Unsupported monomeric stibine oxides (R3SbO) remain undiscovered

Article abstract of DOI:10.1039/d1cc00619c

Attempts to investigate the properties and reactivity of the stiboryl moiety (R₃Sb⁺-O^?or R₃SbO), as in monomeric stibine oxides free of interaction with Lewis acids/bases, led us to conclude that this functional

Full text of DOI:10.1039/d1cc00619c

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Unsupported monomeric stibine oxides (R3SbO)
remain undiscovered†
Cite this: Chem. Commun., 2021,
57, 3484
John S. Wenger and Timothy C. Johnstone
*
Received 3rd February 2021,
Accepted 2nd March 2021
DOI: 10.1039/d1cc00619c
rsc.li/chemcomm
Attempts to investigate the properties and reactivity of the stiboryl reduction in overlap with oxygen-based orbitals. Another pheno-
moiety (R3Sb+–OÀ or R3Sb O), as in monomeric stibine oxides free menon that attends the increase in atom size as the family is
Q
of interaction with Lewis acids/bases, led us to conclude that this descended is an increased propensity of the pnictogen to expand
functional group remains undiscovered. X-ray crystallographic, its coordination sphere. Moreover, the E+–OÀ bond becomes more
computational, and spectroscopic data indicate that previously polarized as E increases in atomic number and decreases in
proposed H-bonded stibine oxide adducts Mes3SbOÁ Á ÁHO3SR are electronegativity, which increases the Lewis acidity of the E(V)
in fact hydroxystibonium salts [Mes3SbOH][RSO3].
center in a pnictine oxide. Alternatively put, the destabilization of
the E–O s bond lowers the energy of the corresponding E–O s*
Pentavalent pnictine oxides (oxo-l5-pnictanes) of the general orbital that serves as the locus of Lewis acidity. This enhanced
form R3EQO 2 R3E+–OÀ (E = N, P, As, Sb, Bi) are of consider- Lewis acidity coupled to a capacity for expanded coordination
able interest to chemists because of the diverse roles that they results in a fundamental change in the chemistry of the pnictine
play across a range of chemical fields. Trimethylamine N-oxide oxides down the group.6,10–12 For example, treatment of either
(Me3NO) has long been used as an oxo transfer reagent in the Ph3P or Ph3As with oxidants such as H2O2 readily affords the
decarbonylation of transition metal complexes through the loss oxides Ph3PO and Ph3AsO (Scheme 1a), but ‘‘triphenylstibine
of CO2.1 Triphenylphosphine oxide (Ph3PO) and triphenylar- oxide’’ is not known to exist as the monomeric Ph3SbO, despite
sine oxide (Ph3AsO) have been used as ligands for lanthanide the use of this structural formula for decades.
ions with applications to catalysis and the reprocessing of
A series of experiments demonstrated that oxidation of Ph3Sb
nuclear fuels.2–4 Ph3PO is perhaps best known as a stable end yields five-coordinate dimers or oligomers (Scheme 1b).13,14 The
product of organic syntheses employing the Wittig, Staudinger, increased bulk of the substituents in Mes3Sb suppresses oligo-
and Mitsunobu reactions. Part of the driving force of these merization upon oxidation with H2O2, but not expansion of the
successful and robust reactions is the stability of the phosphine coordination sphere; the product is the dihydroxystiborane trans-
oxide, which is intimately related to the nature of the pnictogen– Sb(OH)2Mes3 (Scheme 1c).
oxygen bond. The E–O bonding interaction of pentavalent pnicto-
gen oxides has long been a substrate for interesting discussion,
with much of the focus on phosphine oxides.5–8 Although penta-
valent phosphine oxides are typically depicted with a phosphorus–
oxygen double bond (R3PQO), this bonding is now accepted to be
better described as a polar-covalent single bond (R3P+–OÀ) stabi-
lized by p backdonation from the oxide lone pairs into the P–R s*
orbitals.8,9 Descending the pnictogen family, this E+–OÀ bonding
interaction is expected to become less stable because of increased
diffuseness of the pnictogen valence orbitals and a consequent
Department of Chemistry and Biochemistry, University of California Santa Cruz,
Santa Cruz, CA 95064, USA. E-mail: johnstone@ucsc.edu
Scheme 1 (a) Oxidation of Ph3E with H2O2 give monomeric Ph3EO when
E = P or As. (b) Oxidation of Ph3Sb with H2O2 affords dimers/polymers.
(c) Oxidation of Mes3Sb with H2O2 gives the monomeric stiborane trans-
Sb(OH)2Mes3.
† Electronic supplementary information (ESI) available: Experimental proce-
dures, NMR spectra, crystallographic details, and computational data. CCDC
2060339–2060341. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/d1cc00619c
This journal is © The Royal Society of Chemistry 2021
3484 | Chem. Commun., 2021, 57, 3484–3487

View Article Online
Communication
ChemComm
We were interested in exploring the chemistry of the stiboryl
functional group present in monomeric stibine oxides,15 but
careful scrutiny of the literature revealed that these comprise a
largely unexplored class of molecules. Lewis acids can disrupt
stibine oxide dimers (R3SbO)2 and stabilize the Lewis-basic oxo
unit (Scheme 2a).16 Alternatively, rigid scaffolds can poise a
Lewis acid near a stibine to intercept a newly-formed stibine
oxide before it can oligomerize (Scheme 2b).17 In both of these
cases, an interaction of the stiboryl group with a Lewis acid
appears to be essential to stabilize it.
To our knowledge, Huber and co-workers report the most, if
not only, compelling case for an independently stable monomeric
stibine oxide, trimesitylstibine oxide (Mes3SbO). Treatment of
trans-Sb(OH)2Mes3 with either benzenesulfonic acid or trifluoro-
methanesulfonic (triflic) acid was proposed to result in formation
of Mes3SbO (Scheme 2c).18 The strong acids were proposed to
Fig. 1 Relaxed surface scan (BP86/def2-TZVP) of the hydroxytrimesityl-
stibonium benzenesulfonate SbO–H bond length.
protonate an apical hydroxide ligand from the stiborane to effect preliminary insight into the structure of Mes3SbOÁ Á ÁHO3SR by
the loss of a water molecule and the sulfonate anion generated performing a geometry optimization at the PBE0/def2-TZVPP
was then believed to abstract a proton from the remaining Sb–OH level of theory starting from the previously reported crystal-
moiety. The putative stibine oxide was ultimately isolated as a lographic coordinates. The optimization converged smoothly
hydrogen-bonded adduct Mes3SbOÁÁÁHO3SR. Elemental analysis, and bond lengths and angles agreed well between the compu-
1
conductivity, and H NMR and IR spectroscopic data were con- tational and experimental data (see ESI†). Surprisingly, how-
sistent with this proposed structure, but the strongest evidence ever, during the optimization, the hydrogen atom migrates
offered in support of the identity of the Mes3SbOÁÁÁHO3SR species from the sulfonic acid to the stibine oxide, forming the hydro-
was the X-ray crystal structure of the benzenesulfonate adduct, xystibonium benzenesulfonate salt [Mes3Sb(OH)][PhSO3] with
which features a hydrogen bond between the acid and the an O–H bond length of 1.035 Å. A relaxed surface scan (BP86/
stibine oxide.
def2-TZVP) in which the hydrogen atom is systematically
These substances appeared to offer the most direct route to moved from the Sb-bound O atom to the S-bound O atom
an isolated stibine oxide via removal of the hydrogen-bonded revealed that there is only one minimum along this internal
acid. Intrigued by the possibility of studying the properties and coordinate and that it corresponds to the hydroxytrimesitylsti-
reactivity of monomeric stibine oxides, we sought to gain bonium benzenesulfonate (Fig. 1). One chemically intuitive
means of interpreting these results is that PhSO3H is a stronger
Brønsted acid than Mes3SbOH+. This interpretation is con-
firmed by the lower proton affinity computed for the conjugate
base of the former (see ESI†). As discussed by Steiner,19
a
hydrogen bond can be viewed as an incipient proton transfer.
The discrepancy between the proton affinities of the donor and
acceptor results in formal proton transfer and formation of a
predominantly ionic interaction with a moderate hydrogen
bond formed on top.19
Close analysis of the previously published crystallographic
experiments uncovered no reported details regarding the place-
ment or refinement of the PhSO3H hydrogen atom in the
Mes3SbOÁ Á ÁHO3SPh model. Moreover, the previously reported
elemental analysis and spectroscopic data (the HO3SPh
1H NMR resonance was not observed) do not speak directly to
the location of the hydrogen atom in question. To further
interrogate the differences that might be expected between
Mes3SbO and [Mes3SbOH]+, the geometries of these isolated
species were optimized computationally (PBE0/def2-TZVPP). As
expected, the structure of Mes3SbO exhibits an Sb–O bond
length (1.827 Å) shorter than that calculated for [Mes3SbOH]+
(1.932 Å), the latter of which agrees better with the previously
reported experimental Sb–O bond length of 1.894(5) Å.
To resolve the apparent contradictions described above, we
resynthesized the previously reported compounds by adding a
Scheme 2 (a) Product of Lewis acid-induced disaggregation of dimeric
(Ph3SbO)2. (b) Product of the oxidation of triarylstibines held in proximity
by a rigid scaffold. The chelating moiety represents o-C6Cl4O2. (c) Treat-
ment of trans-Sb(OH)2Mes3 with sulfonic acids to putatively afford mono-
meric stibine oxides hydrogen-bonded to Brønsted acids.
This journal is © The Royal Society of Chemistry 2021
Chem. Commun., 2021, 57, 3484–3487 | 3485

View Article Online
ChemComm
Communication
The ionic nature of these species was further supported
by the facile exchange of the triflate counterion in 2 for
[B(C6H3(3,5-CF3)2)4]À, affording salt 3 (Scheme 3). Again, analy-
tical and X-ray crystallographic data support the formulation of
this species as a hydroxystibonium salt (see ESI†).
We grew diffraction-quality crystals of 1 and 2 to redeter-
mine their structures. Solution and refinement of the structure
of 1 proceeded smoothly and afforded a non-H atom model
essentially identical to that reported previously. Inspection of
the difference Fourier synthesis generated using a model in
which all non-H atoms were refined anisotropically and all
C–bound H atoms were included at geometrically calculated
positions, clearly reveals a maximum electron density signal
along the Sb–OÁ Á ÁO–S vector at distance of 0.72 Å from the
Sb-bound O atom (Fig. 2). This maximum indicates that, in
Scheme 3 Synthesis of hydroxystibonium salts 1, 2, and 3. BArF is
4
[B(C6H3(3,5-CF3)2)4]À.
DCM solution of the sulfonic acid to a DCM suspension of contrast to the previous report, the H atom resides on the
1 equiv. of trans-Sb(OH)2Mes3. As the reaction proceeded, all Sb-bound O atom.
solids were solubilized. Addition of hexanes to the reactions with
Refinement of the crystal structure of 2 revealed that it also
PhSO3H and CF3SO3H produced colorless crystals of 1 and 2, contains the [Mes3SbOH]+ cation. The quality of the collected
respectively, which feature analytical characteristics identical to data provides a high degree of confidence in the locations of
those previously reported for the PhSO3H and CF3SO3H adducts these H atoms on the Sb-bound oxygen, despite the well-known
of Mes3SbO. The 1H and 13C NMR spectra of 1 feature a single set difficulties associated with locating H atoms by X-ray crystal-
of mesityl resonances, consistent with maintenance of 3-fold lography. Moreover, the influence of the H atoms on the Sb–O
rotational symmetry. In addition to the remaining phenyl signals bond lengths is unmistakable. The Sb–O bond lengths in 1
of the acid, a broad signal at 9.10 ppm was observed, which we (1.9055(8) Å) and 2 (1.910(1) Å) are in good agreement with the
assign to the hydroxyl proton. The spectra of 2 can be interpreted computationally optimized Sb–O bond length of [Mes3SbOH]+
similarly, but with the hydroxyl proton resonating at 7.51 ppm. (1.932 Å) and are significantly longer than the Sb–O bond
The NMR spectroscopic data are consistent with the formula- length calculated for Mes3SbO (1.827 Å) (Fig. 3).
tion of 1 and 2 as either Mes3SbOÁÁÁHO3SR adducts or
We were also successful in solving and refining the structure
[Mes3SbOH+][RSO3À] salts (Scheme 3). We highlight that in con- of 3ÁC5H12 (Fig. 3). The crystals were weakly diffracting, but the
trast to the resonance structures shown in Scheme 2, the products Sb–O bond length is 1.932(3) Å. Unlike the structures of 1 and 2,
depicted in equilibrium in Scheme 3 are chemically distinct.
there are no strong H-bonds between the cation and anion, but
Varying degrees of either ionic pairing (in the case of the a F atom of one CF3 group is positioned 3.050(5) Å from the
hydroxystibonium salts) or hydrogen bonding (in the case of hydroxyl O atom and the O–HÁÁÁF angle is 159(5)1. The weaker
stibine oxide adducts) could readily account for the difference interaction allows a sharp nOH IR band to be observed at 3601 cmÀ1
;
between the chemical shifts of the hydroxyl 1H NMR reso-
nances in 1 and 2. IR spectroscopy provides an opportunity to
gain more direct insight into the bonding between the Sb and O
atoms in these compounds. The IR spectra of 1 and 2 are
broadly similar and feature strong signals at 612 cmÀ1 and
637 cmÀ1, respectively, which we assign as the nSbO stretching
frequency. These bands were previously interpreted as being
indicative of an SbQO double bond given that they are higher
in energy than the nSbO of species like trans-Sb(OH)2Mes3
(nSbO = 520 cmÀ1).18,20 The previous report of 1 and 2 provides
the only compelling examples of well-characterized monomeric
stibine oxides and so there are no literature comparators. Using
the DFT-optimized structures of Mes3SbO and [Mes3SbOH]+,
frequency calculations afforded nSbO values of 817 cmÀ1 for the
stibine oxide and 647 cmÀ1 for the hydroxystibonium. These
values are consistent with the greater Sb–O bond strength/
stiffness expected for the stibine oxide. The significantly better
agreement of the experimental nSbO with that calculated for
[Mes3SbOH]+ provides a strong confirmation of proposal that 1
and 2 are in fact salts of hydroxystibonium cations that feature
an Sb–O single bond.
Fig. 2 Thermal ellipsoid plot (50% probability, H atoms as spheres of
arbitrary radius) of 1 less the protic hydrogen atom expanded about the
residual electron density maximum (full plot in Fig. 3a). Overlaid in grey
Å
contours (0.075 eÀ À3) is the Fo–Fc map plotted in the plane defined by
Sb1, O1, and O2. Color Code: Sb teal, O red, S yellow, C black, H white.
This journal is © The Royal Society of Chemistry 2021
3486 | Chem. Commun., 2021, 57, 3484–3487

View Article Online
Communication
ChemComm
acid or both a charged Lewis acid (such as a proton) and a charge-
balancing counterion. The interaction of the acid with the Sb-bound
O atom would afford a lengthened Sb–O bond, as observed pre-
viously and in this work.
In summary, our attempts to investigate the properties and
reactivity of the Sb+–OÀ functional group in monomeric stibine
oxides led to the discovery that the only previously reported
examples are hydroxystibonium cations in which a Lewis acid
interacts with the Sb-bound O atom. As such, this work demon-
strates that monomeric stibine oxides unstabilized by inter-
action with a Lewis acid have still yet to be prepared. The lack of
access to such molecules greatly hampers any effort to observe
and understand the systematic variation in the nature
and reactivity of pnictogen–oxygen bonds. Initial attempts to
prepare stibine oxides via deprotonation of 1–3 have been
unsuccessful, but the targeted synthesis of these molecules is
underway.
Conflicts of interest
There are no conflicts to declare.
Fig. 3 Thermal ellipsoid plots (50% probability) of (a) 1, (b) 2, and (c) 3. H
atoms and solvent omitted except Sb–OH, which is shown as a sphere of
arbitrary radius. Ball-and-stick representations of the optimized (PBE0/
def2-TZVPP) structures of (d) [Mes3SbOH]+ and (e) Mes3Sb. Color Code:
Sb teal, O red, S yellow, F green, C black, B pink, H white.
Notes and references
1 T.-Y. Luh, Coord. Chem. Rev., 1984, 60, 255–276.
2 A. W. G. Platt, Coord. Chem. Rev., 2017, 340, 62–78.
3 T. Nemoto, T. Ohshima, K. Yamaguchi and M. Shibasaki, J. Am.
Chem. Soc., 2001, 123, 2725–2732.
the moderate H-bonding in 1 and 2 produces a broad band in this
region (Fig. S16, ESI†).
4 N. N. Patil and V. M. Shinde, J. Radioanal. Nucl. Chem., 1997, 222,
21–24.
5 A. Y. Rogachev and P. Burger, Phys. Chem. Chem. Phys., 2012, 14,
1985–2000.
Our search of the literature uncovered only one other promising
case of a monomeric stibine oxide, tris(2,6-dimethoxyphenyl)stibine
oxide, (2,6-(OMe)2C6H3)3SbO.21 In contrast to the detailed spectro-
scopic data provided in the previous report of 1 and 2, the report of
(2,6-(OMe)2C6H3)3SbO includes only an IR-determined nSbO of
664 cmÀ1. As noted above, this value already provides a strong
indication that the species is not a stibine oxide. The reported
assignment rested heavily on a single-crystal X-ray structural analy-
sis. Although the refinement statistics for the reported structure are
promising, inspection of the deposited data reveals that the struc-
´
´
6 J. A. Dobado, H. Martınez-Garcıa, J. M. Molina and M. R. Sundberg,
J. Am. Chem. Soc., 1998, 120, 8461–8471.
7 L. Zhao, S. Pan, N. Holzmann, P. Schwerdtfeger and G. Frenking,
Chem. Rev., 2019, 119, 8781–8845.
8 T. Yang, D. M. Andrada and G. Frenking, Phys. Chem. Chem. Phys.,
2018, 20, 11856–11866.
9 K. Yamada and N. Koga, J. Comput. Chem., 2013, 34, 149–161.
10 A. P. M. Robertson, S. S. Chitnis, H. A. Jenkins, R. McDonald,
M. J. Ferguson and N. Burford, Chem. – Eur. J., 2015, 21, 7902–7913.
11 Y. Matano, H. Nomura, T. Hisanaga, H. Nakano, M. Shiro and
H. Imahori, Organometallics, 2004, 23, 5471–5480.
¨
12 J. S. Jones and F. P. Gabbaı, Acc. Chem. Res., 2016, 49, 857–867.
ture contains voids that directly abut the Sb-bound O atoms of both 13 J. Bordner, G. O. Doak and T. S. Everett, J. Am. Chem. Soc., 1986, 108,
4206–4213.
non-crystallographically related molecules in the asymmetric unit;
the SQUEEZE algorithm was employed to remove the contribution
14 C. J. Carmalt, J. G. Crossley, N. C. Norman and A. G. Orpen,
Chem. Commun., 1996, 1675–1676.
of the electron density within this region of space to the observed 15 B. Lindquist-Kleissler, J. S. Wenger and T. C. Johnstone,
diffraction intensities.21 In addition to the presence of these voids,
we note that the reported Sb–O bond length of 1.918(3) Å agrees with
Inorg. Chem., 2021, 60, 1846–1856.
16 R. Kather, T. Svoboda, M. Wehrhahn, E. Rychagova, E. Lork,
´
L. Dostal, S. Ketkov and J. Beckmann, Chem. Commun., 2015, 51,
the Sb–O bond lengths that we obtained for the hydroxytrimesityl-
stibonium salts 1, 2, and 3. Following the strategy described above,
we computationally optimized (PBE0/def2-TVZPP) the structures of
5932–5935.
¨
17 C.-H. Chen and F. P. Gabbaı, Dalton Trans., 2018, 47, 12075–12078.
18 F. Huber, T. Westhoff and H. Preut, J. Organomet. Chem., 1987, 323,
173–180.
both the stibine oxide (2,6-(OMe)2C6H3)3SbO and the hydroxystibo- 19 T. Steiner, Angew. Chem., Int. Ed., 2002, 41, 48–76.
nium cation [(2,6-(OMe)2C6H3)3SbOH]+. The Sb–O bond length of
the former was 1.818 Å and that of the latter was 1.920 Å. We suggest
20 T. Westhoff, F. Huber, R. Ru¨ther and H. Preut, J. Organomet. Chem.,
1988, 352, 107–113.
21 I. V. Egorova, V. V. Zhidkov, I. P. Grinishak and N. A. Rodionova,
that the voids in the crystal structure contain either a neutral Lewis
Russ. J. Gen. Chem., 2016, 86, 2484–2491.
This journal is © The Royal Society of Chemistry 2021
Chem. Commun., 2021, 57, 3484–3487 | 3487