Linking of metal centres through boryl ligands: Synthesis, spectroscopic and structural characterisation of symmetrically bridged boryl complexes

Article abstract of DOI:10.1039/b200695m

Dinuclear species containing iron centres linked via various bridging boryl ligands have been synthesised and a combination of crystallographic, computational and spectroscopic (Moessbauer, IR, Raman) techniques have been used to probe the bonding in these complexes.

Full text of DOI:10.1039/b200695m

View Article Online / Journal Homepage / Table of Contents for this issue
Linking of metal centres through boryl ligands: synthesis,
spectroscopic and structural characterisation of symmetrically
bridged boryl complexes
Simon Aldridge,*a Richard J. Calder,a Andrea Rossin,a Anthony A. Dickinson,a
David J. Willock,a Cameron Jones,a David J. Evans,b Jonathan W. Steed,c Mark E. Light,d
Simon J. Colesd and Michael B. Hursthoused
a Dept. of Chemistry, Cardiff University, PO Box 912, Park Place, Cardiff, UK CF10 3TB
b Dept. of Biological Chemistry, John Innes Centre, Norwich Research Park, Colney, Norwich,
UK NR4 7UH
c Dept. of Chemistry, King’s College London, Strand, London, UK WC2R 2LS
d Dept. of Chemistry, University of Southampton, Highfield, Southampton, UK SO17 1BJ
Received 18th January 2002, Accepted 19th February 2002
First published as an Advance Article on the web 27th March 2002
Dinuclear species containing iron centres linked via various bridging boryl ligands have been synthesised and a
combination of crystallographic, computational and spectroscopic (Mössbauer, IR, Raman) techniques have been
used to probe the bonding in these complexes.
Solvents were pre-dried over sodium wire (hexanes, toluene) or
molecular sieves (dichloromethane) and purged with nitrogen
prior to distillation. Hexanes (potassium), toluene (sodium),
and dichloromethane (calcium hydride) were then distilled from
the appropriate drying agent before use. C6D6 (Goss) was
degassed and dried over potassium prior to use; trimethylamine
was dried over sodium wire prior to use. BCl3 (1.0 M solution in
heptane, Aldrich), pentaerythrol (Avocado) and trimethylsilyl
chloride (Aldrich) were used as received, without further purifi-
cation. (η5-C5R4RЈ)Fe(CO)2Na (R = RЈ = H; R = H, RЈ = Me;
and R = RЈ = Me) and hexahydroxybenzene (5) were pre-
pared by minor modification of literature methods;10,11 1,2,4,5-
(Me3SiO)4C6H2 (2a), (ClBO2)2C6H2 (3a) and (HBO2)2C6H2 (3c)
were prepared as described previously.8
1
Introduction
Transition metal boryl complexes (LnM–BX2) have been the
subject of considerable research interest over the past decade,1
partly as a result of their implication in several highly useful
organic transformations, including the metal-catalysed hydro-
and diboration of carbon to element multiple bonds.1,2 More
recently the highly selective functionalization of unactivated
hydrocarbons to boronate esters through the intermediacy of
metal boryl species has also been reported by a number of
groups.3,4 It has been suggested that the unusual regiochemistry
and activity of such systems with regard to C–H activation may
be due to the Lewis acidic properties of the boryl ligand, which
provide favourable kinetics for the formation of boron–carbon
bonds.3d
While the unusual reactivity of transition metal boryl com-
plexes has provided the impetus for much research effort, there
has also been considerable interest in determining the fund-
amental ligand properties of boryl systems through a com-
bination of spectroscopic, crystallographic and computational
approaches.1,5,6 Ultimately a better understanding of the nature
of the metal boron interaction may help provide insight into the
underlying reasons for the unusual reactivity of such systems.
Almost exclusively, however, such studies have focussed on
monodentate boryl ligands (often Bcat, BO2C6H4) adopting a
terminal mode of coordination with respect to the metal centre,
with alternative ligand types (e.g. bridging, chelating, base-
stabilised) being almost totally ignored.6d,7 In this and other
recent work we have sought to expand the coordination chem-
istry of these highly topical ligand systems,5o,6i,6j,8,9 and hereby
report the synthesis, spectroscopic and structural investigation
of a number of dinuclear iron systems featuring bridging boryl
ligands. Such complexes not only offer useful comparison with
analogous terminally bound ligand systems, but also provide a
platform for further investigation of the metal to boryl linkage
through previously under-utilized techniques (DFT, electro-
chemistry, Mössbauer and FT-Raman spectroscopies).
NMR spectra were measured on a Bruker AM-400 or JEOL
Eclipse 300 Plus FT-NMR spectrometer. Residual protons of
1
solvent were used for reference for H and 13C NMR, while a
sealed tube containing a solution of [(nBu4N)(B3H8)] in CDCl3
was used as an external reference for 11B NMR. Infrared spec-
tra were measured for each compound pressed into a disk with
an excess of dried KBr on a Nicolet 500 FT–IR spectrometer.
FT-Raman spectra were measured for powdered samples sealed
in glass ampoules using a LabRam spectrometer. Mass spectra
were measured by the EPSRC National Mass Spectrometry
Service Centre, University of Wales Swansea and by the depart-
mental service. Perfluorotributylamine was used as the standard
for high-resolution EI mass spectra. Mössbauer spectra were
recorded at 77 K on an ES-Technology MS-105 spectrometer
with a 135 MBq 57Co source in a rhodium matrix at ambient
temperature. Spectra were referenced to a 25 µm iron foil at 298
K. Measurements were in zero field on solid samples ground
with boron nitride. Parameters were obtained by fitting the
spectra by Lorenzian curves. Elemental analyses were carried
out both by the departmental analysis service and by Warwick
Analytical Service, University of Warwick.
Abbreviations: s = strong, m = medium, w = weak, sh =
shoulder, s = singlet, m = multiplet, br = broad.
Syntheses
2
Experimental
All manipulations were carried out under a nitrogen or argon
atmosphere using standard Schlenk-line or dry-box techniques.
(Me3SiOCH2)4C (2b). 5.4 g (39.7 mmol) of pentaerythrol (1b)
were suspended in toluene (100 cm3) and 10 equiv. of trimethyl-
2020
J. Chem. Soc., Dalton Trans., 2002, 2020–2026
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b200695m

View Article Online
silyl chloride (50 cm3, 397 mmol) and 10 equiv. (55.3 cm3,
397 mmol) of triethylamine were then added by syringe to the
rapidly stirred reaction mixture. After 12 h at room temperature
the supernatent toluene solution was separated from the
(Et3NH)Cl precipitate by filtration. The precipitate was washed
with toluene (2 × 50 cm3) and the combined washings reduced
to dryness in vacuo yielding 2b as a colourless oil (ca. 70%
CH3), 96.0 (aromatic CH), 145.3 (aromatic quaternary), 214.3
(CO). 11B NMR (96 MHz, toluene, 21 ЊC), δ 51 (br). IR (KBr
disk, cmϪ1) ν(CO) 2001s, 1943s. Mass spectrum (CI): [M ϩ H]ϩ
= 543 (100%), expected isotopic distribution for 2 B, 2 Fe atoms,
exact mass (calc.) m/z 542.9808, (observed) 542.9809. Elemental
analysis: calc. for C22H16B2Fe2O8, C 48.78, H 2.98; observed, C
1
48.85, H 2.94%. 4c: H NMR (400 MHz, C6D6, 21 ЊC), δ 1.89
yield). Examination of the product at this point by 1H and 13
C
(30H, s, η5-C5Me5), 7.08 (s, 2H, (η5-C5Me5)Fe(CO)2BO2C6-
H2O2BFe(CO)2(η5-C5Me5)). 13C NMR (76 MHz, C6D6, 21 ЊC),
δ 10.2 (η5-C5Me5), 94.7 (aromatic CH), 95.8 (C5Me5), 143.8
(aromatic quaternary), 213.3 (CO). 11B NMR (96 MHz, tolu-
ene, 21 ЊC), δ 53 (br). IR (KBr disk, cmϪ1) ν(CO) 2001s, 1943s.
Mass spectrum (EI): [M]ϩ = 654 (30%), expected isotopic
distribution for 2 B, 2 Fe atoms, fragment ion peaks at m/z 626,
598 and 542, corresponding to sequential loss of one, two and
four CO molecules, exact mass (calc.) m/z 654.0982, (observed)
654.0980.
NMR revealed it to be >99% pure and no further purification
1
was therefore attempted. 2b was characterised by H and 13C
NMR and IR spectroscopy, and CI mass spectrometry (includ-
1
ing exact mass determination). H NMR (400 MHz, C6D6, 21
ЊC), δ 0.14 (36H, s, Si(CH3)3), 3.70 (8H, s, C(CH2)4). 13C NMR
(76 MHz, C6D6, 21 ЊC), δ Ϫ0.7 (Si(CH3)3), 47.3 (C(CH2)4), 59.9
(C(CH2)4). IR (neat, cmϪ1): 2959m, 2919m, 2876m, 1475m,
1402w, 1304w, 1249s, 1172m, 1074s, 909s, 883s, 747m, 728m,
694m. Mass spectrum (CI): [M ϩ H]ϩ = 425 (100%), exact mass
(calc.) m/z 425.2395, (observed) 425.2394.
Spiro-[(ꢀ5-C5H5)Fe(CO)2BO2(CH2)2]2C 4d. Complex 4d was
synthesized from 3b in a manner analogous to that described
above for 4c. Purification of the final product was achieved by
recrystallization from toluene at Ϫ30 ЊC and crystals suitable
for X-ray diffraction were grown by layering a toluene solution
with hexanes. Yields of the pale yellow crystalline material are
typically of the order of 40% and 4d has been characterized by
1H, 13C and 11B NMR, IR spectroscopy, EI mass spectrometry
and single crystal X-ray diffraction. 1H NMR (400 MHz, C6D6,
21 ЊC), δ 3.56 (8H, s, C(CH2)4), 4.26 (10H, s, η5-C5H5). 13C
NMR (76 MHz, C6D6, 21 ЊC), δ 36.6 (C(CH2)4), 65.7
(C(CH2)4), 83.3 (η5-C5H5), 216.0 (CO). 11B NMR (96 MHz,
C6D6, 21 ЊC), δ 45.3. IR (KBr disk, cmϪ1) ν(CO) 1998s, 1932s.
Mass spectrum (EI): [M Ϫ CO]ϩ = 480 (weak), expected
isotopic distribution for 2B, 2Fe atoms, fragment ion peaks at
m/z 452 (30%), 424 (weak) and 396 (20%) corresponding to
sequential loss of the three remaining CO ligands.
Spiro-[ClBO2(CH2)2]2C 3b. To a solution of 11.7 g (27.6
mmol) of 2b in hexanes at room temperature was added drop-
wise by syringe 2 equiv. of BCl3 (55 cm3 of a 1.0 M solution in
heptane, 55 mmol). The reaction mixture was warmed to 55 ЊC
and stirred for 12 h, after which the white precipitate so formed
was separated from the supernatent solution by filtration,
washed with hexanes (3 × 30 cm3) and dried in vacuo. The crude
material was then recrystallized from toluene to give 3b as a
white microcrystalline solid in yields of up to 86%. 3b was char-
acterised by 1H, 13C and 11B NMR and IR spectroscopy, and EI
1
mass spectrometry. H NMR (400 MHz, C6D6, 21 ЊC), δ 2.90
(8H, s, C(CH2)4). 13C NMR (76 MHz, C6D6, 21 ЊC), δ 35.2
(C(CH2)4), 65.1 (C(CH2)4). 11B NMR (96 MHz, C6D6, 21 ЊC),
δ 23.1. IR (KBr disk, cmϪ1): 2963m, 2908w (sh), 1490m, 1436m,
1374m, 1262s, 1097s, 1021s, 864w, 801s. Mass spectrum (EI):
[M]ϩ = 225 (weak).
(ꢀ5-C5R4RЈ)Fe(CO)2BO2C6H2O2BFe(CO)2(ꢀ5-C5R4RЈ) (R ؍
RЈ ؍ H 4a; R ؍ H, RЈ ؍ Me 4b; and R ؍ RЈ ؍ Me 4c). The
three complexes were synthesized in a similar manner. To a
suspension of (η5-C5H5)Fe(CO)2Na (0.2 g, 1.0 mmol) in toluene
at Ϫ30 ЊC was added a toluene solution containing 0.5 equiv. of
3a. The reaction mixture was warmed to room temperature and
stirred for one week at which time examination of the solution
by 11B NMR spectroscopy revealed that all of 3a had been
consumed. In each case, the product is only sparingly soluble in
toluene and can therefore be isolated by removal of the super-
natent by filtration, extraction of the residual beige solid with
CH2Cl2 and subsequent crystallisation either by controlled
cooling or by layering with hexanes at Ϫ30 ЊC. This method
generated crystals of 4a and 4b suitable for X-ray diffraction.
Isolated yields of the pale yellow crystalline solids 4a–c are
typically of the order of 50–60% and the compounds have been
characterized by 1H, 13C and 11B NMR, IR spectroscopy, mass
spectrometry, elemental analysis and (in the cases of 4a and 4b)
(Me3SiO)6C6 6. Compound 6 was prepared from hexa-
hydroxybenzene (5)11 using a method analogous to that
described above for 2b. The final product was isolated as a pale
pink crystalline solid in yields of up to 65% and has been char-
acterized by 1H and 13C NMR, EI mass spectrometry (including
exact mass determination) and single crystal X-ray diffraction.
1H NMR (400 MHz, C6D6, 21 ЊC), δ 0.40 (s, SiMe3). 13C NMR
(76 MHz, C6D6, 21 ЊC), δ 0.0 (SiMe3), 134.0 (aromatic quater-
nary). Mass spectrum (EI): [M]ϩ = 606 (100%), exact mass
(calc.) m/z 606.2530, (observed) 606.2536.
Attempts to synthesize (ClBO2)3C6 7. Attempts were made to
synthesise 7 from 6 using a method analogous to that used to
prepare 3a and 3b. Addition of 3 equivalents of BCl3 to a solu-
tion of 6 in hexanes however produced a mixture of two boron-
containing species giving rise to resonances at δ 33.0 and 30.1,
together with unreacted BCl3. The use of elevated reaction
temperatures (50–60 ЊC) or longer reaction times (120 h) did
not lead to significant changes in the product distribution.
Removal of volatiles in vacuo and examination of the resulting
1
by single crystal X-ray diffraction. 4a: H NMR (400 MHz,
C6D6, 21 ЊC), δ 4.95 (10H, s, η5-C5H5), 6.99 (s, 2H, (η5-C5H5)-
Fe(CO)2BO2C6H2O2BFe(CO)2(η5-C5H5)). 13C NMR (76 MHz,
C6D6, 21 ЊC), δ 82.9 (η5-C5H5), 94.7 (aromatic CH), 143.8
(aromatic quaternary), 212.5 (CO). 11B NMR (96 MHz, tolu-
ene, 21 ЊC), δ 48 (br). IR (KBr disk, cmϪ1) ν(CO) 2006s, 1954s.
Mass spectrum (EI): [M]ϩ = 514 (100%), expected isotopic
distribution for 2B, 2Fe atoms, fragment ion peaks at m/z
486, 458, 430, 402 corresponding to sequential loss of four
CO molecules, exact mass (calc.) m/z 513.9417, (observed)
513.9424. Elemental analysis: calc. for C20H12B2Fe2O8, C 46.74,
1
solid by H and 13C NMR revealed the presence of unreacted
trimethylsilyl functions within the product mixture.
General crystallographic method
Data were collected on either an Enraf Nonius Kappa CCD
diffractometer (3a, 3c, 4a, 4b, 4d) or CAD4 four-circle diffract-
ometer (6). For the former structures data collection and cell
refinement were carried out using DENZO and COLLECT,12
and structure solution and refinement using SHELXS-97 and
SHELXL-97, respectively.13 Details of each data collection,
structure solution and refinement can be found in Table 1,
relevant bond lengths and angles are included in figure
captions.
1
H 2.35; observed, C 46.22, H 2.14%. 4b: H NMR (400 MHz,
C6D6, 21 ЊC), δ 1.43 (s, 6H, η5-C5H4CH3 4.14 (4H, m, η5-C5H4-
CH3), 4.26 (4H, m, η5-C5H4CH3), 7.24 (s, 2H, (η5-C5H4CH3)-
Fe(CO)2BO2C6H2O2BFe(CO)2(η5-C5H4CH3)). 13C NMR (76
MHz, C6D6, 21 ЊC), δ 12.7 (η5-C5H4CH3), 82.7, 84.4 (η5-C5H4-
J. Chem. Soc., Dalton Trans., 2002, 2020–2026
2021

View Article Online
Table 1 Crystallographic data for 3a, 3c, 4a, 4b, 4d and 6
3a
3c
4a
4b
4d
6
Empirical formula C6H2B2Cl2O4
C6H4B2O4
161.71
150(2)
0.71073
Orthorhombic
Pbca
10.416(2)
5.2697(11)
12.231(2)
90
90
90
671.3(2)
4
1.600
C20H12B2Fe2O8
513.62
100(2)
0.71073
Monoclinic
P21/n
6.4542(3)
12.2543(4)
12.4180(6)
90
93.604(3)
90
980.22(7)
2
1.740
1.528
516
C22H16B2Fe2O8
541.67
120(2)
C19H18B2Fe2O8
507.66
120(2)
0.71073
Monoclinic
C2/c
26.798(2)
6.0907(6)
12.6373(15)
90
108.968(3)
90
1950.7(3)
4
1.729
1.534
C24H54O6Si6
607.21
150(2)
0.71073
Monoclinic
C2/m
18.520(2)
17.898(2)
12.451(3)
90
96.00(2)
90
4104.3(12)
4
M
T /K
230.60
100(2)
λ/Å
0.71073
Triclinic
P1
0.71073
Crystal system
Space group
a/Å
b/Å
c/Å
α/Њ
β/Њ
Triclinic
¯
¯
P1
4.4108(9)
6.7612(14)
7.5147(15)
105.64(3)
103.88(3)
91.38(3)
208.56(7)
1
6.4039(13)
6.8417(14)
12.547(3)
83.82(3)
86.64(3)
77.38(3)
532.97(19)
1
γ/Њ
V/Å3
Z
Dc/Mg mϪ3
µ/mmϪ1
F(000)
1.836
0.753
114
1.688
1.410
274
0.983
0.230
1320
0.128
328
1032
Crystal size/mm
Reflections
collected
Independent
reflections
Rint
Goodness-of-
fit on F 2
Final R indices
[I > 2σ(I )]
0.40 × 0.10 × 0.05 0.25 × 0.13 × 0.13 0.30 × 0.15 × 0.15 0.44 × 0.28 × 0.12 0.15 × 0.15 × 0.02 0.60 × 0.50 × 0.50
1535
6589
6704
8293
4641
3943
949
768
2243
2401
1649
3817
0.0492
1.099
0.0363
0.991
0.0340
1.029
0.0764
1.051
0.0580
1.089
0.0257
1.104
R1 = 0.0419,
wR2 = 0.0779
R1 = 0.0367,
wR2 = 0.1077
R1 = 0.0419,
wR2 = 0.1065
0.186, Ϫ0.222
R1 = 0.0257,
wR2 = 0.0582
R1 = 0.0315,
wR2 = 0.0605
0.331, Ϫ0.225
R1 = 0.0328,
wR2 = 0.0830
R1 = 0.0340,
wR2 = 0.0840
0.0672, Ϫ0.644
R1 = 0.0427,
wR2 = 0.0886
R1 = 0.0732,
wR2 = 0.1054
0.474, Ϫ0.412
R1 = 0.0689,
wR2 = 0.2193
R1 = 0.1033,
wR2 = 0.2391
0.896, Ϫ0.484
R indices/all data) R1 = 0.0630,
wR2 = 0.0854
Largest diff. peak 0.341, Ϫ0.382
and hole/e ÅϪ3
)
CCDC reference numbers 142393 (see ref. 8), 148084 (see
ref. 6i) and 177260–177263.
See http://www.rsc.org/suppdata/dt/b2/b200695m/ for crys-
tallographic data in CIF or other electronic format.
3
Results and discussion
Synthetic methodology
Dinuclear complexes featuring metal centres linked via bridging
boryl ligands can be synthesised in yields of 50–60% according
to the synthetic route outlined in Scheme 1. This methodology
can be applied both to unsaturated bridging ‘spacer’ groups
such as those based around a 1,2,4,5-tetrasubstituted benzene
framework, and also to saturated aliphatic systems such as the
spiro species derived from pentaerythrol precursors. The bifunc-
tional boron halide reagents (3a, 3b) required in the final
metathesis step are most conveniently prepared from trimethyl-
silyl-substituted precursors (i.e. 2a, 2b) which are freely soluble
in the non-polar organic media employed. By contrast, direct
synthesis from polyhydroxy species and boron trichloride, for
example, suffers from low yields resulting from the difficulty in
adequately drying compounds such as 1a and 1b.
The final metathetical step in the reaction sequence requires
the use of samples of the organometallic anion from which
traces of tetrahydrofuran have been rigorously removed (thf is a
contaminant inherent in the synthetic route used). Yields of the
final products 4a–d are otherwise significantly reduced (con-
ceivably, this may be due to the susceptibility of metal boryl
complexes to decomposition in the presence of nucleophilic/
Lewis basic reagents1a or to the direct reaction of the chloro-
borane precursors with thf ). Not unexpectedly, dinuclear
species 4a–d are significantly less soluble in organic media
than their mononuclear counterparts, with those containing
unsaturated spacer groups (4a–c) being less soluble than those
such as 4d containing saturated bridging units. Attempts to
prepare dinuclear boryl complexes containing other organo-
metallic fragments [e.g. Mn(CO)5 or (η5-C5H5)2M(H) (M = Mo
Scheme 1
or W)] suffered from a lack of solubility of the products in
solvents which did not bring about decomposition. Finally,
attempts to prepare the corresponding trinuclear derivatives
through the intermediacy of the hexa-substituted trimethylsilyl
derivative 6 (Fig. 1) appear to suffer from problems arising from
the isolation of a mixture of products during the synthesis of
the required polyfunctional boron halides (for which the
chemistry outlined in Scheme 2 is proposed).
Structural and spectroscopic studies
Single crystals of metal complexes 4a, 4b and 4d suitable for
X-ray crystallography proved to be accessible by layering with
2022
J. Chem. Soc., Dalton Trans., 2002, 2020–2026

View Article Online
Fig.
2
Molecular structure of (η5-C5H5)Fe(CO)2BO2C6H2O2BFe-
(CO)2(η5-C5H5), 4a. Relevant bond lengths (Å) and angles (Њ): Fe(1)–
B(1) 1.971(2), Fe(1)–C(10) 1.751(2), Fe(1)–Cp 1.721(2), B(1)–O(1)
1.406(2); C(9)–Fe(1)–C(10) 93.97(8), O(1)–B(1)–O(2) 109.15(14), O(1)–
B(1)–Fe(1) 121.94(12), O(2)–B(1)–Fe(1) 125.79(13), O(1)–B(1)–Fe(1)–
Cp 82.2(1). Cp = (η5-C5H5) centroid.
Fig. 1 Molecular structure of one of the two independent molecules
of (Me3SiO)6C6, 6. Ranges of relevant bond lengths (Å) and angles (Њ):
Si–O 1.655(3)–1.673(4), Si–C 1.780(8)–1.936(9), C–O 1.370(6)–
1.389(6), O–Si–C 102.6(2)–110.5(2). Each independent molecule has
crystallographically imposed 2/m symmetry.
Fig. 3 Molecular structure of (η5-C5H4Me)Fe(CO)2BO2C6H2O2BFe-
(CO)2(η5-C5H4Me), 4b. Relevant bond lengths (Å) and angles (Њ):
Fe(1)–B(1) 1.973(2), Fe(1)–C(7) 1.753(2), Fe(1)–Cp 1.718(3), B(1)–O(3)
1.403(2); C(7)–Fe(1)–C(8) 95.56(9), O(3)–B(1)–O(4) 109.07(15), O(3)–
B(1)–Fe(1) 124.67(14), O(4)–B(1)–Fe(1) 126.15(13), O(3)–B(1)–Fe(1)–
Cp 87.7(2).
Fig.
4
Molecular structure of (η5-C5H5)Fe(CO)2BO2C5H8O2BFe-
(CO)2(η5-C5H5), 4d. Relevant bond lengths (Å) and angles (Њ): Fe(1)–
B(1) 2.030(5), Fe(1)–C(7) 1.743(5), Fe(1)–Cp 1.722(6), B(1)–O(3)
1.369(6); C(6)–Fe(1)–C(7) 95.1(2), O(3)–B(1)–O(4) 121.2(4), O(3)–
B(1)–Fe(1) 118.7(3), O(4)–B(1)–Fe(1) 120.2(3), Fe(1)–C(10)–Fe(1Ј)
130.4(4), O(3)–B(1)–Fe(1)–Cp 43.5(3). Each molecule of 4d has
crystallographically imposed two-fold symmetry.
Scheme 2
hexanes of concentrated solutions in either dichloromethane
(4a, 4b) or toluene (4d). In each case the molecular structure
consists of two piano stool (η5-C5H5)Fe(CO)2 fragments linked
in µ2,η1,η1 fashion by a bridging boryl ligand (see Figs. 2–4 and
Table 1). In the case of 4a and 4b the molecule is centro-
symmetric (sitting on a crystallographically imposed inversion
centre) with the geometry of the ligand itself differing little
from that found in chloride or hydride derivatives 3a or 3c
(Figs. 5 and 6). The molecular structures of 4a and 4b are very
similar, with essentially identical Fe–B distances, and a small
(ca. 6Њ) difference in the orientation of the boryl ligand. By
contrast, the molecular structure of 4d is bent [Fe(1)–C(10)–
Fe(1Ј) 130.4(4)Њ] reflecting the conformation of the six-
membered chelate rings, and in particular the angle [132.3(5)Њ]
between the planes defined by C(9), C(10), C(11) and O(3),
B(1), O(4).
Fig. 5 Molecular structure of ClBO2C6H2O2BCl, 3a. Relevant bond
lengths (Å) and angles (Њ): B(1)–Cl(1) 1.737(3), B(1)–O(1) 1.377(3),
B(1)–O(2) 1.377(3), O(1)–B(1)–O(2) 113.4(2), O(1)–B(1)–Cl(1)
123.0(2), O(2)–B(1)–Cl(1) 123.6(2). Each molecule of 3a has
crystallographically imposed inversion symmetry.
Of particular interest is the significantly longer Fe–B bond
length found in 4d [2.030(5) Å] compared to those found in 4a
[1.971(2) Å], 4b [1.973(2) Å] and in the terminally bound Bcat
analogue (η5-C5H5)Fe(CO)2Bcat [1.959(6) Å].6a The latter three
compounds all feature five-membered BO2C2 chelate rings in
J. Chem. Soc., Dalton Trans., 2002, 2020–2026
2023

View Article Online
O2B ligands may well be due to factors such as crystal packing
in the solid state.
For systems of the type (η5-C5H5)Fe(CO)2X, 57Fe Mössbauer
spectroscopy offers an extensively used alternative to crystallo-
graphic methods to probe the electronic properties of a ligand
X.16 In general, a decrease in the Mössbauer effect isotope shift
(i.s.) corresponds to an increase in the s electron density at the
iron nucleus. Such an increase usually results from the presence
of a ligand (X) which has good σ donor or π acceptor
properties.16
The 57Fe Mössbauer spectra of the dinuclear complexes 4a
and 4d at 77 K are reproduced in Fig. 7, and associated spectral
Fig. 6 Molecular structure of HBO2C6H2O2BH, 3c. Relevant bond
lengths (Å) and angles (Њ): B(1)–H(1) 1.100, B(1)–O(1) 1.386(2), B(1)–
O(2) 1.382(2), O(1)–B(1)–O(2) 111.9(1), O(1)–B(1)–H(1) 123.4(4),
O(2)–B(1)–H(1) 124.7(4). Each molecule of 3c has crystallographically
imposed inversion symmetry.
which conjugation of the appropriate symmetry lone pairs of
the oxygen centres [e.g. O(1) and O(2) in 4a] into the aromatic
ring system is possible. In the case of 4d, however, no such
conjugation is possible, and as a consequence O
B π
donation is likely to be more significant. This in turn renders
the boryl boron centre less π acidic, and consequently the
extent of π back donation from the iron centre is reduced.
Hence the Fe–B distance in 4d is similar to that found in
[(η5-C5H4Me)Fe(CO)2]2B3ClN3H3 [mean 2.043(1) Å6d] and
(η5-C5Me5)Fe(CO)2B(NMe2)Cl [2.027(5) Å6e] where little
if any π back bonding is thought to exist. A further con-
sequence is that the B–O distances in 4d [mean 1.365(5) Å] are
significantly shorter than those found in 4a [1.406(2) Å] or 4b
[1.408(2) Å].
The Fe–B distances found in compounds 4a, 4b and the
terminally bound Bcat analogue (η5-C5H5)Fe(CO)2Bcat are
among the shortest known for iron boryl complexes [mean
2.008 Å]. In the case of (η5-C5H5)Fe(CO)2Bcat the short Fe–B
distance [1.959(6) Å] and angle (ϑ) of 7.9Њ between Cp–Fe–B
[Cp = (η5-C5H5) centroid] and BO2 planes have been ascribed to
a π back bonding interaction between the HOMO of the
ϩ
(η5-C5H5)Fe(CO)2 fragment (an aЉ symmetry orbital lying
parallel to the plane of the cyclopentadienyl ligand14) and the
LUMO of the boryl ligand fragment.6a Dinuclear compounds
4a and 4b display a different orientation of the boryl ligand
with respect to the Cp–Fe–B plane [ϑ = 82.2(1) and 87.7(2)Њ,
respectively]. As a consequence any π back bonding interaction
in these complexes must originate not in the metal-based
HOMO but in the HOMOϪ2 orbital of the (η5-C5H5)-
ϩ
Fe(CO)2 fragment (a deeper lying aЈ symmetry orbital per-
pendicular to the HOMO). Furthermore the magnitude of this
π interaction is clearly smaller in 4a and 4b than in (η5-C5H5)-
Fe(CO)2Bcat, as reflected by the significantly lower carbonyl
stretching frequencies for the dinuclear species [ν(CO) = 2006
and 1954; 2001 and 1943; 2024 and 1971 cmϪ1 for 4a, 4b and
(η5-C5H5)Fe(CO)2Bcat, respectively].
Fig. 7 57Fe Mössbauer spectra of (a) 4a and (b) 4d at 77 K.
parameters listed in Table 2. The spectrum of (η5-C5H5)-
Fe(CO)2Br was also measured and spectral parameters are
in good agreement with those reported previously.16b The very
low values of the isomer shifts for both 4a (0.02 mm sϪ1) and 4d
(0.00 mm sϪ1) are comparable to those reported for complexes
containing either good π acceptor ligands (e.g. CNϪ, 0.07 mm
Computational studies carried out using Density Functional
Theory (DFT) for these complexes and others,5o however, reveal
that in general alkoxo-substituted boryl complexes [LnM–
B(OR)2] display a relatively minor π contribution to bonding
[e.g. 10.9% for (η5-C5H5)Fe(CO)2Bcat].5o In addition, an ener-
getic profile for rotation about the Fe–B bond calculated for the
s
Ϫ1 16b), or for those containing good σ donor ligands (e.g.
model compound (η5-C H )Fe(CO) BO C H (featuring a C᎐C
SiMe3, 0.05 mm sϪ1 and CH3, 0.08 mm sϪ1 16c). That the bridg-
ing boryl ligands in 4a and 4d fit into the latter category as
giving low isomer shifts due to strong σ donor rather than
strong π acceptor properties is confirmed by analysis of the IR
data listed in Table 2. The low carbonyl stretching frequencies
found for 4a and 4d (2006, 1954 and 1998, 1932 cmϪ1, respect-
ively) are clearly more consistent with strong σ donor properties
[cf. 2010, 1958 and 1996, 1994 cmϪ1 for (η5-C5H5)Fe(CO)2CH3
and (η5-C5H5)Fe(CO)2SiMe3, respectively16c] than with an
appreciable π acceptor role [2060, 2015 cmϪ1 for (η5-C5H5)-
Fe(CO)2CN]. Indeed, it is noticeable how similar are both
᎐
5
5
2
2
2
2
double bond, rather than an ortho-substituted benzene ring
within the five-membered chelate) shows that (i) the barrier to
rotation about the Fe–B bond is of the order of 3–4 kJ molϪ1
;
and (ii) that near co-planar (ϑ = 8Њ) and near perpendicular
(ϑ = 81Њ) orientations of the boryl ligand plane with respect to
that defined by Cp–Fe–B are of essentially equal energy.5o These
findings are therefore consistent with a description of boryl
ligands as very good σ donors, but relatively poor π acceptors,15
and imply that the differences in orientation of the boryl ligand
between terminally coordinated Bcat and bridging BO2C6H2-
2024
J. Chem. Soc., Dalton Trans., 2002, 2020–2026

View Article Online
Table 2 Comparison of Mössbauer and IR spectral data for complexes of the type (η5-C5H5)Fe(CO)2X (including dinuclear boryl complexes 4a
and 4d)
Complex
Isomer shift/mm sϪ1
Quadrupole splitting/mm sϪ1
ν(CO)/cmϪ1
Ref.
4a
0.02
0.00
0.22
0.23
0.07
0.19
0.05
0.08
0.04
1.91
1.86
1.88
1.87
1.90
1.81
1.77
1.76
1.84
2006, 1954
1998, 1932
2045, 1999
2045, 1999
2060, 2015
2050, 2005
1996, 1944
2010, 1958
2005, 1947
This work
This work
This work
16b
16b
16b
16c
16c
16d
4d
(η5-C5H5)Fe(CO)2Br
(η5-C5H5)Fe(CO)2Br
(η5-C5H5)Fe(CO)2CN
(η5-C5H5)Fe(CO)2SCN
(η5-C5H5)Fe(CO)2SiMe3
(η5-C5H5)Fe(CO)2CH3
(η5-C5H5)Fe(CO)2B5H8
Mössbauer and IR spectral parameters for 4a and 4d to those
reported for (η5-C5H5)Fe(CO)2B5H8.16d This complex contains
a (η5-C5H5)Fe(CO)2 fragment terminally bound to the basal
boron of a square pyramidal cluster, which by implication must
have little or no π acceptor properties.16d
would also like to thank the EPSRC National Mass Spectro-
metry Service for invaluable assistance and Drs I.A. Fallis,
A.J. Amoroso (Cardiff University) and Prof. A.J. Downs
(University of Oxford) for their assistance. We would also like
to thank Dr. P. Low (Durham University) for carrying out some
of the electrochemical measurements. D. J. W. would like to
acknowledge support from the EPSRC, OCF and Synetix for
computing facilities and J. W. S. would like to thank the EPSRC
and the Nuffield Foundation for provision of diffractometer
facilities. S. J. C., M. E. L. and M. B. H. would like to thank the
EPSRC for provision of crystallographic facilities.
By contrast, FT-Raman and electrochemical probes of boryl
complexes 4a–d have proved to be less informative. Raman
spectra of the four dinuclear complexes show little variation,
strong features in the region 520–530 cmϪ1 in each case being
assigned to Fe–B stretching modes. Interestingly however, the
analogous stretching vibration for the more Lewis acidic penta-
fluorophenyl substituted derivative (η5-C5H5)Fe(CO)2B(C6F5)2
is found at 577 cmϪ1. Electrochemical methods have been
used to great effect to assess the possibility for communi-
cation between metal centres through bridging ligand systems.
Consequently, it was hoped that cyclic voltametry (CV) meas-
urements for complexes 4a–c might indicate whether communi-
cation between metal centres through bridging boryl ligands
was possible. CV measurements for all four dinuclear com-
plexes (including complex 4d containing a comparative satur-
ated spacer group) were unfortunately complicated by problems
of irreversibility, which were only partially alleviated by the use
of bulkier ligand systems (η5-C5Me5 vs. η5-C5H5), glassy carbon
electrodes or low temperature. Consequently it proved impos-
sible to definitively assign spectra and therefore to use electro-
chemical methods as a probe of the electronic structure of these
systems.
References
1 (a) G. J. Irvine, M. J. G. Lesley, T. B. Marder, N. C. Norman,
C. R. Rice, E. G. Robins, W. R. Roper, G. R. Whittell and
L. J. Wright, Chem. Rev., 1998, 98, 2685; (b) M. R. Smith,
Prog. Inorg. Chem., 1999, 48, 505; (c) H. Braunschweig, Coord.
Chem. Rev., 2001, 223, 1.
2 See, for example (a) H. C. Brown and B. Singaram, Pure Appl.
Chem., 1987, 59, 879; (b) K. Burgess and M. J. Ohlmeyer,
Chem. Rev., 1991, 91, 1179; (c) C. Iverson and M. R. Smith,
Organometallics, 1996, 15, 5155; (d ) I. Beletskaya and A. Pelter,
Tetrahedron, 1997, 53, 4957; (e) T. B. Marder and N. C. Norman,
Top. Catal., 1998, 5, 63; ( f ) T. Ishiyama and N. Miyaura,
J. Organomet. Chem., 2000, 611, 392.
3 (a) K. M. Waltz, X. He, C. Muhoro and J. F. Hartwig, J. Am. Chem.
Soc., 1995, 117, 11357; (b) K. M. Waltz and J. F. Hartwig,
Science, 1997, 277, 211; (c) K. M. Waltz, C. N. Muhoro and
J. F. Hartwig, Organometallics, 1999, 18, 3383; (d ) K. M. Waltz
and J. F. Hartwig, J. Am. Chem. Soc., 2000, 122, 11358;
(e) K. Kawamura and J. F. Hartwig, J. Am. Chem. Soc., 2001, 123,
8422; ( f ) T. Ishiyama, J. Tagaki, K. Ishida, N. Miyaura, N. R.
Anastasi and J. F. Hartwig, J. Am. Chem. Soc., 2002, 124, 390.
4 (a) C. N. Iverson and M. R. Smith, J. Am. Chem. Soc., 1999, 121,
7696; (b) H. Chen and J. F. Hartwig, Angew. Chem., Int. Ed., 1999,
38, 3391; (c) H. Chen, S. Schlecht, T. C. Semple and J. F. Hartwig,
Science, 2000, 287, 1995; (d ) S. Shimada, A. S. Batsanov, J. A. K.
Howard and T. B. Marder, Angew. Chem., Int. Ed., 2001, 40,
2168; (e) J. Y. Cho, M. K. Tse, D. Holmes, R. E. Maleczka and
M. R. Smith III, Science, 2002, 295, 305.
5 (a) P. R. Rablen, J. F. Hartwig and S. P. Nolan, J. Am. Chem. Soc.,
1994, 116, 4121; (b) D. G. Musaev, A. M. Mebel and K. Morokuma,
J. Am. Chem. Soc., 1994, 116, 10693; (c) A. E. Dorigo and P. v. R.
Schleyer, Angew. Chem., Int. Ed. Engl., 1995, 34, 115; (d ) D. G.
Musaev, T. Matsubara, A. M. Mabal, N. Koga and K. Morokuma,
Pure Appl. Chem., 1995, 67, 257; (e) M. Wagner, N. J. R. v. E.
Hommes, H. Nöth and P. v. R. Schleyer, Inorg. Chem., 1995, 34, 607;
( f ) D. G. Musaev and K. Morokuma, J. Phys. Chem., 1996, 100,
6509; (g) P. R. Rablen and J. F. Hartwig, J. Am. Chem. Soc., 1996,
118, 4648; (h) S. Sakaki and T. Kikuno, Inorg. Chem., 1997, 36,
226; (i) Q. Cui, D. G. Musaev and K. Morokuma, Organometallics,
1997, 16, 1355; (j) Q. Cui, D. G. Musaev and K. Morokuma,
Organometallics, 1998, 17, 742; (k) Q. Cui, D. G. Musaev and
K. Morokuma, Organometallics, 1998, 17, 1383; (l ) S. Sakaki, S. Kai
and M. Sugimoto, Organometallics, 1999, 18, 4825; (m) C. Widauer,
H. Grützmacher and T. Ziegler, Organometallics, 2000, 19, 2097;
(n) K. T. Giju, M. Bickelhaupt and G. Frenking, Inorg. Chem., 2000,
39, 4776; (o) A. A. Dickinson, D. J. Willock, R. J. Calder and
S. Aldridge, Organometallics, 2002, 21, 1146.
4
Conclusions
A series of dinuclear complexes containing iron centres linked
via boryl ligands has been synthesised by metathesis from
bifunctional boron halides. This synthetic approach is applic-
able to different ligand backbones (‘spacer groups’) and relies
heavily on the use of trimethylsilyl substituted precursors to
generate complexes such as 4a–d in reasonable yield. Com-
parative structural studies of 4a and 4d reveal the importance
of the chelating alkoxo backbone of the boryl ligand in deter-
mining its electronic properties. In addition, structural charac-
terization of the bridged boryl systems allows comparison with
terminally bound Bcat analogues. The origins of the different
ligand orientations found have been probed by DFT methods
which suggest (i) that π contributions to bonding and barriers
to Fe–B rotation are small for dialkoxoboryl systems, and (ii)
that crystal packing forces rather than ligand electronics are
almost certainly responsible for the different orientations
observed. Finally, 57Fe Mössbauer spectroscopy has been used
for the first time to probe the bonding in boryl systems, with the
findings being entirely consistent with a description of these
ligands as very good σ donors, but poor π acceptors.
Acknowledgements
We would like to acknowledge funding from the EPSRC
(GR/R04669 and GR/R06458), the BBSRC (DJE), the Royal
Society, the Nuffield Foundation and Cardiff University. We
6 For previous reports of iron boryl complexes see refs. 3b–d together
with (a) J. F. Hartwig and S. Huber, J. Am. Chem. Soc., 1993, 115,
4908; (b) X. He and J. F. Hartwig, Organometallics, 1996, 15, 400;
J. Chem. Soc., Dalton Trans., 2002, 2020–2026
2025

View Article Online
(c) H. Braunschweig, B. Ganter, M. Koster and T. Wagner,
Chem. Ber., 1996, 129, 1099; (d ) H. Braunschweig, C. Kollann and
M. Müller, Eur. J. Inorg. Chem., 1998, 291; (e) H. Braunschweig,
C. Kollann and U. Englert, Eur. J. Inorg. Chem., 1998, 465;
( f ) H. Braunschweig and M. Koster, J. Organomet. Chem., 1999,
588, 231; (g) H. Braunschweig, C. Kollann and K. W. Klinkhammer,
Eur. J. Inorg. Chem., 1999, 1523; (h) T. Yasue, Y. Kawano and
M. Shimoi, Chem. Lett., 2000, 58; (i) S. Aldridge, R. J. Calder,
A. A. Dickinson, D. J. Willock and J. W. Steed, Chem. Commun,
2000, 1377; (j) S. Aldridge, A. Al-Fawaz, R. J. Calder, A. A.
Dickinson, D. J. Willock, M. Light and M. B. Hursthouse, Chem.
Commun., 2001, 1846.
7 D. Curtis, M. J. G. Lesley, N. C. Norman, A. G. Orpen and
J. Starbuck, J. Chem. Soc., Dalton Trans, 1999, 1687.
8 S. Aldridge, R. J. Calder, K. M. A. Malik and J. W. Steed,
J. Organomet. Chem., 2000, 614, 188.
9 S. Aldridge, R. J. Calder, R. E. Baghurst, M. E. Light and
M. B. Hursthouse, J. Organomet. Chem., 2002, 649, 9.
10 (a) R. B. King and M. B. Bisnette, J. Organomet. Chem., 1967, 8,
287; (b) R. B. King, Acc. Chem. Res., 1970, 3, 417.
11 E. Lifshitz, D. Goldfarb, S. Vega, Z. Luz and H. Zimmermann,
J. Am. Chem. Soc., 1987, 109, 7280.
12 Z. Otwinowski and W. Minor, Methods Enzymol., 1996, 276, 307.
13 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467.
14 B. E. R. Schilling, R. Hoffmann and D. Lichtenberger, J. Am. Chem.
Soc., 1979, 101, 585.
15 A similar conclusion has been reached by analysis of the structures
of five-coordinate Rh() Bcat derivatives both in solution and in the
solid state: C. Dai, G. Stringer, T. B. Marder, A. J. Scott, W. Clegg
and N. C. Norman, Inorg. Chem., 1997, 36, 272.
16 (a) R. V. Parish, in The Organic Chemistry of Iron, ed. E. A. Koerner
v. Gustorf, F.-W. Grevels and I. Fischler, Academic Press, 1978, vol.
1; (b) G. J. Long, D. G. Alway and K. W. Barnett, Inorg. Chem.,
1978, 17, 486; (c) K. H. Pannell, C. C. Wu and G. J. Long,
J. Organomet. Chem., 1980, 186, 85; (d ) B. H. Goodreau, L. R.
Orlando, G. J. Long and J. T. Spencer, Inorg. Chem., 1996, 35, 6579.
2026
J. Chem. Soc., Dalton Trans., 2002, 2020–2026