Commun. 2014, 50, 1432–
1434.
[7] a) M. Shieh, C.-H. Ho, W.-S. Sheu, H.-W. Chen, J. Am. Chem. Soc. 2010,
132, 4032–4033; b) C. Cristóbal, E. lvarez, M. Paneque, M. L. Poveda,
Organometallics 2013, 32, 714–717.
[8] H. Leisch, K. Morley, P. C. K. Lau, Chem. Rev. 2011, 111, 4165–4222.
[9] G. J. ten Brink, I. W. C. E. Arends, R. A. Sheldon, Chem. Rev. 2004, 104,
4105–4124.
[10] a) H.-J. Knçlker, J. Prakt. Chem./Chem.-Ztg. 1996, 338, 190–192; b) A.
Albini, Synthesis 1993, 1993, 263–277.
[11] H. Sakurai, M. Kira, M. Kumada, Bull. Chem. Soc. Jpn. 1971, 44, 1167–
1167.
[12] J. A. Soderquist, M. R. Najafi, J. Org. Chem. 1986, 51, 1330–1336.
[13] a) B. Thiele, P. Paetzold, U. Englert, Chem. Ber. 1992, 125, 2681–2686;
b) S. Bae, M. K. Lakshman, J. Org. Chem. 2008, 73, 1311–1319; c) H. P.
Kokatla, P. F. Thomson, S. Bae, V. R. Doddi, M. K. Lakshman, J. Org. Chem.
2011, 76, 7842–7848.
[14] T. Miyada, M. Yamashita, Organometallics 2013, 32, 5281–5284.
[15] A. Tsurusaki, T. Sasamori, N. Tokitoh, Chem. Eur. J. 2014, 20, 3752–3758.
[16] a) P. Kçlle, H. Nçth, R. T. Paine, Chem. Ber. 1989, 122, 423–426; b) H.
Nçth, S. Staude, M. Thomann, R. T. Paine, Chem. Ber. 1993, 126, 611 –
618; c) K.-C. Schwan, A. Y. Timoskin, M. Zabel, M. Scheer, Chem. Eur. J.
2006, 12, 4900–4908; d) H. Nçth, W. Schrägle, Chem. Ber. 1964, 97,
2218–2229.
[17] J. A. Bailey, M. Ploeger, P. G. Pringle, Inorg. Chem. 2014, 53, 7763–7769.
[18] R. B. Bedford, S. L. Hazelwood, P. N. Horton, M. B. Hursthouse, Dalton
Trans. 2003, 4164–4174.
[19] D. B. Grotjahn, X. Zeng, A. L. Cooksy, W. S. Kassel, A. G. DiPasquale, L. N.
Zakharov, A. L. Rheingold, Organometallics 2007, 26, 3385–3402.
[20] D. Vidovic, J. A. Moore, J. N. Jones, A. H. Cowley, J. Am. Chem. Soc. 2005,
127, 4566–4567.
Acknowledgements
We thank Marten Ploeger for his experimental input and the
Engineering and Physical Sciences Research Council for sup-
port.
[21] a) W. Hieber, A. Lipp, Chem. Ber. 1959, 92, 2085–2088; b) T.-Y. Luh,
Coord. Chem. Rev. 1984, 60, 255–276.
[22] a) W. Duczmal, W. Urbaniak, B. Marciniec, J. Organomet. Chem. 1986,
Keywords: oxygen
hydrosilylation · hydroboration · phosphane ligands
insertion
·
phosphinoboranes
·
´
´
317, 85–92; b) B. Marciniec, W. Duczmal, W. Urbaniak, E. Sliwinska, J. Or-
ganomet. Chem. 1990, 385, 319–327; c) K. Ruhland, P. Gigler, E. Herdt-
weck, J. Organomet. Chem. 2008, 693, 874–893; d) B. Natarajan, N.
Jayaraman, J. Organomet. Chem. 2011, 696, 722–730; e) E. Tomµs-Men-
divil, R. García-lvarez, C. Vidal, P. Crochet, V. Cadierno, ACS Catal. 2014,
4, 1901–1910.
[1] a) J. A. Bailey, P. G. Pringle, Coord. Chem. Rev. 2015, DOI: 10.1016/
j.ccr.2015.02.001 and references therein; b) H. H. Karsch, G. Hanika, B.
Huber, K. Meindl, S. Koenig, C. Krueger, G. Mueller, J. Chem. Soc. Chem.
Commun. 1989, 373–375; c) D. C. Pestana, P. P. Power, J. Am. Chem. Soc.
1991, 113, 8426–8437; d) R. T. Paine, H. Noeth, Chem. Rev. 1995, 95,
343–379; e) P. P. Power, Chem. Rev. 1999, 99, 3463–3504; f) R. C. Fischer,
P. P. Power, Chem. Rev. 2010, 110, 3877–3923.
[2] S. J. Geier, T. M. Gilbert, D. W. Stephan, J. Am. Chem. Soc. 2008, 130,
12632–12633.
[3] S. J. Geier, T. M. Gilbert, D. W. Stephan, Inorg. Chem. 2011, 50, 336–344.
[4] J. M. Breunig, A. Hübner, M. Bolte, M, Wagner, H.-W. Lerner, Organome-
tallics 2013, 32, 6792–6799.
[23] B. Marciniec, Coord. Chem. Rev. 2005, 249, 2374–2390.
[24] a) K. Burgess, M. J. Ohlmeyer, Chem. Rev. 1991, 91, 1179–1191; b) I. Be-
letskaya, A. Pelter, Tetrahedron 1997, 53, 4957–5026; c) C. M. Crudden,
D. Edwards, Eur. J. Org. Chem. 2003, 2003, 4695–4712; d) A.-M. Carroll,
T. O’Sullivan, P. Guiry, Adv. Synth. Catal. 2005, 347, 609–631.
[5] A. Amgoune, S. Ladeira, K. Miqueu, D. Bourissou, J. Am. Chem. Soc.
2012, 134, 6560–6563.
Received: January 29, 2015
Published online on February 20, 2015
Chem. Eur. J. 2015, 21, 5360 – 5363
www.chemeurj.org
5363
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim