Ring strain in boroxine rings: Computational and experimental considerations

Article abstract of DOI:10.1016/S0022-328X(01)01060-9

B3LYP/6-311+G(d) calculations indicate that (HBO)₃ (4) and (HBO)₄ (5) possess (zero-point energy corrected) strain enthalpies of 11.4 and 31.6 kJmol^-1, respectively. The absence of eight-membered (RBO)₄ rings is attributed to a combination of ring strain and the lability of the B-O bond. The synthesis, characterization and molecular structure of (PhBO)₃·pyridine (1) are described and chemical phenomena related to the addition of amines to triorganoboroxine rings are rationalized in terms of relief of ring strain in 4.

Full text of DOI:10.1016/S0022-328X(01)01060-9

Journal of Organometallic Chemistry 633 (2001) 149–156
www.elsevier.com/locate/jorganchem
Ring strain in boroxine rings: computational and experimental
considerations
Jens Beckmann a, Dainis Dakternieks a,*, Andrew Duthie a, Allan E.K. Lim a,
Edward R.T. Tiekink b,c
a Centre for Chiral and Molecular Technologies, Deakin Uni6ersity, Geelong, Vic. 3217, Australia
b Department of Chemistry, The Uni6ersity of Adelaide, Adelaide, SA 5005, Australia
c Department of Chemistry, National Uni6ersity of Singapore, 117543, Singapore
Received 2 May 2001; accepted 29 May 2001
Abstract
B3LYP/6-311+G(d) calculations indicate that (HBO)3 (4) and (HBO)4 (5) possess (zero-point energy corrected) strain
enthalpies of 11.4 and 31.6 kJ mol−1, respectively. The absence of eight-membered (RBO)4 rings is attributed to a combination
of ring strain and the lability of the BꢀO bond. The synthesis, characterization and molecular structure of (PhBO)3·pyridine (1)
are described and chemical phenomena related to the addition of amines to triorganoboroxine rings are rationalized in terms of
relief of ring strain in 4. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: DFT analysis; Boroxine rings; Ring strain
1. Introduction
recent publication, Gillespie et al. demonstrated that
the boronꢀoxygen bonds in boron oxides are predomi-
nantly ionic in nature and therefore, the geometries of
these molecules would not be satisfactorily understood
using a covalent model [20].
In 1958, Snyder et al. reported the synthesis of a 1:1
complex between triphenylboroxine and pyridine,
(PhBO)3·pyridine (1), and proposed that the coordina-
tion of the pyridine occurs at one boron atom within
the B3O3 ring (Chart 1) [21].
Over the years, a diverse range of other 1:1 com-
plexes between triorganotriboroxines and nitrogen
donors, (RBO)3·L (R=alkyl, aryl; L=amines, hy-
drazines), have been prepared and some examples were
investigated by X-ray diffraction [22–31]. It is worth
The dehydration of organoboronic acids, RB(OH)2
(R=alkyl, aryl), exclusively produces six-membered
triorganoboroxine rings, (RBO)3, regardless of the na-
ture of the organic substituents [1]. There is one report
describing the synthesis of a four-membered ring (2,4,6-
t-Bu3ꢀC6H3ꢀBO)2 [2], however, the product was not
completely characterized. It has also been suggested
that hydrolysis of triorganoboroxine rings by trace
amounts of water may lead to erroneously low molecu-
lar weight determinations [3]. The preference for six-
membered rings was attributed to resonance
stabilization of a conjugated six p-electron system by
Snyder et al. in 1938 [4] and since then, the degree of
aromaticity in triorganoboroxines has been the subject
of vigorous debate [5–19].
Recent work on the magnetic properties of tri-
organoboroxines and the results of DFT calculations
suggest that the lone pair electrons are mostly localized
at the oxygen atoms and consequently, triorganoborox-
ines possess very little aromatic character [15–19]. In a
* Corresponding author. Fax: +61-35-2272218.
E-mail address: dainis@deakin.edu.au (D. Dakternieks).
Chart 1.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)01060-9

150
J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
We also revisit the synthesis of (PhBO)3·pyridine
(1) and report full details of its molecular structure
and compare it with those of the isoelectronic anion
[Ph2B(OBPh)2O]− (2) [33,34] and the parent triphenyl-
boroxine (PhBO)3 (3) [39,40] (Chart 1).
mentioning that most attempts to prepare such com-
plexes with a stoichiometry other than 1:1 have
failed, the only notable exception being tris(2-(N,N-
dimethylhydrazo)phenyl)boroxine, in which coordina-
tion of two out of three potentially intramolecular
coordinating ligands is found [32]. However, the coor-
dination in the latter compound is weaker than in the
aforementioned 1:1 adducts. According to NMR stud-
ies, complexes between triorganoboroxines and amines
undergo a dissociation–recombination process that is
fast on the NMR time scale at room temperature.
The activation energies associated with this effect
range from 39 to 54 kJ mol−1 [24].
Moreover, the reaction of oxybis(diphenylborane),
Ph2BOBPh2, with four equivalents of phenylboronic
acid, PhB(OH)2, in the presence of tetramethylammo-
nium hydroxide, NMe4OH, provides the anion
[Ph2B(OBPh)2O]− (2) (counterion [NMe4]+) [33],
which is isoelectronic with the amine complex
(PhBO)3·pyridine (1) (Chart 1). The same anion is
formed serendipitously by the hydrolysis of the te-
traphenylborate anion, Ph4B−, in the presence of the
tetrakis[(tri - tert - butylphosphine)gold(I)]phosphonium
cation [{(t-Bu)3PAu}4P]+ [34]. In a similar manner,
the analogous anion [F2B(OBF)2O]− is formed by the
hydrolysis of BF3 in the presence of the hexakis[(tri-
iso-propylphosphane)gold(I)]methanium dication [(i-
Pr3PAu)6C]2+ [35].
2. Discussion
2.1. Differences in reacti6ity between (PhBO)3·pyridine
(1) and (PhBO)3 (3): experimental e6idence of ring
strain?
The reaction of phenylboronic acid, PhB(OH)2,
with pyridine in a 3:1 molar ratio proceeds via com-
plete condensation and provides the 1:1 complex be-
tween
triphenylboroxine
and
pyridine,
(PhBO)3·pyridine (1), in an almost quantitative yield
(Eq. (1)).
(1)
It is important to note that the reaction tempera-
ture during the preparation of (PhBO)3·pyridine (1)
never exceeded 35 °C (reflux temperature of ether). In
sharp contrast, the condensation of phenylboronic
acid, PhB(OH)2, to give triphenylboroxine, (PhBO)3
(3), in the absence of pyridine requires strong desic-
cants or prolonged heating in a high-boiling solvent,
such as toluene or xylene, in a Dean–Stark appara-
tus. Evidently, formation of the six-membered tri-
organoboroxine ring is more facile in the presence of
pyridine. Repeating the reaction between PhB(OH)2
and pyridine in an equimolar ratio under the same
reaction conditions also resulted in the isolation of
(PhBO)3·pyridine (1) as the sole product.
The hydrolysis of triphenylboroxine, (PhBO)3 (3),
which can be considered to be the reverse reaction to
the aforementioned condensation, occurs readily in
organic solvents at room temperature with (traces of)
water to give phenylboronic acid, PhB(OH)2 [41,42].
Triorganoboroxines containing small substituents such
as the methyl group show an even higher reactivity
toward hydrolysis [43]. However, solutions of the
complex (PhBO)3·pyridine (1) remain unchanged un-
der the same conditions. These observations suggest
that the complex (PhBO)3·pyridine (1) possesses a
higher stability towards ring-opening compared to its
parent compound (PhBO)3 (3). One possible explana-
tion is the assumption that ring strain is present in
the parent compound 1. The geometry change of one
boron centre from trigonal planar to tetrahedral upon
complexation may relieve this ring strain. This postu-
late will be examined in detail in the following sec-
tions.
The chemistry of triorganoboroxines is dominated
by ring-cleavage reactions [36]. On the other hand,
1:1 triorganoboroxine–amine complexes are so stable
and easily formed that a convenient literature prepa-
ration of (MeBO)3 from B(OMe)3 involves the initial
formation of (MeBO)3·pyridine followed by removal
of pyridine [37].
In this paper, DFT calculations are applied to in-
vestigate the following questions: (i) Given that tri-
organoboroxines
possess
very
little
aromatic
character, what is the real reason for the exclusive
preference of six-membered (RBO)3 rings over eight-
membered (RBO)4 rings? (ii) Why does the addition
of amines to triorganoboroxines rarely proceed be-
yond 1:1 stoichiometry? (iii) Why are triorganoborox-
ines more susceptible to ring cleavage compared to
their 1:1 amine adducts?
We propose that all three questions may be ad-
dressed using ring strain considerations. To the best
of our knowledge, there are no experimental or com-
putational reports in the literature concerning ring
strain in boron–oxygen rings. The cubic form of the
tetramer (HBO)4 has been studied computationally
[38], however, we have been unable to locate any
computational studies concerning the eight-membered
ring form of (HBO)4.

J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
151
Table 1
2.2. DFT study on ring strain in (HBO)3 (4) and
(HBO)4 (5)
Selected geometric parameters of the B3LYP/6-311+G(d)-optimized
geometries of 4–9
The ring strain in the model compounds (HBO)3 (4)
and (HBO)4 (5) was evaluated using the following
group-equivalent reactions at the B3LYP/6-311+G(d)
level of theory (Eqs. (2) and (3)).
,
,
Molecule
BꢀH (A)
BꢀO (A)
BꢀOꢀB (°)
OꢀBꢀO (°)
4
1.187
1.377
120.6
120.0(6)
138.7
–
119.4
120.0(6)
122.5
expt. a
1.192(7)
1.189
1.193, 1.199 1.352
1.375(2)
1.366
5
6
7
–
(2)
1.190 (B1)
1.352
126.8
117.0
(B1ꢀOH)
1.366
(OꢀB2ꢀO)
119.0
1.193 (B2)
(B2ꢀO)
1.380
(OꢀB1ꢀO)
(3)
(B1ꢀO)
8
9
1.193, 1.196 1.370
1.190 (B1)
125.9
126.4
–
Eqs. (2) and (3) were formed using the group-equiva-
lent procedure of Bachrach [44] in order to minimize
contributions from other chemical changes not associ-
ated with the ring strain.
1.351
116.9
(B1ꢀOH)
1.364
(B2ꢀO1)
1.368
(B2ꢀO2)
1.381
(B1ꢀO1)
(B2ꢀO2ꢀB2) (OꢀB2ꢀO)
127.0 119.0
(B1ꢀO1ꢀB2) (OꢀB1ꢀO)
1.193 (B2)
Selected geometric parameters of the optimized
model compounds 4–9 are listed in Table 1. The calcu-
lated geometric parameters for 4 are found to be in
excellent agreement with experimental values. The atom
numbering scheme for 7 and 9 used in Table 1 is shown
in Chart 2. Comparing the various BꢀOꢀB and OꢀBꢀO
angles between the rings 4 and 5 and the open-chain
compounds 7–9 reveals that while all the OꢀBꢀO an-
gles remain in a reasonably narrow range, the BꢀOꢀB
angles in 4 and 5 are 5.9° lower and 12.2° higher,
respectively, than the average BꢀOꢀB angle of the
open-chain compounds 7–9. This indicates that angular
strain may be present in 4 and 5. The calculated BꢀO
bond lengths for all six model compounds lie in the
a Experimental values from C.H. Chang, R.F. Porter, S.H. Bauer,
Inorg. Chem. 8 (1969) 1689.
,
narrow range between 1.351 and 1.381 A, regardless of
whether they are open-chain or cyclic.
A
puckered ring (5) (framework group D2d
[2SGD(B2H2), X(O4)]) and ‘cubic’ structure (10) (Td
symmetry) were located as minima on the (HBO)4
potential energy surface. The eight-membered ring (5) is
calculated to be 314.2 kJ mol−1 more stable than 10.
The planar eight-membered ring (11) (D4h symmetry) is
calculated to be a first-order saddle-point located 21.3
kJ mol−1 above 5 with an imaginary frequency corre-
sponding to deformation towards the puckered struc-
ture 5. Optimization of a puckered eight-membered ring
structure possessing the D2d [2SGD(O2), X(B4H4)]
framework group leads to the planar conformation.
The structures and selected geometric parameters of the
optimized stationary points on the (HBO)4 potential
energy surface are shown in Fig. 1.
Fig. 1. Structures and selected geometric parameters of the B3LYP/6-
311+G(d)-optimized geometries of the (HBO)4 isomers 5, 10 and 11.
According to Eq. (2), the B3LYP/6-311+G(d)-calcu-
lated strain energy (−DE) of the six-membered ring (4)
is 13.1 kJ mol−1. Incorporating zero-point energy cor-
rections results in a slightly lower strain enthalpy of
11.4 kJ mol−1. It is noteworthy that these values are
within the range of the experimentally measured strain
energy (8–17 kJ mol−1) of the six-membered siloxane
ring (Me2SiO)3 [45]. The dramatic differences in ring
Chart 2.

152
J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
Fig. 2. Molecular structure and crystallographic numbering scheme for (PhBO)3·pyridine (1).
opening polymerization (ROP) behaviour between
(Me2SiO)3 and the practically strainless eight-membered
siloxane ring, (Me2SiO)4 are attributed to this difference
in ring strain [46]. Therefore, it would not be unreason-
able to suggest that six-membered boroxine rings simi-
larly possess a small but chemically significant amount
of ring strain.
There are two reports on the molecular structure of
the parent triphenylboroxine (PhBO)3 (3) [39,40]. The
more precisely determined structure was reported to
have average OꢀBꢀO and BꢀOꢀB angles of 119.3 and
120.3°, respectively, so as to give an overall sum of
internal angles of 719.0°, which indicates that the ring is
nearly planar (sum=720°). The average BꢀO bond
The strain energy of the eight-membered ring (5)
,
distance was reported to be 1.382 A. In the structure of
according to Eq. (3) is calculated to be 34.6 kJ mol−1
.
(PhBO)3·pyridine (1), one boron atom, B(1) reveals a
tetrahedral environment as a result of the coordination
by pyridine. Compared to the parent compound, the
corresponding OꢀB(1)ꢀO bond angle decreases by 6.0°
to 113.3(3)°. In turn, the OꢀBꢀO bond angles of the
This value is reduced to 31.6 kJ mol−1 after zero-point
energy corrections. Thus, the eight-membered ring
(HBO)4 (5) possesses significantly more ring strain than
the six-membered ring (HBO)3 (4). Given the lability of
the BꢀO bond in solution [36], we suggest that the
eight-membered rings, (RBO)4, have not been observed
experimentally to date owing to the fact that their ring
strain would cause disproportionation to the corre-
sponding six-membered triorganoboroxine rings,
(RBO)3, which possess less ring strain. The aforemen-
tioned disproportionation would also be favoured on
entropic grounds.
Table 2
,
Selected bond lengths (A) and bond angles (°) for (PhBO)3·pyridine
(1)
Bond lengths
B(1)ꢀO(1)
B(2)ꢀO(1)
B(3)ꢀO(2)
B(1)ꢀN(1)
B(2)ꢀC(21)
1.454(4)
1.346(4)
1.381(5)
1.635(4)
1.562(5)
B(1)ꢀO(3)
B(2)ꢀO(2)
B(3)ꢀO(3)
B(1)ꢀC(11)
B(3)ꢀC(31)
1.462(4)
1.388(4)
1.348(5)
1.602(5)
1.561(5)
2.3. Molecular structure of (PhBO)3·pyridine (1)
Bond angles
The molecular structure of (PhBO)3·pyridine (1) is
shown in Fig. 2 and selected bond lengths and angles
are listed in Table 2. The structure is essentially molec-
ular with the closest intermolecular contact in crystal
O(1)ꢀB(1)ꢀO(3)
O(1)ꢀB(1)ꢀC(11)
O(3)ꢀB(1)ꢀC(11)
O(1)ꢀB(2)ꢀO(2)
O(2)ꢀB(2)ꢀC(21)
O(2)ꢀB(3)ꢀC(31)
B(1)ꢀO(1)ꢀB(2)
B(1)ꢀO(3)ꢀB(3)
C(2)ꢀN(1)ꢀB(1)
113.3(3)
113.6(3)
111.4(3)
120.7(4)
119.6(3)
118.5(4)
122.4(3)
121.3(3)
123.3(3)
O(1)ꢀB(1)ꢀN(1)
O(3)ꢀB(1)ꢀN(1)
N(1)ꢀB(1)ꢀC(11)
O(1)ꢀB(2)ꢀC(21)
O(2)ꢀB(3)ꢀO(3)
O(3)ꢀB(3)ꢀC(31)
B(2)ꢀO(2)ꢀB(3)
C(2)ꢀN(1)ꢀC(6)
C(6)ꢀN(1)ꢀB(1)
104.8(3)
105.7(3)
107.3(3)
119.7(3)
121.4(3)
120.0(4)
119.2(3)
118.8(3)
117.9(3)
,
lattice indicated by the separation (3.80 A) and super-
imposition (dihedral angle 2.6°) of the pyridine and a
symmetry-related C(31)–C(36) phenyl ring. This ar-
rangement is indicative of a p–p interaction; symmetry
operation: x, 1/2−y,−1/2+z.

J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
153
2.4. Computational estimation of the angular strain
components in (PhBO)3 (3) and (PhBO)3·pyridine (1)
Owing to the strong evidence present that BꢀO bonds
are predominantly ionic in nature and little delocaliza-
tion of electron density from O to B occurs in boroxi-
nes, it should be possible to qualitatively assess ring
strain in boroxines by decomposing it into its angular
strain components. Direct comparison of the sums of
angular strain components of (PhBO)3 (3) and
(PhBO)3·pyridine (1) is possible since both rings are
nearly planar and any torsional contributions to the
ring strain in both molecules will be essentially
identical.
The compounds H2BOBH2 (8) and HB(OH)2 (12)
were selected as model compounds for the BꢀOꢀB and
OꢀBꢀO linkages in (PhBO)3 (3). The conformer of 12
where the two OꢀH bonds are parallel was selected for
study in order to minimize errors resulting from tor-
sional contributions as the two BꢀO bonds flanking any
OꢀBꢀO linkage in (PhBO)3 (3) are also parallel. The
equilibrium BꢀOꢀB angle in 8 was calculated to be
125.9°, whereas the equilibrium OꢀBꢀO angle in 12 was
calculated to be 125.4°.
Both of these equilibrium angles are larger than those
observed in (PhBO)3 (3), providing an explanation to
why the ring does not pucker to reduce ring strain.
Three-dimensional geometry requires that the sum of
endocyclic angles of any six-membered ring should be
less than or equal to 720°, where the maximum sum
corresponds to a planar ring. Puckering would there-
fore necessitate a reduction of the average endocyclic
angle and force both the BꢀOꢀB and OꢀBꢀO linkages
even further away from their preferred values.
Relaxed potential energy scans by varying the
BꢀOꢀB and OꢀBꢀO angles between 115 and 135° were
performed for 8 and 12, respectively; the resulting
potential energy curves are plotted in Fig. 3. Fitting the
curves to polynomials and substituting the experimental
BꢀOꢀB and OꢀBꢀO angles from (PhBO)3 (3) indicate
that the BꢀOꢀB and OꢀBꢀO linkages have angular
strains of 1.9 and 4.6 kJ mol−1, respectively, giving a
total angular strain value of 19.5 kJ mol−1. Upon
complexation of pyridine to one boron atom in 3 to
afford 1, one OꢀBꢀO angle decreases, and the concomi-
tant increases in the remaining five endocyclic angles
cause the sum of angular strain components of the
remaining five linkages to decrease by 6.0 kJ mol−1
from 14.7 to 8.7 kJ mol−1, therefore supporting our
postulate that relief of ring strain occurs upon going
from (PhBO)3 (3) to (PhBO)3·pyridine (1) and
[Ph2B(OBPh)2O]− (2) via widening of the endocyclic
angles. The reduced ring strain of the adduct would
also contribute towards lowering the reactivity of the
boroxine ring towards hydrolysis.
Fig. 3. Energy profiles for 8 and 12 with respect to deformation of the
BꢀOꢀB and OꢀBꢀO angles, respectively.
trigonal boron atoms and the BꢀOꢀB bond angles
increase by an average of 1.8 and 0.7° to average values
of 121.1(2) and 121.0(2)°, respectively, maintaining ap-
proximately the same sum of internal angles (718.3°).
Indeed, like the parent compound (PhBO)3 (3), the
complex (PhBO)3·pyridine (1) is nearly planar with the
,
largest deviation being 0.093 A for B(2). The average
BꢀO bond lengths to the tetrahedral (d4) and trigonal
,
(d3) boron atoms amount to 1.458(3) and 1.366(3) A,
respectively. The increase in the BꢀO bond distances
upon increasing the coordination number at boron
from three to four is well known, having being reported
previously for many other boroxine–Lewis base ad-
ducts [22,24–29,31]. This observed increase has been
attributed to ‘loss of p-bonding in the two BꢀO bonds
involving this boron and a redistribution of the p-elec-
tron density around the remaining atoms of the ring’
[23]. However, Gillespie et al. have demonstrated that
this increase can be rationalized in terms of the close
packing of anion-like ligands about a cation-like central
atom [20]. Indeed, for (PhBO)3·pyridine (1), the ratio
d4/d3 is calculated to be 1.067, which is close to the
theoretical ratio of 1.061 for 4- and 3-coordinate close
packing. The molecular structure of the related borox-
ine anion, [Ph2B(OBPh)2O]− (2) has been investigated
twice (counterions [NMe4]+ and [{(t-Bu)3PAu}4P]+)
[33,34]. A comparison of the more precisely determined
structure with the parent triphenylboroxine, (PhBO)3
(3), produces a similar result. The OꢀBꢀO bond angle
related to the tetrahedral boron atom is decreased by
10.2° to 109.1(8)° as compared to the parent compound
(PhBO)3 (3). In contrast, the average trigonal OꢀBꢀO
bond angles and the mean BꢀOꢀB angles increase by
2.6 and 1.5° to 121.9(9) and 121.8(8)°, respectively. We
suggest that these observations are fully consistent with
the relief of ring strain upon going from (PhBO)3 (3) to
(PhBO)3·pyridine (1) and [Ph2B(OBPh)2O]− (2) and
examine these claims more closely in the following
section.

154
J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
the concomitant widening of the remaining endocyclic
Adduct formation of Lewis bases with the remaining
angles causes a reduction in ring strain in a manner
similar to (PhBO)3·pyridine (1) [49].
boron atoms to form 2:1 and 3:1 adducts would neces-
sitate a decrease in the corresponding OꢀBꢀO angles
and may not be feasible as it would force an unfa-
vourable widening of the BꢀOꢀB angles.
2.5. Thermodynamic and kinetic stability of
(PhBO)3·pyridine (1)
We have attempted to rationalize the (nearly) exclu-
sive formation of 1:1 triorganoboroxine–amine adducts
and the lower propensity of such adducts towards
ring-opening compared to their parent compounds us-
ing ring strain considerations; however, a contrasting
argument is that both of these observations are made
on simple steric grounds, i.e. increased steric crowding
around the 1:1 adduct reduces its susceptibility to at-
tack and there is insufficient space for a second amine
to bind to the ring. While it is true that increased steric
crowding around the boroxine ring would reduce its
reactivity, the steric argument on its own is unconvinc-
ing as higher triorganoboroxine–amine adducts involv-
ing bases as small as ammonia are virtually unknown.
The 1:2 complex (CH3BO)3·2NH3 has been previously
reported, but it has been described as being far less
stable than the corresponding 1:1 adduct [47]. Further-
more, instead of being a formal 1:2 adduct with both
ammonia molecules coordinated to the boroxine ring,
the second molecule of NH3 in this adduct is believed
to be hydrogen-bonded to the first ammonia molecule,
which is coordinated to the ring.
A similar situation to that described above is found
for the six-membered stannasiloxane rings t-
Bu2Sn(OSiPh2)2O [48] and [Me2N(CH2)3]2Sn(OSiPh2)2O
[49]. The former possesses a considerable amount of
ring strain and, consequently, undergoes an ROP upon
crystallization. In contrast, the latter is practically
strain-free due to stabilization by the intramolecular
coordinating amine ligands. The complexation by
amine ligands formally changes the coordination ge-
ometry at tin from a tetrahedral to an octahedral
environment, and consequently the related OꢀSnꢀO
bond angle decreases significantly. We have suggested
that as a result of the narrowing of one OꢀSnꢀO angle,
In order to evaluate how strongly the pyridine is
bonded to the boroxine ring, a thermogravimetric anal-
ysis of 1 was carried out. Endothermic loss of pyridine
was observed between a melting point of 152 and
200 °C.
It is well known that the BꢀN bond in tri-
organoboroxine amine complexes is kinetically labile in
solution [22,23]. Thus, the 13C-NMR spectrum in
CDCl3 of (PhBO)3·pyridine (1) at room temperature
exhibits seven signals implying the magnetic equiva-
lence of all three boron centres. Only recently, an
activation energy of 39 kJ mol−1 was reported for the
dissociation–recombination process of (PhBO)3·pyri-
dine (1) in solution [22].
Molecular weight determinations of (PhBO)3·
pyridine (1) in toluene at 60 and 80 °C at four different
concentrations further suggest that a dissociation pro-
cess takes place and an equilibrium exists between 1
and (PhBO)3 (3) and pyridine. Upon increasing the
temperature or lowering the initial concentration of
(PhBO)3·pyridine (1), this equilibrium shifts from 1
towards (PhBO)3 (3) and pyridine (Fig. 4, Eq. (4)).
(4)
This observation suggests that the pyridine is less
strongly bonded in solution than in the solid state. A
possible explanation for this might be the partial com-
pensation of the ring strain in 3 by an increasing
number of species upon the dissociation process. The
increase of species would directly account to the en-
tropy term of the Gibbs–Helmholtz equation (DG=
DH−TDS). A similar observation was made for the
strained six-membered stannasiloxane ring t-
Bu2Sn(OSiPh2)2O [48], which undergoes ring-opening
polymerization upon crystallization. In solution, how-
ever, the ring strain is compensated by the entropic
favour of a large number of molecular species [48].
2.6. Conclusions
The questions (i)–(iii) posed in Section 1 could be
rationalized using the following ring strain consider-
ations: (i) eight-membered tetraboroxine rings are cal-
culated to possess significantly more ring strain than
six-membered triboroxine rings; (ii) the addition of
amines to triorganotriboroxines stops at 1:1 stoi-
chiometry as maximum relief of ring strain occurs in 1:1
Fig. 4. Molecular weight determination of (PhBO)3·pyridine (1) in
toluene at 60 and 80 °C.

J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
155
adducts; and (iii) triorganoboroxines are more suscepti-
ble towards ring-opening than their 1:1 amine adducts
as they are more strained.
1841 with I]3.0|(I) were used in the subsequent
analysis.
5.1. Crystal data
3. Computational methodology
C23H20B3NO3, MW=390.9, monoclinic, P21/c, a=
,
14.477(4), b=11.507(4), c=12.977(4) A, i=
3
−3
DFT calculations were carried out using the GAUS-
SIAN-98 suite of programs [50]. Geometry optimizations
were performed at the B3LYP/6-311+G(d) level of
theory using standard techniques. The nature of each
stationary point was verified by vibrational frequency
calculations. Unless otherwise stated, each structure is a
minimum on its respective potential energy surface.
,
103.19(3)°, V=2105(1) A , Z=4, Dx=1.233 g cm
,
v(Mo–Ka)=0.79 cm−1, F(000)=816, 271 refined
parameters, zmax=0.14 e A−3. The structure was
,
solved by direct methods [52] and refined by a full-ma-
trix least-squares procedure based on F [51]. The non-
hydrogen atoms were refined with anisotropic
displacement parameters and H atoms were included in
the model at their calculated positions. A weighting
scheme of the form w=1/[|2(F)+0.00001ꢀFoꢀ ] was
2
4. Experimental
employed and at convergence, final R=0.042 and
Rw=0.037. Fig. 2 shows the crystallographic number-
ing scheme which was drawn with ORTEP at the 50%
probability level [53].
Phenylboronic acid and pyridine were commercially
obtained (Aldrich). All solvents were freshly distilled
prior to use. NMR spectra were obtained using a
1
Varian 300 MHz Unity Plus NMR spectrometer. H-
and 13C-chemical shifts l are given in ppm and are
referenced against Me4Si. Molecular weight determina-
tions were performed on a Gonotec Osmomat 070
osmometer. The thermogravimetric analysis was made
on a Perkin–Elmer TGA 7 thermogravimetric analyzer
(with TAC 7/DX controller and gas selector). The
elemental analysis was carried out on an instrument
from Carlo Erba Strumentazione (Model 1106).
6. Supplementary material
Crystallographic data for the structural analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC no. 150673 for compound 1.
Copies of this information may be obtained free of
charge from The Director, CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK (Fax: +44-1223-336033;
e-mail: deposit@ccdc.cam.ac.uk or www: http://
www.ccdc.cam.ac.uk). Calculated energies and the
Cartesian coordinates of the B3LYP/6-311+G(d)-opti-
mized structures of 4–12 are available upon request
from the authors.
4.1. Synthesis of (PhBO)3·pyridine (1)
To a solution of PhB(OH)2 (1.10 g, 9.00 mmol) in
Et2O (100 ml), Py (237 mg, 3.00 mmol) was added at
room temperature. The mixture was stirred for 15 min
before the ether was removed by distillation at normal
pressure. The crude product was recrystallized from
hexane–CH2Cl2, providing colourless crystals of 1 (1.13
g, 2.89 mmol, 96%, m.p. 152 °C).
Acknowledgements
1H-NMR (CDCl3):
l
9.2–7.4. 13C{1H}-NMR
The Australian Research Council is thanked for the
support of the crystallographic facility. Peter Laming
(RMIT University, Melbourne) is gratefully acknowl-
edged for the thermogravimetric analysis. A.E.K.L.
acknowledges an Overseas Postgraduate Research
Scholarship from Deakin University.
(CDCl3): l 143.3, 140.9, 138.7 (broad), 133.6, 129.6,
127.5, 125.5. Anal. Found: C, 70.7; H, 5.1; N, 3.5. Calc.
for C23H20B3NO3 (MW 390.9): C, 70.7; H, 5.2; N,
3.6%.
5. Crystallography
References
Intensity data for a colourless block (0.31 mm3) were
collected at 173 K on a Rigaku AFC7R diffractometer
employing Mo–Ka radiation (u=0.7107 A) and the
ꢀ–2q scan technique such that qmax was 27.5°. Correc-
tions were made for Lorentz and polarization effects
[51] but not for absorption. Of the 5275 reflections
measured, 5062 were unique (Rint=0.041) and of these,
[1] Gmelin Handbook of Inorganic Chemistry, Supplementary
work, vol. 44: Boron Compounds, Part 13: Boron–Oxygen
Compounds, 8th ed., Springer, Heidelberg, 1977.
[2] B. Pachaly, R. West, J. Am. Chem. Soc. 107 (1985) 2987.
[3] G. Nencetti, S. Zanelli, G. Nardini, Ann. Chim. 53 (1963) 1875.
[4] H.R. Snyder, J.A. Kuck, J.R. Johnson, J. Am. Chem. Soc. 60
(1938) 105.
,

156
J. Beckmann et al. / Journal of Organometallic Chemistry 633 (2001) 149–156
[5] M.J. Aroney, R.J.W. Le Fevre, D.S.N. Murthy, J.D. Saxby, J.
[29] M. Yalpani, R. Ko¨ster, Chem. Ber. 121 (1988) 1553.
[30] M.K. Das, J.F. Mariategui, K. Niedenzu, Inorg. Chem. 26
(1987) 3114.
[31] M. Yalpani, R. Boese, Chem. Ber. 116 (1983) 3347.
[32] P.D. Robinson, M.P. Groziak, L. Yi, Acta Crystallogr. Sect. C
52 (1996) 2826.
[33] W. Kliegel, H.W. Motzkus, S.J. Rettig, J. Trotter, Can. J. Chem.
63 (1985) 3516.
[34] E. Zeller, H. Beruda, H. Schmidbaur, Chem. Ber. 126 (1993)
2033.
[35] H. Schmidbaur, B. Brachtha¨user, O. Steigelmann, H. Beruda,
Chem. Ber. 125 (1992) 2705.
[36] I. Haiduc, The Chemistry of Inorganic Ring Systems, Part 1,
Wiley, London, 1970.
[37] D.S. Matteson, J. Org. Chem. 29 (1964) 3399.
[38] M.W. Schmidt, M.S. Gordon, J.A. Boatz, Int. J. Quantum
Chem. 76 (2000) 434.
[39] C.P. Brock, R.P. Minton, K. Niedenzu, Acta Crystallogr. Sect.
C 43 (1987) 1775.
[40] R. Boese, M. Polk, D. Bla¨ser, Angew. Chem. 99 (1987) 239.
[41] L. Santucci, Gazz. Chim. Ital. 113 (1983) 515.
[42] L. Santucci, C. Triboulet, J. Chem. Soc. A (1969) 392.
[43] H.C. Brown, T.E. Cole, Organometallics 4 (1985) 816.
[44] S.M. Bachrach, J. Chem. Educ. 67 (1990) 907.
[45] Y.A. Yuzhelevskii, V.V. Sokolov, L.V. Tagieva, E.G. Kagan,
Vysokomol. Soedin. Ser. B 13 (1971) 95.
[46] J.A. Semlyen, S.J. Clarson, Siloxane Polymers, Prentice–Hall,
Englewood Cliffs, NJ, 1991.
[47] A.B. Burg, J. Am. Chem. Soc. 62 (1940) 2228.
[48] J. Beckmann, K. Jurkschat, D. Schollmeyer, M. Schu¨rmann, J.
Organomet. Chem. 543 (1997) 229.
[49] J. Beckmann, K. Jurkschat, U. Kaltenbrunner, N. Pieper, M.
Schu¨rmann, Organometallics 18 (1999) 1586.
[50] GAUSSIAN-98: Revision A.7, Gaussian, Inc., Pittsburgh, PA,
1998.
[51] TEXSAN: Single crystal structure analysis software, Version 1.04,
Molecular Structure Corporation, The Woodlands, TX, USA,
1997.
[52] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C.
Giacovazzo, A. Guagliardi, A.G.G. Moliterni, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 32 (1999) 115 (SIR-97).
[53] C.K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National
Laboratory, TN, USA, 1976.
Chem. Soc. B (1966) 1066.
[6] D.R. Armstrong, P.G. Perkins, J. Chem. Soc. A (1967) 790.
[7] C.A. Coulson, T.W. Dingle, Acta Crystallogr. Sect. B 24 (1968)
153.
[8] J.F. Labarre, M. Graffeuil, F. Gallais, J. Chim. Phys. Phys.-
Chim. Biol. 65 (1968) 638.
[9] J.F. Labarre, M. Graffeuil, J.P. Faucher, M. Pasdeloup, J.P.
Laurent, Theor. Chim. Acta 11 (1968) 423.
[10] K. Jug, J. Org. Chem. 48 (1983) 1344.
[11] L. Barton, D. Brinza, R.A. Frease, F.L. Longcor, J. Inorg. Nucl.
Chem. 39 (1977) 1845.
[12] W. Biffar, H. No¨th, H. Pommerening, B. Wrackmeyer, Chem.
Ber. 113 (1980) 333.
[13] B. Wrackmeyer, R. Ko¨ster, Chem. Ber. 115 (1982) 2022.
[14] A. Loetz, J. Voitlaender, D. Stephenson, J.A.S. Smith, Z. Natur-
forsch. Teil A 41 (1986) 200.
[15] D.L. Cooper, S.C. Wright, J. Gerratt, P.A. Hyams, J. Chem.
Soc. Perkin Trans. 2 (1989) 719.
[16] E.F. Archibong, A.J. Thakkar, Mol. Phys. 81 (1994) 557.
[17] P.W. Fowler, E. Steiner, J. Phys. Chem. A 101 (1997) 1409.
[18] P.v.R. Schleyer, H. Jiao, N.J.R. van Eikema Hommes, V.G.
Malkin, O. Malkina, J. Am. Chem. Soc. 119 (1997) 12669.
[19] D.W. Lamb, R.I. Keir, G.L.D. Ritchie, Chem. Phys. Lett. 291
(1998) 197.
[20] R.J. Gillespie, I. Bytheway, E.A. Robinson, Inorg. Chem. 37
(1998) 2811.
[21] H.R. Snyder, M.S. Konecky, W.J. Lennarz, J. Am. Chem. Soc.
80 (1958) 3611.
[22] Q.G. Wu, G. Wu, L. Brancaleon, S. Wang, Organometallics 18
(1999) 2553.
[23] M.A. Beckett, D.S. Brassington, P. Owen, M.B. Hursthouse,
M.E. Light, K.M.A. Malik, K.S. Varma, J. Organomet. Chem.
585 (1999) 7.
[24] M.A. Beckett, D.E. Hibbs, M.B. Hursthouse, P. Owen, K.M.
Abdul Malik, K.S. Varma, Main Group Chem. 2 (1998) 251.
[25] M.A. Beckett, G.C. Strickland, K.S. Varma, D.E. Hibbs, M.B.
Hursthouse, K.M.A. Malik, J. Organomet. Chem. 535 (1997) 33.
[26] M.A. Beckett, G.C. Strickland, K.S. Varma, D.E. Hibbs, M.B.
Hursthouse, K.M.A. Malik, Polyhedron 14 (1995) 2623.
[27] G. Ferguson, A.J. Lough, J.P. Sheehan, T.R. Spalding, Acta
Crystallogr. Sect. C 46 (1990) 2390.
[28] J.F. Mariategui, K. Niedenzu, J. Organomet. Chem. 369 (1989)
137.
.