Insertion reactions of dicyclohexylcarbodiimide with aminoboranes, -boryls and -borylenes

Article abstract of DOI:10.1039/b709249k

Insertion reactions of dicyclohexylcarbodiimide with aminoboranes and with aminoboryl and -borylene transition metal complexes have been examined as potential routes to new boron-containing ligand systems. Reactions with systems containing two-coordinate

Full text of DOI:10.1039/b709249k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Insertion reactions of dicyclohexylcarbodiimide with aminoboranes,
-boryls and -borylenes†‡
Glesni A. Pierce,a Natalie D. Coombs,a,b David J. Willock,b Joanna K. Day,a,b
Andreas Staschb,c and Simon Aldridge*a
Received 19th June 2007, Accepted 19th July 2007
First published as an Advance Article on the web 8th August 2007
DOI: 10.1039/b709249k
Insertion reactions of dicyclohexylcarbodiimide with aminoboranes and with aminoboryl and
-borylene transition metal complexes have been examined as potential routes to new boron-containing
ligand systems. Reactions with systems containing two-coordinate boron centres are found to be
significantly more facile than those with three-coordinate substrates. Thus, reaction of
(dicyclohexylamino)boron dichloride (1a) with dicyclohexylcarbodiimide over 36 h at 50 ◦C generates
the (structurally authenticated) guanidinate complex Cy2NC(NCy)2BCl2 (2a) via insertion into the BN
bond. By contrast, the corresponding reaction with the cationic◦aminoborylene complex
[CpFe(CO)2(BNCy2)]+[BArf4]− (4a) proceeds rapidly at ca. −30 C, via initial insertion into the FeB
bond to give [CpFe(CO)2C(NCy)2BNCy2]+[BArf4]− (5a). Consistent with related studies, a key factor in
facilitating such insertion chemistry is thought to be the formation of an initial donor/acceptor
complex between the diimide and the group 13 centre. Thus, DFT studies suggest that
[CpFe(CO)2B(NCy2)(CyNCNCy)]+[BArf4]− is a potential intermediate in the reaction of 4a with
CyNCNCy, and that further reaction to give the observed product, 5a, is strongly exergic
(−183 kJ mol−1). By contrast, DFT calculations for the alternative isomer
[CpFe(CO)2B(CyN)2CNCy2]+[BArf4]− (5aꢀ), formed by BN insertion, suggest that it is 112 kJ mol−1 less
stable than 5a. Such experimental and computational findings imply that under reaction conditions
where a suitable isomerisation pathway is available, cationic complexes such as 5aꢀ, which contain a
four-membered boron-donor heterocycle are likely to be disfavoured with respect to alternative
C-bound isomers.
Introduction
Amidinates and guanidinates have proved to be versatile and
widely exploited ligands for the stabilization of metal complexes
offering a range of potential applications (Scheme 1).1 Among
the elements of group 13 for example, such systems have been
Scheme 1 Amidinate and guanidinate ligands.
exploited in the synthesis of M(I) derivatives of potential interest as
valence isoelectronic analogues of four-membered N-heterocyclic
carbene (NHC) ligands (M = Ga, In, Tl; Scheme 2),2 while
M(III) complexes have been implicated not only as active catalysts
for olefin polymerization (M = Al),3 but also as single-source
precursors for nitride materials (M = Ga).4 However, despite
interest in both the structural and reaction chemistry of these
Scheme 2 Generic metal complexes containing a four membered NHC
ligand (I), a related gallium donor (II), or an (as yet unrealised) boron
analogue (III).
aInorganic Chemistry, South Parks Road, Oxford, UK OX1 3QR. E-mail:
Simon.Aldridge@chem.ox.ac.uk; Fax: +44 1865 272690; Tel: +44 1865
285201
bSchool of Chemistry, Cardiff University, Main Building, Park Place, Cardiff,
UK CF10 3AT
group 13 complexes, analogous boron-containing systems had for
a long time been largely neglected. Thus, amidinate complexes
of boron have been systematically investigated from a structural
perspective only very recently,5,6 and to our knowledge, there exists
only one crystallographically characterized boron guanidinate in
the literature.7
cSchool of Chemistry, Monash University, PO Box 23, Victoria 3800,
Australia
† CCDC reference numbers 651106–651108, 628665. For crystallographic
data in CIF or other electronic format see DOI: 10.1039/b709249k
‡ Electronic supplementary information (ESI) available: Synthetic and
characterizing data for 2b; synthetic, characterizing and crystallographic
data for CpFe(CO)2B(NBz2)Cl; CIFs for all crystal structures [compounds
2a, 2b, 3c, 5a and CpFe(CO)2B(NBz2)Cl]. Cartesian coordinates and
energies for the fully optimized DFT structures of 4a, CyNCNCy,
4a·CyNCNCy, 5a and 5aꢀ. See DOI: 10.1039/b709249k
We have been interested in investigating the potential for
= =
insertion of heteroallenes (such as carbodiimides, RN C NR
8
= =
and isocyanates, RN C O) into BN bonds. Previous reports
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4405–4412 | 4405
©

View Article Online
Scheme 3 Potential routes to cationic metal complexes containing a four-membered boron N-heterocyclic carbene analogue (VII).
of such reactivity date from the 1960s,9 and although relatively
little subsequent exploitation of this chemistry has been reported,6
we perceived that it might offer a simple, one-pot route to
(guanidinate)boron species for structural and reactivity studies.
In addition, we were interested in examining the selectivity for
insertion into BN bonds in systems containing other poten-
tially reactive B–X linkages. In particular, we targeted systems
containing metal–boron bonds of different types, as a potential
methodology for assembling novel ligands at a metal centre,
including, for example, boron counterparts of NHC ligands
(Schemes 2 and 3).2,10 Thus we report on our investigation of
the reactivity of dicyclohexylcarbodiimide with aminoboranes,
and with aminoboryl and -borylene systems, which contain,
mind, we examined the reaction of (dicyclohexylamino)boron
dichloride (1a) with dicyclohexylcarbodiimide (Scheme 4). This
reaction proceeds cleanly (as evidenced by 11B NMR monitoring)
over a period of 36 h in toluene at 50 ◦C to generate the
desired guanidinate complex Cy2NC(NCy)2BCl2 (2a) in ca. 60%
isolated yield. These reaction conditions are somewhat harsher
than those reported by Cowley for the synthesis of the cor-
responding bis(trimethylsilyl)amino derivative, and the isolated
yield is slightly lower.6b We have also explored an alternative salt
elimination approach to the synthesis of these (guanidinate)boron
dihalides (see Scheme 4 and ESI for experimental details‡);6
i
thus Pr2NC(NCy)2BCl2 (2b) can be synthesized by a two-step
i
process via the guanidine Pr2NC(NCy)(NHCy). In practice this
=
respectively, M–B single and M B double bonds. Ultimately
methodology was found not only to involve significantly more
manipulation, but also to result in a lower overall yield of the
desired guanidinate complex.
these studies reveal that, in conjunction with the cationic metal
fragments employed, B-donor four-membered carbene analogues
are likely to be unstable with respect to isomeric C-bound systems.
Results and discussion
Although isolated previous examples of carbodiimide insertion
into the BN bonds of three-coordinate boranes have been reported
previously,6,7,9 to our knowledge only one crystallographically
characterized guanidinate complex of boron has been described
to date,7 and no examples of dihalo substituted systems (IV)
which might potentially act as useful borane precursors e.g. for
novel boron ligand systems (Scheme 3, Route 1). With this in
Scheme 4 Respective syntheses of (guanidinate)boron dihalides 2a and
2b by insertion and salt elimination methodologies.
The structures of 2a and 2b as determined by single crystal
X-ray diffraction are shown in Fig. 1. The two compounds possess
Fig. 1 Molecular structures of Cy2NC(NCy)2BCl2 (2a; left) and iPr2N◦C(NCy)2BCl2 (2b; right); hydrogen atoms omitted for clarity and ORTEP ellipsoids
˚
plotted at the 50% probability level. Key bond lengths (A) and angles ( ): (for 2a) B(1)–Cl(1) 1.854(2), B(1)–N(1) 1.554(3), C(1)–N(1) 1.354(2), C(1)–N(2)
1.348(3), N(1)–B(1)–N(1ꢀ) 84.4(2), N(1)–C(1)–N(1ꢀ) 100.9(2), torsion between N(1)C(1)N(1ꢀ) and C(8)–N(2)–C(8ꢀ) planes 28.3; (for 2b) B(1)–Cl(1)
1.852(3), B(1)–N(1) 1.544(4), B(1)–N(2) 1.554(4), C(1)–N(1) 1.356(3), C(1)–N(2) 1.350(3), C(1)–N(3) 1.349(3), N(1)–B(1)–N(2) 84.4(2), N(1)–C(1)–N(2)
100.5(2), torsion between N(1)C(1)N(2) and C(14)–N(3)–C(17) planes 30.4.
4406 | Dalton Trans., 2007, 4405–4412
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
superficially similar geometries, each featuring the expected four-
membered BN2C heterocycle and exocyclic NR2 moiety (R = Cy,
iPr). In both cases the exocyclic NC2 unit is twisted significantly
out of the plane of the (guanidinate)B ring (torsion angles of
28.3 and 30.4◦ for 2a and 2b, respectively). While the geometries
of the BN2C rings themselves resemble closely that of the
compound Me2NC(NPh)2B(CF3)2 reported by Bu¨rger,7 the greater
steric demands of the exocyclic NR2 groups in 2a and 2b (cf.
NMe2) are presumably responsible for the wider torsion angles
[cf. 13.0◦ for Me2NC(NPh)2B(CF3)2].7 Consistent with this, and
with a consequently smaller degree of NC p bonding, the C–
Scheme 5 Reactions of (dialkyl)aminoboryl complexes 3a (Rꢀ = Cy) and
i
3b (Rꢀ = Pr) with dicyclohexylcarbodiimide.
with BN insertion of the carbodiimide at some stage during the
course of the reaction. Postulating that the forcing conditions
employed might lead to decomposition of the putative species
CpFe(CO)2B(guanidinate)Cl, we targeted aminoboryl complexes
which might allow insertion of carbodiimide under milder
conditions. With this in mind, diarylamino substituents were
targeted, given the weaker BN bonds reportedly associated with
aryl substituents.14 Thus the complex CpFe(CO)2B(NPh2)Cl (3c)
was synthesized from the corresponding dichloride, Ph2NBCl2
(1c), using standard salt elimination chemistry (Scheme 6) and
characterized by conventional spectroscopic and crystallographic
techniques.
˚
NR2 distances found for 2a and 2b [1.348(3) and 1.349(3) A,
respectively] are significantly longer that those found for the
two independent molecules of Me2NC(NPh)2B(CF3)2 [1.315(3),
7
˚
1.316(3) A]. A similar planar BN2C motif is found in the
(amidinate)boron dihalides reported by Cowley, with the caveat
that the ring C–N distances in 2a and 2b are slightly longer than
those typically found for amidinate derivatives.6 Thus the CN
distances measured for (2,4,6-tBu3C6H2)C(NCy)2BCl2 [1.331(4)
6b
˚
and 1.332(4) A] can be compared with the corresponding bond
˚
lengths found for 2a [1.354(2) A], and rationalized on the basis
of an additional resonance form for guanidinates (cf. amidinates)
featuring C–N single bonds within the four-membered ring and a
=
C N double bond exo to it.
While the chemistry represented by Scheme 4 clearly confirms
the viability of carbodiimide insertion into BN bonds as a synthetic
methodology, the utility of systems 2a and 2b in further chemistry
(such as that outlined in Scheme 3) appears to be limited. Thus,
the inability to carry out halide substitution with organometallic
anions at the four-coordinate boron centre effectively eliminates
species of this type as viable precursors to novel boron ligand
systems. In common with four-coordinate boron dihalides incor-
porating aminopicoline substituents,11 little evidence of M–B bond
formation is forthcoming from the reaction of (guanidinate)boron
dihalides with nucleophilic reagents such as Na[CpFe(CO)2]. This
inertness contrasts with the ready halide substitution observed
for related amidinate derivatives containing aluminium, gallium
or indium dihalide fragments,12 and is presumably a result of
the smaller and less electrophilic nature of the boron centre.
Consequently, it was decided to investigate the possibility for direct
insertion of a carbodiimide molecule into the BN linkage of an
existing aminoboryl or -borylene complex (Scheme 3, Routes 2
and 3). The viability of such an approach is clearly dependent on
achieving selectivity for insertion into the BN bond, and issues of
migratory aptitude of the various boron-bound substituents were
therefore of interest.
Scheme 6 Synthesis of (diphenyl)aminoboryl complex 3c and attempted
reaction with dicyclohexylcarbodiimide.
The structure of 3c determined by single crystal X-ray diffrac-
tion (Fig. 2) is similar to those of related (dialkyl)aminoboryl
i
complexes, CpFe(CO)2B(NR2)Cl (R = Cy, Pr, Bz), featuring an
˚
˚
Fe–B distance of ca. 2.0 A [2.022(3) A for 3c, cf. 2.053(3), 2.054(4)
i
˚
and 2.033 A (mean) for R = Cy, Pr and Bz respectively] and
a boryl ligand plane [i.e. that defined by B(1), Cl(1) and N(1)]
orientated approximately perpendicular to that defined by the
Cp centroid, iron and boron atoms (85.5◦ for 3c).13 The two
phenyl substituents are orientated at angles of 79.8 and 67.6◦ with
The reactions of the previously reported aminoboryl complexes
i
CpFe(CO)2B(NR2)Cl (3a: R = Cy; 3b: R = Pr)13 with dicyclo-
hexylcarbodiimide in toluene at 65 ◦C can be shown by 11B NMR
spectroscopy to proceed along similar lines, in each case generating
a predominant boron-containing product giving rise to a signal at
dB 4–5. Crystallographic studies and comparison of spectroscopic
data for the isolated compounds with those measured for authentic
samples reveal, however, that the boron-containing product in
each case is the (guanidinate)boron dichloride R2NC(NCy)2BCl2
Fig. 2 Molecular structure of CpFe(CO)2B(NPh2)Cl (3c); hydrogen
atoms omitted for clarity and ORTEP ellipsoids plotted at the 50%
◦
˚
probability level. Key bond lengths (A) and angles ( ): Fe(1)–B(1) 2.022(3),
B(1)–N(1) 1.412(3), B(1)–Cl(1) 1.824(3), Fe(1)–B(1)–Cl(1) 116.2(1),
Fe(1)–B(1)–N(1) 130.2(2), N(1)–B(1)–Cl(1) 113.5(2), B(1)–N(1)–C(8)
124.7(2), B(1)–N(1)–C(14) 122.3(2), C(8)–N(1)–C(14) 113.0(2), torsion be-
tween Cp centroid–Fe(1)–B(1) and N(1)–B(1)–Cl(1) planes 85.5, torsions
between N(1)–B(1)–Cl(1) and phenyl ring planes 67.6 and 79.8.
i
(2a: R = Cy; 2b: R = Pr). The only identifiable organometallic
product of each reaction is the [CpFe(CO)2]2 dimer (Scheme 5).
The products isolated using these alkylamino substituted boryl
complexes, while unexpected (but reproducible), are consistent
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4405–4412 | 4407
©

View Article Online
respect to the plane defined by the boryl ligand, presumably on
steric grounds, and there is no statistically significant difference
˚
in the BN distance [1.412(3) A] compared to those measured for
(dialkyl)amino derivatives [1.396(4), 1.389(5) and 1.398 (mean)
i
for R = Cy, Pr, Bz).13 Given these structural similarities and, if
anything, the marginally weaker N→B p donation implied by the
slightly higher carbonyl stretching frequencies measured for 3c
(2011, 1951 cm−1, cf. 2000, 1939; 2001, 1941; and 1996, 1933 cm−1
i
for R = Cy, Pr and Bz, respectively)13 we were surprised to find
that 3c is essentially unreactive towards carbodiimide substrates
even under relatively forcing conditions (120 h at 60 ◦C). On this
basis it would appear that insertion of a carbodiimide molecule
into the BN bond of aminoboryl complexes is not facile, such
that even in cases where this chemistry does appear to occur,
the forcing conditions required lead to the isolation of species
presumably resulting from decomposition of the initial insertion
product, CpFe(CO)2B(guanidinate)Cl. Conceivably, the level of
steric crowding at the putative four-coordinate boron centre
renders such species labile with respect to substituent redistri-
bution or elimination processes. As a result, we sought to probe
the reactivity of two-coordinate amino-borylene systems towards
insertion chemistry, where presumably such steric problems would
be reduced.
Fig. 3 Molecular structure of the cationic component of [CpFe(CO)2-
C(NCy2)BNCy2]+[BArf4]− (5a); hydrogen atoms and counter-ion omitted
for clarity and ORTEP ellipso◦ids plotted at the 50% probability level. Key
˚
bond lengths (A) and angles ( ): Fe(1)–C(8) 1.938(3), C(8)–N(1) 1.358(4),
C(8)–N(2) 1.357(4), B(1)–N(1) 1.506(2), B(1)–N(2) 1.491(2), B(1)–N(3)
1.371(4), N(1)–C(8)–N(2) 98.3(2), N(1)–B(1)–N(2) 86.5(2).
Reactions of dicyclohexylcarbodiimide with the cationic
aminoborylene complexes [CpFe(CO)2(BNR2)]+[BArf4]− (4a: R =
bond lengths in the four membered heterocyclic ring with those
i
˚
Cy; 4b: R = Pr) are indeed much more facile, proceeding in
measured for 2a and 2b [e.g. B(1)–N(1) 1.506(2) A, C(8)–N(1)
each case at ca. −30 ◦C in dichloromethane to give a product
characterized by a single 11B NMR resonance at dB ca. 25.
The relatively high field positions of these signals, however, are
inconsistent with a cationic system containing a three-coordinate
boron donor ligand directly bonded to a metal centre. Thus,
signals at dB 56.9 and 53.7 have been reported for the N-base
stabilized aminoborylene adducts [CpFe(CO)2{BNCy2(4-pic)}]+
and [CpFe(CO)2{BNiPr2(iPrNCMe2)]+, respectively.13a,15 In the
case of 4a, the product formed by insertion of a single molecule
of dicyclohexylcarbodiimide can be isolated as a crystalline solid
and its structure established crystallographically (Fig. 3). Of key
importance in establishing the course of this reaction is the
unequivocal identification of the positions of the boron and
quaternary guanidinate carbon atoms in the product. In essence,
this assignment reflects whether such chemistry proceeds via FeB
or BN insertion (to give the isomeric products 5a or 5aꢀrespectively;
Scheme 7). Thus, despite the difficulties in differentiating between
elements differing by a single electron on the basis of X-
ray diffraction data, a number of factors point to a structure
[CpFe(CO)2C(NCy)2BNCy2]+[BArf4]− (i.e. 5a) for this insertion
product: (i) both 11B (vide supra) and 13C NMR data are consistent
with a connectivity implying the presence of a M–C (but no M–
B) bond {dC 251.5, cf . dC 236.1 for CpFe(CO)2C(O)(NPh)ZrL,
where L = [(2-FC6H4)NSiMe2]3CH};16 and (ii) comparisons of
˚
˚
1.358(4) A, cf . analogous values of 1.554(3) and 1.354(2) A for
2a] are informative in establishing the positions of the BN and CN
linkages.
˚
The Fe–C bond in 5a [1.938(3) A] is relatively short for
a single bond between the [CpFe(CO)2] fragment and an sp2
hybridized carbon atom, being similar to that measured for
˚
CpFe(CO)2CS2ZrL {1.934(11) A; L = [(2-FC6H4)NSiMe2]3CH},
a compound which has been described as possessing partial
16
=
carbenoid (Fe C) character. Despite this, and the fact that the
metalla-amidinate carbon in 5a resonates at a low field more akin
to that measured for CpFe(CO)2C(O)(NPh)ZrL (dC 236.1) than
those typically found for quaternary amidinate or guanidinate
carbons (dC ca. 160–170),17 DFT calculations carried out for the
Fe–C linkage in 5a imply that the overall contribution from a
similar carbenoid resonance form (Scheme 8) is small. Thus, the
calculated r: p ratio for the FeC bonding density (85 : 15) reflects a
significantly reduced p component than that obtained for a model
+
=
=
Fe C double bond {64 : 36 for [CpFe(CO)2( CH2)] } using the
same precedented methodology.17
DFT calculations can also be exploited to shed some light on
the likely nature of the intermediate species formed in the reaction
of 4a with dicyclohexylcarbodiimide. The N→B donor/acceptor
complex CyNCN(Cy)·B(Ph)Cl2 is known to precede insertion
of dicyclohexylcarbodiimide into the B–C bond of PhBCl2,18
Scheme 7 Potential insertion reactions of dicyclohexylcarbodiimide into either the BN or the FeB bonds of cationic borylene complex 4a.
4408 | Dalton Trans., 2007, 4405–4412 This journal is The Royal Society of Chemistry 2007
©

View Article Online
Scheme 8 Possible contributing resonance structures for the cationic component of metalla-amidinate complex 5a.
and related B-bound adducts of aminoborylenes with ketones,
imines and phosphine oxides have been characterized in pre-
vious studies.13,19 As such, a potential intermediate is the 1 :
1 adduct [CpFe(CO)2B(NCy2)(CyNCNCy)]+[BArf4]− formed by
coordination of one of the diimide nitrogens at the borylene
centre of 4a. In order to assess the viability of the interme-
diate [CpFe(CO)2B(NCy2)(CyNCNCy)]+ and its potential for
conversion into the isomeric products 5a and 5aꢀ, a series of
DFT calculations was carried out, salient results of which are
summarized in Scheme 9 (x, y, z files corresponding to the fully
optimized geometries are included in the ESI‡).
A number of features are apparent from these computational
studies: (i) an excellent match of calculated and experimentally
determined structural parameters is found for the cationic com-
ponents of 4a and 5a (the greatest discrepancies being the 2–3%
over-estimate in bond lengths to iron are typical of this method);17
(ii) in contrast to the coordination of MeNCNMe to PhBCl2,18 for-
mation of the adduct [CpFe(CO)2B(NCy2)(CyNCNCy)]+[BArf4]−
from 4a and CyNCNCy is energetically uphill (+48 kJ mol−1),
reflecting, even for this three-coordinate adduct, a high degree
of steric crowding between the boron-bound [CpFe(CO)2], NCy2
and CyNCNCy fragments; (iii) the overall insertion processes
leading to the formation of both 5a and 5aꢀ(from 4a and
CyNCNCy) are exergic; and (iv) the energy change associated
with the formation of the experimentally realised C-bound isomer
5a (−135 kJ mol−1) is significantly more favourable than that
calculated for the B-bound isomer 5aꢀ(−23 kJ mol−1). Such data
provide a convincing thermodynamic rationale for the formation of
5a over 5aꢀ, presumably with the retention of the strong BN linkage
in 5a being a key energetic factor (an adiabatic bond dissociation
energy of 585 kJ mol−1 having been calculated for the BN bond
20
=
in H2N BH2). It is also conceivable that kinetic factors may
be important. Hetero-allene insertions into B–C, Al–C or Al–N
bonds have been shown to involve a two step mechanism, with
the barriers associated with migration of the boron/aluminium-
bound substituent and with formation of the final four-membered
=
EN2C (E = B, Al) chelate ring (by rotation about a C N bond)
being substituent dependent.18,21 In terms of migratory aptitude,
previous work with related isocyanate substrates has shown that
B to C migration is typically favoured by substituents possessing
weaker B–X bonds and more nucleophilic fragments at X.22 In
the case of the intermediate [CpFe(CO)2B(NCy2)(CyNCNCy)]+
and related complexes of the type [CpFe(CO)2B(NR2)L]+ (L =
imine, picoline, phosphine oxide donor),23 DFT-based parti-
tioning of bonding densities is consistent with a significant
N→B p donation {e.g. a 64 : 36 r : p breakdown to the B–
NCy2 bond in [CpFe(CO)2B(NCy2)(CyNCNCy)]+}, but with a
relatively minor p component to the Fe–B bond {e.g. 88 : 12 for
[CpFe(CO)2B(NCy2)(CyNCNCy)]+}. On this basis, it seems likely
that migration of the [CpFe(CO)2] (rather than the NCy2) fragment
would be expected not only to lead to the formation of the more
thermodynamically stable product, but also to be more facile.
Experimentally, therefore it is hardly surprising that no trace
of the competing BN insertion product 5aꢀis observed under
any of the reaction conditions examined. BN insertion can be
observed in these systems, but only after initial FeB insertion.
Scheme 9 DFT (BLYP) calculated structures and energies (kJ mol−1) relevant to the insertion of dicyclohexylcarbodiimide into the Fe B or B N bonds
of 4a.
=
=
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4405–4412 | 4409
©

View Article Online
Thus, in the presence of excess carbodiimide at (or slightly below)
room temperature, 5a reacts rapidly to generate the bis(insertion)
product [CpFe(CO)2C(NCy)2B(NCy)2CNCy]+ [BArf4]− (6a), the
synthesis and structure of which (along with the corresponding
NiPr2 derivative) has recently been reported (Scheme 10).8
NMR spectra were measured on a Bruker AM-400 or JEOL
300 Eclipse Plus FT-NMR spectrometer. Residual signals of
1
solvent were used as reference for H and 13C NMR; 11B and
19F NMR spectra were referenced with respect to Et2O·BF3 and
CFCl3, respectively. Infrared spectra were measured for each
compound pressed into a disk with an excess of dried KBr or
as a solution in an appropriate solvent on a Nicolet 500 FT-
IR spectrometer. Mass spectra were measured by the EPSRC
National Mass Spectrometry Service Centre, University of Wales,
Swansea, or by the departmental service. Perfluorotributylamine
was used as a standard for high resolution EI mass spectra.
Elemental microanalyses were carried out by Medac Ltd. or by
the departmental service. Abbreviations: b = broad, s = singlet,
d = doublet, t = triplet, q = quintet, m = multiplet, fwhm =
frequency width at half maximum.
Scheme 10 Insertion of a second equivalent of carbodiimide into the BN
bond of 5a.
Conclusions
(b) Syntheses
Insertion reactions of dicyclohexylcarbodiimide with amino-
boranes and with aminoboryl and -borylene complexes have
been examined. Reactions with systems containing two-coordinate
boron centres are found to be significantly more facile than those
with three-coordinate precursors. Previous studies have shown that
a key factor in facilitating such insertion chemistry is the formation
of an initial donor/acceptor complex between the diimide and the
group 13 centre,18 a reaction likely to be favoured by the lower coor-
dination number and more electrophilic boron centre in borylene
systems. Furthermore, DFT studies suggest that, from the po-
tential intermediate [CpFe(CO)2B(NCy2)(CyNCNCy)]+[BArf4]−,
further reaction to give the observed C-bound product,
[CpFe(CO)2C(NCy)2BNCy2]+[BArf4]− (5a) is strongly exergic. By
contrast, DFT calculations for the alternative isomer
[CpFe(CO)2B(CyN)2CNCy2]+[BArf4]− (5aꢀ)—which features a
four-membered boron guanidinate ligand with parallels in NHC
chemistry—reveal that it is ca. 112 kJ mol−1 less stable than 5a. As
such, it seems likely that under reaction conditions where a suitable
isomerisation pathway is available, cationic complexes containing
such B-donor NHC analogues will be disfavoured with respect
to alternative structures which maximize the number of BNR2
linkages.
Synthesis of Cy2NC(NCy)2BCl2 (2a) by insertion of dicyclohexyl-
carbodiimide into the BN bond of Cy2NBCl2 (1a). To a solution
of 1a (0.20 g, 0.8 mmol) in toluene (10 cm3) was added a solution
of CyNCNCy (0.20 g, 1.0 mmol) also in toluene (10 cm3). The
reaction mixture was heated at 50 ◦C and progress monitored by
11B NMR spectroscopy. After 36 h, the reaction was judged to
be complete by quantitative conversion of the signal at dB 28.8
due to 1a to that at dB 4.3 characteristic of 2a. Volatiles were
removed in vacuo, the residue extracted with toluene, layered with
hexanes and cooled to −30 ◦C yielding crystals suitable for X-ray
1
diffraction. Yield of crystalline product: 0.22 g, 58%. H NMR
(300 MHz, C6D6) d 0.65–2.31 (overlapping m, 40H, CH2 of Cy),
3.14 (overlapping m, 4H, CH of Cy). 13C NMR (76 MHz, C6D6) d
24.2, 24.4 (4-CH2 of Cy), 25.1, 25.5 (3-CH2 of Cy), 30.6, 32.3 (2-
CH2 of Cy), 54.9, 58.2 (CH of Cy), 160.9 (guanidinate quaternary).
11B NMR (96 MHz, C6D5CD3) d 4.3 (s, fwhm ca. 19 Hz). IR
(benzene soln, cm−1) m(CN) 1618. EI-MS, m/z (%): 467 (weak) M+,
432 (72%) (M − Cl)+; exact mass (calc. for M+) m/z 466.3036,
(obs.) 466.3033. Elemental analysis (calc. for C25H44BCl2N3) C
64.11, H 9.47, N 8.97, (obs.) C 63.88, H 9.28, N 8.78.
Synthesis of CpFe(CO)2B(NPh2)Cl (3c).
A solution of
Ph2NBCl2 (0.77 g, 3.09 mmol) in diethyl ether (20 cm3) was
added to a suspension of Na[CpFe(CO)2] (0.62 g, 3.09 mmol)
in diethyl ether (20 cm3) and the mixture stirred for 12 h. Volatiles
were removed in vacuo and the residue extracted into hexanes,
concentrated (to ca. 5 cm3) and cooled to −30 ◦C yielding 3c
as a colourless microcrystalline solid which was washed with
cold hexane (2 × 30 cm3). Crystals suitable for X-ray diffraction
were obtained by recrystallization from a hexanes–toluene mixture
Experimental
(a) General considerations
All manipulations were carried out under a nitrogen or argon
atmosphere using standard Schlenk line or dry box techniques.
Solvents werepre-dried over sodiumwire (hexanes, toluene, diethyl
ether) or molecular sieves (dichloromethane) and purged with ni-
trogen prior to distillation from the appropriate drying agent (hex-
anes: potassium; toluene, diethyl ether: sodium; dichloromethane:
CaH2). Benzene-d6, toluene-d8 and dichloromethane-d2 (Goss)
were degassed and dried over potassium (benzene-d6, toluene-d8)
or molecular sieves (dichloromethane-d2) prior to use. Dicyclo-
hexylcarbodiimide (Aldrich) was dried under high vacuum prior
1
(ca. 50 : 1). Yield of crystalline product: 0.62 g, 51%. H NMR
(300 MHz, CD2Cl2) d 4.67 (s, 5H, Cp), 6.85–7.23 (m, 10H, Ph). 13
C
NMR (76 MHz, CD2Cl2) d 84.1 (Cp), 125.1, 126.2 (ortho-C of Ph),
126.6, 127.8 (para-C of Ph), 128.1, 128.5 (meta-C of Ph), 146.4,
149.2 (ipso-C of Ph), 213.7 (CO). 11B NMR (96 MHz, CD2Cl2) d
61.2 (b, fwhm ca. 210 Hz). IR (CD2Cl2 soln, cm−1) m(CO) 2011,
1951. EI-MS, m/z (%): 363 (weak) (M − CO)+, 335 (26) (M −
2CO)+, 169 (100) (Ph2N)+; exact mass (calc.) m/z 363.0290, (obs.)
363.0285. A similar procedure can be used to give access to the
related complex CpFe(CO)2B(NBz2)Cl (Bz = benzyl, CH2Ph; see
ESI‡).
n
to use; BuLi (1.6 M solution in hexanes, Acros Organics) was
used as received. Starting materials Na[CpFe(CO)2], R2NBCl2
i
(1a: R = Cy; 1b: R = Pr; 1c: R = Ph), CpFe(CO)2B(NR2)Cl
(3a: R = Cy; 3b: R = Pr) and [CpFe(CO)2(BNCy2)]+[BArf4]− (4a)
i
were prepared according to literature methods.13,24–27
4410 | Dalton Trans., 2007, 4405–4412
This journal is The Royal Society of Chemistry 2007
©

View Article Online
i
Reactions of CpFe(CO)2B(NR2)Cl (3a: R = Cy; 3b R = Pr)
imide in dichloromethane (5 cm3) to a solution of 4a also
in dichloromethane (0.61 g, 0.05 mmol in 20 cm3) with
11B NMR monitoring allowed for quantitative conversion
of 4a (dB 93.1) to 5a (dB 26.5); under such conditions
only very minor quantities of the double insertion product
[CpFe(CO)2C(NCy)2B(NCy2)CNCy2]+ [BArf4]− (6a) are detected.8
Filtration of the reaction mixture, concentration (to ca. 5 cm3),
layering with hexanes and storage at −30 ◦C led to the formation of
single crystals suitable for X-ray diffraction. Isolated yield: 0.17 g,
with CyNCNCy. The reactions were carried out in an analogous
manner, exemplified for 3b. To a solution of 3b (0.035 g, 0.1 mmol)
in toluene (5 cm3) was added a solution of CyNCNCy (0.05 g,
0.24 mmol) in toluene (5 cm3) and the reaction mixture heated
at 65 ◦C. After 168 h the reaction was judged to be complete
by 11B NMR spectroscopy (conversion of the signal at dB 54.8
due to 3b to that at dB 4.6 characteristic of 2b); volatiles
were removed in vacuo and the residue extracted with hexanes,
concentrated to ca. 10 cm3 and cooled to −30 ◦C yielding 2b
as a microcrystalline solid. Isolated yield: 0.013 g, 31%. The
identity of 2b was confirmed by comparison of spectroscopic
data (see above) with an authentic sample, and by crystallographic
studies carried out on single crystals obtained from hexanes. The
predominant iron containing product was subsequently confirmed
1
24%. H NMR (400 MHz, CD2Cl2) d 1.02–1.77 (m, 40H, CH2
of Cy), 3.26 (m, 2H, CH of Cy), 3.42 (m, 2H, CH of Cy), 4.99
(s, 5H, Cp), 7.46 (s, 4H, para-H of BArf4−), 7.63 (s, 8H, ortho-H
of BArf4−). 13C NMR (76 MHz, CD2Cl2), d 25.1, 25.2 (4-CH2 of
Cy), 25.9, 27.1 (3-CH2 of Cy), 33.1, 35.4 (2-CH2 of Cy), 58.1,
58.7 (CH of Cy), 86.3 (Cp), 117.5 (para-CH of BArf4−), 124.7 (q,
1JCF = 274 Hz, CF3 of BArf4−), 129.0 (2JCF = 32 Hz, meta-C of
BArf4−), 134.9 (ortho-CH of BArf4−), 161.9 (q, 1JCB = 50 Hz, ipso-
C of BArf4−), 209.6 (CO), 251.5 (metalla-amidinate quaternary).
11B NMR (96 MHz, CD2Cl2) d 26.5 (br, fwhm 430 Hz), −7.6
(BArf4−). 19F NMR (283 MHz, CD2Cl2) d −62.7 (CF3). IR
(CD2Cl2 soln, cm−1) m(CO) 2049, 2003. Mass spec. (positive
1
as [CpFe(CO)2]2 by a combination of IR, H NMR and mass
spectroscopic techniques.28 The analogous reaction carried out
between CyNCNCy and (diphenylamino)boryl complex 3c led to
no appreciable conversion over a period of 120 h at 60 ◦C in
toluene, as judged by 11B NMR spectroscopy.
Stoichiometric reaction of [CpFe(CO)2(BNCy2)]+[BArf4]− (4a)
with CyNCNCy: synthesis of [CpFe(CO)2C(NCy)2BNCy2]+-
[BArf4]− (5a). Addition of 1 equiv. of dicyclohexylcarbodi-
ion nanoelectrospray): m/z 518.1 (M − CO+, weak), correct
−
isotope pattern, (negative ion nanoelectrospray): 863.1 (BArf4
100%).
,
Table 1 Crystallographic data for 2a, 2b, 3c and 5
2a
2b
3c
5
Empirical formula
CCDC deposition number
Formula weight
C25H44BCl2N3
651106
468.34
C19H36BCl2N3
651107
388.22
150(2)
0.71073
Orthorhombic
Pna21
C19H15BClFeNO2
651108
391.43
150(2)
0.71073
Monoclinic
P21/c
C67H68B2F24FeN3O2
628665
1480.71
150(2)
0.71069
Monoclinic
P21/c
Temperature/K
150(2)
˚
Wavelength/A
0.71073
Monoclinic
C2/c
Crystal system
Space group
Unit cell dimensions:
˚
a/A
12.981(3)
16.321(3)
12.556(6)
90
103.50(3)
90
2586.7(9)
4
13.087(3)
9.288(2)
17.990(4)
90
90
90
2186.8(8)
4
7.708(2)
15.878(3)
14.297(3)
90
91.75(3)
90
1749.0(6)
4
1.487
19.098(5)
18.691(5)
23.205(5)
90
112.45(1)
90
7623(3)
4
1.290
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
3
˚
Volume/A
Z
Density(calc.)/Mg m−3
Absorption coefficient/mm−1
F(000)
1.203
0.269
1016
1.179
0.304
840
1.026
800
0.298
3036
Crystal size/mm3
0.40 × 0.10 × 0.10
3.00 to 26.00
−15 to 16, −19 to 20, −15
to 15
0.25 × 0.15 × 0.12
2.92 to 25.00
−14 to 15, −11 to 11, −21
to 18
0.30 × 0.20 × 0.06
0.30 × 0.15 × 0.15
h range for data collection/◦
Index ranges (h, k, l)
2.94 to 26.99
3.02 to 25.00
−9 to 9, −20 to 20, −18 to −22 to 22, −22 to 20, −27
18
to 27
No. of reflections collected
No. of indep reflns/Rint
Completeness to hmax (%)
Absorption correction
Max. and min. transmission
Refinement method
4812
2533 (0.0382)
99.5
11 678
3606 (0.0739)
99.6
7380
3811 (0.0347)
99.6
20 582
13 257 (0.0342)
98.6
Sortav
Sortav
Sortav
Sortav
0.975 and 0.844
Full matrix least squares
(F2)
0.984 and 0.653
Full matrix least squares
(F2)
3606/1/231
1.026
0.943 and 0.696
Full matrix least squares
(F2)
3811/0/226
1.050
0.956 and 0.834
Full matrix least squares
(F2)
13 257/96/1089
1.063
No. of data/restraints/params 2533/0/142
Goodness-of-fit on F2
Final R indices [I>2r(I)]
R indices (all data)
1.038
R1 = 0.0427, wR2 = 0.0951 R1 = 0.0415, wR2 = 0.0953 R1 = 0.0363, wR2 = 0.0747 R1 = 0.0687, wR2 = 0.1711
R1 = 0.0619, wR2 = 0.1036 R1 = 0.0502, wR2 = 0.1000 R1 = 0.0552, wR2 = 0.0816 R1 = 0.1026, wR2 = 0.1858
Largest diff. peak and
0.410 and −0.239
0.188 and −0.202
0.306 and −0.331
0.525 and −0.304
−3
˚
hole/e A
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4405–4412 | 4411
©

View Article Online
(c) Crystallographic method†
6 (a) Z. Lu, N. J. Hill, M. Findlater and A. H. Cowley, Inorg. Chim.
Acta, 2007, 360, 1316; (b) M. Findlater, N. J. Hill and A. H. Cowley,
Polyhedron, 2006, 25, 983; (c) N. J. Hill, M. Findlater and A. H. Cowley,
Dalton Trans., 2005, 3229.
7 (a) D. J. Brauer, S. Bucheim-Spiegel, H. Bu¨rger, R. Gielen, G. Pawelke
and J. Rothe, Organometallics, 1997, 16, 5321; (b) A. Ansorge, D. J.
Brauer, H. Bu¨rger, F. Do¨rrenbach, T. Hagen, G. Pawelke and W. Weuter,
J. Organomet. Chem., 1991, 407, 283.
8 For a preliminary communication of part of this work see: G. A. Pierce,
S. Aldridge, C. Jones, T. Gans-Eichler, A. Stasch, N. D. Coombs and
D. J. Willock, Angew. Chem., Int. Ed., 2007, 46, 2043.
9 R. Jefferson, M. F. Lappert, B. Prokai and B. P. Tilley, J. Chem. Soc.
A, 1966, 1585.
Data for 2a, 2b, 3c, 5a and CpFe(CO)2B(NBz2)Cl were collected
on an Bruker Nonius Kappa CCD diffractometer. Data collection
and cell refinement were carried out using DENZO and COL-
LECT; structure solution and refinement used SHELXS-97 and
SHELXL-97, respectively; absorption corrections were performed
using SORTAV.29 With the exception of CpFe(CO)2B(NBz2)Cl,
details of each data collection, structure solution and refinement
can be found in Table 1. Relevant bond lengths and angles are
included in the figure captions and complete details of each
structure have been deposited with the CCDC (numbers as listed
in Table 1).
10 E. Despagnet-Ayoub and R. H. Grubbs, J. Am. Chem. Soc., 2004, 126,
10198.
11 (a) S. Aldridge, R. J. Calder, D. L. Coombs, C. Jones, J. W. Steed, S.
Coles and M. B. Hursthouse, New J. Chem., 2002, 26, 677; (b) D. L.
Coombs, PhD thesis, Cardiff University, 2003.
CCDC reference numbers 651106–651108, 628665.
For crystallographic data in CIF or other electronic format see
DOI: 10.1039/b709249k
12 C. Jones, S. Aldridge, T. Gans-Eichler and A. Stasch, Dalton Trans.,
2006, 5357.
13 (a) S. Aldridge, C. Jones, T. Gans-Eichler, A. Stasch, D. L. Kays (ne´e
Coombs), N. D. Coombs and D. J. Willock, Angew. Chem., Int. Ed.,
2006, 45, 6118; (b) D. L. Kays (ne´e Coombs), S. Aldridge, J. K. Day
and L.-L. Ooi, Angew. Chem., Int. Ed., 2005, 44, 7457; (c) For a related
(dimethylamino)boryl derivative see: H. Braunschweig, C. Kollann and
U. Englert, Eur. J. Inorg. Chem., 1998, 465.
(d) Computational method
Geometry optimizations and estimates of r/p bonding density
using DFT methods employed a recently reported methodology.17
14 G. J. Bullen and N. H. Clark, J. Chem. Soc. A, 1970, 992.
15 D. L. Kays (ne´e Coombs), J. K. Day, S. Aldridge, R. W. Harrington
and W. Clegg, Angew. Chem., Int. Ed., 2006, 45, 3513.
16 L. H. Gade, H. Memmler, U. Kauper, A. Schneider, S. Fabre, I.
Bezougli, M. Lutz, C. Galka, I. J. Scowen and M. McPartlin, Chem.–
Eur. J., 2000, 6, 692.
Acknowledgements
We would like to acknowledge financial support from the EPSRC
and assistance from the EPSRC National Mass Spectroscopy
Service Centre, University of Wales, Swansea.
17 (a) S. Aldridge, A. Rossin, D. L. Coombs and D. J. Willock, Dalton
Trans., 2004, 2649; (b) A. A. Dickinson, D. J. Willock, R. J. Calder and
S. Aldridge, Organometallics, 2002, 21, 1146.
18 N. J. Hill, J. A. Moore, M. Findlater and A. H. Cowley, Chem. Commun.,
References
2005, 5462.
1 For relevant review articles, see for example: (a) P. C. Junk and M. L.
Cole, Chem. Commun., 2007, 1579; (b) F. T. Edelmann, Coord. Chem.
Rev., 1994, 137, 403; (c) J. Barker and M. Kilner, Coord. Chem. Rev.,
1994, 133, 219.
2 (a) S. P. Green, C. Jones and A. Stasch, Inorg. Chem., 2007, 46, 11;
(b) C. Jones, P. C. Junk, J. A. Platts and A Stasch, J. Am. Chem. Soc.,
2006, 128, 2206.
3 (a) R. J. Meier and E. Coglin, J. Phys. Chem. A, 2001, 105, 3867;
(b) S. Dagorne, I. A. Guzei, M. P. Coles and R. F. Jordan, J. Am.
Chem. Soc., 2000, 122, 274; (c) G. Talarico and P. H. M. Budzelaar,
Organometallics, 2000, 19, 5691; (d) S. Dagorne, R. F. Jordan and V. G.
Young, Organometallics, 1999, 18, 4619; (e) M. P. Coles, D. C. Svenson,
R. F. Jordan and V. G. Young, Organometallics, 1998, 17, 4042; (f) M. P.
Coles, D. C. Svenson, R. F. Jordan and V. G. Young, Organometallics,
1997, 16, 5183; (g) R. Duchateau, A. Meetsma and J. H. Teuben, Chem.
Commun., 1996, 223.
4 J. Barker, N. C. Blacker, P. R. Phillips, N. W. Alcock, W. Errington and
M. G. H. Wallbridge, J. Chem. Soc., Dalton Trans., 1996, 431.
5 (a) C. Ergezinger, F. Weller and K. Dehnicke, Z. Naturforsch., B:
Chem. Sci., 1988, 43, 1621; (b) P. Blais, T. Chivers, A. Downward
and M. Parvez, Can. J. Chem., 2000, 78, 10; (c) M. R. Terry, L. A.
Mercando, C. Kelley, G. L. Geoffroy, P. Nombel, N. Lugen, R.
Mathieu, R. L. Ostrander, B. E. Owens-Whitmire and A. L. Rheingold,
Organometallics, 1994, 13, 843.
19 S. De, P. Parameswaran and E. D. Jemmis, Inorg. Chem., 2007, 46,
6091.
20 D. J. Grant and D. A. Dixon, J. Phys. Chem. A, 2006, 110, 12955.
21 C. N. Rowley, G. A. DiLabio and S. T. Barry, Inorg. Chem., 2005, 44,
1983.
22 R. H. Cragg and T. J. Miller, J. Organomet. Chem., 1983, 255, 143.
23 (a) S. Aldridge and D. L. Kays, Main Group Chem., submitted; (b) N. D.
Coombs and S. Aldridge, unpublished results.
24 (a) R. B. King and M. B. Bisnette, J. Organomet. Chem., 1967, 8, 287;
(b) R. B. King, Acc. Chem. Res., 1970, 3, 417.
25 J. Casanova and M. Geisel, Inorg. Chem., 1973, 13, 2783.
26 W. Gerrard, H. R. Hudson and E. F. Mooney, J. Chem. Soc., 1960,
5168.
27 W. Maringgele, M. Noltemeyer and A. Meller, Organometallics, 1997,
16, 2276.
28 R. D. Fischer, A. Vogler and K. Noack, J. Organomet. Chem., 1967, 7,
135.
29 (a) Denzo Z. Otwinowski and W. Minor, in Methods in Enzymology,
ed. C. W. Carter and R. M. Sweet, Academic Press, New York, 1996,
vol. 276, p. 307; (b) R. Hooft, Data Collection Software, B. V. Nonius,
Delft, 1998; (c) G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, A46,
467; (d) G. M. Sheldrick, SHELX97: Programs for Crystal Structure
Analysis (Release 97–2), University of Gottingen, Gottingen, 1997;
¨
¨
(e) Sortav R. H. Blessing, Acta Crystallogr., Sect. A, 1995, A51, 33.
4412 | Dalton Trans., 2007, 4405–4412
This journal is
The Royal Society of Chemistry 2007
©