Structure-reactivity studies on hypervalent square-pyramidal dithieno[3,2-b:2′,3′-d]phospholes

Article abstract of DOI:10.1039/d1dt00062d

A series of neutral pentacoordinate dithieno[3,2-b:2′,3′-d]phosphole compounds were synthesized by [4 + 1] cycloaddition witho-quinones. Counter to the expected trigonal bipyramidal geometry, the luminescent hypervalent dithienophospholes exhibit square p

Full text of DOI:10.1039/d1dt00062d

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Structure-reactivity studies on hypervalent square-
pyramidal dithieno[3,2-b:2’,3’-d]phospholes†
Cite this: Dalton Trans., 2021, 50,
2243
Nayanthara Asok, Joshua R. Gaffen,
Thomas Baumgartner
Ekadashi Pradhan,
Tao Zeng * and
*
A series of neutral pentacoordinate dithieno[3,2-b:2’,3’-d]phosphole compounds were synthesized by
[4 + 1] cycloaddition with o-quinones. Counter to the expected trigonal bipyramidal geometry, the lumi-
nescent hypervalent dithienophospholes exhibit square pyramidal geometry with inherently Lewis acidic
phosphorus center that is stabilized via supramolecular π-stacking interactions in the solid state and in
solution. Due to their Lewis-acid character, the compounds react with nucleophiles, suggesting their
potential as mediator in organic transformations. The new species thus present an intriguing structural
plaform for the design of neutral P(V) Lewis acids with useful reactivities.
Received 7th January 2021,
Accepted 19th January 2021
DOI: 10.1039/d1dt00062d
rsc.li/dalton
the comprehensive investigation of dithienophospholes as
potential material for application in organic electronics.3b
Introduction
Organic conjugated material research is a popular field owing
While most of the previous work focused on amendment of
to the practical application of these materials in various the electronics by exocyclic substitution reactions, including
devices, such as organic light-emitting devices (OLEDs), the functionalization of the phosphorus lone pair,3b the altera-
organic photovoltaic devices (OPVs), organic field-effect tran- tion of the geometry of the P center as a methodology to influ-
sistors (OFETs), and sensors.1 While numerous high-perform- ence the energies of the frontier molecular orbitals (FMOs) is
ing p-type materials have been developed and are utilized in relatively less explored.4–6 An interesting structural platform
these devices, the scarcity of corresponding n-type materials for this purpose are hypervalent phosphorus compounds with
demands innovative molecular designs for improved coordination numbers of five or six.7
efficiency.2
Since trivalent phosphorus species are known to undergo [4
One particularly intriguing area of research towards this + 1] cycloaddition with o-quinones to form pentacoordinate
goal deals with the development of main group element-con- molecules,8 we were interested in exploring if this approach
taining building blocks.2c The unique structural elements and could be extended to dithienophospholes as well. The Valence
intrinsic electronics of the main group components provide a Shell Electron Pair Repulsion (VSEPR) model-predicted trigo-
convenient avenue for effective property tuning, particularly nal bipyramidal (3BP) geometry of the proposed molecule II
with a view toward organic electronics. A valuable scaffold that (Fig. 1b) with pentavalent phosphorus exists by virtue of a
is extensively studied in this context is dithieno[3,2-b:2′,3′-d]- 3-center-4-electron bond (3c–4e bond, shown in red), with its
phosphole (I, Fig. 1a).3 The electron-accepting ability of conju-
gated phospholes arises from σ*–π* hyperconjugation, in
which the σ* orbital of the exocyclic P–R bond interacts with
the π* orbital of the butadiene moiety to generate a stabilized
LUMO level (Fig. 1a).4 The ease of alteration of the LUMO
energy by modification of the P center (R) or the main frame-
work (R′), along with strong fluorescence properties has driven
Department of Chemistry, York University, 4700 Keele Street, Toronto, ON M3J 1P3,
Canada. E-mail: tzeng@yorku.ca, tbaumgar@yorku.ca
†Electronic supplementary information (ESI) available: Synthesis and character-
ization of the compounds; theoretical calculation details; Reactivity and stability
studies, crystallographic data; CCDC 2044241(A), 2044242 (B), 2044243(C),
2044244(D) and 2044245 (A·H2O). For ESI and crystallographic data in CIF or
Fig. 1 (a) σ*–π* Hyperconjugation in dithienophosphole; (b) 3c–4e
other electronic format see DOI: 10.1039/d1dt00062d
bond in the proposed hypervalent molecules.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 2243–2252 | 2243

View Article Online
Paper
Dalton Transactions
three molecular orbitals (MOs) of which the bonding and non- suitable quinones, while the phenyl- and perfluorophenyl-sub-
bonding MOs are filled.9 This non-classical bond would stituted 13a and 214 were our dithienophospholes of choice.
impart the electron-accepting properties by providing a vacant
An equimolar solution of compound 1 and o-chloranil in
σ* orbital that can interact with the π* orbital of butadiene degassed, dry DCM (MBraun SPS) afforded compound A,
moiety (Fig. 1b). Importantly, the σ* orbital is oriented per- which was isolated after column chromatography (Silica,
fectly orthogonally to the butadiene moiety and should thus DCM) in 63% yield (Scheme 1). The same reaction conditions
lead to better overlap with the π-system. We expected the worked well for the perfluorophenyl phosphole
2 with
hypervalency to be effectively leveraged in generating neutral o-chloranil and phenanthrene, resulting in the isolation of
and stable pentavalent phosphorus compounds with enhanced compounds C in 62% yield and D in 80% yield, respectively.
electron-accepting abilities compared to that of their parent A similar equimolar reaction of compound 1 and phenan-
trivalent analogue.
threne quinone was initially conducted in refluxing THF as
To our surprise, however, many unexpected observations, the solubility of the quinone was better in this solvent. Even
chiefly the generation of square pyramidal (4SQ) structures, though the reaction exhibited a new phosphorus species after
arose during the course of this study, and we were afforded 20 hours, later revealed to be compound B, upon isolation
with an intriguing opportunity to shift our focus with these only the starting material 1 and corresponding oxide 3 were
new species. Pentacoordinate phosphorus in a 4SQ geometry detected. We thus revisited synthesis of B with meticulously
is uncommon, but not unprecedented.10 However, literature dried DCM (degassed DCM, dried by passing through acti-
largely only provides a nebulous description of what dictates vated neutral alumina, followed by three cycles of freeze-
the geometry of such species. Structural assignments are gen- pump-thaw). Using these conditions, compound B was suc-
erally based on a competing mixture of steric forces and unde- cessfully synthesised in a glovebox and isolated by crystalliza-
fined electronic arguments, using the withdrawing nature of tion from toluene in 61% yield. The successful generation of
the phosphorus substituents.10a–c,11 While the reported 4SQ the 5-coordinate species was indicated by the shift in fluo-
systems do result in Lewis-acidic neutral phosphorus rescence emission (see ESI†) and a significant highfield shift
species,12 the studies do not strictly afford an understanding of the 31P NMR resonances, consistent with a 5-coordinate
of the electronic requirements for a 4SQ P(V) center. For better phosphorus(V) center.15
insight into that question, one can look into the heavier conge-
ners of stibanes. Antimony pentahalides are well known Lewis
acids, but several typical pentacoordinate Sb(V) species have
nebulous geometric configurations.13 Even the simple SbPh5
exhibits a distorted 4SQ geometry.13a,b Counterintuitively, the
perhalogenated species Sb(C6X5)5 exhibit a fairly regular 3BP
geometry, as typically a more withdrawing substituent is pre-
sumed to enforce a 4SQ geometry in phosphorus species.13c
The SbPh5 geometry is proposed to exist in the solid state due
to the isoergic nature of the 3BP and 4SQ geometries, yet this
similarity in energy is not quite understood.13a,b Finally, a
related series of heterobicycles reported by Holmes et al.
strictly exhibits 4SQ geometry, highly reminiscent of the new
series of structures reported herein.13d,e
With this contribution, we investigate what ultimately dic-
tates the geometric orientation of these species, and what
impact does that have on the inherent Lewis acidity of the
phosphorus center. To this end, we report the synthesis, struc-
tures, properties, and reactivity of the new hypervalent dithie-
nophospholes that provide an exciting opportunity for under-
standing these types of system more deeply.
Results and discussion
Synthetic procedures
For our study, we focused on species that would potentially be
able to (a) stabilize the hypervalent state and/or (b) enhance
the electron-acceptor properties of the target species. To this
Scheme 1 Synthesis of quinone-functionalized hypervalent dithieno-
end, we identified o-chloranil and phenanthrene quinone as phospholes.
2244 | Dalton Trans., 2021, 50, 2243–2252
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Solid state characterization
phosphole. The bite angle of the 1,2-diol system is geometri-
cally restrictive. The resultant 5-membered ring with the phos-
phorus center exhibits a nearly 90° O–P–O bond angle.
However, while catecholates form strong chelates with P(V)
centers, they are often labile, detaching and reattaching to
afford the preferred geometric isomers;19 and even allow for
the generation of reactive sites at either the unbound P or O
atoms. These available sites are short lived, however, and are
typically observed only if captured, such as by a silyl group or
cation for O− or a base such as F− for P+, to prevent the re-
association of the oxygen with the phosphorus center.19 With
respect to the phosphole unit, while breaking a P–C bond
within the phosphole is not entire impossible, it is a less labile
substituent than the catecholate. In addition, the geometry of
the C–P–C unit of the phosphole ring is not dictated by, for
example, like above the strain of a 4-membered ring system,
but the electronic interactions between the phosphorus center,
the diene system and the exocyclic substituents. Our system
would therefore, intuitively, generate the most favorable
isomer in solution due to the lability of the catecholate and
the geometry at the phosphole being dictated by the associ-
ation of said catecholate, when compared to these structures
of precedent.19 Still, the small bite angle of the catecholate
prohibits the formation of the most favourable 3c–4e bond, an
O–P–O bond, in which the optimized 3BP structure is formed
by the axial oxygen donating a lone pair into the opposing
σ*PC orbital, requiring instead a weaker 3c–4e C–P–O bond
between one catecholate oxygen and one ipso-carbon to retain
a 3BP structure. To better understand the complex interplay
between these factors, we turned our focus to computational
chemistry. A detailed analysis of the 3BP vs. 4SQ structural pre-
ference is given in section 9(i) of the ESI† using compound A
as a representative species.
Single crystals suitable for X-ray diffraction were readily grown
from compounds A, C, and D and with some difficulty of com-
pound B. Rather unexpectedly, all the structures exhibited a
distorted square pyramidal (4SQ) geometry (Fig. 2) as opposed
to the more typical, and VSEPR supported 3BP geometry con-
taining a 3c–4e bond (Fig. 1b).
The central phosphorus atom is moved out of the plane of
the π-conjugated scaffolds by approximately 0.4 Å, with the aryl
groups (phenyl for A and B, perfluorophenyl for C and D)
arranged nearly orthogonally to the plane containing the
quinone unit and dithienophosphole moiety (angles are 103.7
(7)°, 103.6(5)°, 105.6(2)° and 103.2(4)° A, B, C, and D, respect-
ively) to form the apex of the square pyramid.
While the average P–C bond lengths in these structures
(1.8 Å) are very close to usually observed P–C bond lengths of
1.84 Å,16a the P–O bonds are elongated to 1.7 Å from normally
observed P–O bond lengths of 1.63 Å,16b but consistent with a
catecholate chelating to a P(V) center.10f,g,i
Precedent for these structures, albeit uncommon, is found
in the literature,10b but previous studies do not afford a clear
explanation of what constrains such systems to this 4SQ geo-
metry. While the related 3BP structures predominantly orient
the substituents about the phosphorus center to afford the
most favorable 3c–4e bonds, a similar argument is not
obviously applicable to the 4SQ equivalent species.10c While it
has been suggested that a strong electron-withdrawing ligand
will instil 4SQ geometry,10a,11 that is not often the case; oxa-
lates and tetrachlorocatecholates readily afford 3BP structures,
as found in chirality resolution agents.17 The 4SQ geometry
can also be enforced via geometric constraints;10a,b,11 Holmes
et al. leveraged the strain of a 4-membered ring to offset the
presumably higher-energy 4SQ geometry.10f Gilhula and
Radosevich employ tetradentate or scorpionate ligands to
enforce a 4SQ geometry in order to make use of the reactivity
of the 4SQ phosphorus center in chemical transformations.18
Our species, however, consist instead of a catecholate and
The natural bond orbital (NBO) population analysis20
shows expected electron transfer from phosphorus to both
oxygen atoms, with natural atomic charges of +1.90 and −0.75
for P and O atoms, respectively. Both oxygen atoms, somewhat
unexpectedly, are available to donate a lone pair to their oppo-
site σ*PC orbitals to generate two 3c–4e bonds. In the 3BP
structure, one strong formal 3c–4e bond exists between the
axial O and the axial C(ipso), while in the 4SQ structure, two
symmetrically equivalent 3c–4e bonds persist in the square
base. Our calculation shows that the conversion from the 3BP
to the 4SQ structure involves the process of weakening the 3c–
4e bond of the 3BP isomer, relaxing the 120° bond angles
between the equatorial substituents, and developing a second
3c–4e bond. This energy compensation results in comparable
stabilities of the 3BP and 4SQ structures, and thus their
dynamic sampling in solution.7b,10b,21 Since the 4SQ structure
is in-between two 3BP structures (each having one or the other
O atom at the axial position), the sampling results in an aver-
aged 4SQ structure, which is observed in our structural charac-
terizations. In the solid state, 4SQ structures are exclusively
observed because their parallel arrangements of the extended
conjugated scaffolds between adjacent molecules also
enhances their π–π stacking interactions (vide infra). As such
Fig. 2 Molecular structures of compounds A, B, C, and D in the solid
state. Ellipsoids are set at 50% probability.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 2243–2252 | 2245

View Article Online
Paper
Dalton Transactions
the ability to generate stable 4SQ species relies on generating
Consequently, we were interested in the FIAs25 of the new
two 3c–4e bonds of intermediate strength that afford the pentacoordinate dithienophospholes of this series to deter-
system similar stability as the strong 3c–4e bond of the 3BP mine the Lewis acidity of the hypervalent P center. The calcu-
isomer. This process may be performed either electronically, lated FIAs for compounds B, A, D, and C (gas phase) are 306,
geometrically, or both as long as the resultant 3c–4e bonds of 334, 348, and 381 KJ mol−1 in increasing order (compared to
the 4SQ isomer are sufficiently strong so as to afford an isoer- the isodesmic reaction Me3SiF− → Me3Si+ + F−), respectively,
gic isomer to the 3BP structure. This explanation can be seen that are correlated with the electron-withdrawing effects of
in the computational analysis of structures A–D, whose energy both the o-chloranil and perfluorophenyl groups; the latter at
differences between the two isomers are −0.4, 0.1, 1.7, and the axial position has an expectedly stronger impact on enhan-
1.1 kcal mol−1, respectively (negative means 4SQ being more cing the FIA. However, while the phosphorus center in these
stable). With the small energy differences, the molecules in compounds is Lewis acidic, clearly none of the new hyperva-
fact likely dynamically rock between the 3BP and 4SQ struc- lent species are Lewis Super Acids. It is noteworthy, however,
tures (akin to Berry pseudorotation),7b,10b,21 affording some that the FIA values for A–D are significantly higher than that of
insight into the aforementioned dynamic solutions state the isoelectronic (IIIB) (71 KJ mol−1).12d This highlights the
nature of these compounds. This dynamical rocking reflects effect of their neutral character in contrast to the anionic
the lability of the bidentate interaction between P and the cate- nature of (IIIB) with its already present fluoride substituent.
cholate O atoms, despite their rigid O–P–O angle. Such an in-
Fluoride abstraction reaction. With the computational data
depth analysis of the preference towards 4SQ structure of a suggesting moderate Lewis acidity, we followed up with fluor-
P(V) structure has not been previously explored. Note that, as ide abstraction experiments. Compound A was initially reacted
expected, compounds C and D exhibit slightly more favorable with one equivalent of potassium fluoride (KF) in dry toluene,
3BP geometric isomers. But this is not due to the more electro- but the 31P{1H} NMR spectrum of the reaction mixture after
negative C6F5 group, which always occupies an equatorial posi- one hour confirmed the formation of compound 3 instead of
tion in the 3BP structures.
K+[A–F−]. Similarly, the reaction of A with one equivalent of
tetrabutylammonium fluoride (TBAF) showed weak doublets
with J(P–F) coupling for TBA+[A–F−] (1JP–F ∼ 780 Hz)15b after
1
Reactivity studies of the hypervalent compounds
one hour of reaction time, but with 3 as the major product.
The results were overall inconclusive regarding the Lewis
acidity of this series of compounds, but suggested a different
reaction was clearly occurring, at least for compound A
(Scheme 2; ESI, section S3†).
Solvatochromism study of compound A. Due to the polar
nature of its scaffold (as well as the presence of the sigma
hole), we surmised that compound A would exhibit solvato-
chromism. Solutions of A (c = 10−4 M) in DCM and toluene
showed an absorption maximum λmax = 388 nm that is blue-
shifted to 358 nm in coordinating solvents such as acetone,
THF, and acetonitrile (Fig. 4). The change in λmax was
accompanied by a change in solution color from yellow in
toluene (and DCM) to colorless in coordinating solvents,
indirectly supporting the Lewis acidity of A (see ESI, section
4†).
We surmised that the unexpected 4SQ geometry could also
impart intriguing reactivity to these hypervalent compounds,
similar to related species in the literature,22 as suggested by
the electron-density mapping (Quantum Theory of Atoms in
molecules; QTAIM) representatively on compound A.23 An
area of low electron density on the P center, often referred to
as sigma hole,24 is a potential coordination site for nucleo-
philes such as OH− or F−, which would make these hyperva-
lent compounds neutral Lewis acids (Fig. 3a).12 Recent
examples of such rare neutral Lewis acids, include the bis
(perchlorocatecholato)silane (IIIA) with a Lewis-acidic Si(IV)
center, reported by Greb et al.12d,e This compound has been
classified as Lewis Super Acid by virtue of its high fluoride
ion affinity (FIA; 507 KJ mol−1) and acquires similar square
pyramidal geometry on abstraction of a fluoride anion (IIIB)
(Fig. 3b).12d
Reactivity with water. The fluoride abstraction studies
suggested that presence of trace amounts of water trigger the
transformation of A to 3 by a nucleophilic addition,19 rather
than a simple retro-[4 + 1] cycloaddition to the starting phosp-
hole 1. Notably, a solution of A in THF that was rigorously
Fig. 3 (a) Electron density mapping showing the sigma hole in A; (b)
silicon-based neutral Lewis acid.12d
Scheme 2 Transformation of A to 3 via fluoride abstraction reaction.
2246 | Dalton Trans., 2021, 50, 2243–2252
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
crystals after two weeks, yet maintained suitable single-crystal
characteristics for XRD analysis (Fig. 5b). X-ray crystallography
on the transformed crystals confirmed the formation of phosp-
hole oxide (3) with 3,4,5,6-tetrachloro-1,2-benzenediol (4) as
the other by-product (Fig. 5c; ESI†). On the contrary to reac-
tions happening exclusively on the surface, i.e., the direct
interface with the atmospheric moisture, the crystals are com-
pletely transformed to 3 and 4 showing that the decomposition
of the outer layer opens the core of the unreacted crystal to a
cascade of reactions. Notably, the crystallinity is preserved
even as entropy increases. While the hydrolysis of P(V) catecho-
lates is not unknown,10e,26 the intriguing crystal-to-crystal
transformation suggests that the corresponding mechanism
for A may not necessarily require the breaking of a P–O cate-
cholate bond (vide supra) as the initial step.
Fig. 4 UV-vis spectra of
A in coordinating (THF, acetonitrile, and
acetone) and non-coordinating (toluene and DCM) solvents.
To gain more insights into the detailed mechanism of the
reaction with water in general, we again turned our focus to
computational chemistry. We first focused on the reaction in
solution. The calculations employ a THF solvent model with
one explicit THF molecule coordinated to H2O via a hydrogen
bond. The whole reaction is calculated to be exoergic by
20.0 kcal mol−1, and there is essentially only one 20.3 kcal
mol−1 barrier (TS1), which is feasible to overcome at room
temperature (see ESI†). This rate-determining step involves
concerted migrations of four electron pairs (Fig. 6a): (1) the
addition of a H2O lone pair to the acidic P center, forming
the new O–P bond; (2) the heterolytic breakage of one original
quinone O–P bond and the transfer of the bond electrons
to the O atom; (3) this O atom donating its lone pair to a
proton of the H2O molecule to form an OH bond; and (4) het-
erolytic breakage of one OH bond in H2O (see ESI for more
details†).
dried over neutral alumina and freeze-pump-thawed, does
indeed not show any formation of 3 by 31P{1H} NMR spec-
troscopy (ESI, section 4†). To further verify the adventitious
water hypothesis, compound A was exposed to a controlled
quantity of water in solution. Upon addition of 50% degassed
water to a similarly dried THF solution of A, its blue-green
fluorescence quickly changed to the bright blue emission
characteristic of dithienophosphole oxide 3 (Fig. 5a). The for-
mation of 3 was further confirmed by a downfield shift of the
31P{1H} NMR signal from −39.4 ppm for A (THF) to 13.9 ppm
for 3 (THF/H2O) (see ESI†). While compound 3 was clearly
identified by NMR spectroscopy, the nature of the C6Cl4O2-
fragment, however, was unclear due to the ambiguity of the
13C{1H} NMR spectroscopic data.
In our experiments, alteration of the electronegativities of
the axial group (C6H5 vs. C6F5) and the side group (chloranil
vs. phenanthrene quinone) did not indicate a clear trend of
reaction rates. Logic dictates that an electronegative axial
group favors electron pair migration (1) but disfavors (2). An
electronegative side group favors (2) but disfavors (3). The sto-
chastic rates of the transformation of A, C, and D are reflec-
tions of these contradictory features. Once the H2O-addition
intermediate (IM1, Fig. 6b) is formed, the subsequent steps (in
essence, second O–P breakage and proton transfer) are all
thermodynamically and kinetically favorable (see Fig. S28(a) in
ESI†). In the whole process, the phosphorus atom acts as a
Lewis acid in electron pair migrations (1) and (5) and the
aryloxy oxygen atom acts as a Lewis base in (3). The calcu-
lations show that an essentially similar mechanism is also
Intriguingly, the rod-shaped golden yellow crystals of A,
when left on the bench and exposed to atmospheric moisture
in a loosely closed vial, transformed into pale yellow hexagonal
Fig. 5 (a) Solution-phase transformation; (b) crystal-to-crystal trans-
formation; (c) products 3 and 4 of the crystal-to-crystal transformation
(A·H2O).
Fig. 6 Concerted electron-pair migrations for the reaction with water.
Dalton Trans., 2021, 50, 2243–2252 | 2247
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Dalton Transactions
applicable to the solid-state transformation of the crystals of A
in the presence of water (see Fig. S28(b) in ESI†). This pro-
posed mechanism is uniquely supported by our observed reac-
tivity, as the typical alternate reaction, requiring the breakage
of a P–O bond first, instead of the acidity of the phosphorus
center, relies on the basicity or electron-withdrawing nature of
the substituent of which the O–P bond is broken, dictating a
more intuitive ordering of the rates of transformation.17a,c
Supramolecular stabilization of the hypervalent compounds
Despite the Lewis acidity of the phosphorus center and the
resulting increased reactivity to trace amounts of water, these
kinetically labile compounds can nonetheless be crystallized
from a concentrated solution on the bench. We assumed that
their stability in the crystalline state would likely arise from
intermolecular stabilization effects, given the fairly planar
scaffold that allows for π–π interactions. The crystal packing of
A shows molecules arranged in a parallel conformation where
the exposed P center of one molecule is protected by the plane
of a second molecule (3.5 Å) (Fig. 7a). Similar π–π interactions
were also observed in single crystals of compounds B and C,
while D forms π-stacked tetramers, likely due to the presence
of both highly electron-withdrawing and -donating groups that
complement each other intermolecularly (see ESI†).
The supramolecular dimerization interaction provides
kinetic stability (at least for the medium term and the absence
of nucleophiles), while the close intermolecular distance may
afford electronic transitions across the dimers. To investigate
both this, as well as the fate of these dimers in the solution
state, the luminescence of these emissive compounds was
studied as a function of concentration in dry toluene.
Notably, the interactions were found to be considerably
strong, as excimer-type emissions27 were observed, even at con-
centrations as low as 10−13 M. The excimer of compound A
emits at 459 nm, while lowering the concentration to 10−7
M
revealed a peak with vibronic structure centered around
409.5 nm that can be assigned to the monomeric species
(Fig. 7b; we were able to exclude Raman scattering for these
peaks by varying the excitation wavelength). Similarly, the
monomer for compound B is detected at 396.5 nm (ESI†).
Compound C does not show clear monomer peaks even at a
concentration of 10−6 M, which indicates extremely strong
supramolecular interaction due to the increased Lewis acidity
of the P(V) center hosting two highly electron-withdrawing sub-
stituents (Fig. 7c; here, the peak at 450 nm corresponds to
Raman scattering). As already suggested by its solid-state struc-
Fig. 7 (a) Crystal packing in
A
showing supramolecular π-stacked
dimers; (b) fluorescence emission spectra showing monomer and
excimer for A (excitation at 365 nm); (c) fluorescence emission spectra
showing only excimers in C (excitation at 397 nm).
ture, compound D also behaves differently in solution. The (dimethylamino) sulfur (trimethylsilyl) difluoride (5) was
emission spectrum resolves into multiple peaks on dilution, as employed as a dry fluoride source for the anion abstraction
several different π–π interactions may persist, while exhibiting reaction.12d Reactions of one equivalent of 5 with the hyperva-
aggregation caused quenching (ACQ)28 (ESI, section 5†).
lent compounds A–D in meticulously dried solvent under
argon, after one hour each give rise to highfield-shifted 31P
{1H} NMR doublets with strong P–F coupling (as corroborated
Revisiting the fluoride abstraction reaction
As the initial fluoride abstraction reactions using typical fluor- by 19F{1H} NMR spectroscopy). The shifts are consistent with
ide salts ultimately led to the formation of 3 in all cases octahedral P(V) species, and align well with related systems in
(Scheme 2), we concluded that this was due to the presence of the literature (Scheme 3).29 Coupling constants (1JP–F) of 776
adventitious water (vide supra). To eliminate this scenario, tris Hz (A–F−), 752 Hz (B–F−), 795 Hz (C–F−), and 764 Hz (D–F−),
2248 | Dalton Trans., 2021, 50, 2243–2252
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Scheme 4 Transformation of B–F– by deoxyaryloxylation reaction.
for a similar oxygen-transfer reaction in the literature, but the
authors were unable to determine the fate of the corres-
ponding diolate subsituent.19a
Given the highly symmetric nature of the phenanthrene-
based moiety that was confirmed via 13C{1H} and 1H NMR
spectroscopy after its separation (see ESI†), formation of
diphenanthro[9,10-b:9′,10′-e][1,4]dioxin (6) is the only plausible
outcome for this reaction. The observed reaction could then be
classified as an double deoxyaryloxylation, where one of the
quinone oxygen atoms is transferred to phosphorus while the
remaining forms an ether bond between the aryl groups. This
intriguing observation indicates the potential utility of the new
species as mediator in other organic transformations.31,32
To lend more credence to this proposed reaction, we again
turned to computational chemistry using compound A as
representative species for determination of a suitable reaction
Scheme 3 Fluoride abstraction using 5 as dry source.
respectively, confirm the abilities of the pentavalent molecules mechanism. The mechanism of the deoxyaryloxylation reac-
to expand their coordination sphere to generate hexacoordi- tion is shown in Fig. 8. The first step is the abstraction of F−
nate species that are indeed reasonably stable under inert from 5 by A. This step is barrierless and exoergic by 8.7 kcal
conditions.15b,19,29,30
mol−1. The abstraction results in a hexacoordinate compound
Notwithstanding, also within one hour of reaction time, we IM5 that aligns with the experimental observations (vide
observed the additional formation of a noticeable amount of supra). Subsequently, IM5 undergoes heterolytic O–P bond
dithienophosphole oxide 3 (or its C6F5-substituted relative 2-O) breakage to return to the pentacoordinate IM6, whose dan-
by 31P{1H} NMR spectroscopy, along with the fluoride adducts gling O− is stabilized by the [S(NMe2)3]+ countercation. The
(ESI, section 6†).
barrier for this step (TS3) is calculated to be 7.2 kcal mol−1
.
Despite the anaerobic conditions under which the fluoride Compared to the addition of H2O to A (TS1 in Fig. S26a, ESI†),
abstraction reactions were conducted, the presence of dithie- only the migration arrows (2) and (3) are applicable here, and
nophosphole oxide suggests that further reactions beyond that the barrier is lower by two thirds (vs. 20.3 kcal mol−1 of TS1).
of the fluoride abstraction must occur. This also indicated a This is because: (1) the hexacoordinate phosphorus center
fleeting nature for the corresponding fluoride adducts. with formal −1 charge in IM5 has a stronger tendency to break
Importantly however, given that there is no available external the O–P bond and release the charge to the more electronega-
oxygen source for the formation of 3 (or its C6F5-substituted tive oxygen center; (2) the [S(NMe2)3]+ countercation is readily
relative), the reaction pathway for this transformation must be available to stabilize the anionic oxide. Both aspects are proto-
clearly distinct from the one observed in the presence of water. typical for the reactivity of octahedral P(V) catecholates.33
Reaction of 5 with B shows the formation of B–F− and 3 after Without the rigid ∼90° O–P–O bite angle, the pentacoordinate
one hour in the 31P{1H} NMR spectrum in a roughly 11 : 7 ratio IM6 adopts a 3BP structure, consistent with the discussion
(see ESI, section 6†). The reaction mixture was then refluxed above, and described in more detail in the ESI (section 9).† An
for 17 days to fully convert B–F− into 3. After completion, the ipso-C atom of the dithieno-unit and the most electronegative
13C{1H} NMR spectrum of the reaction mixture was recorded fluoride occupy the axial positions. In the third step, electron
to establish the fate of the former phenanthrene quinone unit pair migrations (10) and (11) occur in concerted fashion
(see ESI†). In addition to the eight peaks for 3, seven other between a dimer of IM6; the remaining P–O–C unit is broken
peaks suggested the presence of a symmetric, 9,10-dioxo-phe- across the O–C bond in order to form a new O–C bond across
nanthrene fragment (Scheme 4 and ESI†). The anaerobic reac- the two o-chloranil groups. Since it is computationally costly to
tion conditions imply that the PvO unit of 3 must have been search for the transition state (TS4) for this large reaction
formed by breaking a C–O bond of B–F−. There is precedence complex, we estimate the barrier to be about 34.0 kcal mol−1
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 2243–2252 | 2249

View Article Online
Paper
Dalton Transactions
Fig. 8 Energy landscape for the deoxyaryloxylation reaction of A. In this figure two curves are used to schematically represent the o-chloranil and
dithieno groups for better clarity. When necessary, the two carbon atoms connected to oxygen in the o-chloranil group are explicitly shown, while
the other atoms of the group are omitted for clarity. Electron pair movements are shown in blue. Key structures are shown in Fig. S29 (ESI).† Energy
differences between structures connected by dashed lines are given in kcal mol−1. The energies for the first two dashed lines are for each A unit,
while the others are for the whole system with two A units. The “x 2” suggests that the given energies are for each A unit and should be multiplied by
two for the total system with two A units. IM and TS are numbered consecutively for all mechanism studies. IM1–4 and TS1–2 are shown in section
9 of ESI† for A + H2O reaction mechanism. Similarly, details on pair migrations (1)–(9) are provided in section 9 of the ESI.† The energies were
obtained using CAM-B3LYP/MA-DEF2-SVP/CPCM(THF) calculations and with zero-point energy corrections.
without running the search (vide infra). The barrier-less pair final products 3 and perchloro dibenzodioxin for more than a
migrations (12) and (13) convert the pentacoordinate anion of month. Similarly, persistent mixtures are observed for C and D
P1 (the two P atoms are labelled as P1 and P2 in Fig. 8 for and for all the three species, the final products (−17.5, −20.9,
better clarity) to 3, whose oxygen atom weakly coordinates to −27.5 kcal mol−1 for A, C, and D, respectively) are slightly
the Si atom of an SiMe3F molecule (shown as R3PvO⋯SiMe3F, lower or even substantially higher in energy than IM5 (−17.4,
with a 1.9 kcal mol−1 binding energy), and release an F− to the −29.9, −24.5 kcal mol−1 for A, C, and D, respectively), leading
solution (shown as S(NMe2)3F). The product IM7 of this step to the prolonged observation of IM5 in solution. Compound B
has a similar electronic structure as IM6, with a dangling O− is the only species that undergoes a complete deoxyaryloxyla-
stabilized by [S(NMe2)3]+. The difference between IM6 and IM7 tion, since the fluorination is much less exoergic than the
is that the C6Cl4 unit has been replaced by the C6Cl4–O–C6Cl4 deoxyaryloxylation (−11.2 vs. −28.1 kcal mol−1). This lowest
dimer. IM7 then undergoes the concerted pair migrations (14) exoergicity in fluorination among A–D is consistent with the
and (15) to form perchlorodibenzodioxin (the o-chloranil con- least electronegative side and axial groups of B. Still, the two
gener to 6). The remaining pentacoordinate anion of P2 under- approx. 30 kcal mol−1 barriers of TS4 and TS5 are in line with
goes barrier-less migrations (12) and (13) to form another the experimentally observed deoxyaryloxylation using B that
molecule of 3 (also coordinated to SiMe3F) and returns an F− took 17 days to complete.
to the solution. The transition state for this step (TS5) was
The mechanism also relies on the dynamic nature of the
found to be 34.0 kcal mol−1 above IM6. Since the pair geometric structure present in these species. The fluoride
migrations (14) and (15) resemble (10) and (11), the barrier of abstraction requires the available coordination site of the 4SQ
TS4 is expected to be similar to that of TS5. We thus estimated phosphorus center. The subsequent P–O bond breaking is,
the energy of TS4 to be about 34.0 kcal mol−1 (vide supra). again, not uncommon among octahedral P(V) catecholates as a
However, one should note that this conjecture is for the acti- means to optimize the three 3c–4e bonds present in the struc-
vation energy. Considering that TS4 is in a bimolecular inter- ture. The less favorable F–P–C bond drives the ring opening
action, while TS5 is in a intramolecular interaction, the and the subsequent reactivity along with the thermodynamic
reduction of entropy would give a higher Gibbs free energy stability of a tetrahedral P(V) oxide. The generation of 3 during
barrier for TS4 than TS5. But this does not change the overall the reaction with more typical fluoride ion sources is also
picture of the reaction, which is held as a mixture of IM5 and readily explained via this mechanism. Instead of a simple com-
IM6 for a long time.
petition between F− and water reacting at the Lewis acidic
The proposed mechanism suggests F− to act as a catalyst phosphorus center, the decomposition occurs from IM5. Upon
and is exoergic by 17.5 kcal mol−1. However, the two >30 kcal ring opening, the water reacts with the now accessible phos-
mol−1 barriers severely slow down the process. Also, the reac- phorus center instead of either rebinding the catechol or gen-
tion may be trapped at the first step of forming IM5, which is erating the dimer IM6. This would explain the presence of
isoergic with the final product. In fact, the addition of 5 to A both IM5 and 3 in solution from these reactions, as well as the
results in a long-lasting mixture of the fluorinated IM5 and the slow rate of hydrolysis.
2250 | Dalton Trans., 2021, 50, 2243–2252
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
gated main group species. Consequently, these species may
have the potential to be exploited as neutral and stable hyper-
Conclusions
A facile method toward hypervalent dithieno[3,2-b:2′,3′-d]- valent main group catalysts (or mediators) in organic chem-
phospholes via the [4 + 1] cycloaddition of ortho-quinones to istry. Now with a model to design these uniquely reactive com-
trivalent species is reported. Our approach alleviates the pounds in hand, we are currently looking into expanding the
requirements for additional kinetic stabilization, as often scope of application of these reactive species.
observed in previously reported systems. The structure of the
obtained pentacoordinate derivatives deviates from the VSEPR
theory-dictated trigonal bipyramidal geometry to form square Conflicts of interest
pyramidal molecules, that in turn create a well-exposed Lewis
There are no conflicts to declare.
acidic phosphorus center, evident in the presence of a sigma
hole. This structural preference arises from the shallow poten-
tial energy surface connecting the two types of structure,
which respectively balance two intermediate 3c–4e bonds on
Acknowledgements
the one side, and both a strong and a weak 3c–4e on the other
Financial support by the Natural Sciences and Engineering
side. While we do not and cannot strictly rule out steric con-
Research Council of Canada (NSERC) and the Canada
straints about the phosphorus center, through this series of
Foundation for Innovation is gratefully acknowledged. T.B.
compounds, we were able to establish a thermodynamic model
thanks the Canada Research Chairs program for support. T.Z.
thanks Calcul Quebéc and Compute Canada for computational
resources.
of the restrictions on forming a λ5-P(V) center with 4SQ geome-
try. Additionally, based on the limited Lewis acidity of these
compounds, we can propose the constraints of this model on
Lewis acidity: (i) the stability of dimers formed even in solu-
tion; (ii) the dynamic geometry of these structures due to the
extremely low energy difference between 3BP and 4SQ geome-
Notes and references
tries; and (iii) the alternative reactivity of these species upon
reacting with the Lewis base. The latter potentially affords this
series of compounds an alternate use in the synthesis of other
organic molecules. We can ascribe its Lewis acidity to these
properties arising from the competing electronic effects of the
system, in which the favorability of one strong 3c–4e bond is in
direct competition of the generation of two 3c–4e bonds. By
understanding the difference between the phosphorus
bonding in, for example catechol versus dithienophosphole,
this model may allow for the specific generation of λ5-P(V)
species strictly with either 4SQ or 3BP geometries. This would
in turn enable targeted reactivities by adjusting the favorability
of one 3c–4e bond to be much stronger or weaker than, seren-
dipitously, this series of compounds in which both geometries
are nearly isoergic. The Lewis acidic character of the 4SQ phos-
phorus center imparts high electrophilicity to water. The
ensuing rapid transformation of the pentacoordinate com-
pounds to dithienophosphole oxide, and an intriguing crystal-
to-crystal transformation can be ascribed to the unique reactiv-
ity of these compounds as well as the significance of the geo-
metric state toward their reactivity. Despite their high electro-
philicity, the peculiar geometry of these compounds allows for
kinetic intermolecular stabilization by supramolecular inter-
actions in the solid state under ambient conditions.
Hexacoordinate P-species are accessible by fluoride abstrac-
tion, which reveals the potential of these derivatives to facili-
tate interesting organic transformations such as the deoxyary-
loxylation reaction that involves a generally challenging C–O
1 S. Ogawa, Organic Electronics Materials and Devices,
Springer, Japan, 2015.
2 (a) M. Stolar and T. Baumgartner, Phys. Chem. Chem. Phys.,
2013, 15, 9007; (b) M. Stolar and T. Baumgartner, Chem. –
Asian J., 2014, 9, 1212; (c) T. Baumgartner and F. JäkleMain
group strategies towards functional hybrid materials, Wiley,
Hoboken, 2018.
3 (a) T. Baumgartner, T. Neumann and B. Wirges, Angew.
Chem., Int. Ed., 2004, 43, 6197; (b) C. Romero-Nieto and
T. Baumgartner, Synlett, 2013, 920.
4 T. Baumgartner, Acc. Chem. Res., 2014, 47, 1613.
5 Z. Wang, N. Asok, J. Gaffen, Y. Gottlieb, W. Bi, C. Gendy,
R. Dobrovetsky and T. Baumgartner, Chem, 2018, 4, 2628.
6 (a) Y. Ren, W. H. Kan, M. A. Henderson, P. G. Bomben,
C. P. Berlinguette, V. Thangadurai and T. Baumgartner,
J. Am. Chem. Soc., 2011, 133, 1701; (b) Y. Ren and
T. Baumgartner, Inorg. Chem., 2012, 51, 2669.
7 (a) C. Y. Wong, D. K. Kennepohl and R. G. Cavell, Chem.
Rev., 1996, 96, 1917; (b) V. Iaroshenko, Organophosphorus
Chemistry: From Molecules to Applications, Wiley-VCH,
Weinheim, Germany, 2019.
8 F. H. Osman and F. A. El-Samahy, Chem. Rev., 2002, 102,
629.
9 (a) J. F. Berry, Dalton Trans., 2012, 41, 700;
(b) M. L. H. Green and G. Parkin, Dalton Trans., 2016, 45,
18784; (c) S. Noury and B. Silvi, Inorg. Chem., 2002, 41,
2164; (d) C. R. Landis and F. Weinhold, Inorg. Chem., 2013,
52, 5154.
bond activation.34 The synthesized class of hypervalent phos- 10 (a) J. A. Howard, D. R. Russell and S. Trippett, J.C.S. Chem.
phorus compounds presents an innovative structural platform
while hosting desirable chemical properties that showcase
once more the intriguing opportunities arising from conju-
Commun., 1973, 1306, 856; (b) R. R. Holmes and
J. A. Deiters, J. Am. Chem. Soc., 1977, 99, 3318;
(c) W. Althoff, R. O. Day, R. K. Brown and R. R. Holmes,
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 2243–2252 | 2251

View Article Online
Paper
Dalton Transactions
Inorg. Chem., 1978, 17, 3265; (d) S. Sase, N. Kano and 21 J. P. Foster and F. Weinhold, J. Am. Chem. Soc., 1980, 102,
T. Kawashima, J. Org. Chem., 2006, 71, 5448; (e) M. Kaupp, 7211.
Angew. Chem., Int. Ed., 2001, 40, 3534; (f) T. E. Clark, 22 Y. Lin, J. C. Gilhula and A. T. Radosevich, Chem. Sci., 2018,
R. O. Day and R. R. Holmes, Inorg. Chem., 1979, 18, 1653; 9, 4338.
(g) T. E. Clark, R. O. Day and R. R. Holmes, Inorg. Chem., 23 T. A. Keith, AIMAll (Version 19.10.12), TK Gristmill Software,
1979, 18, 1660; (h) T. E. Clark, R. O. Day and R. R. Holmes, Overland Park KS, USA, 2019, (aim.tkgristmill.com).
Inorg. Chem., 1979, 18, 1668; (i) H. R. Allcock and 24 J. Y. C. Liam and P. D. Beer, Chem, 2018, 4, 731.
E. C. Bissell, J. Am. Chem. Soc., 1973, 95, 3154.
11 R. R. Holmes, Acc. Chem. Res., 1979, 12, 257.
12 (a) K. B. Dillon, R. N. Reeve and T. C. Waddington, J. Chem.
25 (a) L. Greb, Chem. – Eur. J., 2018, 24, 17881; (b) J. Zhou,
L. L. Liu, L. L. Cao and D. W. Stephan, Angew. Chem., Int.
Ed., 2019, 58, 5407.
Soc., Dalton Trans., 1978, 1465; (b) L. A. Mück, 26 (a) R. R. Holmes, Chem. Rev., 1996, 96, 927;
A. Y. Timoshkin and G. Frenking, Inorg. Chem., 2012, 51,
640; (c) V. Fasano, J. H. W. LaFortune, J. M. Bayne,
M. J. Ingleson and D. W. Stephan, Chem. Commun., 2018,
54, 662; (d) R. Maskey, M. Schädler, C. Legler and L. Greb,
Angew. Chem., Int. Ed., 2018, 57, 1717; (e) H. Ruppert and
L. Greb, Organometallics, 2020, 39, 4340.
(b) P. G. Mashaffy, J. D. Wieser and L. K. Montgomery,
J. Am. Chem. Soc., 1977, 99, 4515; (c) P. W. Siu and
D. P. Gates, Organometallics, 2009, 28, 4491;
(d) T. Kawashima, R. Okazaki and R. Okazaki, Angew.
Chem., Int. Ed. Engl., 1997, 36, 2500.
27 J. R. Lakowicz, Principles of Fluorescence Spectroscopy,
Kluwer Academic/Plenum, New York, 1999.
13 (a) P. J. Wheatley, J. Chem. Soc., 1964, 3718;
(b) A. L. Beauchamp, M. J. Bennett and F. A. Cotton, J. Am. 28 R. Jakubiak, C. J. Collison, W. C. Wan and L. J. Rothberg,
Chem. Soc., 1968, 90, 6675; (c) M. A. Garcίa-Monforte, J. Phys. Chem. A., 1999, 103, 2394.
P. J. Alonso, I. Ara, B. Menjόn and P. Romero, Angew. 29 (a) T. Kaukorat, P. G. Jones and R. Schmutzler, Heteroat.
Chem., Int. Ed., 2012, 51, 2754; (d) R. R. Holmes, R. O. Day,
Chem., 1991, 2, 81; (b) J. A. Gibson, G. Roschenthaler and
V. Chandrasekhar and J. M. Holmes, Inorg. Chem., 1987,
R. Schmutzler, J. Chem. Soc., Dalton Trans., 1975, 918.
26, 157; (e) R. R. Holmes, R. O. Day, V. Chandrasekhar and 30 (a) A. F. Janzen, Coord. Chem. Rev., 1994, 130, 355;
J. M. Holmes, Phosphorus, Sulfur Silicon, 1995, 99, 13;
(f) T. T. Bamgboye, M. J. Begley and D. B. Sowerby,
J. Organomet. Chem., 1989, 362, 77; (g) A. P. M. Robertson,
(b) M. Gallagher, J. Chem. Soc., Chem. Commun., 1976, 321;
(c) J. J. H. M. Font Freide and S. Trippett, J. Chem. Soc.,
Chem. Commun., 1980, 157.
N. Burford, R. McDonald and M. J. Ferguson, Angew. 31 (a) A. E. Waked, S. S. Chitnis and D. W. Stephan, Chem.
Chem., Int. Ed., 2014, 53, 3480; (h) B. Pan and F. P. Gabbai,
J. Am. Chem. Soc., 2014, 136, 9564.
Commun., 2019, 55, 8971; (b) T. Stahl, H. F. T. Klare and
M. Oestreich, ACS Catal., 2013, 3, 1578; (c) P. Spies,
G. Erker, G. Kehr, K. Bergander, R. Fröhlich, S. Grimme
and D. W. Stephan, Chem. Commun., 2007, 5072;
(d) R. Dobrovetsky, K. Takeuchi and D. W. Stephan, Chem.
Commun., 2015, 51, 2396; (e) M. C. Hilton, X. Zhang,
B. T. Boyle, J. V. Alegre-Requena, R. S. Paton and
A. McNally, Science, 2018, 362, 799; (f) N. Đorđevic,
M. Q. Y. Tay, S. Muthaiah, R. Ganguly, D. Dimić and
14 Y. Ren, Y. Dienes, S. Hettel, M. Parvez, B. Hoge and
T. Baumgartner, Organometallics, 2009, 28, 734.
15 (a) P. Narayanan, H. M. Berman, F. Ramirez, J. F. Marecek,
Y. Chaw and V. A. V. Prasad, J. Am. Chem. Soc., 1977, 99,
3336; (b) M. Koenig, A. Klaebe, A. Munoz and R. Wolf,
J. Chem. Soc., Perkin Trans. 2, 1979, 40.
16 (a) T. S. Cameron and B. Dahlén, J. Chem. Soc., Perkin
Trans. 2, 1975, 1737; (b) J. A. Dobado, H. Martínez-García,
J. M. Molina and M. R. Sundberg, J. Am. Chem. Soc., 1998,
120, 8461.
17 (a) J. Lacour and A. Londez, J. Organomet. Chem., 2002, 643,
392; (b) J. Lacour, A. Londez, D. Tran, V. Desvergnes-Breuil,
S. Constant and G. Bernardinelli, Helv. Chim. Acta, 2002,
85, 1364; (c) J. Lacour, C. Ginglinger, C. Grivet and
G. Bernardinelli, Angew. Chem., Int. Ed., 1997, 36, 608.
18 J. C. Gilhula and A. T. Radosevich, Chem. Sci., 2019, 10,
7177.
D.
Vidović,
Inorg.
Chem.,
2015,
54,
4180;
(g) D. W. A. Stephan, Angew. Chem., Int. Ed., 2017, 56, 5984.
32 (a) J. J. Jennings, B. W. Wigman, B. M. Armstrong and
A. K. Franz, J. Org. Chem., 2019, 84, 15845; (b) J. M. Bayne
and D. W. Stephan, Chem. Soc. Rev., 2016, 45, 765;
(c) S. Postle, V. Podgorny and D. W. Stephan, Dalton Trans.,
2016, 45, 14651.
33 (a) A. Munoz, M. Gallagher, A. Klabe and R. Wolf, Tetrahedron
Lett., 1976, 9, 673; (b) F. Ramirez, M. Nawakowski and
J. F. Marecek, J. Am. Chem. Soc., 1977, 99, 4515.
19 (a) F. Ramirez, V. A. V. Prasad and J. F. Marecek, J. Am. 34 (a) K. E. Torraca, X. Huang, C. A. Parrish and
Chem. Soc., 1974, 96, 7269; (b) S. Trippett and
R. E. L. Waddling, Tetrahedron Lett., 1979, 20, 193.
20 R. R. Holmes, J. Am. Chem. Soc., 1975, 97, 5379.
S. L. Buchwald, J. Am. Chem. Soc., 2001, 123, 10770;
(b) M. Palucki, J. P. Wolfe and S. L. Buchwald, J. Am. Chem.
Soc., 1997, 119, 3395.
2252 | Dalton Trans., 2021, 50, 2243–2252
This journal is © The Royal Society of Chemistry 2021