Intramolecular base-stabilised adducts of main group halides

Article abstract of DOI:10.1039/b110692a

The coordination chemistry of alkoxo and amido ligands bearing pendant nitrogen or sulfur donors has been developed to include examples of main group complexes. The structural implications of amido ligand coordination augmented by a tethered pyridyl base have been investigated for group 13 and 14 derivatives {[6-(Me₃SiN)-2-Me-C₅H₃N]SnCl₃ and [6-(Me₃SiN)-2-Me-C₅H₃N]BBr₂}. Furthermore, the scope, kinetics and mechanism of the synthesis of alkoxo and amido derivatives from trimethylsilyl substituted precursors have been investigated in depth. The methodology is found to be applicable to a range of ligand frameworks of varying flexibility, stabilised by one or two "tether" arms.

Full text of DOI:10.1039/b110692a

View Article Online / Journal Homepage / Table of Contents for this issue
Intramolecular base-stabilised adducts of main group halides
Simon Aldridge,*a Richard J. Calder,a Deborah L. Coombs,a Cameron Jones,a
Jonathan W. Steed,b Simon Colesc and Michael B. Hursthousec
a
Department of Chemistry, Cardiff University, PO Box 912, Park Place, Cardiff, UK CF103TB.
E-mail: AldridgeS@cardiff.ac.uk
Department of Chemistry, King’s College London, Strand, London, UK WC2R 2LS
Department of Chemistry, University of Southampton, Highfield, Southampton, UK SO17 1BJ
b
c
Received (in Montpellier, France) 21st November 2001, Accepted 18th January 2002
First published as an Advance Article on the web 22nd May 2002
The coordination chemistry of alkoxo and amido ligands bearing pendant nitrogen or sulfur donors has been
developed to include examples of main group complexes. The structural implications of amido ligand
coordination augmented by a tethered pyridyl base have been investigated for group 13 and 14 derivatives
{[6-(Me3SiN)-2-Me-C5H3N]SnCl3 and [6-(Me3SiN)-2-Me-C5H3N]BBr2}. Furthermore, the scope, kinetics and
mechanism of the synthesis of alkoxo and amido derivatives from trimethylsilyl substituted precursors have
been investigated in depth. The methodology is found to be applicable to a range of ligand frameworks of
varying flexibility, stabilised by one or two ‘‘tether’’ arms.
Intramolecular coordination of a tethered base is an exten-
sively used strategy for the stabilisation of complexes con-
taining electronically unsaturated centres, unusual oxidation
states or reactive bonds.1,2 The coordination of such a base in
effect provides a sizeable kinetic barrier to attack by external
nucleophiles and also reduces the thermodynamic incentive for
oligomerisation via bridging ligands.
Within the field of transition metal chemistry alkoxo and
amido ligands have attracted widespread interest,1,3,4 for
example in the synthesis of non-metallocene olefin poly-
merisation catalysts.5 The use of a tethered pyridyl base within
the amido ligand framework as in I reduces the likelihood of
oligomerisation, leading to tractable mononuclear catalyst
precursors.6 The electronic properties of the donor function
and the conformational rigidity of the tether have been shown
to be vital in controlling the reactivity of catalytic systems.1a
Related ligands have also found use in the synthesis of ear-
ly=late transition metal bimetallic systems,7 in MOCVD pre-
cursors8 and in electroluminescence.9 The use of such ligands
in main group chemistry is, however, much less widely
exploited, despite several structural studies detailing the
interesting effects of ligand architecture on the coordination
geometry of post-transition metals.8–18
ligand I are rare, with Group 13 derivatives in general being
limited to a single 2 : 1 ligand : metal complex for aluminium
and none as yet being reported for boron.12,13 In addition,
aside from the structural interest, the group 13 derivatives offer
themselves as potential precursors to base-stabilised transition
metal complexes of the types LnMER and LnMER2 (e.g.,
borylene and boryl derivatives for E ¼ B) formed by meta-
thetical reaction with organometallic anions. Despite intense
interest in the structure and bonding of such complexes, the
synthetic utility of a stabilising tethered base has only recently
been touched upon.19,20
From a synthetic standpoint we have sought to exploit the
utility of trimethylsilyl substituted amine=ether precursors as
mild sources of the amido or alkoxo ligands.14,21,22 The scope
of this methodology, both in terms of the type of ligand fra-
mework available and the number of amido/alkoxo linkages
that can be introduced, has been probed, as has the mechanism
of intramolecular base-assisted elimination of the trimethylsilyl
halide.
Experimental
We have therefore sought to extend the coordination
chemistry of ligands I–III (Chart 1) by the synthesis of deri-
vatives in which the presence of a Lewis base function within
the ligand framework brings about the isolation of stable
mononuclear group 13 and 14 derivatives. Such a range of
ligand frameworks allows the investigation of the roles of
donor function and tether architecture on the geometry of com-
plexes produced. Unlike their transition metal counterparts,
structurally characterised species from groups 13–15 containing
All manipulations were carried out under a nitrogen or argon
atmosphere using standard Schlenk line or dry box techniques.
Solvents were pre-dried over sodium wire and purged with
nitrogen prior to distillation. Tetrahydrofuran (thf) and hex-
anes were distilled from potassium; benzene and toluene were
distilled from sodium before use. Triethylamine was dried over
sodium wire prior to use. Starting materials Me3SiCl, SnCl4 ,
BF3ÁOEt2 , BCl3 , BBr3 , N-methyldiethanolamine, 2,20-thio-
diethanol and 2-amino-6-picoline were used as received
(Aldrich) without further purification. 6-[Bis(trimethylsilyl)-
amino]-2-picoline and 2-[bis(trimethylsilyl)aminomethyl]pyr-
idine were prepared by minor modification of literature
methods.14,23
NMR spectra were measured on a Bruker AM-400 or Jeol
300 Eclipse Plus FT-NMR spectrometer. Residual protons of
1
solvent were used for reference for H and 13C NMR, while a
sealed tube containing [(nBu4N)(B3H8)] was used as an external
Chart 1
DOI: 10.1039/b110692a
New J. Chem., 2002, 26, 677–686
677
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

View Article Online
reference for 11B NMR. Infrared spectra were measured for
each compound pressed into a disk with an excess of dried KBr
on a Nicolet 500 FT-IR spectrometer. Mass spectra were
measured by the EPSRC National Mass Spectrometry Service
Centre, University of Wales, Swansea. Perfluorotributylamine
and polyethylenimine were used as the standards for high-
resolution EI and CI mass spectra, respectively. Elemental
analyses were carried out both by the departmental analysis
service and by Warwick Analytical Service, University of
Warwick.
allowed to warm to room temperature and stirred for 12 h.
Filtration, removal of volatiles in vacuo and recrystallisation
from warm hexanes led to the formation of 4a–c in isolated
yields of 28.0% (X ¼ F), 29.4% (X ¼ Cl) and 26.1% (X ¼ Br).
The adducts have been characterised by multinuclear NMR,
mass spectrometry and elemental analysis.
4a 1H NMR (300 MHz, C6D6) d 0.00 [18H, s, Si(CH3)3],
4.91 (2H, s, CH2), 6.19 (1H, t, J ¼ 6.6 Hz, aromatic), 6.86 (1H,
t, J ¼ 7.6Hz, aromatic), 7.67 (1H, d, J ¼ 8.0 Hz, aromatic),
8.44 (1H, d, J ¼ 5.5 Hz, aromatic). 13C NMR (76MHz, C 6D6)
d 1.2 [Si(CH3)3], 46.4 (CH2), 122.3, 124.2, 141.5, 143.7 (aro-
Abbreviations: st ¼ strong, m ¼ medium, w ¼ weak, sh ¼
shoulder, s ¼ singlet, d ¼ doublet, t ¼ triplet, q ¼ quartet, pcq ¼
partially collapsed quartet, mp ¼ multiplet, b ¼ broad.
matic CH). 11B NMR (96MHz, C D6) 0.35 (q, JBF ¼ 9.4 Hz).
6
19F NMR (283 MHz, C6D6) À145.5 (pcq, JBF ¼ 9.4 Hz). MS
(EI): [M À BF3]þ ¼ 253 (100%), [M À BF3 À Me]þ ¼ 237 (20%).
Anal. calcd. for C9H15BF3N2Si2: C 44.99%, H 7.57%, N
8.74%; found: C 44.61%, H 7.34%, N 8.99%.
Syntheses and kinetic studies
4b 1H NMR (300 MHz, C6D6) d 0.02 [18H, s, Si(CH3)3],
5.44 (2H, s, CH2), 6.06 (1H, t, J ¼ 6.8 Hz, aromatic), 6.77 (1H,
t, J ¼ 7.3 Hz, aromatic), 7.87 (1H, d, J ¼ 8.0 Hz, aromatic),
9.35 (1H, mp, aromatic). 11B NMR (96MHz, C 6D6) 7.78. MS
(EI): [M À BCl3 À Me]þ ¼ 237 (100%). Anal. calcd. for
C9H15BCl3N2Si2: C 38.98%, H 6.56%, N 7.57%; found: C
38.55%, H 6.35%, N 7.59%.
Synthesis of 6-[(trimethylsilyl)amido]-2-picolylboron di-
bromide (1c). Boron tribromide (2.65 g, 1.0 cm3, 10.6mmol)
was added slowly to a solution of 6-[bis(trimethylsilyl)]amino-
2-picoline (2.68 g, 10.6 mmol) in hexanes (40 cm3) at À78 ꢀC
and stirred at this temperature for 30 min. Volatiles were then
removed in vacuo, yielding a white precipitate that was
recrystallised from hot hexanes to yield 1c as a white crystal-
line material (1.32 g, 35.5%). Single crystals suitable for X-ray
crystallography were grown by cooling a hexanes solution to
4c 1H NMR (300 MHz, C6D6) d 0.03 [18H, s, Si(CH3)3], 5.59
(2H, s, CH2), 5.97 (1H, t, J ¼ 6.6 Hz, aromatic), 6.67 (1H, t,
J ¼ 7.6Hz, aromatic), 7.87 (1H, d, J ¼ 8.0 Hz, aromatic), 9.67
(1H, mp, aromatic). 13C NMR (76MHz, C 6D6) d 1.1
[Si(CH3)3], 49.2 (CH2), 122.0, 125.3, 141.5, 148.1 (aromatic
1
À30 ꢀC for 1 week. 1c has been characterised by H, 13C and
11B NMR, IR, mass spectroscopy, elemental analysis and
single crystal X-ray diffraction. 1H NMR (300 MHz, C6D6)
d 0.22 [9H, s, Si(CH3)3], 1.94 (3H, s, CH3), 5.43 (1H, d,
J ¼ 7.5 Hz, aromatic), 5.50 (1H, d, J ¼ 8.5 Hz, aromatic), 6.54
(1H, t, J ¼ 8.1 Hz, aromatic). 13C NMR (76MHz, C 6D6) d 0.0
[Si(CH3)3], 17.1 (CCH3), 104.2, 111.6, 144.9 (aromatic CH),
150.8, 163.1 (aromatic quaternary). 11B NMR (96MHz, C 6D6)
d À9.9. IR (KBr, cmÀ1): 2958 m, 1614 st, 1492 st, 1128 m,
1083 m, 1038 m, 976 w, 848 m, 669 m. MS (EI):
[M þ Br]þ ¼ 430 (100%). Anal. calcd. for C9H15N2SiBBr2: C
30.89%, H 4.33%, N 8.00%; found: C 31.14%, H 4.26%, N
8.38%. 1a and 1b were prepared in similar fashion from
BF3ÁOEt2 and BCl3 , respectively; spectroscopic data are in line
with those reported previously.14
CH). 11B NMR (96MHz, C D6) À9.45. MS (EI): [M]þ ¼ 503
6
(weak), [M À BBr3 À Me]þ ¼ 237 (100%). Anal. calcd. for
C9H15BBr3N2Si2: C 28.65%, H 4.82%, N 5.57%; found: C
28.32%, H 4.83%, N 5.38%.
For each of the compounds 4a–c observation of the parent
ion using EI-MS (of solid samples) proved difficult, with ready
fragmentation of the adducts occurring. The use of softer
ionisation techniques was frustrated by the high sensitivity of
the complexes to air and moisture.
Kinetic studies: formation of 5a from 4a. A solution con-
taining 0.15 mmol of 4a dissolved in 0.8 cm3 of C6D6 was
1
transferred to a Young’s NMR tube, the H NMR spectrum
Synthesis of 6-[(trimethylsilyl)amido]-2-picolyltin trichloride
(3). Tin tetrachloride (0.89 g, 0.4 cm3, 3.42 mmol) was added
slowly to a solution of 6-[bis(trimethylsilyl)]amino-2-picoline
(0.78 g, 3.09 mmol) in hexanes (25 cm3) at À78 ꢀC and the
reaction mixture allowed to warm slowly to room temperature.
After stirring for 10 h, the pale yellow solution was filtered,
concentrated to approximately 10 cm3 and cooled slowly to
À30 ꢀC to yield pale yellow crystals of 3 in ca. 75% yield.
Single crystals suitable for X-ray crystallography were grown
by cooling a hexanes solution to À30 ꢀC for 1 week. 3 has been
measured at room temperature and the sample then warmed to
the desired thermolysis temperature (either 55 or 75 ꢀC). The
1H NMR spectrum of the reaction mixture was then measured
periodically with the concentrations of 4a and 5a being eval-
uated on the basis of the integration of the methylene signals.y
Cyclisation was judged to be complete after 23 days at 55 ꢀC or
3 days at 75 ꢀC. NMR data for 5a: 1H NMR (300 MHz, C6D6)
d 0.11 [9H, s, Si(CH3)3], 4.34 (2H, s, CH2), 6.61 (1H, t,
J ¼ 5.3 Hz, aromatic), 7.14 (1H, t of d, J ¼ 7.2, 1.8 Hz, aro-
matic), 7.28 (1H, d, J ¼ 7.8 Hz, aromatic), 8.47 (1H, d, J ¼ 4.3
Hz, aromatic). 13C NMR (76MHz, C 6D6) d 1.7 [Si(CH3)3],
51.1 (CH2), 120.2, 120.9, 135.5, 149.1 (aromatic CH), 164.4
(aromatic quaternary). 11B NMR (96MHz, C 6D6) d 7.3
1
characterised by H, 13C and 119Sn NMR, IR, mass spectro-
scopy, elemental analysis and single crystal X-ray diffraction.
1H NMR (300 MHz, C6D6) d 0.22 [9H, s, Si(CH3)3], 2.02 (3H,
s, CH3), 5.80 (1H, d, J ¼ 7.2 Hz, aromatic), 5.98 (1H, d,
J ¼ 8.5 Hz, aromatic), 6.69 (1H, t, J ¼ 7.9 Hz, aromatic). 13C
NMR (76MHz, C D6) d 1.4 [Si(CH3)3], 19.6(C CH3), 106.3,
6
y Estimates of the relative concentrations of 4a and 5a in solution were
made by integration of the 1H NMR signal due to the methylene
protons in each compound. A plot of ln(mole fraction of 4a) vs. time
(s) (Figs. 4 and 5) could be fitted to the linear relationship:
114.5, 143.0 (aromatic CH), 153.8, 169.3 (aromatic qua-
ternary). 119Sn NMR (112 MHz, C6D6) d À338.2. IR (KBr,
cmÀ1): 2966 w, 1639 m b, 1506 w, 1460 w, 1393 w, 1327 vw,
1265 m, 1173 w, 840 st, 789 st, 712 w, 549 w. MS (EI):
[M]þ ¼ 404 (15%), [M À CH3]þ ¼ 389 (100%). Anal. calcd. for
C9H15Cl3N2SnSi: C 26.73%, H 3.74%, N 6.92%; found: C
26.60%, H 3.76%, N 6.73%.
ln(mole fraction) ¼ (À4 Â 10À6)t þ 4.411 (R2 ¼ 0.9954) at 75 ꢀC
or
ln(mole fraction) ¼ (À7 Â 10À7)t þ 4.509 (R2 ¼ 0.9765) at 55 ꢀC
Reaction of 2-[bis(trimethylsilyl)aminomethyl]pyridine with
boron trihalides. In a typical reaction BX3 (X ¼ Cl, Br) or
BF3ÁOEt2 (1.1 mmol) was added dropwise to a solution con-
taining 1 equiv of 2-[bis(trimethylsilyl)aminomethyl]pyridine in
ca. 40 cm3 of hexanes at À78 ꢀC. The reaction mixture was
from which it is possible to obtain a value for the first order rate
constant, k1 , of 4 Â 10À6 sÀ1 at 75 ꢀC and 7 Â 10À7 sÀ1 at 55 ꢀC. A
similar treatment of the product 5a gives values of k1 of 7 Â 10À6 and
9 Â 10À7 sÀ1 at 75 and 55 ꢀC, respectively. Graph fitting was carried out
using Microsoft Excel and SigmaPlot.
678
New J. Chem., 2002, 26, 677–686

View Article Online
(t, JBF ¼ 35.4 Hz). 19F NMR (283 MHz, C6D6) À157.1 (pcq,
JBF ¼ 35.4 Hz).
then heated to 55 ꢀC in the NMR spectrometer and maintained
at that temperature for a period of 96h, during which the 11B
and 19F spectra were measured hourly. During the course of
the thermolysis reaction resonances characteristic of 8a and the
intermediate 9a were observed, both sets of signals ultimately
disappearing to leave only those characteristic of the final
product 10a. NMR data for 8a: 1H NMR (300 MHz, C6D6) d
0.04 [18 H, s, Si(CH3)3], 2.58 (3H, s, CH3), 3.24 (4H, m, CH2),
3.74 (4H, m, CH2). 13C NMR (76MHz, C 6D6) d 0.1
[Si(CH3)3], 44.2 (CH3), 56.4 (CH2), 61.3 (s, CH2). 11B NMR
Synthesis of (Me3SiOCH2CH2)2NMe (6). N-Methyldietha-
nolamine (20 cm3, 0.18 mol) was dissolved in toluene (150 cm3)
and 4 equiv of triethylamine (100 cm3, 0.72 mol) and 4 equiv of
chlorotrimethylsilane (91 cm3, 0.72 mol) were then added via
syringe at room temperature. After 24 h the supernatent
toluene solution was filtered and the (Et3NH)Cl precipitate
washed with toluene (2 Â 15 cm3). After removal in vacuo of
volatiles from the combined washings, 6 was isolated as a very
(96MHz, C D6) d À0.3 (q, JB–F ¼ 16.2 Hz). 19F NMR (283
6
MHz, C6D6) d À156.0 (q, JF–B ¼ 16.1 Hz). The intermediate 9a
is only ever formed in very small quantities in the presence of
much larger amounts of 8a and 10a; no attempts to isolate the
intermediate have been made and we merely report the posi-
tions of NMR resonances determined during kinetic runs: 11B
NMR (96MHz, C 6D6) d 3.9 (t, JB–F ¼ 21.1 Hz). 19F NMR
(283 MHz, C6D6) d 157.1 (q, JF–B ¼ 21.0 Hz).
1
pale brown liquid in yields of ca. 50%. At this point H and
13C NMR indicated that the compound was >99% pure and
no further purification was therefore attempted. Character-
isation of compound 6 was achieved by 1H and 13C NMR and
mass spectrometry (including exact mass determination). 1H
NMR (300 MHz, C6D6) d 0.02 (18 H, s, SiMe3), 2.15 (3H, s,
CH3), 2.43 (4H, m, CH2), 3.51 (4H, m, CH2). 13C NMR (76
MHz, C6D6) d 0.1 (SiMe3), 44.0 (CH3), 61.2 (CH2), 61.6
(CH2). MS (CI): [M þ H]þ ¼ 264 (55%). Exact mass calcd. for
C11H30NO2Si2 [M þ H]þ: 264.1815; found: 264.1813.
Reaction of 6 with BCl3 . To a sample of 2.26g (8.6mmol)
of 6 in toluene at room temperature was added by syringe 1
3
equiv of BCl3 (8.6cm of a 1 M solution in heptane); the 11B
NMR spectrum measured at this point revealed that the sole
boron containing species present was the adduct Me(Me3-
SiOCH2CH2)2NÁBCl3 (8b). Removal of all volatile species
in vacuo, followed by extraction with hexanes (2 Â 40 cm3),
filtration and removal of volatiles in high vacuum (10À4 Torr)
led to the isolation of 8b as a colourless oil in 80–90% yield
and >99% purity. Thermolysis of an approximately 1 M
solution of 8b in toluene or benzene led to the evolution of
Me3SiCl and the formation of a mixture of the mono- and di-
cyclised species 9b and 10b. Despite extensive variation in
temperature, concentration and reaction time it proved
impossible to drive the reaction to completion, such that the
reaction mixture at this point, although predominantly con-
sisting of 10b, was always contaminated with the intermediate
9b. Removal of volatiles in vacuo, extraction of the residual
solid with hexanes (5 Â 20 cm3) and recrystallisation from
concentrated hexanes solution at À30 ꢀC led to the isolation of
the intermediate 9b as a white solid in 10–15% isolated yield.
The remaining solid, insoluble in hexanes, was found to consist
predominantly of the di-cyclised species 10b; despite many
attempts it proved impossible to obtain samples of 10b pre-
pared this way free from traces of 9b. 8b and 9b have been
characterised by 1H, 13C and 11B NMR and mass spectro-
metry, and the 1H, 13C and 11B signals due to 10b deduced
from impure product contaminated with traces of 9b.
Synthesis of (Me3SiOCH2CH2)2S (7). Compound 7 was
prepared from 2,20-thiodiethanol by an analogous method to
that described above for 6 and was isolated as a colourless oil
and characterised by 1H and 13C NMR and mass spectrometry
(including exact mass determination). 1H NMR (300 MHz,
C6D6) d 0.02 (18 H, s, SiMe3), 2.53 (4H, t, J ¼ 7.2 Hz, CH2),
3.57 (4H, t, J ¼ 7.2 Hz, CH2). 13C NMR (76MHz, C 6D6) d 0.0
(SiMe3), 35.6( CH2), 63.5 (CH2). MS (CI): [M þ H]þ ¼ 267
(45%). Exact mass calcd. for C10H27O2SSi2 [M þ H]þ:
267.1270; found: 267.1269.
Synthesis of MeN(CH2CH2O)2BF (10a). To a solution of
9.2 g (35 mmol) of (Me3SiOCH2CH2)2NMe (6) in benzene
(200 cm3) at room temperature was added slowly via syringe 1
equiv of BF3ÁOEt2 . The reaction mixture was heated to 55 ꢀC
and monitored by 11B NMR over a period of 96h until the
signals due to intermediate species 8a and 9a had disappeared.
At this point the colourless solution was separated from a
small amount of brown oily precipitate by filtration. Cooling
of the resultant colourless solution to room temperature led to
the precipitation of 10a as a white crystalline solid of >95%
purity. Extremely slow cooling from 55 ꢀC led to the formation
of X-ray quality crystals. 10a can be further purified by
recrystallisation from benzene or toluene, or by sublimation at
ca. 150 ꢀC (10À4 Torr) onto a cold finger held at À78 ꢀC. Yields
8b 1H NMR (300 MHz, C6D6) d 0.01 [18H, s, Si(CH3)3],
2.67 (3H, m, CH3), 3.36(4H, m, C H2), 3.62 (4H, m, CH2). 13
C
1
are typically in excess of 90%. 10a was characterised by H,
13C, 11B and 19F NMR, IR, mass spectrometry (including
exact mass determination), elemental analysis and single
crystal X-ray diffraction. 1H NMR (300 MHz, C6D6) d 1.8
(3H, d, JH–F ¼ 1.9 Hz, CH3), 1.9 (4H, m, CH2), 3.5 (4H, b m,
CH2). 13C NMR (76MHz, C 6D6) d 42.2 (d, JC–F ¼ 4.9 Hz,
CH3), 58.4 (d, JC–F ¼ 2.0 Hz, CH2), 58.6(s, CH2). 11B NMR
(96MHz, C 6D6) d 8.8 (d, JB–F ¼ 25.3 Hz). 19F NMR (283 MHz,
C6D6) d À157.2 (q, JF–B ¼ 25.3 Hz). IR (cmÀ1): 1102 st [n(B–
F)]. MS (EI): Mþ ¼ 147 (100%). Exact mass calcd. for
C5H11BFNO2: 147.0867; found: 147.0868. Anal. calcd. for
C5H11BFNO2: C 40.87, H, 7.55, N 9.53; found: C 40.44, H
7.37, N 9.51%.
NMR (76MHz, C 6D6) d 0.1 [Si(CH3)3], 45.7 (CH3), 58.6
(CH2), 60.7 (CH2). 11B NMR (96MHz, C D6) d 11.1 (s). MS
6
(EI): Mþ ¼ 379 (weak), [M À BCl3]þ ¼ 263 (100%).
9b 1H NMR (300 MHz, C6D6) d 0.27 [9H, s, Si(CH3)3], 2.02
(3H, s, CH3), 2.38 (2H, t, J ¼ 5.9 Hz, CH2), 2.53 (2H, t,
J ¼ 7.1 Hz, CH2), 3.24 (2H, t, J ¼ 7.1 Hz, CH2), 3.86(2H, t,
J ¼ 5.9 Hz, CH2). 13C NMR (76MHz, C 6D6) d 0.9 [Si(CH3)3],
41.5 (CH3), 42.3 (CH2), 58.5 (CH2), 59.5 (CH2), 61.6 (CH2).
11B NMR (96MHz, C 6D6) d 17.2 (s). MS (EI): Mþ ¼ 271
(100%).
1
10b H NMR (300 MHz, C6D6) d 2.01 (3H, s, CH3), 2.30
(4H, m, CH2), 3.08 (4H, m, CH2). 13C NMR (76MHz, C D6)
6
d 42.7 (CH3), 57.4 (CH2), 58.9 (CH2). 11B NMR (96MHz,
C6D6) d 13.0 (b s).
Kinetic studies: formation of 10a from 6 and BF3ÁOEt2 . A
sample of ca. 100 mg of (Me3SiOCH2CH2)2NMe (6) was dis-
solved in C6D6 and transferred to a Young’s NMR tube. One
equivalent of BF3ÁOEt2 was added by syringe at room tem-
perature and the 11B NMR spectrum measured at this point
revealed that the sole boron containing species present was the
adduct Me(Me3SiOCH2CH2)2NÁBF3 (8a). The solution was
Synthesis of S(CH2CH2O)2BF (12a). Compound 12a was
prepared from (Me3SiOCH2CH2)2S (7) by a method analo-
gous to that described for 10a. Purification of the crude pro-
duct by recrystallisation from hot toluene or benzene resulted
in the isolation of 12a as an off-white crystalline solid in ca.
New J. Chem., 2002, 26, 677–686
679

View Article Online
1
73% yield. Compound 12a has been characterised by H, 13C
atoms N(1), C(2), C(1), O(1) and B(1). There appeared to be a
positional disorder across the mirror plane whereby the site
labelled F(1) was 50% occupied by the atom O(2) and vice
versa. The positional and thermal parameters of F(1) and O(2)
were successfully refined together using the EXYZ and EADP
commands of SHELX 97. In addition, the atom C(4) was
disordered over the mirror plane in a 50 : 50 ratio and therefore
the atom C(4) was refined with 50% occupancy. Only one of
the sites occupied by C(4) is shown in the ORTEP. Hydrogens
were only added to C(1) and C(2) in calculated positions using
the riding model. As a result of the disorder in the molecule
and the fact that F(1) and O(2) were refined using the same
positional and thermal parameters no comments can be made
about the bond lengths and angles within the molecule.
However, the gross molecular framework of the molecule is
unequivocal and is fully supported by the spectroscopic data.
CCDC reference numbers 183227–183229. See http://www.
rsc.org/suppdata/nj/b1/b110692a/ for crystallographic data in
CIF or other electronic format
and 11B NMR, IR, mass spectrometry (including exact mass
determination) and elemental analysis. 1H NMR (300 MHz,
C6D6) d 2.6(4H, m, C H2), 3.9 (4H, m, CH2). 13C NMR (76
MHz, C6D6) d 33.7 (s, CH2), 62.2 (s, CH2). 11B NMR (96
MHz, C6D6) d 17.3 (b). 19F NMR (283 MHz, C6D6) d À150.0.
IR (cmÀ1): 1068 st [n(B–F)]. MS (EI): Mþ ¼ 150 (100%). Exact
mass calcd. for C4H8BFO2S: 150.0322; found: 150.0321. Anal.
calcd. for C4H18BFO2S: C 32.02, H, 5.37%; found: C 31.86, H
5.14%.
General crystallographic method
Single crystals of compounds 1c, 3 and 10a were mounted on
an Enraf Nonius Kappa CCD diffractometer equipped with an
+
Mo-Ka radiation source (l ¼ 0.71073 A). Data collection was
carried out at 100(2) K (1c) or at 150(2) K (3 and 10a). Data
collection and cell refinement were carried out using DEN-
ZO,24a and structure solution and refinement (by full-matrix
least-squares) using SHELXS-97 and SHELXL-97,24b respec-
tively. Fuller details of each data collection, structure solution
and refinement can be found in Table 1, with relevant bond
lengths and angles for compounds 1c and 3 being listed in
Tables 2 and 3, respectively. In the case of compound 10a, the
molecule was found to sit on a mirror plane containing the
Results
The work outlined herein shows that the reaction between
Lewis acidic main group halides and trimethylsilyl substituted
ether or amine reagents is a versatile route to a range of
intramolecular base-stabilised main group halides (as shown in
Chart 2 and Schemes 1 and 2 below).
Table 1 Crystallographic data for 1c, 3 and 10a
This synthetic approach with concomitant evolution of the
trimethylsilyl halide, has several advantages over possible
alternative routes, such as the corresponding reaction with the
parent amine or alcohol, or with alkali metal alkoxides or
amides. Firstly, the solubility of the trimethylsilyl derivatives
in weakly coordinating organic solvents allows the reaction to
be carried out in a single-phase toluene or hexanes medium.
This was found not only to lead to improved reaction rates,
but also to a much cleaner product distribution, especially in
the synthesis of complexes such as 10 and 12 that require the
formation of two O–B linkages. In addition, formation of the
trimethylsilyl derivatives, by reaction of the corresponding
alcohol or amine with a large excess of trimethylsilyl chloride
and extraction into hexanes, effectively constitutes a simple
method for removal of water. Any moisture present is simply
converted by the excess of trimethylsilyl chloride into hexa-
methyldisiloxane, which is removed in vacuo. By contrast
physical methods of drying polyhydroxy species such as N-
methyldiethanolamine are much less efficient, such that direct
reaction with a boron trihalide, for example, inevitably leads to
contamination of the product with B(OH)3 and other species
formed by hydrolysis of B–X bonds.
1c
3
10a
Empirical
formula
Formula
weight
Crystal system
C9H15BBr2N2Si C9H15Cl3N2SiSn C5H11BFNO2
349.95
.94604.36146
Triclinic
Monoclinic
P21/n
6.8246(7)
13.9297(17)
14.2729(17)
90
93.113(9)
90
1354.8(3)
4
100(2)
Orthorhombic
Cmc21
9.800(2)
12.045(2)
5.9433(12)
90
90
90
701.6(2)
4
ꢀ
P1
Space group
+
a/A
+
9.2498(18)
9.708(19)
17.738(4)
93.62(3)
90.61(3)
104.26(3)
1540.1(5)
4
b/A
+
c/A
a/ꢀ
b/ꢀ
g/ꢀ
+
U/A3
Z
T/k
150(2)
2.235
23521
150(2)
0.118
1698
m mmÀ1
Reflections
collected
Independent
reflections
Rint
6.043
6753
2643
6924
628
0.1559
0.0809
0.0307
0.0769
0.0549
0.0790
0.2346
Final R1 [I > 2s(I)] 0.0622
Final wR2 [I > 2s(I)] 0.1337
R1 (all data)
wR2 (all data)
0.08260.0386 0.1072
0.14460.0812
0.2531
+
Table 2 Selected bond lengths (A) and angles (ꢀ) for 1c
Br1–B1
Si1–N2
Si1–C8
Si1–C7
Si1–C9
N1–C5
1.993(7)
1.747(5)
1.857(7)
1.857(8)
1.860(7)
1.345(8)
N1–C1
N1–B1
C1–N2
B1–N2
B1–Br2
1.367(8)
1.577(9)
1.368(8)
1.547(9)
2.004(7)
N2–Si1–C8
N2–Si1–C7
C8–Si1–C7
N2–Si1–C9
C8–Si1–C9
C7–Si1–C9
C5–N1–B1
C1–N1–B1
N1–C1–N2
107.8(3)
N2–B1–N1
N2–B1–Br1
N1–B1–Br1
N2–B1–Br2
N1–B1–Br2
Br1–B1–Br2
C1–N2–B1
C1–N2–Si1
B1–N2–Si1
84.1(5)
116.3(4)
113.3(4)
115.6(4)
113.4(4)
111.5(3)
88.6(5)
109.5(3)
110.2(3)
104.6(3)
113.6(3)
110.9(3)
148.1(5)
87.4(5)
132.8(5)
138.2(4)
99.9(5)
Chart 2
680
New J. Chem., 2002, 26, 677–686

View Article Online
We have sought to establish the scope of this synthetic
methodology for the formation of main group halides con-
taining either alkoxo or amido ligands, and stabilised by the
coordination of a tethered nitrogen or sulfur donor (see
Schemes 1 and 2).
Synthesis of main group halides stabilised by pyridyl donors
Although the chelating properties of amido ligands bearing
pendant pyridyl donors have been employed extensively in the
chemistry of early transition metals,6,7,25–33 their use in stabi-
lising analogous main group derivatives remains little exploi-
ted.8–18 We have therefore sought to extend the coordination
chemistry of the 6-amido-2-picolyl ligand framework by using
it to stabilise otherwise highly Lewis acidic main group halides
(exemplified by halides of boron and tin).
Fig. 1 Molecular structure of 1c, with ellipsoids drawn at the 50%
probability level.
Synthesis of amido boron dihalide complexes 1a–c, in which
the coordination sphere of the boron centre is augmented by
interaction with a tethered pyridyl donor, can be accomplished
by the reaction of the corresponding bis(trimethylsilyl) sub-
stituted ligand framework 6-[(Me3Si)2N]-2-Me-C5H3N with
the appropriate boron trihalide (Scheme 1). Low temperature
(À78 ꢀC) and slow addition of the boron trihalide are required
in order to prevent further reaction leading to elimination of a
second equivalent of trimethylsilyl halide and the formation of
bifunctional adducts of the type reported by Raston, Junk and
colleagues.14 In the case of the fluoride and chloride derivatives
1a and 1b, spectroscopic and analytical data are in agreement
with the proposed formulation, although it proved impossible
to grow single crystals of sufficient quality of either compound
in order to probe the structural properties of the complexes.14
In the case of the bromide derivative, 1c, the EI mass spectrum
rather surprisingly features a predominant peak corresponding
to the ion [M þ Br]þ. In order to ascertain whether this relates
to some degree of oligomerisation or significant secondary
interaction in the solid state (rather than being an artefact of
the EI-MS experiment), an X-ray crystal structure analysis was
therefore desirable. Single crystals of 1c proved to be accessible
via slow cooling to À30 ꢀC of solutions in hexanes, and the
crystal structure, depicted in Fig. 1, shows the compound to be
monomeric, featuring no unusual secondary intermolecular
interactions. 1c possesses a rather strained four-membered ring
composed of one boron, one carbon and two nitrogen atoms.
That the chelate ring is planar is demonstrated by the sum of
the internal angles being 360ꢀ. In addition, the two N–C dis-
tances within the ring are virtually identical [N(1)–
Fig. 2 Proposed ‘diaza-allyl’ formulation for 1c.
a diaza-allyl formulation (see Fig. 2). This is also reflected in
the scarcely significant differences between N–B distances
+
+
[N(1)–B(1) ¼ 1.577(9) A, N(2)–B(1) ¼ 1.547(9) A]. A similar
situation is found for the Li–N and N–C distances in the
lithium compound [6-(Me3SiN)-2-Me-C5H3N]Li(tmeda).10 By
means of comparison, the N–B distances found in compound
2a synthesised by Burger and co-workers and featuring dis-
crete amido and amino functionalities within an analogous
+
+
four-membered ring are 1.535(8) A and 1.635(8) A (for the
N–B single and coordinative bonds, respectively).34 To our
knowledge 1c represents the first structurally characterised
example of a boron complex containing ligand I and the first
example of a 1 : 1 complex for any of the group 13 elements.13a
Boron dihalide fragments stabilised by a nitrogen donor as
part of a four-membered chelate ring are also rare; bond
lengths and angles for the four-membered ring are similar to
those found for the amidinate derivative CpMn(CO)2=C=C
+
+
C(1) ¼ 1.367(8) A, N(2)–C(1) ¼ 1.368(8) A], indicating that the
structure can almost certainly be described in terms of near
equal contributions from two resonance forms, consistent with
(Me)C(NtBu)2BCl2 and for [2-(Cl2BN)-6-Me-C5H3N]BCl2
35
synthesised by Raston et al., which contains a similar picolyl
framework to 1c.14
Similar chemistry can also be used to give access to the
tethered-pyridyl stabilised amidotin trichloride species,
3
(Scheme 1). As in the case of 1, the reaction proceeds smoothly
at À78 ꢀC, and the product is obtained as a pale yellow crys-
talline solid in yields of up to 75%. Attempts to synthesise the
tin dichloride derivative [6-(Me3SiN)-2-Me-C5H3N]2SnCl2
either by reaction of SnCl4 with two or more equivalents of 6-
[(Me3Si)2N]-2-Me-C5H3N, or by reaction of
3 with 6-
[(Me3Si)2N]-2-Me-C5H3N merely led to the recovery of 3.
Interestingly this chemistry is indicative of the mild nature of
trimethylsilyl substituted amine precursors as sources of amido
ligands. Analogous chemistry using the corresponding lithium
reagent 6-[Li(Me3Si)N]-2-Me-C5H3N as the source of ligand I,
leads exclusively to the 2 : 1 complex [6-(Me3SiN)-2-Me-
C5H3N]2SnCl2 . In this way appropriate choice of reagent can
be used to selectively give access to the required main group
complex. Spectroscopic data for 3 are in accord with the
proposed formulation, and are confirmed by the results of a
single crystal X-ray diffraction study (Fig. 3 and Table 3).
Scheme 1
New J. Chem., 2002, 26, 677–686
681

View Article Online
Fig. 4 Plot of ln(mole fraction of 4a) vs. time for the thermolysis of
4a at 75 ꢀC, as determined from integration of 1H NMR signals.y
Fig. 3 Molecular structure of one of the two independent molecules
of 3, with ellipsoids drawn at the 50% probability level.
There are two independent molecules of 3 within the
asymmetric unit, which differ only in minor details; the
molecular structure consists of a five-coordinate tin centre
bonded to three chloride ligands and one 6-(trimethylsilyl-
amido)-2-picolyl ligand. The geometry at the tin centre is best
thought of as being a distorted trigonal bipyramid, with the
sum of the angles subtended at the tin centre by the three
equatorial substituents [Cl(4), Cl(6) and N(3)] being 354.3(1)ꢀ.
The geometry of the amidopicolyl chelate enforces an angle of
62.8(1)ꢀ for the N(3)–Sn(2)–N(4) unit (as opposed to 90ꢀ for an
idealized trigonal bipyramid) and the same constraint is
responsible for bending of the N(4)–Sn(2)–Cl(5) angle from
linearity [161.5(1)ꢀ]. Interestingly, although the NCNSn che-
late ring is planar (as was also found for 1c) the bond lengths
within the ring are more consistent with discrete amido and
picolyl donors than with the delocalised diaza-allyl formula-
tion proposed for 1c. Thus, the N–C distances with the chelate
Fig. 5 Plot of ln(mole fraction of 4a) vs. time for the thermolysis of
4a at 55 ꢀC, as determined from integration of 1H NMR signals.y
of the ligand but also the occupation of distinct axial and
equatorial sites by the picolyl and amido donors, with the
bulky trimethylsilyl substituent attached to the amido function
ring (which were identical in the case of 1c), are found to be
1.376(3) and 1.354(3) A for N(3)–C(10) and N(4)–C(10), resp-
+
ensuring that it occupies the equatorial site. Thus, for example,
36
even in [(Ph3P=N)SnCl3]2
37
and [Me3SiNC(Ph)NSiMe3]-
SnCl3 which contain equivalent nitrogen donors occupying
ectively. In addition, the distance from the tin centre to the
amido nitrogen N(3) [2.075(2) A] is markedly shorter than that
+
+
both axial and equatorial sites within a trigonal biyramidal
complex, the axial Sn–N bond is longer than the equatorial
to the picolyl nitrogen N(4) [2.209(2) A]. Presumably the
widely differing Sn–N distances reflect not only the asymmetry
+
linkage {e.g., 2.092 and 2.039 A for [(Ph3P=N)SnCl3]2}. As
expected there is also a small but significant lengthening of the
+
Table 3 Selected bond lengths (A) and angles (ꢀ) for 3
+
Sn–Clax distance [2.346(1) A] compared to the Sn-Cleq bonds
[2.324(1) and 2.328(1) A]. In many respects the structure of 3
+
C10–N4
C10–N3
C10–Sn2
C14–N4
C16–Si2
C17–Si2
C18–Si2
1.354(3)
1.376(3)
2.612(3)
1.353(3)
1.861(3)
1.853(3)
1.856(3)
N3–Si2
1.770(2)
2.075(2)
2.209(2)
2.3279(9)
2.3464(10)
2.3244(10)
resembles that of [6-(Cl2AsN)-2-Me-C5H3N]AsCl2 , which is
described by Raston and co-workers as having an analogous
trigonal bipyramidal geometry at As(1) (with one of the
equatorial sites occupied by the As lone pair), and which
possesses similar asymmetry in the As(1)–N distances.12
In contrast to the chemistry outlined above, reaction of
bis(trimethylsilyl)aminomethyl pyridine, 2-[(Me3Si)2NCH2]-
C5H4N, with BX3 (X ¼ F, Cl, Br) at room temperature leads
not to the pyridyl stabilised amidoboron dihalides 5, but to the
simple adducts 4 (Scheme 1). The structures of 4a–c have been
deduced from spectroscopic data, with pyridyl rather than
amine coordination at the boron centre being implied by 11B
and 1H NMR data that are remarkably similar to those found
for the analogous (pyridine)ÁBX3 adducts.38 Cyclisation from 4
to 5 with concomitant loss of trimethylsilyl halide requires
extremely forcing conditions. Monitoring of the thermolysis of
N3–Sn2
N4–Sn2
Cl4–Sn2
Cl5–Sn2
Cl6–Sn2
N4–C10–N3
N4–C10–C11
N3–C10–C11
N4–C10–Sn2
N3–C10–Sn2
C11–C10–Sn2
C10–N3–Si2
C10–N3–Sn2
Si2–N3–Sn2
C14–N4–C10
C14–N4–Sn2
C10–N4–Sn2
N3–Sn2–N4
N3–Sn2–Cl6
N4–Sn2–Cl6
N3–Sn2–Cl4
N4–Sn2–Cl4
109.8(2)
119.0(2)
131.2(2)
57.70(13)
52.14(12)
176.35(19)
132.88(18)
96.30(16)
130.69(11)
123.3(2)
145.54(19)
91.09(15)
62.76(8)
126.03(6)
92.58(6)
Cl6–Sn2–Cl4
N3–Sn2–Cl5
N4–Sn2–Cl5
Cl6–Sn2–Cl5
Cl4–Sn2–Cl5
N3–Sn2–C10
N4–Sn2–C10
Cl6–Sn2–C10
Cl4–Sn2–C10
Cl5–Sn2–C10
N3–Si2–C17
N3–Si2–C18
C17–Si2–C18
N3–Si2–C16
C17–Si2–C16
C18–Si2–C16
108.33(3)
98.82(6)
161.45(6)
97.40(4)
97.33(4)
31.56(8)
31.20(8)
111.25(6)
110.03(6)
130.32(6)
109.02(12)
111.36(12)
109.63(16)
108.36(13)
110.46(15)
108.00(14)
1
a solution of 4a in benzene by H NMR, for example, reveals
that the cyclisation process is only complete after 3 days at
75 ꢀC or 23 days at 55 ꢀC. Furthermore, integration of the
119.98(6)
94.28(6)
1
intensities of the H NMR signals due to the methylene pro-
tons of 4a and 5a at intervals over the period of the reaction
682
New J. Chem., 2002, 26, 677–686

View Article Online
(Figs. 4 and 5) demonstrates that the reaction is first order in
4a and that the rate constant, k1 , for the cyclisation process is
of the order of 4 Â 10À6 sÀ1 at 75 ꢀC and 7 Â 10À7 sÀ1 at 55 ꢀC.
For the analogous chloride and bromide species 4b and 4c it
proved impossible to drive the cyclisation reaction to com-
pletion, such that thermolysis invariably leads to inseparable
mixtures of products.
Furthermore, although the pyridyl donor stabilised amido-
boron and tin halides 1 and 3 can be synthesised in good to
moderate yields by this route the analogous reactions to form
alkoxoboron dihalides from 6-(Me3SiO)-2-Me-C5H3N are
much less straightforward. More forcing conditions are
required in order to bring about the formation of the O–B bond
(with elimination of the trimethylsilyl halide) and as a con-
sequence inseparable mixtures of products are always formed.
We have therefore sought alternative ligand frameworks for the
stabilisation of alkoxo-substituted main group halides.
Synthesis of alkoxoboron halides stabilised by amine or
thioether donors
Alkoxoboron mono- and di-halides containing a nitrogen or
sulfur donor tethered by a two carbon aliphatic bridge (e.g., 10
and 12) can be synthesised from the bis(trimethylsilyl) ether
derivatives of N-methyldiethanolamine or 2,20-thiodiethanol
according to the chemistry outlined in Scheme 2. Initial for-
mation of the adduct 8 is followed by stepwise cyclisation to
give successively the base-stabilised di- and mono-halides 9
and 10. For X ¼ F and E ¼ NMe, formation of the adduct 8a
proceeds cleanly at room temperature, whereas the subsequent
cyclisation steps require heating to 55 ꢀC over a period of 48–
72 h. Monitoring the reaction by 11B and 19F NMR it is
possible to observe signals due to the intermediate 9a and the
final product 10a even at relatively short reaction times (Fig. 6).
Over a period of 96h signals due to 8a and 9a completely
disappear, leaving only the di-cyclised product 10a. By mea-
suring the intensities of the 11B signals due to 8a, 9a and 10a at
hourly intervals it proved possible to obtain plots of con-
centration vs. time for these three species (see Fig. 7).z
Fig. 6 11B{1H} NMR signals for 8a, 9a and 10a in benzene at 55ꢀ C.
After an initial induction period in which 8a is formed from
6 and BF3ÁOEt2 , the concentration of 8a then displays the
expected decay by a first-order process. A plot of ln(mole
fraction) vs. time (for 5 h < t < 25 h) shows a linear relationship
from which measurement of the gradient gives a first-order
rate constant of k1 ¼ 2.7 Â 10À5 sÀ1 for the conversion of 8a
into the mono-cyclised species 9a.x Complex 9a shows the
concentration/time behaviour expected of an intermediate
species, with the final, di-cyclised product 10a becoming the
predominant species in solution for t > 6h. The concentration
vs. time behaviour for 9a and 10a could also be modelled using
graph-fitting analysis.{
x A plot of ln(mole fraction) vs. time (h) for the adduct 8a at 55 ꢀC
over the period 5 h < t < 25 h (Fig. 8) could be fitted to the linear
relationship:
ln(mole fraction) ¼ À0.0984t À 0.3587 (R2 ¼ 0.9982)
from which it is possible to obtain a value for the first-order rate
constant, k1 , of 9.8 Â 10À2 hÀ1 or 2.7 Â 10À5 sÀ1. Data for t < 5 h were
omitted from the calculation as NMR measurements indicated the
presence of small quantities of BF3ÁOEt2 in the reaction mixture during
this period. Consequently, for t < 5 h, 8a is still being formed from
BF3ÁOEt2 and 6 and the concentration of 8a will be a more complex
function of time over this period. Graph fitting was carried out using
the program SigmaPlot.
Scheme 2
{ A plot of mole fraction vs. time (h) for the intermediate 9a at 55 ꢀC
could be fitted to the expression:
z Estimates of the concentrations of 8a, 9a and 10a in solution were
made by integration of the respective signals in the 11B NMR spectrum
of the reaction mixture (Fig. 6). This approach makes the assumption
of equal relaxation times for the 11B nuclei in each of the three species.
The error likely to be introduced by making this assumption is small;
the similar chemical environments of the 11B nuclei in each of the three
species mean that the relaxation times in each case are likely to be
similar (and short). There is considerable precedent for this type of
approach in determining the concentrations of boron-containing
species in solution.39 A similar approach making use of the 19F
NMR data for the reaction mixture was complicated by the
overlapping nature of two of the resonances.
mole fraction ¼ 4.813(eÀ0.1759t À eÀ0.1879t) (R2 ¼ 0.9778)
from which values of 0.1879 and 0.1759 hÀ1 can be calculated for k1
and k2 . Conversion to sÀ1 gives k1 ¼ 5.2 Â 10À5 sÀ1 and k2 ¼ 4.9 Â 10À5
s
À1. Similarly, the concentration vs. time (h) behaviour for the product
10a is modelled by the expression
mole fraction¼ 0.5451(1 À eÀ0.1538t ) À 0.5177(1 À eÀ0.1524t) (R2 ¼ 0.9959)
which yields k1 ¼ 4.3 Â 10À5 sÀ1 and k2 ¼ 4.3 Â 10À5
s
À1. Graph fitting
was carried out using the program SigmaPlot.
New J. Chem., 2002, 26, 677–686
683

View Article Online
Fig. 7 Plot of mole fraction vs. time for compounds 8a (diamonds),
9a (squares) and 10a (triangles) during the thermolysis of 8a at 55 ꢀC
(as determined from integration of 11B NMR signals).x
Fig. 9 Molecular structure of 10a, with ellipsoids drawn at the 30%
probability level.
temperatures are needed to complete the first cyclisation step,
and despite prolonged reaction times (3–4 weeks) at elevated
temperatures (120 ꢀC) it proved impossible to drive the final
cyclisation step to completion. Furthermore, the cyclisation
steps, which required somewhat forcing conditions for X ¼ F,
and which could not be driven fully to completion for X ¼ Cl,
appear to be less favourable still for X ¼ Br. In this case
similar chemistry leads to a mixture of products in which the
di-cyclised analogue 10c is thought to be a very minor
component.
By contrast, the analogous reaction chemistry involving
stabilisation by a thioether (rather than amino) donor is much
more facile. The thioether adduct [(Me3SiOCH2CH2)2SÁBF3]
(11a) can be identified in the 11B spectrum of the reaction
mixture (formed by adding BF3ÁOEt2 to 7) after short times at
room temperature (t < 1 h). However, cyclisation appears to
occur rapidly under these conditions and after 6h the only
boron-containing species present in solution is the thioether
stabilised dialkoxoboron fluoride 12a. 12a has been char-
acterised by spectroscopic and analytical data, and although
single crystals suitable for X-ray diffraction could not be
obtained, a structure analogous to that found for 10a is
proposed.
Fig. 8 Plot of ln(mole fraction of 8a) vs. time for the thermolysis of
8a at 55 ꢀC, as determined from integration of 11B NMR signals.x
The time dependence of the concentration of the inter-
mediate 9a could be fitted to a difference of exponentials
expression for which values of 5.2 Â 10À5 sÀ1 and 4.9 Â 10À5
sÀ1 were obtained for the rate constants k1 and k2 for the two
cyclisation steps. An analogous treatment of the concentration
of the final product 10a gives k1 ¼ 4.3 Â 10À5 sÀ1 and
k2 ¼ 4.3 Â 10À5
s
À1. The kinetic data for the cyclisation reac-
tions are therefore consistent with two-first order steps for
which the rate constants are of the order of 4 Â 10À5 sÀ1 and
are identical within the error limits of the experiment.
Discussion
Spectroscopic and analytical data for 10a are in accord with
the proposed structure, and these inferences were confirmed by
the results of a single crystal X-ray diffraction study (see
Fig. 9). The molecule is disordered about a crystallographically
imposed mirror plane and although this was successfully
modelled by assuming 50% occupancy for the disordered
atoms, detailed discussion of the structural parameters for this
part of the molecule is not possible. 10a represents the first
crystallographically characterised example of an alkoxo boron
halide stabilised by the coordination of an amine nitrogen
attached through a one- or two-atom tether. This form of
chelate does find precedent, however, amongst aryl derivatives
of group 13 elements, and the structural parameters for 10a are
broadly in agreement with those found for such analogous
five-membered ring systems.40
Synthesis of intramolecular base-stabilised main group halides
has been accomplished by the reaction of trimethylsilyl sub-
stituted amines or ethers with the corresponding halide pre-
cursor. This route has been shown to be applicable to a range
of amido- and alkoxoboron and -tin halides stabilised by
nitrogen or sulfur donors attached by one- or two-atom
tethers. Structural details have been elucidated for compounds
stabilised by bases tethered through either one (1c and 3) or
two arms (10a). Studies of kinetics and reactivity have also
allowed us to reach several conclusions concerning the
mechanism of such reactions.
(i) The synthesis of amido-substituted halides is more facile
than the corresponding reaction used to make alkoxo deriva-
tives, that is
RnE–SiMe3 þ BX3 ! RnE–BX2 þ Me3SiX
Similar reaction chemistry can be used to give access to
analogous derivatives containing other halides (e.g., the
chloride complexes 8b, 9b and 10b). Formation of the adduct
8b is easily accomplished from 6 and BCl3 , but the cyclisation
processes leading to the dichloride 9b and monochloride 10b
proceed much less readily than in the fluoride case. Higher
occurs more readily for RnE ¼ R2N than for RnE ¼ RO.
Hence, for example, elimination of Me3SiX from a reaction
mixture containing 6-[(Me3Si)2N]-2-Me-C5H3N and BX3
occurs readily at À78 ꢀC, with careful control of tempera-
ture being required to prevent further reaction, whereas the
684
New J. Chem., 2002, 26, 677–686

View Article Online
and the Nuffield Foundation for the provision of computing
equipment.
References
1
See, for example: (a) L. H. Gade, Chem. Commun., 2000, 173; (b)
R. R. Schrock, Acc. Chem. Res., 1997, 30, 9; (c) J. G. Verkade,
Acc. Chem. Res., 1993, 26, 483.
2
(a) A. H. Cowley, F. P. Gabbaı, D. A. Atwood, C. J. Carrano,
L. M. Mokry and M. R. Bond, J. Am. Chem. Soc., 1994, 116,
1559; (b) L. Contreras, A. H. Cowley, F. P. Gabbaı, R. A. Jones,
C. J. Carrano and M. R. Bond, J. Organomet. Chem., 1995, 489,
C1; (c) H. S. Isom, A. H. Cowley, A. Decken, F. Sissingh, S.
Corbelin and R. J. Lagow, Organometallics, 1995, 14, 2400.
See, for example: (a) M. H. Chisholm and I. P. Rothwell, in
Comprehensive Coordination Chemistry, eds. G. Wilkinson, R. D.
Gillard and J. A. McCleverty, Pergamon, Oxford, 1987, vol. 2,
p. 161; (b) M. F. Lappert, P. P. Power, A. R. Sanger and R. C.
Srivastava, Metal and Metalloid Amides, Ellis Horwood-Wiley,
Chichester, 1980
See, for example: (a) D. C. Bradley, Chem. Rev., 1989, 89, 1317;
(b) R. C. Mehrotra, in Advances in Inorganic Chemistry and
Radiochemistry, eds. H. G. Emeleus and A. G. Sharpe, Academic
Press, London, 1983, vol. 26, pp. 269–335; (c) D. C. Bradley, R. C.
Mehrotra and P. D. Gaur, Metal Alkoxides, Academic Press,
London, 1978
3
4
5
Scheme 3
corresponding reactions with 6-(Me3SiO)-2-Me-C5H3N have
proved impossible to drive to completion.
(ii) The synthesis of alkoxoboron halides by the elimination
of trimethsilyl halide is more facile in the case of fluoride than
for chloride or bromide, that is
(a) G. J. P. Britovsek, V. C. Gibson and D. F. Wass, Angew.
Chem.,Int. Ed. , 1999, 38, 428; (b) J. D. Scollard and D. H.
McConville, J. Am. Chem. Soc., 1996, 118, 10008.
6See, for example: H. Fuhrmann, S. Brenner, P. Arndt and R.
Kempe, Inorg. Chem., 1996, 35, 6742.
RO–SiMe3 þ BX3 ! RO–BX2 þ Me3SiX
proceeds more readily in the case of X ¼ F than for X ¼ Cl
7
8
A. Spannenberg, M. Oberthur, H. Noss, A. Tillack, P. Arndt and
R. Kempe, Angew. Chem.,Int. Ed. , 1998, 37, 2079.
A. M. Arif, D. C. Bradley, H. Dawes, D. M. Frigo, M. B.
Hursthouse and B. Hussain, J. Chem. Soc.,Dalton T rans. , 1987,
2159.
or Br.
(iii) Cyclisation by elimination of trimethylsilyl halide occurs
more readily for halides stabilised by sulfur donors than in the
analogous amino systems, that is
9
(a) W. Liu, A. Hassan and S. Wang, Organometallics, 1997, 16,
4257; (b) J. Ashenhurst, L. Brancaleon, S. Gao, W. Liu, H.
Schmider, S. Wang, G. Wu and Q. G. Wu, Organometallics, 1998,
17, 5334; (c) Q. Wu, M. Esteghamatian, N.-X. Hu, Z. Popovic, G.
Enright, S. R. Breeze and S. Wang, Angew. Chem.,Int. Ed. , 1999,
38, 985.
occurs more readily for E ¼ S than for E ¼ NMe.
(iv) Analysis of kinetic data for several systems has shown
that the trimethylsilyl halide evolving cyclisation step is a first-
order process, and that in the case of systems offering the
possibility for two step-wise cyclisation reactions, there is no
significant difference in the rates of these two steps.
10 L. M. Engelhardt, G. E. Jacobsen, P. C. Junk, C. L. Raston,
B. W. Skelton and A. H. White, J. Chem. Soc.,Dalton T rans. ,
1988, 1011.
11 L. M. Engelhardt, G. E. Jacobsen, W. C. Patalinghug, B. W.
Skelton, C. L. Raston and A. H. White, J. Chem. Soc.,Dalton
T rans., 1991, 2859.
12 (a) C. L. Raston, B. W. Skelton, V.-A. Tolhurst and A. H. White,
Polyhedron, 1998, 17, 935; (b) C. L. Raston, B. W. Skelton, V.-A.
Tolhurst and A. H. White, J. Chem. Soc.,Dalton T rans. , 2000,
1279.
13 (a) C. Jones, P. C. Junk, S. G. Leary and N. A. Smithies, Main
Group Metal Chem., 2001, 24, 383; (b) C. Jones, P. C. Junk, S. G.
Leary, N. A. Smithies and J. Steed, Inorg. Chem. Commun.,
in press.
14 L. M. Engelhardt, G. E. Jacobsen, P. C. Junk, C. L. Raston and
A. H. White, J. Chem. Soc.,Chem. Commun. , 1990, 89.
15 Y. Zhou and D. Richeson, Organometallics, 1995, 14, 3558.
16L. M. Engelhardt, M. G. Gardiner, C. Jones, P. C. Junk, C. L.
Raston and A. H. White, J. Chem. Soc.,Dalton T rans. , 1996, 3053.
17 H. Gonitzka and D. Stalke, Eur. J. Inorg. Chem., 1998, 311.
18 Q. Zhao, H.-S. Sun, W.-Z. Chen, Y.-J. Liu and X.-Z. You,
J. Organomet. Chem., 1999, 556, 159.
19 (a) H. Braunschweig, J. Muller and B. Ganter, Inorg. Chem., 1996,
35, 7443; (b) G. J. Irvine, C. E. F. Rickard, W. R. Roper, A.
Williamson and L. J. Wright, Angew. Chem.,Int. Ed. , 2000, 39,
948.
These observations are consistent with
a cyclisation
mechanism in which the rate-determining step involves
intramolecular nucleophilic attack at the boron centre by the
nitrogen or oxygen centre of the trimethylsilyl substituted
amine or ether (Scheme 3). That the rate of attack is faster
for the formation of amido rather than alkoxo linkages
[(i) above] is consistent with the greater nucleophilicity of
amine rather than ether donors. In addition, the reduced rates
observed for boron centres coordinated by nitrogen rather
than sulfur donors [(iii) above] is consistent with the greater
stabilisation of the pre-cyclised adduct by the more strongly
electron-donating nature of the N ! B interaction compared
to S ! B. The observation of enhanced rates for fluorides
(compared to chlorides or bromides) may be due in part to
the more favourable thermodynamics associated with the
elimination of trimethylsilyl fluoride (rather than Me3SiCl or
Me3SiBr).
20 T. Yasue, Y. Kawano and M. Shimoi, Chem. L ett., 2000, 58.
21 S. Aldridge, R. J. Calder, K. A. M. Malik and J. W. Steed,
J. Organomet. Chem., 2000, 614, 188.
22 S. Aldridge, R. J. Calder, A. A. Dickinson, D. J. Willock and
J. W. Steed, Chem. Commun., 2000, 1377.
23 M. Westerhausen, T. Bollwein, N. Makropoulous, T. M. Rotter,
T. Habereder, M. Suter and H. Noth, Eur. J. Inorg. Chem., 2001,
851.
24 (a) Z. Otwinowski and W. Minor, in Methods in Enzymology, eds.
C. W. Carter and R. M. Sweet, Academic Press, New York, 1996,
vol. 276, p. 307; (b) G. M. Sheldrick, SHELX97, Programs for
Acknowledgements
We would like to acknowledge the support of the EPSRC, the
Royal Society and Cardiff University for the funding of this
project. The assistance of the EPSRC National Mass Spec-
trometry Service Centre, University of Wales Swansea, is also
gratefully acknowledged. JWS thanks the EPSRC and King’s
College London for the provision of an X-ray diffractometer
New J. Chem., 2002, 26, 677–686
685

View Article Online
Crystal Structure Analysis (release 97-2), University of Gottingen,
Germany, 1998
33 R. Kempe, S. Brenner and P. Arndt, Organometallics, 1996, 15,
1071.
25 A. Spannenberg, A. Tillack, P. Arndt, R. Kirmse and R. Kempe,
Polyhedron, 1998, 17, 845.
26A. Spannenberg, H. Fuhrmann, P. Arndt, W. Baumann and R.
Kempe, Angew. Chem.,Int. Ed. , 1998, 37, 3363.
27 H. Noss, M. Oberthur, C. Fischer, W. P. Kretschmer and R.
Kempe, Eur. J. Inorg. Chem., 1999, 2283.
28 H. K. Lee, Y.-L. Wong, Z.-Y. Zhou, Z.-Y. Zhang, D. K. P. Ng
and T. C. W. Mak, J. Chem. Soc.,Dalton T rans. , 2000, 539.
29 A. Spannenberg, P. Arndt and R. Kempe, Angew. Chem.,Int. Ed. ,
1998, 37, 832.
30 G. Hillebrand, A. Spannenberg, P. Arndt and R. Kempe, Orga-
nometallics, 1997, 16, 5585.
34 A. Ansorge, D. J. Brauer, H. Burger, F. Dorrenbach, T. Hagen,
G. Pawelke and W. Weuter, J. Organomet. Chem., 1991, 407, 283.
35 M. R. Terry, L. A. Mercando, C. Kelley, G. L. Geoffroy, P.
Nombel, N. Lugan, R. Mathieu, R. L. Ostrander, B. E. Owens-
Waltermire and A. L. Rheingold, Organometallics, 1994, 13, 843.
36H. W. Roesky, U. Seseke, M. Noltemeyer and G. M. Sheldrick, Z.
Naturforsch.,B , 1988, 43, 1130.
37 C. Ergezinger, F. Weller and K. Dehnicke, Z. Naturforsch.,B ,
1988, 43, 1621.
38 G. R. Eaton and W. N. Lipscomb, NMR Studies of Boron Hy-
drides and Related Species, Benjamin, New York 1969
39 See, for example: S. Aldridge, M. Shang and T. P. Fehlner, J. Am.
Chem. Soc., 1998, 120, 2586.
31 M. Oberthur, G. Hillebrand, P. Arndt and R. Kempe, Chem. Ber.,
1997, 130, 789.
32 R. Kempe and P. Arndt, Inorg. Chem., 1996, 35, 2644.
40 See, for example: S. Caron and J. M. Hawkins, J. Org. Chem.,
1998, 63, 2054.
686
New J. Chem., 2002, 26, 677–686