Homogeneous CO hydrogenation: Ligand effects on the lewis acid-assisted reductive coupling of carbon monoxide

Article abstract of DOI:10.1021/om100638d

Structure-function studies on the role of pendent Lewis acids in the reductive coupling of CO are reported. Cationic rhenium carbonyl complexes containing zero, one, or two phosphinoborane ligands (Ph₂P(CH ₂)_nB(C₈H₁₄), n = 1-3) react with the nucleophilic hydride [HPt(dmpe)₂]⁺ to reduce [M-CO] ⁺ to M-CHO; this step is relatively insensitive to the Lewis acid, as both pendent (internal) and external boranes of appropriate acid strength can be used. In contrast, whether a second hydride transfer and C-C bond forming steps occur depends strongly on the number of carbon atoms between P and B in the phosphinoborane ligands, as well as the number of pendent acids in the complex: shorter linker chain lengths favor such reductive coupling, whereas longer chains and external boranes are ineffective. A number of different species containing partially reduced CO groups, whose exact structures vary considerably with the nature and number of phosphinoborane ligands, have been crystallographically characterized. The reaction of [(Ph₂P(CH ₂)₂B(C₈H₁₄))₂Re(CO) ₄]⁺ with [HPt(dmpe)₂]⁺ takes place via a hydride shuttle mechanism, in which hydride is transferred from Pt to a pendent borane and thence to CO, rather than by direct hydride attack at CO. Addition of a second hydride in C₆D₅Cl at -40 °C affords an unusual anionic bis(carbene) complex, which converts to a C-C bonded product on warming. These results support a working model for Lewis acid-assisted reductive coupling of CO, in which B (pendent or external) shuttles hydride from Pt to coordinated CO, followed by formation of an intramolecular B-O bond, which facilitates reductive coupling.

Full text of DOI:10.1021/om100638d

Organometallics 2010, 29, 4499–4516 4499
DOI: 10.1021/om100638d
Homogeneous CO Hydrogenation: Ligand Effects on the Lewis
Acid-Assisted Reductive Coupling of Carbon Monoxide
Alexander J. M. Miller, Jay A. Labinger,* and John E. Bercaw*
Arnold and Mabel Beckman Laboratories of Chemical Synthesis, California Institute of Technology,
Pasadena, California 91125, United States
Received July 1, 2010
Structure-function studies on the role of pendent Lewis acids in the reductive coupling of CO are
reported. Cationic rhenium carbonyl complexes containing zero, one, or two phosphinoborane
ligands (Ph2P(CH2)nB(C8H14), n=1-3) react with the nucleophilic hydride [HPt(dmpe)2]þ to
reduce [M-CO]þ to M-CHO; this step is relatively insensitive to the Lewis acid, as both pendent
(internal) and external boranes of appropriate acid strength can be used. In contrast, whether a
second hydride transfer and C-C bond forming steps occur depends strongly on the number of
carbon atoms between P and B in the phosphinoborane ligands, as well as the number of pendent
acids in the complex: shorter linker chain lengths favor such reductive coupling, whereas longer
chains and external boranes are ineffective. A number of different species containing par-
tially reduced CO groups, whose exact structures vary considerably with the nature and
number of phosphinoborane ligands, have been crystallographically characterized. The reaction
of [(Ph2P(CH2)2B(C8H14))2Re(CO)4]þ with [HPt(dmpe)2]þ takes place via a “hydride shuttle”
mechanism, in which hydride is transferred from Pt to a pendent borane and thence to CO, rather
than by direct hydride attack at CO. Addition of a second hydride in C6D5Cl at -40 °C affords an
unusual anionic bis(carbene) complex, which converts to a C-C bonded product on warming. These
results support a working model for Lewis acid-assisted reductive coupling of CO, in which B
(pendent or external) shuttles hydride from Pt to coordinated CO, followed by formation of an
intramolecular B-O bond, which facilitates reductive coupling.
Introduction
reaction intermediates, such as formyl5 and hydroxy-
methyl5d,6 groups, have been synthesized by reaction of main
group metal hydrides with metal carbonyls; in some cases
sequential reduction and protonation of mid to late transi-
tion metal carbonyls resulted in the release of C2 and higher
organics.4c,7 Unfortunately, those hydrides are not readily
obtainable from H2 and are incompatible with the strong
acids required, precluding catalysis. Some early transition
metal species (and their f-block congeners) that can be
formed directly from H2,8 such as Cp*2ZrH2,4d can effect
both CO reduction and C-C coupling, but the extremely
strong M-O bonds thus generated will again preclude
catalysis. A few examples of homogeneous catalytic conver-
sion of syngas to methanol and C2 products (EtOH, ethylene
Many of the approaches under active investigation for
obtaining fuels and commodity chemicals from coal, meth-
ane, or biomass follow an indirect route via synthesis gas
(syngas, CO/H2), typified by the Fischer-Tropsch (F-T)
process.1,2 Selective conversion of syngas to discrete multi-
carbon hydrocarbons or oxygenates could offer substantial
improvements over nonselective F-T.3 The possibility that
homogeneous catalysis might provide more promising oppor-
tunities for selective transformations has intrigued organo-
metallic chemists for several decades.4 A number of plausible
*To whom correspondence should be addressed. E-mail: jal@
caltech.edu; bercaw@caltech.edu.
(1) Khodakov, A. Y.; Chu, W.; Fongarland, P. Chem. Rev. 2007,
107, 1692.
(2) Rofer-DePoorter, C. K. Chem. Rev. 1981, 81, 447.
(3) United Nations Development Program. World Energy Assess-
ment Report: Energy and the Challenge of Sustainability; United Nations,
New York, 2000.
(4) (a) Gladysz, J. A. Adv. Organomet. Chem. 1982, 20, 1. (b) Herrmann,
W. A. Angew. Chem., Int. Ed. 1982, 21, 117. (c) Cutler, A. R.; Hanna, P. K.;
Vites, J. C. Chem. Rev. 1988, 88, 1363. (d) Wolczanski, P. T.; Bercaw, J. E. Acc.
Chem. Res. 1980, 13, 121.
(5) (a) Collman, J. P.; Winter, S. R. J. Am. Chem. Soc. 1973, 95, 4089.
(b) Casey, C. P.; Neumann, S. M. J. Am. Chem. Soc. 1976, 98, 5395. (c) Tam,
W.; Wong, W.-K.; Gladysz, J. A. J. Am. Chem. Soc. 1979, 101, 1589.
(d) Casey, C. P.; Andrews, M. A.; McAlister, D. R.; Rinz, J. E. J. Am. Chem.
Soc. 1980, 102, 1927.
(6) Sweet, J. R.; Graham, W. A. G. J. Am. Chem. Soc. 1982, 104, 2811.
(7) Brown, S. L.; Davies, S. G. J. Chem. Soc., Chem. Commun.
1986, 84.
(8) (a) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw,
J. E. J. Am. Chem. Soc. 1978, 100, 2716. (b) Fagan, P. J.; Moloy, K. G.;
Marks, T. J. J. Am. Chem. Soc. 1981, 103, 6959. (c) Gambarotta, S.; Floriani,
C.; Chiesivilla, A.; Guastini, C. J. Am. Chem. Soc. 1983, 105, 1690.
(d) Evans, W. J.; Grate, J. W.; Doedens, R. J. J. Am. Chem. Soc. 1985,
107, 1671. (e) Miller, R. L.; Toreki, R.; Lapointe, R. E.; Wolczanski, P. T.;
Vanduyne, G. D.; Roe, D. C. J. Am. Chem. Soc. 1993, 115, 5570.
(f) Werkema, E. L.; Maron, L.; Eisenstein, O.; Andersen, R. A. J. Am.
Chem. Soc. 2007, 129, 2529. (g) Shima, T.; Hou, Z. J. Am. Chem. Soc.
2006, 128, 8124.
r
2010 American Chemical Society
Published on Web 09/29/2010
pubs.acs.org/Organometallics

4500 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Scheme 1
Scheme 2
of organic carbonyls14 and alkyl migratory insertions of
coordinated CO15 is well known; participation of Lewis
acids in homogeneous catalytic CO reduction9b and hetero-
geneous F-T chemistry16 has also been proposed. The addi-
tional advantage of intramolecular attachment could make
conversion possible with a relatively weak acidic center,
which might not be sufficiently activating as a separate pro-
moter. A demonstration of this concept was achieved by
positioning a trialkylborane center in the secondary coor-
dination sphere of a carbonyl cation, affording the transfor-
mations shown in Scheme 2.10a Analogous chemistry is not
observed in the absence of pendent Lewis acids.
While alkylborane-based systems are probably not viable
catalysts because of strong B-O bonding, they do provide an
excellent framework for investigating some of the key issues
arising from this approach. In particular, which factors affect
how the chelate effect operates? How does reactivity depend on
the length of the tether connecting the borane to the metal
center? Can an external borane promote the same transforma-
tions? What happens if there is only one Lewis acid center
attached instead of two? We would also like to learn something
about the fundamental mechanism of this transformation: how
is hydride transferred from Pt to C, and how does C-C bond
formation take place? We report here on our findings.
glycol) have been reported, but extremely high pressures and
temperatures are required.9
Our renewed interest in this approach10 was inspired by
the discovery that late transition metal hydrides [HM(PP)2]þ
(M = Ni, Pd, Pt; PP = chelating diphosphine ligand), which
may be formed directly from H2 by heterolytic activation,11
can form C-H bonds by facile nucleophilic attack on metal
carbonyl complexes.12 Because these metals should not form
prohibitively strong M-O bonds, we proposed that a dual
system, consisting of the late transition metal hydrogen acti-
vator along with a relatively electrophilic metal carbonyl
complex (probably from the middle of the transition series,
where C-H bond formation is well documented4a,c,5,12a),
could lead to catalytic conversion, perhaps as shown in
Scheme 1. We10b and others13 have demonstrated stoichio-
metric versions of Scheme 1, but the various steps operate
under conditions that are mutually incompatible; in parti-
cular, some steps require very strong acids or other electro-
philes that would react with hydride sources.
In order to overcome the problems posed by strong
hydride reductants, we designed metal carbonyl complexes
with pendent Lewis acids. Lewis acid promotion of reductions
Results and Discussion
(9) (a) Dombek, B. D. Adv. Catal. 1983, 32, 325. (b) Demitras, G. C.;
Muetterties, E. L. J. Am. Chem. Soc. 1977, 99, 2796.
(10) (a) Miller, A. J. M.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem.
Soc. 2008, 130, 11874. (b) Elowe, P. R.; West, N. M.; Labinger, J. A.; Bercaw,
J. E. Organometallics 2009, 28, 6218. (c) Miller, A. J. M.; Labinger, J. A.;
Bercaw, J. E. J. Am. Chem. Soc. 2010, 132, 3301.
(11) The hydride used in this work, [HPt(dmpe)2][PF6], is more easily
isolated when prepared from NaBH4. This synthetic route was used
rather than direct formation from H2, as detailed in the Supporting
Information.
(12) (a) Miedaner, A.; DuBois, D. L.; Curtis, C. J.; Haltiwanger,
R. C. Organometallics 1993, 12, 299. (b) Curtis, C. J.; Miedaner, A.; Ellis,
W. W.; DuBois, D. L. J. Am. Chem. Soc. 2002, 124, 1918. (c) DuBois, M. R.;
DuBois, D. L. Chem. Soc. Rev. 2009, 38, 62.
Reduction of [(PPh3)2Re(CO)4]þ ([1-Ph2]þ). The chemistry
of[(PPh3)2Re(CO)4]þ ([1-Ph2]þ) with[HPt(dmpe)2]þ ([HPt]þ;
dmpe = 1,2-bis(dimethylphosphino)ethane; see Chart 1 for
the system used for abbreviations) in the absence and presence
of added external Lewis acids is summarized in Scheme 3.
(15) (a) Butts, S. B.; Holt, E. M.; Strauss, S. H.; Alcock, N. W.;
Stimson, R. E.; Shriver, D. F. J. Am. Chem. Soc. 1979, 101, 5864.
(b) Butts, S. B.; Strauss, S. H.; Holt, E. M.; Stimson, R. E.; Alcock, N. W.;
Shriver, D. F. J. Am. Chem. Soc. 1980, 102, 5093. (c) Labinger, J. A.; Miller,
J. S. J. Am. Chem. Soc. 1982, 104, 6856. (d) Grimmett, D. L.; Labinger,
J. A.; Bonfiglio, J. N.; Masuo, S. T.; Shearin, E.; Miller, J. S. J. Am. Chem.
Soc. 1982, 104, 6858. (e) Labinger, J. A.; Bonfiglio, J. N.; Grimmett, D. L.;
Masuo, S. T.; Shearin, E.; Miller, J. S. Organometallics 1983, 2, 733.
(16) Maitlis, P. M.; Zanotti, V. Chem. Commun. 2009, 1619.
(13) Crawford, E. J.; Bodnar, T. W.; Cutler, A. R. J. Am. Chem. Soc.
1986, 108, 6202.
(14) Yamamoto, H. Lewis Acids in Organic Synthesis; Wiley-VCH:
Weinheim, 2000.

Article
Organometallics, Vol. 29, No. 20, 2010 4501
Chart 1
Scheme 3
Scheme 4
A solution of [1-Ph2][BPh4] and [HPt][PF6] in C6D5Cl shows
no discernible reaction over 24 h, although it should be noted
that [1-Ph2][BPh4] is only slightly soluble in C6D5Cl.10a In
addition, the choice of anion can play an important role in
reductions using [HPt]þ: as previously reported, the equilib-
rium shown in Scheme 4 lies far toward the left side (un-
to the Re-H products dominates, preventing formation of
more than a very small amount of formyl complex.
reduced [1-E2]þ), but can be driven to the right by precipita-
Addition of isopropyl pinacol borate, a weak Lewis acid, has
no effect on the above reaction: only traces of the characteristic
1H NMR formyl signal for 2-Ph2 are observed, at an essentially
unchanged chemical shift (ReCHO, δ 15.21), which indicates
no interaction between the formyl oxygen and the boron center.
In contrast, a solution of [1-Ph2][BF4] with [HPt][PF6] and 2
10c
.
tion of [Pt][BF4]x[PF6]2-x
Suspensions of [1-Ph2][BF4]
(also fairly insoluble) and [HPt][PF6] in C6D5Cl or THF
slowly convert to a mixture of (PPh3)2Re(CO)3(CHO) (2-
Ph2; 1H NMR ReCHO, δ 15.22 in C6D5Cl), (PPh3)2Re-
(CO)3(H) (4-Ph2), and(PPh3)Re(CO)4(H) (4-Ph1).17 The half-
life for consumption of [HPt]þ is 2-4 days, and the yield of
2-Ph2 never exceeds ∼5% (by NMR, relative to [HPt]þ).
Therefore, while the unassisted transfer of a hydride from
[HPt]þ to [1-Ph2]þ can be observed, it appears to be thermo-
dynamically unfavorable, unless driven by precipitation of
the [Pt]2þ salt, and is kinetically so slow that decarbonylation
t
equiv of Bu(CH2)2B(C8H14) in C6D5Cl shows clean conver-
sion to a formyl species in a couple of hours.18 The Re-CHO
(18) When [1-Ph2][BArF4] was treated with [HPt][PF6] in the pres-
ence of 2 or 10 equiv of BEt3 (1.0 M in hexanes), only ∼1% boroxy-
carbene was detected, and no further reaction was observed over ∼24 h,
consistent with inhibition of reduction by small amounts of [Pt][BArF
]
4 2
(17) See Supporting Information for full details.
in solution.

4502 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
bind to (PPh3)2Re(CO)3(CHO) overTHF(Keq =0.19M-1 10c
by roughly 3 orders of magnitude.
)
The BEt3-stabilizedformyl2-Ph2 BEt3 issomewhat longer-
3
lived than 2-Ph2, but still decomposes over the course
of a few days to Re-H species 4-Ph2 and 4-Ph1 (Scheme 3;
decomposition times tend to vary widely, as has been
observed for other formyls21). No further reduction was
observed when 2-Ph2 BEt3 was treated with additional
3
NaHBEt3 or [HPt][PF6], even in the presence of excess
trialkylborane.
The stronger Lewis acid B(C6F5)3 does not promote
C-H bond formation from [1-Ph2][BF4] and [HPt][PF6].
Instead, hydride is transferred from [HPt]þ to B, forming the
stable salt [1-Ph2][HB(C6F5)3] along with precipitated [Pt]2þ
.
B(C6F5)3 forms a stable boroxycarbene when added to
preformed (PPh3)2Re(CO)3(CHO),10b but gentle heating of
2-Ph2 B(C6F5)3 provides the same tetracarbonyl boro-
3
hydride salt [1-Ph2][HB(C6F5)3],22 suggesting that [HB-
(C6F5)3]- is a weaker hydride donor than 2-Ph2. Similar to 2-
Ph2 BEt3, 2-Ph2 B(C6F5)3 shows no further reaction with
3
3
hydride sources.22
In summary, while small amounts of an unstable formyl
can be generated slowly from [1-Ph2]þ and [HPt]þ alone,
high yields of a boroxycarbene (stabilized formyl) species
were generated rapidly from [1-Ph2]þ and [HPt]þ in the
presence of the appropriate external Lewis acid. Trialkyl-
borate Lewis acids are too weak to have an effect; the strong
Lewis acid B(C6F5)3 diverts the hydride transfer to make a
stable borohydride; intermediate acid strength trialkyl-
boranes greatly accelerate the first hydride transfer reaction
and increase the lifetime of the reduced product. However,
none of the external Lewis acids promote the further reduction
or C-C coupling chemistry achieved by the pendent Lewis
acid in [1-E2]þ.
Synthesis of Complexes with Pendent Boranes. Given the
importance of intramolecular interactions, a structure-
function study on pendent Lewis acid assistance was de-
signed, requiring complexes in which the number of phos-
phinoborane ligands and the length of the hydrocarbon
chain connecting P and B are varied. Complexes with one
or two phosphinoborane ligands with (CH2)1-3 linkers were
synthesized, along with a mixed complex containing one
phosphinoborane and one simple tertiary phosphine ligand.
A system that identifies the phosphine ligand(s) and the class
of complex will be used, as depicted in Chart 1. The ligands
will be identified by the specific hydrocarbon linker (methy-
lene, 1,2-ethanediyl, 1,3-propanediyl): M = Ph2PCH2B-
(C8H14), E = Ph2P(CH2)2B(C8H14), P = Ph2P(CH2)3B-
(C8H14), and Ph = PPh3. The cationic rhenium carbonyl
fragment is designated [1]þ, with the first reduction Re-
CHO products designated 2, and the doubly reduced C-C
coupled products designated [3]-. The number of phosphine
ligands dictates the number of carbonyl ligands on the metal
complexes: for [1-M1]þ, [1]þ = [Re(CO)5]þ; for [1-M2]þ,
[1]þ = [Re(CO)4]þ. Full synthetic details for all new species
are provided in the Supporting Information.
Figure 1. Structural representation of (2-Ph2 BEt3) (C6H5CH3)
3
3
with thermal ellipsoids at 50% probability. Only one of the two
independent molecules in the asymmetric unit is portrayed; the
bond lengths and angles are similar. H atoms (except on the
carbene) and two toluene solvent molecules of crystallization are
˚
omitted for clarity. Selected bond lengths (A) and angles (deg):
Re2-C4B 2.126(3), Re-CO(av) 1.985, C4B-O4B 1.252(3),
O4B-B1B 1.638(3), Re2-C4B-O4B 125.2(2), C4B-O4B-
B1B 129.2(2).
resonance (δ 14.60) is shifted well upfield from that observed
for the Lewis acid-free reaction, and the 31P NMR resonance
also differs (δ 14.5, vs δ 15.8 in the absence of borane),
indicating significant interaction between the borane and the
formyl ligand. Treatment of [1-Ph2]þ with NaHBEt3 gave a
spectroscopically similar product, which was structurally char-
acterized by X-ray diffraction (XRD) as the boroxycarbene
(PPh3)2Re(CO)3(CHOBEt3) (2-Ph2 BEt3) (Figure 1). In com-
3
parison to analogous adducts of the stronger Lewis acids BF3
and B(C6F5)3,10b the boroxycarbene moiety in 2-Ph2 BEt3 has
3
longer B-O and Re-C bonds and a shorter C-O bond,
consistent with less carbene character19 and a weaker B-O
interaction. Accordingly, the BEt3 can be removed under
vacuum to afford 2-Ph2.
On titration of (PPh3)2Re(CO)3(CHO) with a solution of
1
tBu(CH2)2B(C8H14) in THF-d8, the formyl H NMR reso-
nance showed a steady upfield shift, with no change in line
shape, consistent with fast, reversible adduct formation. The
equilibrium constant for adduct formation was estimated
from a Benesi-Hildebrand plot20 (Figure S11) as Keq = 100
M
-1, corresponding to a free energy of B-O bond formation
t
of 11.4 kcal/mol.17 Notably, Bu(CH2)2B(C8H14) prefers to
(19) Such species can be regarded either as borane-stabilized formyls
or as boroxycarbenes. There is a continuum, and M-CHO-BR3 species
are probably best considered intermediate between formyl and boroxy-
carbene. We choose to describe these as boroxycarbenes, due to the lack
of observed CdO stretch in the IR and the relatively downfield 13C
chemical shift of Re-CHO. Solid-state structures appear to be inter-
mediate between carbene and formyl.
(20) (a) Benesi, H. A.; Hildebrand, J. H. J. Am. Chem. Soc. 1949, 71,
2703. (b) Rose, N. J.; Drago, R. S. J. Am. Chem. Soc. 1959, 81, 6138.
(c) Baldwin, S. M.; Bercaw, J. E.; Brintzinger, H. H. J. Am. Chem. Soc. 2008,
130, 17423.
Two routes are available for bis(phosphinoborane) com-
plexes: prior generation of the entire ligand followed by
(21) (a) Narayanan, B. A.; Amatore, C. A.; Kochi, J. K. Organo-
metallics 1984, 3, 802. (b) Sumner, C. E.; Nelson, G. O. J. Am. Chem. Soc.
1984, 106, 432. (c) Narayanan, B. A.; Amatore, C.; Kochi, J. K. Organo-
metallics 1987, 6, 129.
(22) Elowe, P. R. Ph.D. Thesis, California Institute of Technology,
Pasadena, CA, 2009.

Article
Organometallics, Vol. 29, No. 20, 2010 4503
Scheme 5
coordination, or hydroboration of a precoordinated alkenyl-
phosphine (Scheme 5). The ethylene-linked cation [1-E2]-
[BF4] was previously reported;10a the Mn analogue [1-E2-
Mn][BF4] was obtained in good yield by a similar procedure,
as was the complex with a propylene-linked phosphinoborane
ligand ([1-P2][BF4]). Hydroboration of the allylphosphine
with 9-BBN (9-borabicyclo[3.3.1]nonane) proceeded signifi-
cantly more readily than the vinyl analogue, possibly reflecting
reduced conjugation of the double bond with the electron-
deficient metal center.
The ligand with a methylene spacer is inaccessible by
hydroboration; the reaction of Ph2PCH2Li and ClB(C8H14)
in cold pentane affords an insoluble white powder from
which Ph2PCH2B(C8H14) is obtained.23 Treatment of Re-
(CO)5Br with 2 equiv of Ph2PCH2B(C8H14) produced trans,
mer-(Ph2PCH2B(C8H14))2Re(CO)3Br, from which halide
abstraction could not be achieved with Agþ (which reacts
with trialkylboranes24) or Tlþ salts; but addition of
[Et3Si][B(C6F5)4]25 to a C6H5Cl solution under 1 atm of
CO gave the desired cation, trans-[(Ph2PCH2B(C8H14))2Re-
(CO)4][B(C6F5)4] ([1-M2][B(C6F5)4]), whose structure was
confirmed by XRD (Figure 2). As in [1-E2][BF4], no intra-
Figure 2. Structural representation of [1-M2][B(C6F5)4], with
ellipsoids shown at 50% probability. One of the B(C8H14) groups
was disordered by a rotation around the B-C bond; only the
major component (66%) is shown (represented as light blue
isotropic atoms).17 All H atoms and a disordered dichloro-
ethane solvent molecule of crystallization are omitted for clarity.
molecular CO B interaction is observed (although there
3 3 3
are close contacts between the [B(C6F5)4] anion and nearby
˚
˚
Selected bond lengths (A) and angles (deg): Re-P1 2.4521(5),
Re-P2 2.4509(5), Re-C(av) = 1.99, P1-Re-P2 172.34(2).
˚
CO’s: O3-F19 2.867 A, O4-F19 3.042 A).
Reaction of [(PPh3)Re(CO)5]OTf with Ph2P(CH2)2B-
(C8H14) seemed an obvious route to an asymmetric cation
with only one Lewis acid, but under both thermal and
photolytic conditions only intractable mixtures were ob-
(23) This insoluble material is likely an oligomer of the desired
phosphinoborane. The crude material was treated with pyridine, which
led to dissolution of monomeric adduct Ph2PCH2B(C8H14) pyridine,
3
which was fully characterized. Treatment with BF3 Et2O prompted
3
26
reprecipitation of Ph2PCH2B(C8H14) as an analytically pure white
powder. See Supporting Information for full details.
tained. Instead, the reaction of trans-Re(PPh3)(I)(CO)4
with AgOTf followed by Ph2P(CH2)2B(C8H14) under gentle
heating gives trans-[(Ph2P(CH2)2B(C8H14)(PPh3)Re(CO)4]-
[OTf] (1-E1Ph1) in good yield (Scheme 6).
ꢀ
(24) (a) Brown, H. C.; Hebert, N. C.; Snyder, C. H. J. Am. Chem. Soc.
1961, 83, 1001. (b) Brown, H. C.; Snyder, C. H. J. Am. Chem. Soc. 1961, 83,
1002. (c) Brown, H. C.; Verbrugge, C.; Snyder, C. H. J. Am. Chem. Soc.
1961, 83, 1001.
(25) Lambert, J. B.; Zhang, S.; Ciro, S. M. Organometallics 1994,
13, 2430.
(26) Ingham, W. L.; Coville, N. J. Organometallics 1992, 11, 2551.

4504 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Scheme 6
Scheme 7
Figure 3. Structural representation of [1-M1][OTf], with ellip-
soids at 50% probability. A highly disordered solvent molecule
was omitted for clarity. The two phenyl groups were disordered
by rotation about the P-C bond, and only the major con-
tributor is included for clarity.17 Selected bond lengths (A)
˚
and angles (deg): Re-C1 2.027(2), Re-C2 2.000(2), Re-C3
2.016(2), Re-C4 2.021(2), Re-C5 1.973(2), Re-P 2.4846(5),
P-Re-C5 176.54(6), C1-Re-C3 176.96(8).
(by NMR), with concomitant precipitation of [Pt][BF4]x-
[PF6]2-x. At room temperature the 1H NMR spectrum of the
product exhibits a diagnostic triplet at δ 13.89 (JPH = 5.9
Hz) but is otherwise broad and nondescript, while the
31P{1H} NMR spectrum shows two very broad resonances.
(Similar reactivity and spectroscopy was observed in THF-d8.)
Variable-temperature NMR studies show fluxional behav-
ior; at low temperatures both the 1H and 31P NMR spectra
display sharp doublets indicating nonequivalent P and CH2
groups. These results strongly indicate the boroxycarbene
structure 2-M2 (Scheme 8), although we were unable to
obtain an analytically pure sample. (See SI for details of
the NMR and assignment; a related intramolecular boroxy-
carbene has been obtained for a CpFe system.28)
The fluxional process in 2-M2 interconverts the two
phosphinoborane ligands (ΔGq = 13.6 ( 0.2 kcal/mol, esti-
mated from the coalescence temperature of 323 K). A
sequence of B-O bond breaking, rotation, and B-O bond
formation with the other ligand seems a plausible mecha-
nism. 13C{1H} NMR spectroscopy is not entirely consistent
with this, however, as the CO carbon signals are very broad
at room temperature, suggesting their involvement in the
exchange process. Transfer of a hydride from a formyl to an
adjacent carbonyl ligand (Scheme 8, bottom) has been ob-
served previously (ΔGq = 11.7 kcal mol-1);29 that may be
more consistent with our observations. Furthermore, treat-
ment of 2-M2 with up to 100 equiv of pyridine leads to no
B-O cleavage; only coordination of pyridine to the free B
center is observed (Scheme 8, top), resulting in sharp NMR
resonances expected for a nonfluxional product.17 If rever-
sible B-O cleavage were occurring rapidly on the NMR time
scale, pyridine binding might be expected to intervene. This
observation stands in stark contrast to B-O bonds involving
intermolecular Lewis acids as well as phosphinoboranes with
Complexes with a single phosphinoborane ligand were
synthesized starting with Re(CO)5OTf (Scheme 7). Both
Ph2PCH2B(C8H14) and Ph2P(CH2)2B(C8H14)27 displace
triflate from Re(CO)5OTf at 60 °C to give [(Ph2PCH2B-
(C8H14))Re(CO)5][OTf] ([1-M1][OTf]), which was character-
ized crystallographically (Figure 3), and [(Ph2P(CH2)2B-
(C8H14))Re(CO)5][OTf] ([1-E1][OTf]), respectively. The
propylene-linked analogue was obtained by reaction
of Ph2PCH2CHCH2 with Re(CO)5OTf at 60 °C, afford-
ing [(Ph2PCH2CHCH2)Re(CO)5][OTf]; hydroboration
with 9-BBN gave [(Ph2P(CH2)3B(C8H14))Re(CO)5][OTf]
([1-P1][OTf]).
Reduction of [(Ph2PCH2B(C8H14))2Re(CO)4]þ ([1-M2]þ).
[1-M2]þ was readily synthesized only as the [B(C6F5)4]- salt,
which complicated reductions with [HPt]þ: reduction with
[HPt][PF6] proceeded very slowly to ∼90% conversion (by
NMR) over 6 days. Partially soluble [Pt][B(C6F5)4]2, visible
by NMR in small amounts throughout the reaction, may
inhibit reduction according to the equilibrium in Scheme 4
(precipitation of [Pt]2þ cannot drive the reaction). To cir-
cumvent this problem, [1-M2][B(C6F5)4] was treated with
both [N(heptyl)4][BF4] (or NaBF4) and [HPt][PF6] in
C6D5Cl, which led to complete consumption of [1-M2]þ
and formation of a new Re species over ∼3 h in >95% yield
(28) Kubo, K.; Nakazawa, H.; Nakahara, S.; Yoshino, K.; Mizuta,
T.; Miyoshi, K. Organometallics 2000, 19, 4932.
(29) Luan, L.; Brookhart, M.; Templeton, J. L. Organometallics
1992, 11, 1433.
(27) Fischbach, A.; Bazinet, P. R.; Waterman, R.; Tilley, T. D.
Organometallics 2008, 27, 1135.

Article
Organometallics, Vol. 29, No. 20, 2010 4505
Scheme 8
Scheme 9
longer linkers, which are readily broken in the presence of
pyridine (vide infra). Similarly, carbene 2-M2 is remarkably
stable in comparison to 2-Ph2 and 2-Ph2 BEt3: no decom-
3
position is observed over short periods at 95 °C or weeks at
room temperature.
Treatment of [1-M2][B(C6F5)4] with 2 equiv each of
[N(heptyl)4][BF4] and [HPt][PF6] in THF-d8 led to formation
of a single new Re-containing product (∼70% yield by
NMR, relative to [N(heptyl)4]þ), again with precipitation
of [Pt]2þ. Because the presence of soluble salts such as
[N(heptyl)4][B(C6F5)4] prevented isolation and structural
determination by XRD, the product was characterized spec-
troscopically. Two sharp doublets are found in the 31P{1H}
NMR spectrum at δ 22.2 and 37.3 (JPP=127 Hz), and two
slightly broadened doublets are also found in the 1H NMR
spectrum, δ 3.12 and 4.63 (JHH = 13.0 Hz). The downfield
31P NMR shift is consistent with a five-membered ring,30 as
expected for a structure analogous to the C-C coupled
product [3-E2]- (Scheme 2),10a but a number of other spectro-
scopic features rule out this assignment, including the absence
of a downfield 13C NMR signal for a carbene carbon and the
mer-tricarbonyl geometry assigned by 13C NMR and IR.
A simple alkyl is also ruled out by the five-membered-ring
structure, no observed P coupling to the CH2 group, and no
three-coordinate 11B NMR signals. All the spectroscopic
data are consistent with the doubly reduced “confused” alkyl
species [5]- (Scheme 9), in which the CH2 is bound to one B,
and O is bound to Re and the other B, forming a bicyclic ring
system. The very broad C signal and slightly broad H signals
of the CH2 group point to coordination to quadrupolar B,
the B-O-Re fragment would provide the expected down-
field 31P NMR shift due to a five-membered ring, and two
11B NMR signals consistent with four-coordinate BR4- and
BR3(OR0)- are observed. In addition, high-resolution mass
spectrometry (HRMS) gives the expected mass for [5]-. Thus
although a second reduction was facilitated by the pendent
Lewis acid, an entirely different product was obtained than
in the complex supported by ethylene-linked ligands: no
C-C bond formation occurred, and an unusual isomeriza-
tion was instead observed, presumably driven by formation
of four-coordinate B in favorable ring sizes. The pendent
acid is clearly required for this transformation: whereas 2-M2
(preformed in situ) reacted with [HPt][PF6] and [N(heptyl)4]-
[BF4] to give [5]-, similar treatment of 2-M2-py produced no
observable reaction.
Reduction of [(Ph2P(CH2)2B(C8H14))2Re(CO)4]þ ([1-E2]þ).
As previously reported,10a treatment of [1-E2][BF4] in C6D5Cl
with 1 equiv of NaHBEt3 or [HPt][BF4] affords boroxy-
carbene 2-E2 (Scheme 2). The carbene is identified by a char-
acteristic downfield signal in the 1H NMR spectrum (δ 13.96)
and a closely separated pair of doublets in the 31P{1H} NMR
spectrum (δ 9.9). Boroxycarbene 2-E2 spontaneously crystal-
lizes from CD2Cl2, and XRD revealed a binuclear structure
in which the pendent borane from one Re center acts as an
intermolecular Lewis acid for the other; the two carbene
10a
˚
carbons are separated by ∼5 A.
Pulsed gradient spin-
echo (PGSE) diffusion NMR experiments31 on 2-E2, using
(PPh3)2Re(CO)3Br (6) as an internal standard of similar size
and shape to the proposed monomer, suggest that signif-
icant intermolecular interactions are maintained in solution.
Initial measurements provided a ratio of hydrodynamic
volumes of 2-E2:6; then excess pyridine was added to the
NMR tube, and the experiment was repeated. If 2-E2 were
strictly monomeric in solution, the ratio would be expected
to increase, since pyridine binding would increase the hydro-
dynamic volume; if 2-E2 were strictly dimeric, the second
ratio should be close to 0.5 times the first, as the monomer
should have roughly half the hydrodynamic volume of the
dimer. The values of the ratio were 2.34(7) before pyridine
addition and 1.6(1) after; the decrease by a factor of 0.69(5)
could reflect a monomer-dimer equilibrium, giving a
(31) Li, D. Y.; Keresztes, I.; Hopson, R.; Williard, P. G. Acc. Chem.
Res. 2009, 42, 270.
(30) Garrou, P. E. Chem. Rev. 1981, 81, 229.

4506 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Scheme 10
Figure 4. Structural representation of 8 (C6H6)2.5(C5H12)0.5
,
3
with ellipsoids at 50% probability. H atoms and solvents of
crystallization are omitted, and Ph groups are truncated for
clarity. One site of solvent cocrystallization was a disordered
mixture of benzene and pentane.17 Selected bond distances (A)
˚
time-averaged PGSE value.32 The broad 1H and 31P NMR
resonances of 2-E2 at 298 K support such a conclusion,
although full variable-temperature NMR experiments could
not be carried out due to the thermal instability (vide infra).17
Perhaps owing to its fluxional solution behavior, 2-E2 is
quite reactive: C6D5Cl solutions of 2-E2 disproportionate
(at a rate dependent on the concentration of 2-E2) to a 1:1
mixture of [1-E2]þ and [3-E2]- (Scheme 10).
and angles (deg): Re1-CO(av) 1.988, Re1-P1 2.4437(6),
Re1-P2 2.4490(6), B1-O10 1.609(4), B2-O9 1.597(3), B1-
O10-B4 143.0(2). B2-O9-B3 141.7(2).
shows two doublets, δ 12.6 and 39.9 (JPP= 124.8 Hz); the
uncommonly downfield-shifted doublet is consistent with
1
a five-membered phosphorus-containing ring.30 In the H
NMR spectrum, the downfield carbene resonance has been
replaced by a new broad resonance at δ 5.85. These observa-
tions along with 2-D NMR experiments support structure 7
(Scheme 10), in which a B-C bond has been cleaved and the
ethylene linker has added to the carbene carbon, giving a
(boroxyalkyl)metal complex.17
The same reduction chemistry is operative in THF,
although the carbene is quite short-lived in this solvent: dis-
proportionation is complete after a few minutes at room tem-
perature, in stark contrast to the extremely stable methylene-
linked analogue 2-M2. This rate enhancement might be due
to the more polar solvent facilitating transformation to the
ionic products [1-E2]þ and [3-E2]-, the enhanced solubility
of NaBF4 in THF (with Naþ being incorporated into the
product), or the ability of THF to break intermolecular B-O
bonds in 2-E2. The 31P{1H} NMR signal for 2-E2 in THF
appears as a singlet (δ 10.3), probably a result of competitive
borane binding by THF and formyl, implying relatively weak
B-O bonds in 2-E2 when compared to 2-M2. The aforemen-
tioned PGSE experiment is consistent with this description,
as reaction of pyridine with carbene 2-E2 (in C6D5Cl) dis-
places the formyl group from the borane, forming an un-
Boroxycarbene 2-E2 acts as a strong hydride donor,
liberating H2 with reagents such as H2O and [HNMe3]-
[BPh4] (Scheme 10). In the latter case, the parent cationic
tetracarbonyl [1-E2]þ re-forms; with water the binuclear
B-O(H)-B-bridged complex 8, which contains a 20-membered
ring, was formed and characterized by XRD (Figure 4).
[tBu3PH][HB(C6F5)3], which is a potential hydride source as
well as a protic reagent (and which can be formed directly,
without any metal complex required, from H233), was tested
to see if further reduction and C-C coupling could be
obtained with such a weak hydride. Instead, H2 release was
observed again, and the ion pair [1-E2][HB(C6F5)3] was
obtained, in addition to free PtBu3. Thus [HB(C6F5)3]- is
too weak a hydride to reduce the rhenium carbonyl cation,10b
consistent with the stability of [1-Ph2][HB(C6F5)4] (Scheme 3)
and despite the presence of a pendent acid group.
1
stabilized formyl (31P{1H} singlet at δ 11.7, H Re-CHO
signal at δ 15.21) that decomposes without disproportionation
to give Re-H by decarbonylation (the normal pathway for
formyl decomposition).4a
While 2-E2 is stable for a number of hours at room
temperature in C6D5Cl, it decomposes as the temperature
is raised: up to 40 °C, disproportionation is the main path
followed, but heating at 90 °C for a few minutes converts
2-E2 to a single new product whose 31P{1H} NMR spectrum
Carbene 2-E2 can be directly reduced (before dispropor-
tionation occurs) to [3-E2]- with a second equivalent of
NaHBEt3 or [HPt][BF4].10a In contrast to 2-E2, [3-E2]- is
thermally stable up to 70 °C in THF solution, as well as
(32) (a) Macchioni, A.; Ciancaleoni, G.; Zuccaccia, C.; Zuccaccia, D.
Chem. Soc. Rev. 2008, 37, 479. (b) Cohen, Y.; Avram, L.; Frish, L. Angew.
Chem., Int. Ed. 2005, 44, 520.
(33) (a) Welch, G. C.; Stephan, D. W. J. Am. Chem. Soc. 2007, 129,
1880. (b) Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46.

Article
Organometallics, Vol. 29, No. 20, 2010 4507
Scheme 11
Scheme 12
toward further reduction with excess NaHBEt3, presumably
because a dianionic product would be disfavored. The B-O
bonds of [3-E2]- appear to be much stronger than those of
2-E2; they are not cleaved by addition of pyridine. Unlike
2-E2, treatment of [3-E2]- with [Pt][BArF4]2 did not lead to
any reaction, implying that its formation is essentially irre-
versible. [3-E2]- reacts with strong acids (HBF4, HOTf) to
give a single product before going on to multiple species, and
addition of MeOTf cleanly yields a stable neutral product,
but none of these have been reliably characterized.
Mechanism of C-H Bond Formation: Evidence for Hydride
Shuttling. Plausible pathways for the transfer of hydride
from [HPt]þ to Re-CO in [1-E2]þ include (Scheme 11) the
following: (path A) hydride transfer to the pendent borane,
followed by intramolecular hydride transfer from boron to
carbon; (path B) hydride transfer from [HPt]þ directly to the
Re-CO fragment, yielding a formyl, which is trapped by the
pendent borane; (path C) transient borane coordination to
a Re-CO oxygen, activating the carbon for nucleophilic
attack.
1
conspicuously lacking Pt satellites (δ -1.6). The H NMR
spectrum showed the hydride signal expected for [HPt]þ
(δ -12.02, pent, 2JPH = 29.3 Hz, 1JPtH = 694 Hz) along with
another broad doublet with Pt satellites at δ -3.25 (JPH
=
When the reaction was monitored by 31P{1H} NMR spectros-
copy at -40 °C, a new broad resonance was observed at
δ -6, near the sharper signal for [HPt]þ at δ -7. Both signals
disappear as the reaction goes on, although the broad
resonance is consumed first. The appearance of a broad
Pt-H signal suggests the possibility of a Pt-H-B-bridged
intermediate, which would be most consistent with pathway
A. In further support, reaction of [HPt][PF6] with BEt3 led to
166 Hz, JPtH = 958 Hz), downfield of most Pt hydrides, but
closer to the chemical shift expected for [HBEt3]- (δ ≈ -0.5)
or [Et3BHBEt3]- (δ ≈ -2.7).34 11B NMR at this temperature
showed signals for free BEt3 and another close to that of
[HBEt3]-.
These observations strongly indicate that H has been
transferred (partly or entirely) to B; we tentatively assign
the broad Pt-Hsignalto [(dmpe)2Pt-H-BEt3]þ or the borane
adduct [(dmpe)2Pt(HBEt3)(BEt3)]þ (Scheme 12), since the
large JPH would be consistent with a phosphine trans to the
hydride35 and the broadening with quadrupolar interaction
with boron. The added bulk of the coordinated BEt3 group
could favor a trigonal-bipyramidal geometry; phosphines in
such systems have been observed to undergo rapid inter-
change, leading to loss of Pt satellites in the 31P NMR.36
precipitation of [Pt][PF6]2 and the appearance of a broad 31
P
resonance at δ -2.4, along with a broadened resonance for
[HPt]þ at δ -7.1. Fluxionality was also apparent in a broad
1H NMR signal in the hydride region at δ -11.7 and a 11
B
NMR signal at 14.8 (compare Li[Et3BHBEt3], δ 8.734). At
-40 °C the signals sharpened and separated into two sets.
The 31P NMR spectrum showed two singlets, one corre-
1
sponding to [HPt]þ (δ -7.4, JPtP = 2240 Hz) and one
(35) Miedaner, A.; Raebiger, J. W.; Curtis, C. J.; Miller, S. M.;
DuBois, D. L. Organometallics 2004, 23, 2670.
(36) Lin, W. B.; Wilson, S. R.; Girolami, G. S. Inorg. Chem. 1997, 36,
2662.
(34) DuBois, D. L.; Blake, D. M.; Miedaner, A.; Curtis, C. J.;
DuBois, M. R.; Franz, J. A.; Linehan, J. C. Organometallics 2006, 25,
4414.

4508 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Precipitation of [Pt]2þ is consistent with the stoichiometry
required to form [Et3BHBEt3]-. The reaction of preformed
[Li][Et3BHBEt3] with [Pt]2þ afforded multiple products; the
major product was Pt(dmpe)2, but some of the trans hydride
was also observed. Similar chemistry has been previously
reported for HRh(dmpe)2.34
Scheme 13
While hydride transfer between B and Pt appears facile
when the [PF6] anion is employed, no reaction is observed
when [HPt][BArF4] is treated with BEt3. This difference is
consistent with our previous observation that [1-E2][BArF
]
4
is not reduced by [HPt][BArF4] in C6D5Cl10c and suggests
that an equilibrium involving hydride transfer from Pt to B
could be driving the chemistry of Scheme 4. Furthermore,
we recently reported that the combination of a suitably
strong and bulky base and H2 is sufficient to directly reduce
[1-E2]þ;10c that reduction must go through a pendent boro-
hydride intermediate. All of this evidence supports the
“hydride shuttle” mechanism of path A (Scheme 11), where-
in formation of a small amount of trialkylborohydride is
followed by rapid intramolecular carbonyl reduction. Note
also that [1-Ph2]þ does react with [HPt]þ to give an (unstable)
formyl (vide supra), but that reaction is very slow and is
strongly accelerated by the addition of BEt3, suggesting that
direct transfer of hydride to carbonyl carbon as in path B is
possible but much slower than the shuttle route.
Mechanism of C-C Bond Formation: Evidence for a Bis-
(carbene) Intermediate. Reaction of [1-E2][BF4] with 2 equiv
of NaHBEt3 or [HPt][PF6], or of preformed 2-E2 with
1 equiv of either reductant, affords [3-E2]- rapidly at room
temperature. The reaction is irreversible and, in contrast to
the first hydride transfer, is not driven by precipitation
([3-E2]- does precipitate from C6D5Cl, but remains in solu-
tion in THF-d8), as evidenced by the lack of observed reac-
tion when [3-E2]- is treated with [Pt][BArF4]2. The second
reduction and subsequent C-C coupling necessarily in-
volves a number of elementary steps. The mononuclear
form of 2-E2 (accessible according to the PGSE NMR data,
vide supra), containing intramolecular borane coordination,
is proposed to be an intermediate in the transformation to
mononuclear product [3-E2]-. Two reasonable pathways for
the formation of [3-E2]- are shown in Scheme 13: path A,
reduction at the carbene center to give an anionic boroxyalkyl
species, which undergoes Lewis acid-assisted migratory
insertion; or path B, reduction at a second CO to give an
anionic bis(carbene) (or doubly stabilized bis(formyl):
[CpRe(NO)(CHO)2]- has been previously reported5c,d) spe-
cies, followed by carbene coupling and hydrogen shift to give
the observed product. Related hydrogen shifts have been
previously observed in Fe37 and Mn10b systems.
Evidence for pathway B is provided by low-temperature
NMR studies. When 1 equiv of NaHBEt3 is added to an
NMR tube containing a frozen solution of 2-E2 in C6D5Cl
and the solution is warmed to -40 °C, clean conversion to a
new symmetric species is observed. This species features a
sharp singlet at δ 10.0 in the 31P{1H} NMR spectrum, a new
downfield singlet at δ 16.43 in the 1H NMR, which integrates
as two protons, and only one CO resonance and one boroxy-
carbene resonance in the 13C{1H} NMR, all consistent with
the bis(carbene) structure [9]-. Upon slowly warming the
sample, no new species were observed by NMR spectros-
copy; instead, the signal-to-noise decreased until essentially
no signals attributable to Re species were visible, and a large
amount of white precipitate was found upon removal of the
tube from the probe. The precipitate was isolated and
dissolved in THF-d8, and NMR showed it to be [3-E2]-.
It should be noted that the observation of [9]- in C6D5Cl
does not necessarily rule out a migratory insertion mech-
anism, since bis(carbene) formation could be reversible.
Migratory insertion of a carbonyl into a Re-C bond has
been reported to proceed with assistance from a (much
stronger) Lewis acid.38 The reduction of [1-M1]þ (vide infra)
appears unambiguously to proceed via migratory insertion
as well, but that reaction is quite slow.
Reduction of [(Ph2P(CH2)2B(C8H14))2Mn(CO)4]þ ([1-E2-
Mn]þ). Treatment of [1-E2-Mn][BF4] with 1 equiv of NaH-
BEt3 in C6D5Cl showed significant conversion to a Mn-
CHO species, with broad singlets at δ 13.81 (1H NMR) and δ
61.0 (31P NMR). When [1-E2-Mn][BF4] was similarly treated
with 1 equiv of NaHBEt3 in THF-d8, a mixture of products
formed, showing no downfield 1H NMR signal;39 however,
addition of 2 equiv of NaHBEt3 in THF-d8 led to rapid
and clean conversion to a single species that exhibits NMR
signals—in particular, two doublets each in the 1H and 31P-
{1H} NMR spectra and correlation between the two proton
doublets (δ 4.32 and 4.82) and a carbene carbon signal (δ 332)
by 1H-13C gHMBC spectroscopy—very similar to those of
(38) Lindner, E.; von Au, G. Angew. Chem., Int. Ed. 1980, 19, 824.
(39) The reactivity in THF is consistent with the initially formed
carbene undergoing fast disproportionation chemistry, just as the Re
analogue.
(37) Crawford, E. J.; Lambert, C.; Menard, K. P.; Cutler, A. R.
J. Am. Chem. Soc. 1985, 107, 3130.

Article
Organometallics, Vol. 29, No. 20, 2010 4509
Scheme 14
consistent with the shift of the formyl proton40 and the
broadness of the 31P resonance (attributed to moderately
fast B-O bond breaking/forming). Complex [2-P2]n precip-
itates as an oligomer; addition of pyridine redissolves it
by coordinating to boron and breaking the B-O bonds,
generating complex 2-P2 2py. Because the “bare” formyl
3
2-P2 2py is not stabilized by any Lewis acid interaction,
3
it decomposes over a period of hours, resulting in two new
Re-H 1H NMR signals, one a triplet and the other a
doublet, which can be assigned to 4-P2 2py and 4-P1 py
3
3
(Scheme 15), respectively. These products were not isolated;
the pyridine adduct of the free ligand (Ph2P(CH2)3B-
(C8H14) pyridine) was also observed by 31P{1H} NMR
spectroscopy.
3
Treatment of [1-P2][BF4] with 2 equiv of NaHBEt3 led to
NMR spectra that indicated the formation of multiple asym-
metric species, possibly including one or more with new C-C
bonds, but they could not be separated or fully characterized.
Treatment of [1-P2][BF4] with 2 equiv of [HPt][PF6] gave only
boroxycarbene [2-P2]n and unreacted [HPt][PF6]. This failure
to undergo a second hydride addition may support our
hypothesis that intramolecular facilitation is required for
multiple reductions by [HPt]þ: an intramolecular B-O inter-
action in 2-P2 would require an eight-membered ring, which is
probably too large for effective stabilization.
Overview of (Ph2P(CH2)nB(C8H14))2Re Complexes: Ring
Size Effects. Only methylene-linked boroxycarbene 2-M2
clearly exhibits intramolecular interaction of the pendent
borane with the [Re-CHO] moiety, forming a favorable
six-membered ring. The corresponding seven- and eight-
membered rings that would result from intramolecular co-
ordination in the ethylene- and propylene-linked analogues
are not observed;41 instead intermolecular coordination leads
to a dimer or an insoluble oligomer, respectively. In the
absence of significant intramolecular interactions in the
ethylene- and propylene-linked systems, the two boroxy-
carbenes are expected to exhibit similar stability and hydride
strength. Indeed, mixing preformed 2-E2 with [1-P2]þ (or
2-P2 with [1-E2]þ) gives an equilibrium mixture (Scheme 16),
with Keq ≈ 1. Similarly, addition of carbene 2-E2 to simple
phosphine complex [1-Ph2]þ, or of simple formyl complex
2-Ph2 to B-linked carbonyl complex [1-E2]þ, establishes an
equilibrium, also with Keq ≈ 1.17
the Re C-C coupling product [3-E2]-. Accordingly we assign
the two products as 2-E2-Mn and [3-E2-Mn]- (Scheme 14).
2-E2-Mn is presumed to be dimeric like the Re analogue,
although we have no direct evidence for that.17 The JPP value
of 27 Hz in [3-E2-Mn]- is much smaller than that in [3-E2]-,
suggesting the former has cis phosphines (one possible
isomer is drawn in Scheme 14). Both 2-E2-Mn and [3-E2-
Mn]- appear to be considerably less stable than their Re
analogues; attempts to isolate [3-E2-Mn]- by concentration
led only to decomposition.
Treatment of [1-E2-Mn][BF4] with 2 equiv of [HPt][PF6] in
THF-d8 also led to formation of [3-E2-Mn]-, although the
reaction did not go to completion. Manganese, an inex-
pensive first-row alternative to rhenium, is thus capable of
effecting the same reductive coupling chemistry, albeit with
decreased product stability.
Reduction of [(Ph2P(CH2)3B(C8H14))2Re(CO)4]þ ([1-P2]þ).
Treatment of [1-P2][BF4] with 1 equiv of NaHBEt3 in C6D5Cl
appears to afford a single product in good yield by NMR
spectroscopy. The 1H NMR spectrum includes a singlet at δ
14.10, characteristic of a formyl or boroxycarbene proton.
The 31P{1H} NMR spectrum shows a broad singlet (δ 6.6), in
contrast to the AB pattern of 2-E2 and the widely spaced
broad resonances in 2-M2. The same species is obtained using
[HPt]þ, with essentially identical NMR signals, indicating
that free BEt3 does not play a role in the broadening of the
31P signal.
A white solid precipitates from this solution over a few
hours; it is insoluble in common organic solvents and water,
but can be redissolved by addition of excess pyridine. The
resulting solution exhibits a [Re-CHO] resonance, shifted
downfield by ∼1 ppm (δ 14.99) from the original position,
while the 31P{1H} NMR resonance is slightly shifted (δ 6.2)
and significantly sharpened; the 11B{1H} NMR shows a
single resonance at δ 2.2, consistent with nonfluxional four-
coordinate boron.
In contrast, mixing preformed 2-E2 with the methylene-
linked [1-M2]þ results in complete hydride transfer, giving
only the intramolecularly stabilized carbene 2-M2, which
does not react with [1-E2][BF4]. Furthermore, 2-M2 is able to
accept a second hydride from 2-E2, affording the “confused”
alkyl [5]- (Scheme 9) as the major product. Clearly the
favorability of intramolecular coordination with the methy-
lene linker effects a substantial difference in stability and
reactivity. The fact that ethylene-linked boroxycarbene 2-E2
can undergo a second hydride addition and C-C coupling,
which in general does not appear to be well facilitated by
external Lewis acids, is further evidence that 2-E2 exists in
solution as an equilibrium between the dimer and an intra-
molecularly coordinated monomer.
(40) As the linker length is increased, the chemical shift of the formyl
in C6D5Cl moves downfield, which may indicate less B-O interaction. A
similar trend is observed upon titration of (PPh3)2Re(CO)3(CHO) with
trialkylboranes, as the formyl shifts upfield with added borane.
(41) The large size of Re and P in these rings is expected to favor
smaller ring sizes than organic counterparts, which suggests that seven-
and eight-membered rings are more destablized than organic systems.
On the basis of these results structure [2-P2]n is assigned
to the initially formed soluble species (Scheme 15). This
species presumably contains only weak B-O interactions,

4510 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Scheme 15
reaction of [HPt]þ with the solvent gives some CD2HCl.17 In
both solvents [Pt][OTf]x[PF6]2-x precipitates, but that is not a
Scheme 16
requisite driving force: treatment of [1-M1][BArF ] with
4
[HPt][BArF ] (which forms soluble [Pt][BArF ] ) led to com-
4
4 2
plete reduction as well (albeit more slowly, as with the
[1-Ph1]þ salts above; also see below). 2-M1 exhibits the char-
acteristic downfield 1H NMR signal (δ 14.07, d, 3JPH = 2.5 Hz
in C6D5Cl); a broad 13C{1H} resonance at δ 283 is attributed to
the carbene carbon. In contrast to most of the other boroxy-
carbenes studied, which show either broad or no 11B resonances,
2-M1 displays a relatively sharp 11B{1H} signal at δ 11.2, con-
sistent with four-coordinate boron.45
Reduction of [(PPh3)Re(CO)5]þ ([1-Ph1]þ). In general
we would expect monophosphine rhenium pentacarbonyl
cations to be more electrophilic than their bis(phosphine)
counterparts, rendering them better hydride acceptors, and
indeed, incontrast to[1-Ph2]þ, [(PPh3)Re(CO)5]þ ([1-Ph1]þ) is
readily reduced by [HPt]þ in the absence of any Lewis acid
assistance.10b Furthermore, precipitation of a [Pt]2þ salt is not
required to drive the reaction, although reduction of [1-Ph1]-
[OTf] with [HPt][PF6] (which results in precipitation of
Reaction of 2-M1 with pyridine led after ∼30 min to
complete disappearance of the carbene resonance and
growth of a new doublet at δ 4.91 (JPH = 5.8 Hz, 2H) in
1
the H NMR spectrum; two equal-intensity singlets were
observed in the 31P{1H} NMR spectrum. These spectro-
scopic characteristics are quite similar to those of the product
of [HPt]þ reduction of the ethylene-linked analogue 1-E1
(vide infra), as well as of the minor side product that
accompanied formation of 2-M1; accordingly we assign it
[Pt][OTf]x[PF6]2-x) is more rapid than that of [1-Ph1][BArF
]
4
with [HPt][BArF4] (in which case [Pt][BArF
] remains in
4 2
solution).42 The resultant monophosphine formyl species
(PPh3)Re(CO)4(CHO) (2-Ph1) is much less stable than
2-Ph2;43 it quickly (t1/2 ≈ 30 min) deinserts to give Re-H21
and cannot easily be further reduced (unless a strong electro-
phile such as MeOTf is added, to afford an electrophilic
methoxycarbene10b).
as the bridging alkyl-carbene dirhenium species 10-M1 py
3
(Scheme 18), formed via disproportionation. It is noteworthy
that the B-O bond of 2-M1 can be cleaved by pyridine, in
contrast to that in 2-M2 (Scheme 8); perhaps the more
electrophilic nature of the monophosphine complexes places
more of the negative charge at C instead of O.
Reduction of [(Ph2PCH2B(C8H14))Re(CO)5]þ ([1-M1]þ).
Reaction of [1-M1][OTf] with 1 equiv of [HPt][PF6] in
C6D5Cl gives boroxycarbene 2-M1 as the major (∼90%
isolated yield) product in ∼10 min at room temperature
(Scheme 17).44 The minor (∼10%) side product was tenta-
tively assigned as an alkyl species bridging two Re centers, a
hypothesis later solidified by spectroscopic analogy to related
reduced monophosphine-ligated species (vide infra). Similar
results were obtained in CD2Cl2, although a slow background
By itself, 2-M1 is quite stable in solution: there is no observ-
able decarbonylation over a month, indicating a particularly
strong B-O interaction that inhibits the normal formyl
deinsertion path. This stability is far greater than 2-Ph2,
2-E2, and 2-Ph1, in line with that of 2-M2. There is eventually
some decomposition after two months to a new species
(Scheme 19), identified by XRD as the alkoxy-boroxy-
carbene complex 11 (Figure 5), whose formation likely
involves B-C cleavage to extrude 9-BBN, coupling of the
methylene fragment to the formyl carbon, and addition of
the resulting alkoxy oxygen to a carbonyl carbon on another
Re center, although the exact mechanistic sequence is not
known. 11 exhibits an unusually upfield 31P{1H} NMR sig-
nal, δ -54.0, typical of P in a four-membered ring;30 the alk-
oxyboroxycarbene 13C{1H} resonance is observed at δ 210.2
(d, JPC = 9.4 Hz).
(42) One possibility for this kinetic discrepancy could be tighter ion
pairing for the smaller anions, which may aid in the transfer of hydride
between two cationic species.
(43) (a) Tam, W.; Lin, G.-Y.; Gladysz, J. A. Organometallics 1982, 1,
525. (b) Gibson, D. H.; Owens, K.; Mandal, S. K.; Sattich, W. E.; Franco, J. O.
Organometallics 1989, 8, 498.
(44) Stronger hydrides such as NaHBEt3 generally resulted in much
more hydride formation and were thus avoided. These reactions were
sometimes accompanied by intense color changes consistent with reduc-
tion processes, which would labilize the CO ligands and lead to reactivity
at Re itself.
(45) Eaton, G. R. J. Chem. Educ. 1969, 46, 547.

Article
Organometallics, Vol. 29, No. 20, 2010 4511
Scheme 17
Scheme 18
Scheme 19
Figure 5. Structural representation of 11 with ellipsoids at 50%
probability. Most H atoms are omitted for clarity. Selected bond
˚
distances (A) and angles (deg): Re1-C9 2.20(1) C9-O9 1.33(1),
C9-O10 1.25(1), O10-B1 1.59(2), O9-C32 1.49(1), Re2-C32
2.28(1), P2-Re2-C8 159.0(5), P1-Re1-C2 172.2(4).
Addition of a second equivalent of [HPt][PF6] to 2-M1
produces two new species over a period of hours. At early
stages the major species exhibits a 1H doublet at δ 4.54
Scheme 20
(3JPH = 2.6 Hz), which in H-13C and H-31P gHMBC
NMR experiments showed three-bond coupling (through
Re) to CO and P ligands; these and other multinuclear and
multidimensional NMR experiments support assignment
as the mononuclear Re alkyl complex 12.17 Over 48 h 12
is converted quantitatively (by NMR) to the second species,
assigned as Re-acyl [3-M1]- (Scheme 20), based on (inter
1
1
alia) the following NMR evidence: (1) an AB pattern (JHH
=
16 Hz) corresponding to geminal CH2 protons, further
coupled to an acyl carbon at 275 ppm (by 1H-13C-gHMBC,
which also shows that there is no longer any two- or three-
bond coupling between the CH2 group and CO ligands); (2)
a single 31P{1H} NMR resonance with a downfield shift, δ
38.1, consistent with a five-membered ring structure,30 which
is weakly coupled to one of the CH2 protons.
It is unclear whether the migratory insertion leading from
12 to [3-M1]- is Lewis acid-assisted (which would require
(46) (a) Mckinney, R. J.; Kaesz, H. D. J. Am. Chem. Soc. 1975, 97,
3066. (b) Casey, C. P.; Scheck, D. M. J. Am. Chem. Soc. 1980, 102, 2723.
(c) Casey, C. P.; Jones, J. J. Am. Chem. Soc. 1980, 102, 6154. (d) Warner,
K. E.; Norton, J. R. Organometallics 1985, 4, 2150. (e) Martin, B. D.;
Warner, K. E.; Norton, J. R. J. Am. Chem. Soc. 1986, 108, 33.
B-O cleavage). Migratory insertion on Re in the absence of
a Lewis acid is very rare and usually slow,38,46 but the

4512 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Scheme 21
Figure 6. Structural representation of [Na(THF)3][3-M1] with
ellipsoids at 50% probability. Most H atoms, and C atoms of
THF molecules, are omitted for clarity. Selected bond lengths
˚
(A) and angles (deg): Re-C1 1.935(5), Re-C2 1.849(5), Re-C3
1.951(5), Re-C5 2.130(4), Re-O4 2.206(3), O4-C4 1.440(4),
O4-B 1.575(5), C4-C5 1.532(6), C5-O5 1.218(5), O5-Na
2.238(3), O2#1-Na 2.411(3), Re-C5-O5 143.2(3).
Two pathways can be envisioned for disproportionation
of 2-E1 to 10-E1: (1) hydride donation from one Re-CHO
group (exposed by B-O cleavage) to a second RedCHOBR3
to give a Re-CH2O- fragment capable of nucleophilic
attack on the M-CO of re-formed [1-E1]þ or (2) nucleophilic
attack by the O of the Re-CHO group (again exposed by
B-O cleavage) on another carbene, followed by a hydride
shift. The first would require breaking a bond between
Re-CH2O- and B, which is expected to be quite strong
(recall the inability of pyridine to break the B-O bond in
[3-E2]-), perhaps favoring the second. The two alternatives
can be distinguished by reaction of 2-M1 with 2-P1, in which
the routes beginning with hydride transfer (left, Scheme 22)
and O attack (right, Scheme 22) would produce different pro-
ducts. Only one mixed-ligand dinuclear complex, 10-M1P1,
in which the alkyl group is bound to the Re ligated by the
propylene-linked ligand, is observed by NMR (>90%),17
supporting the preferred nucleophilic O attack mechanism.
The same mechanism is likely operating in the reaction of
methylene-linked 2-M1 with pyridine (vide supra), with the
Lewis base inducing B-O bond cleavage, exposing a formyl O
that attacks unperturbed 2-M1 to afford the related 10-M1.
Very similar mechanisms have been proposed for the formation
of related metalloesters (lacking borane stabilization).5d,47,48
Reduction of [(Ph2P(CH2)2B(C8H14))(Ph3P)Re(CO)4]þ
([1-E1Ph1]þ). When [1-E1Ph1][OTf] was treated with NaHBEt3
in C6D5Cl, a single species formed immediately, which exhib-
reaction hereis quite slow compared to the insertions observed
following reduction of 2-E2 and 2-E2-Mn. (The analogous
reaction in the 2-E1Ph1 system is also faster; vide infra.)
Reaction of 2-M1 with the stronger hydride NaHBEt3 led to
[3-M1]- much more rapidly. Acceleration could be due to
BEt3- or Naþ-assisted migratory insertion or product stabi-
lization, but separate experiments probing any additive
effect of BEt3 or NaBF4 (on [HPt]þ-mediated reductions)
were inconclusive.
Reduction of [1-M1][BArF4] using [HPt][BArF4] resulted
in a change in the rate-limiting step, as reduction of the
carbene was slower than migratory insertion. Therefore, no
alkyl intermediate was observed, and the reaction took about
1 week to reach completion.
The presence of the Naþ counterion in the synthesis of
[3-M1]- allowed for crystallization of single crystals of
[Na(THF)3][3-M1] from a cold THF/pentane mixture, with
an XRD study confirming the structure assigned by NMR
(Figure 6). A B-O bond is formed between the cyclic alk-
oxide and the borane in the secondary coordination sphere;
the bond distances are consistent with an acyl functionality,
as illustrated, withþa relatively short CdO distance of
˚
1.218(5) A. The Na counterion bridges an acyl O of one
molecule and a carbonyl O of another molecule, forming a
zig-zagging network in the crystal lattice.
Reduction of [(Ph2P(CH2)2B(C8H14))Re(CO)5]þ ([1-E1]þ).
The reaction of [1-E1][OTf], with 1 equiv of [HPt][PF6] in
C6D5Cl proceeds readily at ambient temperature, but the
expected boroxycarbene product is not observed; instead a
single species is formed in good yield, whose NMR proper-
ties (in particular, a 1H doublet at δ 5.06 (3JPH = 6.9 Hz, 2H);
1
ited a downfield H NMR singlet, δ 13.74, and two 31P{1H}
NMR doublets, δ 9.3 and 15.4 (JPP = 97 Hz), but after a few
minutes solids precipitated, which could be redissolved by
adding pyridine, suggesting that the initial product is 2-E1Ph1
(Scheme 23), which rearranges to an insoluble oligomer (similar
to propylene-linked 2-P2). No evidence of disproportionation
or any other transformation of [1-E1Ph1]þ was observed before
precipitation.
1
two 31P peaks in a 1:1 ratio, one of which correlates to H
1
δ 5.06 in H-31P gHMBC; a 13C resonance at δ 209.6) are
reminiscent of the dioxycarbene linkage in 11 and suggest the
related structure 10-E1 (Scheme 21). The expected carbene
species 2-E1 can be observed as an intermediate at low
temperature: NMR spectroscopic monitoring of the reaction
in a thawing C6D5Cl solution at -40 °C showed nearly
complete conversion to a carbene species, which at 25 °C is
cleanly and quickly converted to 10-E1 (Figure S157).
Treatment of [1-E1Ph1][OTf] with 2 equiv of NaHBEt3 in
THF cleanly afforded a new reduced species whose NMR
(47) Casey, C. P.; Andrews, M. A.; McAlister, D. R. J. Am. Chem.
Soc. 1979, 101, 3371.
(48) Berke, H.; Weiler, G. Angew. Chem., Int. Ed. Engl. 1982, 21, 150.

Article
Organometallics, Vol. 29, No. 20, 2010 4513
Scheme 22
acyl whose O interacts with the sodium counterion, which
also interacts with a CO of a neighboring molecule, forming
long chains of repeating units in the crystal lattice. These
chains are more linear than the zigzag network of [Na-
(THF)3][3-M1], due to Naþ bridging trans-disposed ligands
rather than cis-disposed ones. The metrical parameters are
also quite similar to [Na(THF)3][3-M1].
Reaction of [3-E1Ph1]- with MeOTf gives the neutral
methoxycarbene complex 13 (Figure 7). A smooth trend in
the Re-C and C-O bond distances, reflecting decreased
acyl and increased carbene character, can be seen on com-
paring the “free” acyl group in the sodium salt of [3-E1Ph1]-,
˚
2.133(1) and 1.251(2) A, the boroxycarbene in the sodium
-
salt of [3-E2] , 2.0960(9) and 1.271(1) A,10a and the methoxy-
˚
˚
carbene in 13, 2.0250(6) and 1.3079(7) A, respectively. (The
Re-C and C-O bonds in [3-E1Ph1]- are only ∼0.02 A
˚
shorter and longer, respectively, than that of structurally similar
anionic Re-acyl species,49 indicating that the O-Na inter-
action does not much perturb the acyl bonding.) Notably,
the B-O distances in these three species, which presumably
are related to B-O bond strength, are quite similar—1.593-
-
˚
˚
˚
(1) A (boroxycarbene) and 1.580(1) A in [3-E2] , 1.574(2) A
-
˚
in [3-E1Ph1] , and 1.5897(8) A in 13—and shorter than
the (easily broken) boroxycarbene B-O bond in 2-E2,
1.612(1) A.
˚
In contrast to the result when 2 equiv of NaHBEt3 is
employed, treatment of [1-E1Ph1][OTf] with 2 equiv [HPt]-
[PF6] in THF-d8 does not afford C-C bond formation: after
initial rapid formation of the carbene species 2-E1Ph1, no
further reaction was observed before precipitation of the
carbene (which occurs over a few hours under these con-
ditions). This may be compared to the reaction of [1-E2][BF4]
with [HPt][PF6], which proceeds to [3-E2]- within a few
minutes, suggesting that a major role of the second pendent
Lewis acid in the latter system is to promote hydride transfer
from the weaker hydride source [HPt]þ, besides aiding the CO
insertion step. Further comparison can be made to 2-M2-py,
which also effectively has only one Lewis acid and also is not
reduced by [HPt]þ. It is somewhat surprising that CO coupling
proceeds so readily in the formation of [3-E1Ph1]- from
[1-E1Ph1][OTf] and 2 equiv of NaHBEt3, as unassisted migra-
tory insertion is not generally observed in simple alkyl species
of rhenium under such mild conditions (especially without
added ligand);38,46 perhaps BEt3 (the byproduct of reduction
by [HBEt3]-) provides the needed assistance as an external
Lewis acid.
Scheme 23
Reduction of [(Ph2P(CH2)3B(C8H14))Re(CO)5]þ ([1-P1]þ).
Treatment of 1,3-propanediyl-linked phosphinoborane
cation [1-P1][OTf] with 1 equiv of [HPt][PF6] in C6D5Cl
gives a new species with a 1H NMR resonance at δ 13.74 (s),
a
31P{1H} resonance at δ 1.7, and a broad 11B resonance at
δ 85. The 11B chemical shift indicates three-coordinate
boron,45 suggesting the unstabilized formyl structure 2-P1
(Scheme 24). Consistent with this formulation, and unlike all
of the other [Re-CHO] species with pendent boranes re-
ported here, 2-P1 undergoes CO loss and Re-H formation
over a period of several hours, similar to the behavior of
(PPh3)Re(CO)4(CHO) (2-Ph1). A small amount (<5% by
NMR) of an additional species, possibly an alkyl, formed
and IR data are consistent with C-C bond formation ana-
logous to that found in [3-E2]-, except for an IR band at 1572
cm-1, suggesting structure [3-E1Ph1]- (Scheme 23), with a
free acyl group (in contrast to the boroxycarbene function-
ality in [3-E2]-, Scheme 2). [Na(THF)3][3-E1Ph1] was crystal-
lized and the structure confirmed by XRD (Figure 7); the
single borane is bound to the alkoxide oxygen, leaving a free
(49) (a) Lenhert, P. G.; Lukehart, C. M.; Sotiropoulos, P. D.; Srinivasan,
K. Inorg. Chem. 1984, 23, 1807. (b) Djukic, J. P.; Dotz, K. H.; Pfeffer, M.;
De Cian, A.; Fischer, J. Inorg. Chem. 1998, 37, 3649. (c) Padolik, L. L.; Gallucci,
J. C.; Wojcicki, A. J. Am. Chem. Soc. 1993, 115, 9986.

4514 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
Figure 7. Structural representation of [Na(THF)3][3-E1Ph1] (left) and 13 (C6H5Cl)0.5 (right), with thermal ellipsoids at 50% proba-
3
bility. Most H atoms, C atoms of THF molecules, and solvent molecules of crystallization are omitted for clarity. The chlorobenzene
˚
molecule cocrystallized with 13 was disordered and sat near a center of symmetry. Selected bond distances (A) and angles (deg):
[Na(THF)3][3-E1Ph1]: Re-C1 1.9428(6), Re-C2 1.8799(6), Re-C3 2.1361(6), Re-O4 2.2470(5), O4-C4 1.4379(8), O4-B1 1.5731(9),
C4-C3 1.5181(9), C3-O3 1.2416(8), O3-Na 2.2935(6), Re-O3-C3 143.20(5). 13 (C6H5Cl)0.5: Re-C1 1.9902(6), Re-C2 1.8772(7),
3
Re-C4 2.0250(6), Re-O3 2.2263(5), O3-C3 1.4359(8), O3-B1 1.5897(8), C3-C4 1.496(1), C4-O4 1.3079(7), O4-C5 1.457(1),
Re-C4-O4 145.15(6).
Scheme 24
initially and remained through the reaction, but no C-C
coupling was observed before complete decomposition, even
with excess [HPt][PF6]. Apparently an eight-membered
ring is too large for intramolecular stabilization of the
formyl; the absence of any effective intermolecular sta-
bilization may reflect lower basicity of the formyl O
for this monophosphine complex, compared to the bis-
(phosphine) complexes where dimeric and oligomeric species
are observed.
the potential eight-membered ring in 2-P1 is not at all
favorable, and the species behaves as a simple, unstabi-
lized formyl.
Conclusions
The precise nature of Lewis acid additives has a large
impact on the reductive coupling of CO. External trialkyl-
borane additives can facilitate reduction of [(PPh3)2Re-
(CO)4]þ by [HPt]þ and stabilization of the resultant formyl,
but no further reduction or C-C bond formation is ob-
served. To achieve the latter transformation, intramolecular
interaction between boron and oxygen appears to be re-
quired, which is possible when the phosphine ligand contains
a pendent alkylborane. We hypothesize that the size of the
ring formed by intramolecular B-O interaction determines
speciation, as shown in Scheme 26: shorter chain lengths give
smaller, more stable rings, giving monomeric boroxy-
carbenes, while longer chain lengths give larger rings, which
are destabilized, leading to either more formyl character
(little interaction) or intermolecular interactions being domi-
nant (with the pendent borane from one molecule binding
oxygen on another). The monomeric species that contain
intramolecular borane coordination appear to be readily
reduced by [HPt]þ (although bis(phosphine) complexes re-
quire a second pendent acid to shuttle the Pt-H), while
oligomeric or externally borane-coordinated species do not
show reductive coupling chemistry. This could be due to a
geometric change of the boroxycarbene, imposed by the rigid
ring structure (Figure 8). When the geometry is constrained
Overview of (Ph2P(CH2)nB(C8H14))Re Complexes: Ring
Size Effects. The effect of changing chelate ring size is sum-
marized in Scheme 25; remarkably, each system exhibits dif-
ferent reduction chemistry. With a methylene spacer, boroxy-
carbene 2-M1 is stable and can be further reduced by [HPt]þ
to undergo C-C coupling; with an ethylene spacer, the less
stabilized 2-E1 rapidly disproportionates to dinuclear alkyl
10-E1 (which does not undergo spontaneous C-C coupling);
with a propylene spacer, neither stabilization nor novel reac-
tivity of the Re-CHO species is observed, and the formyl
rapidly decarbonylates, similar to borane-free systems.43b
Addition of pyridine to 2-M1 disrupts the B-O bond and
leads to 10-M1 py, a product very similar to that obtained in
3
the ethylene-linked case, showing that under appropriate
conditions the selectivity can be tailored.
The chain-length effect may be attributed to the relative
stability of the intramolecular B-O interaction; the six-
membered ring in 2-M1 is quite favorable, and the result-
ing stable boroxycarbene readily undergoes further reduc-
tion. For 2-E1 there is an apparently rapid equilibrium
between free formyl and boroxycarbene (an intramolec-
ular seven-membered ring), leading to disproportionation;

Article
Organometallics, Vol. 29, No. 20, 2010 4515
Scheme 25
Scheme 26
in the P-Re plane, the back-bonding is competing with three
CO ligands, whereas the more favored geometry, essentially
90° out of plane, competes with only two CO ligands. This
subtle change would render the chelated boroxycarbene
more electrophilic and perhaps more prone to nucleophilic
attack.
The number of Lewis acids is also important. Reductive
coupling in systems with only one Lewis acid was observed in
only two cases ([1-M1]þ and [1-E1Ph1]þ), although in the case
of the less electrophilic bis(phosphine) complex, strong
hydride sources were required. This implies that the C-C
bond forming step does not require Lewis acid assistance

4516 Organometallics, Vol. 29, No. 20, 2010
Miller et al.
of the key steps of Lewis acid-facilitated CO hydrogenation.
The number of Lewis acids and their strength are both
important factors, and having a favorable ring size for
intramolecular B-O coordination is essential. Work still
remains in understanding the mechanistic subtleties of this
system and in closing the cycle on rhenium by liberating the
organic fragment.
Acknowledgment. Larry Henling and Mike Day assisted
with crystallography. The Bruker KAPPA APEXII X-ray
diffractometer was purchased via an NSF CRIF:MU
award to the California Institute of Technology (CHE-
0639094). David VanderVelde assisted with 2D and
PGSE NMR spectroscopy. Dr. Daniel L. DuBois and
Dr. Nathan M. West contributed enlightening discus-
sions. This work was funded by BP (MC2 Program) and
the Moore Foundation.
Figure 8
and may be driven more by ring-strain arguments. A second
acid is required to shuttle a hydride from Pt to B in the
case of bis(phosphine) complexes, if [HPt]þ is used as the
reductant.
We have identified key requirements for successful integra-
tion of pendent boranes and have a preliminary understanding
Supporting Information Available: This material is available
free of charge via the Internet at http://pubs.acs.org.