Two Independent Orthogonal Stereomutations at a Single Asymmetric Center: A Narcissistic Couple

Article abstract of DOI:10.1002/chem.201604080

The energy barriers in our recently discovered Walden-type inversion of chlorophosphonium salts are similar to those for Cope rearrangements of caged cyclic hydrocarbons. Therefore, we have designed a molecular system that integrates the two processes, thereby producing the first embodiment of a chemical species that can undergo two entirely different and independent stereomutation mechanisms at the same nominal asymmetric center. Thus, the energy barrier to the rearrangement of 9-phenyl-9-phosphabarbaralane oxide, which is easily prepared by a new high-yielding synthesis, was found to be roughly 11 kcal mol^?1. This value is in contrast to the parent barbaralane (7.3 kcal mol^?1) but in good agreement with our computational results for the rearrangement barriers. Crucially, in the corresponding chlorophosphonium derivative, two stereomutations occur simultaneously: a fast Cope rearrangement (barrier ≈12 kcal mol^?1) and a slow Walden-type inversion of the phosphorus center (barrier ≈21 kcal mol^?1). The computational model also revealed a relationship between the Cope rearrangement barrier and the bridgehead distance. The phenomenon of two independent and geometrically orthogonal stereomutations at a single asymmetric center provided important general insights into reaction pathway bifurcation, microscopic reversibility, and dynamic stereochemistry. This first example of coexisting alternative mechanisms that involve covalent bonds may encourage the design of new types of dynamic molecular structures.

Full text of DOI:10.1002/chem.201604080

A Journal of
Accepted Article
Title: Two independent orthogonal stereomutations at a single
asymmetric centre: a narcissistic couple.
Authors: Malachi W. Gillick-Healy, Elizabeth V Jennings, Helge Muller-
Bunz, Yannick Ortin, Kirill Nikitin, and Declan G. Gilheany
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201604080
Link to VoR: http://dx.doi.org/10.1002/chem.201604080
Supported by

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
Two Independent Orthogonal Stereomutations at a Single
Asymmetric Centre: A Narcissistic Couple
Malachi W. Gillick-Healy, Elizabeth V. Jennings, Helge Müller-Bunz, Yannick Ortin, Kirill Nikitin*, and
Declan G. Gilheany*
Abstract: The energy barriers in our recently discovered Walden inversion of chlorophosphonium salts are similar to those in Cope
rearrangements of caged cyclic hydrocarbons. We have therefore designed a molecular system integrating the two processes, thereby
producing the first embodiment in a chemical species of two entirely different and independent stereomutation mechanisms at the same
nominal asymmetric centre. Thus, the energy barrier to the rearrangement of 9-phenyl-9-phosphabarbaralane oxide, easily prepared by a
new high-yielding synthesis, was found to be ca. 11 kcal/mol. This value is in contrast to the parent barbaralane (7.3 kcal/mol) but in good
agreement with our computational results for the rearrangement barriers. Crucially, in the corresponding P-chlorophosphonium derivative, two
stereomutations occur simultaneously: fast Cope rearrangement (barrier ca. 12 kcal/mol) and slow Walden-type inversion of the phosphorus
(barrier ca. 21 kcal/mol). The computational model also revealed a relationship between the Cope rearrangement barrier and the bridgehead
distance. The phenomenon of two independent and geometrically orthogonal stereomutations at a single asymmetric centre provides
important general insights into reaction pathway bifurcation, microscopic reversibility and dynamic stereochemistry. This first example of co-
existing alternative mechanisms involving covalent bonds may encourage the design of new types of dynamic molecular structures.
Introduction
caged structures, semibullvalene (5),[15] barbaralane (4)[16,17] and
dihydrobullvalene (3)[18]
have lower barriers to the Cope
truly
rearrangement encouraging the search for
a
The lifting and reinstatement of molecular symmetry has been a
key tool for the study of reactivity and mechanism.[1] Commonly,
fast interconversion of less symmetrical entities via a more
symmetrical transition state (TS) leads to dynamic equivalence
of molecular fragments. One particular type of dynamic
phenomena is narcissistic reactions[ 2 ] where reactant(s) and
product(s) are mirror images with respect to a fixed plane. In
general, such processes lead to inversion of stereochemistry.
For example, in a Walden-type inversion at a tetrahedral centre,
one antisymmetric coordinate changes sign during the reaction.
In the Cope rearrangement,[ 3 ] two such coordinates can be
identified.[4] Other dynamic processes (some of them occur even
in solid state) involving covalent bonds are known for many
bishomoaromatic[19,20] ground state neutral molecule. In recent
years, such compounds were investigated by Borden and
Schleyer and their co-workers.[21]
Our interest in fluxional processes arises from observation of
the dynamic behaviour of completely different systems
–
tetrahedral P-stereogenic phosphonium halides (XPS, X=Cl, Br).
These systems undergo Walden-type inversion of the reaction
centre[22a] as shown in Scheme 1c and are at the heart of our
dynamic kinetic resolution manifold for efficient stereoselective
preparation of organophosphorus compounds.[22b,c]
a)
fluxional organic[1] and organometallic[5,6]
compounds;
on
the other hand, conformational switching in synthetic molecules
can be controlled and used, for example, to mimic more complex
biomolecules.[7] In this report we present the first example of a
1
2
4
5
3
“narcissistic
couple”:
two
independent
narcissistic
slow
fast
stereomutation reactions occurring simultaneously within the
same molecule.
The classical Cope process is very slow in acyclic 1,5-
hexadiene (1) (activation barrier over 30 kcal/mol) for which
b)
c)
7.3 kcal/mol
early
calculations
indicated
a
possible
biradicaloid
R1
R1
P
intermediate.[8] Later, nucleus independent chemical shift[9] and
perturbation theory calculations showed that the TS has a
R3
R3
ca. 10 kcal/mol
Cl
Cl
significant aromatic contribution.[10]
The truly fast dynamic
P
Cl
R2
R2
behaviour of tricyclic systems 2-5 (Scheme 1), e.g. bullvalene
(2), was first postulated by W. von E. Doering.[11] However, it
was not until it was synthesized by Schröder that its fluxional
As predicted, it continually
rearranges with the barrier only about 13 kcal/mol.[13,14] Other
Cl
Scheme 1. a) Cope rearrangement: 1,5-hexadiene[10,
(1);
23
]
12
]
properties were confirmed.[
bullvalene[12,13,14
(
2
)
(
with
ca. 10 kcal/mol; barbaralane[16,24
) 5.2–5.5 kcal/mol. b) fast narcissistic
c) narcissistic P-configuration inversion in
chlorophosphonium species has barrier comparable to
a
barrier of ca. 13 kcal/mol;
]
dihydrobullvalene[18
3
)
(4) 7.3
]
]
kcal/mol; semibullvalene[24,25,26]
(5
rearrangement of
4.
3
.
[a]
Mr M.W. Gillick-Healy, Dr E.V. Jennings, Dr H. Muller-Bunz, Dr K.
Nikitin, Prof D.G. Gilheany
School of Chemistry, University College Dublin
E-mail: kirill.nikitin@ucd.ie; decaln.gilheany@ucd.ie
Despite the formation and cleavage of covalent P-X bonds,
the process is surprisingly fast (typical barriers ca. 10-11
kcal/mol) which, importantly, is comparable to dynamic Cope
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
rearrangement reaction barriers. This suggests that both
dynamic processes combined in the same molecule could be
observed simultaneously.
However, our key finding is that the salt 11Cl does indeed
combine the electrocyclic stereomutation with an alternative
stereomutation via a nucleophilic attack on the phosphonium
centre – making this species possibly the first known compound
undergoing two distinct, and independent, stereomutation
processes on the same chirality centre. We note that there are
known systems which are considered to involve two
Molecules incorporating aspects of dynamic pericyclic
chemistry with organophosphorus chemistry are known. One of
the earliest reported 9-phosphabarbaralanes was 12 (Scheme 2)
reported by Katz and co-workers[27] who found a low barrier of
7.2 kcal/mol for its Cope rearrangement, close to that of the
parent barbaralane 4. This appeared to suggest that the different
size and electronic parameters of phosphorus had little effect on
the Cope rearrangement. In contrast, a later study by Cowley
and co-workers[28] focused on the aminochlorophosphonium salt
(9) for which they found a notably higher barrier of 12.0 kcal/mol.
This stark difference implies either one measurement to be
incorrect or, unexpectedly, that the fairly similar phosphorus
entity caused a significant increase in the barrier for 9 – in this
case, we deemed it a finding deserving of further investigation.
The structures 9-12 also appealed to us because they are P-
stereogenic and fluxional. A change of P-configuration in 9, for
example, occurs during the Cope rearrangement due to
reorganisation of the carbon skeleton. We considered that an
analogue of 9 could also undergo a fast nucleophilic inversion of
the same phosphorus chirality centre, opening the possibility of
second simultaneous stereomutation[29] processes on the same
nominal chiral centre. We are adopting the term stereomutation
because neither of the terms epimerization or racemisation
seem to adequately describe this situation. The co-existence of
the two processes would appear to combine the phosphorus
chirality centre with a virtual chirality centre pertaining to the
whole molecular species. We therefore considered that the
fluxional dynamics of such phosphonium species would be of
general interest also to the stereochemistry community.
stereomutation pathways[30]
,
e.g. Berry pseudorotation and
threefold cyclic permutation in pentacoordinate complexes.[30c]
However, in all of those cases the alternative pathways are
closely related electronically and topologically.
Results and Discussion
Synthesis and structure of 9-phosphabarbaralanes
Scheme
2 outlines the previous approaches to the 9-
phosphabarbaralane skeleton. The original synthesis of oxide 12
by Katz and co-workers[27] involves oxidation of phosphine 10,
generated photochemically from bicyclic 8 – the product of the
reaction of PhPCl2 with dianion 7. The route used by Cowley and
co-workers is much more direct but, in their account, appeared
to depend for its success on the generation of a phosphenium
ion intermediate, stabilized by the amino-substituent on
phosphorus.[31,32] However, it was previously known that
aluminium trichloride assisted various cycloadditions of
alkenes[33] and dienes[34] to organochlorophosphines. We were
therefore gratified to find that treatment of cyclooctatetraene 6
with dichlorophenylphosphine in the presence of an equimolar
amount of aluminium trichloride gave one product, 11AlCl4 δP =
65 ppm, as evidenced by 31P NMR.
Ph
P
K metal
Ph-PCl2
2
7
8
1
5
hv
6
Cl
Ph
Me2N
P
P
Me2N-PCl2
AlCl3
Figure 1. Crystal structure of rac-9-phenyl-9-phospha-barbaralane oxide 12: a)
space-fill fragment of crystal packing showing ordered arrangement of the two
enantiomers; b) individual SP-12.
[O]
10
9 AlCl4
Ph
O
Ph
Cl
OH-
Ph-PCl2
P
P
The
reaction
of
dichlorophenylphosphine
with
AlCl3
X = AlCl4
cyclooctatetraene does not follow the classical McCormack
cycloaddition route characteristic of conjugated dienes.[35] This is
because a planar arrangement of the 1,3-diene array is not
available as both the D4h conformation of 6 and its valence
(COCl)2
X = Cl
11 X
12
this work
isomer, bicyclo[4.2.0]octa-2,4,7-triene, are unstable.[
36
, 37 ]
Scheme 2. Approaches to 9-phosphabarbarlanes:
9

AlCl4[28]; oxide 12
in this work: oxide 12, which can be converted to
Cl, was prepared via the stage of 11 AlCl4
Formation of caged cations 9 and 11 is more likely to be a result
of 1,5-cycloaddition to 6 similar to the sulfur analogue,[38] but
other possible mechanisms31 cannot be ruled out. Upon careful
alkaline hydrolysis and purification, the salt 11AlCl4 furnished in
high yield (96%) the phosphine oxide 12 whose identity was
confirmed by X-ray crystallography (Figure 1). The 1H NMR
clearly demonstrated that 12 continually undergoes Cope
rearrangement in solution (vide infra).
]
via phosphine 8[27
chloride 11
;


In this work we revisit and further explore the fluxional
dynamics of P-analogues of barbaralane, both experimentally
and computationally. Our detailed study shows that the barrier to
the Cope rearrangement in phospha-derivatives 11 and 12 is (as
in cation 9) certainly very different from the parent barbaralane 4.
The molecule of 12 shown in Figure 1 has well-defined
single bonds (C3–C4, 1.48 Å) adjacent to the cyclopropane ring,
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
and normal length double bonds (C4–C5, 1.34 Å). The structure
is very far from being a putative pericyclic transition state
whereby the bonds on the cyclopropane side and “open ring”
side of the molecule would be equalized. The open ring carbons
overlap at low temperatures and very wide peak separation is
probably why the much lower barrier, ca. 7 kcal/mol, was
incorrectly measured by Katz and co-workers.[27] Importantly,
we confirm that oxide 12, like the aminophosphonium salt 9, has
(C5–C7) are 2.42 Å apart which indicates no bonding interaction. a much higher Cope barrier than the parent barbaralane 4.
This is closely comparable to derivatized barbaralanes which
have open ring distance of 2.32–2.40 Å and a closed ring
cyclopropane distance of 1.56–1.67 Å.[39] The observation that
phosphorus-bridged C1-C6 are 2.75 Å apart, markedly farther
than in barbaralane at 2.49 Å, turns out to be quite significant as
will be discussed below. The distinctly asymmetric structure of
12 and the fact that at 100K the two enantiomers of 12 form
ordered racemic crystals (Figure 1a) reveal that the Cope
process is frozen – a situation quite different from that reported
for salt 9AlCl4 where the structure in the solid state is close to
mirror-symmetrical.[28] We believe that a fluxional process in 12
would result in a disordered crystal, which, after flash freezing,
should remain disordered (i.e. where each molecule adopts a
random enantiomeric form). Consistent with this, crystals of 12
examined at 298K displayed the same ordering. Therefore, this
Cope rearrangement, clearly visible in solution, is not occurring
in the solid state even at room temperature.
Because of these complex coalescence patterns in the
variable temperature NMR spectra of 12 we were anxious for
independent confirmation. We therefore ran a series of 2D
EXSY spectra of 12 at low temperature. The analysis of
exchange rates at -70 °C (see ESI for details) gave a slightly
higher barrier of 11.6 ± 0.2 kcal/mol (vide infra).
Stereomutation of chlorophosphonium species 11Cl by
Cope rearrangement
Turning now to the corresponding halophosphonium salts, we
have previously reported the facile conversion of phosphine
oxides to such salts under the action of oxalyl halides.[40] Indeed,
both chloro- or bromophosphonium species 11X could be
formed upon reaction of 12 with oxalyl halides. While this
reaction proceeds efficiently with oxalyl chloride, it works only
sluggishly with oxalyl bromide.
The 31P NMR of the chloride 11Cl in DCM, δP = 65.9 ppm,
shows that it is ionic and tetracoordinate, rather than
pentacoordinate; and this situation persisted even in benzene[41]
,
presumably due to the higher strain imposed by the carbon
backbone in the pentacoordinate form. The barrier to the Cope
rearrangement for 11Cl (11.9 ± 0.3 kcal/mol) was ascertained
by dynamic 1H NMR using the coalescences of the three sets of
peaks as shown in Figure 3. As in the case of 12, the pattern is
again heavily overlapped and complicated in this case by
significant peak drift. Therefore, an additional verification by
EXSY was again sought. At -60 °C (see ESI for details), the
barrier was 12.9 ± 0.3 kcal/mol, again 1 kcal/mol higher than the
coalescence result. This effect has previously been reported for
Cope rearrangement.[42] Such positive entropy of activation is
H3/H2
H1
H6
–70 °C
– 50 °C
–30 °C
0 °C
H5 H7
H2/H7
H1/H6
H3/H5
H4 H8
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
H (ppm)
Figure 2. Variable temperature 1H NMR of 12: signal of H2 (overlapped
with H3 at -70 °C) coalesces with H7 at ca. -15 °C indicating a Cope
barrier of 10.9
 0.3 kcal/mol (spectra stack angle 6 degrees).
H3
H5
H2
H1
H7
H6
70 °C
–
Stereomutation of oxide 12 by Cope rearrangement
–
20 °C
0 °C
1
Figure 2 shows the H NMR spectra of a 0.1 M solution of the
phosphine oxide 12 in DCM recorded at various temperatures
(see also ESI). It transpired that the coalescence pattern was
rather complex in that the peaks assigned to H2 and H3
(numbers as in Figure 1) directly overlapped at low temperature.
As the temperature is raised both H2 and H3 signals become
very broad and eventually coalesce with their corresponding H7
and H5. The significant shift differences at low temperature
between the pairs H2/H7 and H3/H5 and the existence of a
certain amount of signal drift with temperature both complicate
the discernment of the coalescence temperatures. On the other
hand, the signals of H1 and H6 (frequency separation ~500Hz)
coalesce in a more straightforward manner. Standard Eyring
analysis of all three coalescences (see Figure 2 and ESI for
more detail) gave a barrier of 10.9 ± 0.3 kcal/mol. The unusual
+
20 °C
H3/H5
H1/H6
H4 H8
H2/H7
6.4 6.1 5.8 5.5 5.2 4.9 4.6 4.3 4.0 3.7 3.4 3.1 2.8
H (ppm)
Figure 3. Variable temperature 1H NMR of 11
Cl: signal of H2 (3.3 ppm
at -70 °C) coalesces with H7 at ca. 5 °C indicating a barrier of 11.9
 0.3
kcal/mol (spectra stack angle 6 degrees).
likely due to the reduced bond orders in the TS. Notwithstanding
this, clearly the barrier for 11Cl found by both methods is
consistently higher (by ca.
1 kcal/mol) than that for the
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
corresponding parent oxide 12 at comparable temperatures.
Most significantly, both results are much higher than that of
barbaralane 4 (7.3 kcal/mol).
convenient situation arises from the relative rigidity of tricyclic
system 4 where options for variation are limited to the remaining
five bonds involved in rearrangement. These factors enable us
to successfully prospect for how the TS might be approached
during the progress of the reaction. Figure 4 shows how
stretching of the d1 distance in 4 in 0.02 Å increments is
accompanied by increase in energy reaching the bond-breaking
point, E = 7.6 kcal/mol, whereby both the stretched and the
forming bond distances are nearly equal, 2.09 Å and 2.07 Å,
respectively. The other, pinching pathway is characterized by
significantly different d1 and d2 at the point of highest energy.
The origin of the minor difference in Cope barrier observed
for the oxide 12 and chloride 11 may be related both to
differences in bond angles in phosphine oxides and
phosphonium salts[43] and to differences in the bond polarity (P-
Cl > P-Br > P=O) caused by P-heteroatom back-bonding
interaction[44] (as evidenced by proton NMR[22a]). Also, as will be
shown in the section below, the preferred orientation of the
phenyl group attached to the phosphorus may also affect the
height of the rearrangement barrier.
Computational modelling of the Cope rearrangement
Recent calculations[45] on Cope rearrangement indicated that 4
has a more homoaromatic TS when compared to 5 or 1,5-
methanosemibullvalene. Despite this, the rearrangement barrier
is lower in 5 and thus it can be concluded that other factors
control the barriers for instance, that allylic interactions and ring
strain also play a major role.[24]
To rationalize our experimental observations: i) the
rearrangement in the cyclic phosphine oxide 12 is much slower
than in the structurally analogous hydrocarbon 4; ii) the barrier is
appreciably higher for the ionic 11Cl than neutral molecule 12,
we turned to electronic structure calculations. As an
encouragement, the experimental barriers in the all-carbon
systems 2–5 have been calculated using the B3LYP/6-31G*
level of theory with good success.[24] Our approach to finding the
stationary point corresponding to the TS geometry involved
single-coordinate driving followed by TS optimization. Having
confirmed that this basic approach works well on b we used it to
find the barriers in 12 and the ionic 11•Cl in dichloromethane
solution (Table 1). As an example, let us consider the
rearrangement of 4 – the parent compound to 11 and 12. One
can select either of two pathways from the fully relaxed
geometry towards the TS: one is the bond-breaking (stretching)
of the closed cyclopropane bond and other bond-making
Figure 4. Coordinate driving in
4 viewed along the energy axis: a)
stretching of d1 leads to shortening of d2; b) rapid change of d2 after
reaching d1 = 2.09 Å, E = 7.6 kcal/mol; c) optimized TS (grey circle): d1 =
d2 = 2.07 Å, E = 7.5 kcal/mol.
This general situation whereby the stretching pathway is
lower energetically is observed for all structures (see ESI for
details). The optimized TS energies (e.g. point c in Figure 4)
correspond closely to the reported experimental barriers and the
trends observed for systems 2–5 are also well replicated by our
calculations (Table 1).
Having obtained promising computational results on all-
carbon systems, we applied our computations to
phosphabarbaralane derivatives, 11 and 12, and found that the
calculated energy values were in good agreement with
experimentally determined barriers in DCM (Table 1). Moreover,
the calculated ground state structure of 12 is a very close match
to the X-ray structure. As expected, the calculated C–P–O
angles in 12 (111.16°, 114.24°, 115.92°) and the respective C–
P–O bond angles in solid-state (110.71°, 114.71°, 113.22°) were
larger than the calculated C–P–Cl in 11Cl (109.27°, 107.47°,
110.28°). This is consistent with C–P–O bond angles being
wider than C–P–X in triphenylphosphine oxide and
Table 1. Calculated and experimental barriers (kcal/mol) for Cope
rearrangement in molecules 2–5, 11 and 12.
Bond-
Bond-
Optimized
TSa
Exp.
Barrierb
Species
breaking
making
2
3
13.4
9.6
15.2
10.0
8.8
13.6 c
9.6 c
13.3
10.0
4
7.6
7.5 c
7.3
5
5.5
6.8
5.5 c
5.2
11
12
11.7
10.3
13.3
11.2
11.5 d
10.3 d
11.9 (12.9)e
10.9 (11.6)e
halophosphonium salts (Figure 5a)[43]
.
[a] DFT B3LYP 6-31G*; [b] data for 2-5 summarised in ref. 24; [c] in vacuum;
[d] in DCM; [e] in DCM; VT-NMR result at a higher temperature (EXSY at a
lower temperature).
It can be concluded that the Cope rearrangement barrier in
the phosphorus-containing system 12 is significantly higher than
in the parent hydrocarbon 4. We propose that their different
geometries, as illustrated in Figure 5b, strongly affect the barrier
in the sequence 2 > 11 > 12 > 4. Thus the 9-phospha-analogues
11 and 12, having d(C1-C6) = 2.75–2.80 Å, are more similar to
(pinching) between the non-bonded carbons in the forming
cyclopropane ring by gradual (0.02 Å increments) variation of
the selected interatomic distance. Each can be seen as an
approximate reaction path resulting in roughly estimated TS
geometries and barriers (Figure 4). Interestingly, even this
bullvalene
2 (3.0 Å) in terms of rearrangement barrier.
Importantly, geometry can also explain the higher
rearrangement barrier in the ionic 11Cl having slightly longer
d(C1-C6) = 2.80 Å when compared to neutral species 12. The
straightforward single-coordinate technique gives
a
lucid
physical picture of the potential approaches to the TS. This
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
source of this elongation is a complex interplay of electronic and
steric differences between the oxide 12 and chlorophosphonium
salt 11Cl: the larger covalently bound Cl atom, as well as the
impact of the closely associated chloride anion; and the
conformation of the phenyl group affecting distance d(C1-C6) in
11Cl.
ca. 19.6 kcal/mol for 11Cl in good agreement with these 2D
EXSY experiments.
It has earlier been found that interallylic bonding present in
molecules like 2-5 has such a large effect on the barrier as to
render the aromaticity of the transition state insignificant.[24] The
systems 11 and 12 are interesting examples of fluxional
molecules where the electronic and the geometry effects, to a
certain extent, can be separated. Although the substituents at
the 9-position are not involved in the forming/breaking bonds
they still exert electronic and geometric influence over the whole
molecular entity. Unlike the all-carbon systems, where the
number of bridging carbons is different in each, our phospha-
analogues 11 and 12 differ only by a single atom attached to the
phosphorus and the effect of that atom can be explicitly
observed.
a
Figure 6. The 1H 2D EXSY spectrum showing the exchanging chemical
environments of the alkene protons H4 and H8 due to Walden inversion
of the phosphorus centre in 11Cl (CDCl3, 50 °C, 500 ms mixing time,
volume integration shown in green).
b
O
Cl
Ph
Ph
P
P
3.0
2
2.75
12
2.80
11
2.5
4
Implications of the Dual Stereomutation
Chlorophosphonium salt 11Cl, has been shown to have two
separate dynamic processes leading to stereomutation of its
phosphorus centre. These two reactions have very different
rates and mechanisms as shown in Scheme 3. This new
phenomenon of dual stereomutation via two different
mechanisms is more than a curiosity – it has several significant
implications. The first of these is for our understanding of
reaction pathway bifurcation.[46] Thus, on any one molecule,
during the slower Walden inversion process, many individual
Cope rearrangement events will have occurred, each of which
has proceeded by a distinct reaction pathway. Furthermore, in a
thought experiment, one can easily conceive of an analogue
molecule engineered in such a fashion that the two processes,
Cope rearrangement and Walden inversion, have activation
barriers less than kT apart (typically 12–13 kcal/mol). In such a
situation, certain intriguing questions arise that may require
adaptation of the application of standard TS theory: (i) to what
extent would the two mechanisms interact statistically and
energetically?; (ii) would a concerted occurrence of the two
processes (necessarily in nominally opposite stereochemically
directions) require more energy or less?; (iii) how would the
relative probability of each pathway depend on the nature of
excitation event (direction of impact; vibration mode) which, in
turn, may depend on the surrounding medium?[46c]
Figure 5. a) experimental bond angles in triphenylphosphine derivatives
(from ref. 43) showing wider C-P-C angles in halophosphonium salts; b)
calculated interatomic distances (Å) showing similarity of 11 and 12 to 2.
Stereomutation of chlorophosphonium 11Cl: Walden-type
inversion of P-centre
Recently, we have described
a degenerate nucleophilic
inversion of acyclic chlorophosphonium salts,[22a] which exhibited
barriers to Walden-type nucleophilic inversion in the range 10–
15 kcal/mol. Therefore, a second stereomutation process of the
system 11Cl is possible due to direct nucleophilic attack on
phosphorus. This proposed attack, if it took place, would now
dynamically interchange diastereotopic positions H4 and H8,
which are unaffected in the 1H VT-NMR observations of the
Cope rearrangement (Figure 3). However, little change was
observed in those signals on raising the temperature as far as
50 °C, signifying a significantly higher barrier than for the acyclic
phosphonium structures.
Gratifyingly, in the 2D EXSY experiments the low rate of
exchange interconversion at 50 °C could be detected by the
appearance of in-phase cross peaks between positions H4 and
H8 (Figure 6). This unequivocally confirmed that the salt 11Cl
also undergoes stereomutation of the P-centre by nucleophilic
attack at phosphorus with a barrier of 21 ± 0.5 kcal/mol. A
computational examination of the nucleophilic process in 11Cl
was carried out by using an approach recently adopted by us for
estimating such barriers.[22a] This method, performed with ground
state tetrahedral structure and pentacoordinate structure as
approximation of the TS (see ESI for details), gave a barrier of
Another aspect of this dual stereomutation process is related
to the application of the principle of microscopic reversibility.[47]
Certain inaccurate formulations of this principle[48] could suggest
its violation by the dynamic system 11Cl. On the other hand, its
correct interpretation[47c] for this case could serve as an easily
recognizable example of the validity of the principle. Adapting
Tolman’s original statement[47a] “the number [of molecules per
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
unit time] leaving by any one particular path shall be equal to the
number arriving by the reverse of that particular path” it
becomes clear that different mechanistic pathways can lead
from (S)-11Cl to (R)-11Cl and vice versa. It is also easy to see
that over the ensemble of molecules the average rates for every
mechanistic path could be equal to their reverse without violating
the principle. Therefore, the dual stereomutation process in 11
and its analogues (vide supra) is a new illustration of the
principle.
Finally, from a stereochemistry point of view, the dual
stereomutation has yet another significant implication:
interpretation of epimerization and racemisation processes. An
epimerization typically occurs between diastereomers by
inverting one chirality centre (or another element) and, in that
sense, racemisation is a particular type of epimerization. Let us
consider now the two different mechanisms operating in system
11Cl: the respective transition states are mirror-symmetrical,
but the orientations of the mirrors are at ca. 90 degrees to one
another i.e. orthogonal (Scheme 3). This symmetry analysis
appears to suggest that there are two chirality centres in 11: an
apparent phosphorus chirality centre and another implicit
chirality centre borne of the barbaralane cage. Accordingly, the
dual orthogonal independent stereomutation of the two centres
that we have described could be termed degenerate
epimerization – indeed, should the carbon atoms of 11 be
uniquely labelled, we would be looking at interconversion of four
diastereomers rather than two enantiomers of 11.
Scheme 3. Phosphonium system 11

Cl undergoing degenerate
epimerization (see text): nucleophilic attack via a pentacoordinate
transition state (top) and Cope rearrangement of the carboskeleton
(bottom) showing both putative transition states which are mirror-
symmetrical with orthogonal orientation of the respective mirror planes
m1 and m2.
Conclusions
In this work two entirely different and independent
stereomutation mechanisms at the same nominal asymmetric
centre were identified for the first time. The operation and
dynamic co-existence of two narcissistic reactions has general
implications for our understanding of reaction pathway
bifurcation, microscopic reversibility and chirality.
Our study, therefore, proves that interconversion of
molecular entities by entirely different covalent chemistry
mechanisms is
a reality. This is highly relevant to our
understanding of the functioning of natural and synthetic
Brownian machines[49] and motors[50], whereby precise control
over alternative mechanistic pathways is exploited to achieve
the required function under equilibrium or non-equilibrium
conditions. Such systems, typically, incorporate controlled “gate”
moieties and their components, designed to slide over high or
low steric barriers, accomplish the required dynamic function
without breaking covalent bonds. In this report, we have
presented the first example in which two alternative reaction
pathways involving rearrangement of covalent bonds co-exist in
a single molecule. This finding opens perspectives of designing
new types of fluxional and/or directional molecular machines.
We have described a straightforward preparation of a 9-
phospha-analogue of barbaralane, the oxide 12, in a single
synthetic step via the stage of phosphonium derivative 11•AlCl4.
The Cope rearrangement barrier in 12 was unambiguously
determined by VT and EXSY techniques to be at least ca. 11
kcal/mol – significantly higher than in the parent barbaralane,
thereby resolving the previous contradicting data on such 9-
phospha-analogues in the literature. Based on DFT calculations,
we propose that the barrier to a Cope rearrangement increases
with the length of the bridgehead connection as drawn in Figure
5. The variation of bridgehead distances between 11 and 12 is
also responsible for slight variation in their Cope barriers (ca. 1
kcal/mol). Moreover, the barrier is further increased in the solid
state, possibly due to cooperative interactions in the crystal
lattice, as revealed by X-ray diffractometry of 12.
The corresponding chlorophosphonium salt 11Cl undergoes
two simultaneous dynamic processes: Cope rearrangement and
nucleophilic Walden-type inversion of the phosphorus chirality
centre, the barrier for the inversion process being much higher
than for the Cope rearrangement. Therefore, two classes of
fluxional processes – degenerate electrocyclic rearrangement
and degenerate nucleophilic attack on a tetrahedral centre –
have been found to have an intersection in the dynamic
behaviour of 11Cl.
Experimental Section
A detailed description of the experimental procedures, characterization of
compounds, X-ray, VT-NMR, EXSY data, and DFT calculation details
can be found in the electronic supporting information (ESI) and is
available free-of-charge from the publisher.
Acknowledgements
We are very grateful to UCD Centre for Synthesis and Chemical
Biology (CSCB) and the UCD School of Chemistry for access to
their extensive analysis facilities. Helpful discussion with Prof K.
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
Lammertsma and Dr A.W. Ehlers of VU Amsterdam; Prof G.
Lloyd-Jones of Edinburgh University; and Dr E. McGarrigle of
University College Dublin is highly appreciated. This work was
supported by Science Foundation Ireland through Principal
Investigator Grant 09/IN.1/B2627 and by the Irish Research
[18] G. Schroeder, J. F. Oth, Angew. Chem. 1967, 79, 458–467; Angew.
Chem. Int. Ed. 1967, 6, 414-423.
[19] S. Winstein, J. Am. Chem. Soc. 1959, 82, 6524-6525.
[20] R. V. Williams, Chem. Rev. 2001, 101, 1185–1204.
[21] a) E.C. Brown, D.K. Henze, W.T. Borden, J. Am. Chem. Soc. 2002, 124,
14977–14982; b) H. Wu, H. Jiao, Z. Wang, P. v. R. Schleyer, J. Am.
Chem. Soc. 2003, 125, 10524–10525.
Council through
a
Government of Ireland Postgraduate
Scholarship to MGH (GOIPG/2013/1081).
[22] a) E. V. Jennings, K. Nikitin, Y. Ortin, D. G. Gilheany, J. Am. Chem.
Soc. 2014, 136, 16217–16226; b) K. Nikitin, K. Rajendran, H. Müller-
Bunz, D. G. Gilheany, Angew. Chem. 2014, 126, 1937–1940; Angew.
Chem. Int. Ed. 2014, 126, 1906-09; c) K. Rajendran, K. Nikitin, D. G.
Gilheany, J. Am. Chem. Soc. 2015, 137, 9375–9381.
Keywords: microscopic reversibility
electrocyclic reactions • nucleophilic substitution • cycloaddition
•
stereomutation
•
[23] R. Hoffmann, R. B. Woodward, J. Am. Chem. Soc. 1965, 87, 4389-
4390.
[1]
[2]
a) M. McGlinchey, Symmetry 2014, 6, 622-654; b) T. Bauert, L. Merz,
D.Bandera, M. Parschau, J.S. Siegel, K.H. Ernst, J. Am. Chem. Soc.
2009, 131, 3460-3461.
[24] D. A. Hrovat, E. C. Brown, R. V. Williams, H. Quast, W. T. Borden, J.
Org. Chem. 2005, 70, 2627-2632.
Narcissistic reactions: reactions where reactant(s) and product(s) are
mirror images with respect to a fixed plane, with at least one non-zero
antisymmetric coordinate which changes sign during the reaction – see
ref 4.
[25] A. K. Cheng, F. A. L. Anet, J. Mioduski, J. Meinwald, J. Am. Chem. Soc.
1974, 96, 2887-2891.
[26] D. Moskau, R. Aydin, W. Leber, H. Guenther, H. Quast, H. D. Martin, K.
Hassenrueck, K. Grohmann, L. S. Miller, Chem. Ber. 1989, 122, 925-
931.
[3]
[4]
a) A. C. Cope, E. M. Hardy, J. Am. Chem. Soc. 1940, 62, 441; b) R. V.
Williams, Eur. J. Org. Chem. 2001, 227-235.
[27] (a) T. J. Katz, J. C. Carnahan, G. M. Clarke, N. Acton, J. Am. Chem.
Soc. 1970, 92, 734-735. (b) E. W. Turnblom, T. J. Katz, J. Am. Chem.
Soc. 1973, 95, 4292-4311.
a) L. Salem, J. Durup, G. Bergeron, D. Cazes, X. Chapuisat, H. Kagan,
J. Am. Chem. Soc. 1970, 92, 4472-4474; b) L. Salem, Acc. Chem. Res.
1971, 4, 322-328.
[28] S. A. Weissman, S. G. Baxter, A. M. Arif, A. H. Cowley, J. Am. Chem.
Soc. 1986, 108, 529-531.
[5]
[6]
(a) A. J. Campbell, C. A. Fyfe, J. Am. Chem. Soc. 1972, 94, 8387-8391;
(b) C. A. Fyfe, A. J. Campbell, E. Jr. Maslowsky, J. Chem. Soc.
D: Chem. Commun. 1971, 1032-1034; (c) C. A. Fyfe, B. A. Dunell, J.
Ripmeester, Can. J. Chem. 1971, 49, 3332-3335; d) F. A. Cotton, Acc.
Chem. Res. 1968, 1, 257-265.
[29] see: IUPAC Compendium of Chemical Terminology, 2nd ed. Compiled
by A. D. McNaught, A. Wilkinson. 1997,
Blackwell Scientific
Publications, Oxford.
[30] a) A.R. Bassindale, J. C.-Y. Lau, P.G. Taylor, J. Organomet. Chem.
1995, 499, 137-141; b) J. Kong, P. v. R. Schleyer, H.S. Rzepa, J. Org.
Chem. 2010, 75, 5164–5169; c) E. P. A. Couzijn, J. C. Slootweg, A. W.
Ehlers, K. Lammertsma, J. Am. Chem. Soc. 2010, 132, 18127–18140.
[31] a) C. K. SooHoo, S. G. Baxter, J. Am. Chem. Soc. 1983, 105, 7443-
7444; b) A. H. Cowley, R. A. Kemp, N. C. Lasch, C. A. Stewart, B. R.
Wittlesey, T. C. Wright, Inorg. Chem. 1986, 25, 740-749.
a) R.T. Edidin, J.R. Norton, K. Mislow, Organometallics 1982, 1, 561–
562; b) M. Mlekuz, P. Bougeard, B. G. Sayer, M.J. McGlinchey, S.
Peng, A. Marinetti, J.-Y. Saillard, J. Ben Naceur, B. Mentzen, G.
Jaouen, Organometallics 1985, 4, 1123–1130; c) A. Borchardt, A.
Fichicello, K. V. Kilway, K.K. Baldridge, J. S. Siegel, J. Am. Chem. Soc.
1992, 114, 1921–1923; d) M. Stradiotto, M.J. McGlinchey, Coord.
Chem. Rev. 2001, 311–378.
[32] a) A. H. Cowley, R. A. Kemp, Chem. Rev. 1985, 85, 367-382; b) C.
Hering, A. Schulz, A. Villinger, Inorg. Chem. 2013, 52, 5214−5225; c) M.
H. Holthausen, K-O. Feldman, S. Schulz, A. Hepp, J. J. Weigand, Inorg.
Chem. 2012, 51, 3374-3387.
[7]
[8]
M. De Poli, W. Zawodny, O. Quinonero, M. Lorch, S.J Webb, J.
Clayden, Science 2016, 352, 575-580.
a) M. J. S. Dewar, Angew. Chem. 1971, 83, 859–875; Angew. Chem.
Int. Ed. 1971, 10, 761-776; b) M. J. S. Dewar, E. F. Healy, Chem. Phys.
Lett. 1987, 141, 521-524.
[33] a) C. Symmes, L. D. Quin, J. Org. Chem. 1978, 43, 1250-1253; b) A.
Marinetti, F. X. Buzin, L. Ricard, J. Org. Chem. 1997, 62, 297-301.
[34] a) Y. Kashman, Y. Menachem, E. Benary, Tetrahedron 1973, 29, 4279-
4283; b) A. Rudi, Y. Kashman, Tetrahedron 1981, 37, 4269-4275.
[35] a) W. B. McCormack, US 2663736, 2663737, 1953; Chem. Abstr. 1955,
49, 7601; b) L. D. Quin, The Heterocyclic Chemistry of Phosphorus,
Chapter 2. Wiley-Interscience: New York, 1981.
[9]
a) Z. Chen, C. S. Wannere, C. Corminboeuf, R. Puchta, P. v. R.
Schleyer, Chem. Rev. 2005, 105, 3842-3888. b) P. v. R. Schleyer,
Chem. Soc. Rev. 2014, 43, 4909-4921.
[10] a) D. A. Hrovat, K. Morokuma, W.T. Borden, J. Am. Chem. Soc. 1994,
116, 1072-1076; (b) B. R. Beno, D. A. Hrovat, H. Lange, H.-Y. Yoo, K.
N. Houk, W. T. Borden, J. Am. Chem. Soc. 1999, 121, 10529-10537.
[36] a) D. A. Hrovat, W. T. Borden, J. Am. Chem. Soc. 1992, 114, 5879-
5881; b) X. Bao, D. A. Hrovat, W. T. Borden, J. Org. Chem. 2012, 956-
965.
[11]
W. von E. Doering, W. R. Roth, Angew. Chem. 1963, 75, 27–35;
Angew. Chem. Int. Ed. 1963, 2, 115-122.
[12] a) G. Schroeder, Angew. Chem. 1963, 87, 693–707; Angew. Chem. Int.
Ed., 1963, 2, 481-482; b) G. Schroeder, Chem. Ber., 1964, 97, 3140-
3149; c) J.F.M. Oth, K. Müllen, J.-M. Gilles, G. Schröder, Helv. Chim.
Acta 1974, 57, 1415–1433.
[37] a) L. A. Paquette, Acc. Chem. Res. 1993, 26, 57-62; b) F.-G. Klarner,
Angew. Chem. 2001, 113, 4099–4103; Angew. Chem. Int. Ed.
2001, 40, 3977-3981; c) T. Nishinaga, T. Ohmae, M. Iyoda, Symmetry
2010, 2, 76-97.
[13] A. Allerhand, H. S. Gutowsky, J. Am. Chem. Soc. 1965, 87, 4092- 4096.
[14] a) T. Koritsanszky, J. Buschmann, P. Luger, J. Phys. Chem. 1996, 100,
10547–10553; b) S. M. Johnson, J. S. McKechnie, B. T. S. Lin, I. C.
Paul, J. Am. Chem. Soc. 1967, 89, 7123; c) R. Poupko, H. E.
Zimmermann, Z. Luz, J. Am. Chem. Soc. 1984, 106, 5391-5394.
[38] L. A. Paquette, U. Jacobsson, M. Oku, J. Chem. Soc. Chem. Commun.
1975, 115-116.
[39] H. Quast, C. Becker, M. Witzel, E.-M. Peters, K. Peters, H.G. von
Schnering, Liebigs Ann. 1996, 985-997.
[40] a) K. V. Rajendran, D. G. Gilheany, Chem. Commun. 2012, 48, 817-
819; b) K. V; Rajendran, D. G. Gilheany, Chem. Commun. 2012, 48,
10040-10042; c) P. A; Byrne, K. V Rajendran, J. Muldoon, D. G.
Gilheany, Org. Biomol. Chem. 2012, 10, 3531-3537.
[15]
a) H. E. Zimmerman, R. M. Paufler, J. Am. Chem. Soc. 1960, 82,
1514-1515. b) H. E. Zimmerman, G. E. Grunwald, J. Am. Chem. Soc.
1966, 88, 183-184.
[16] W. v. E. Doering, B. M. Ferrier, E. T. Fossel, J. H. Hartenstein, M. Jr.
Jones, G. W. Klumpp, R. M. Rubin, M. Saunders, Tetrahedron 1967, 13,
3943-63
[41] a) S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard, J. M. Sheffield, G. M.
Thompson, J. Chem. Soc., Dalton Trans., 1997, 4823-4827; (b) S. M.
Godfrey, C. A. McAuliffe, I. Mushtaq, R. G. Pritchard, J. M. Sheffield, J.
Chem. Soc., Dalton Trans., 1998, 3815-3818; (c) S. M. Godfrey, C. A.
McAuliffe, R. G. Pritchard, J. M. Sheffield, Chem. Commun., 1996,
2521-2522; d) N. C. Gonnella, C. Busacca, S. Campbell, M. Eriksson,
[17] H. Gunther, J. Runsink, H. Schmickler, P. Schmitt, J. Org. Chem. 1985,
50, 289-293.
This article is protected by copyright. All rights reserved.

10.1002/chem.201604080
Chemistry - A European Journal
FULL PAPER
N. Grinberg, T. Bartholomeyzik, S. Ma, D. L. Norwood, Magn. Reson.
Chem., 2009, 47, 461-464.
[42] For an example of ΔG‡ drift with temperature see H. Gunther, J. Ulmen,
Tetrahedron, 1974, 30, 3871-3877.
[43] (a) K. A. Al-Farhan, J. Crystallogr. Spectrosc. Res., 1992, 22, 687-689;
(b) K. Nikitin, D. G. Gilheany, H. Müller-Bunz, Chem. Commun., 2013,
49, 1434-1436.
[44] D. G. Gilheany, Chem. Rev. 1994, 94, 1339-1374.
[45] Y. Ichikawa, S. Sakai, J. Phys. Org. Chem. 2012, 25, 409–419.
[46] a) B.K. Carpenter, J.N. Harvey, D.R. Glowacki, Phys. Chem. Chem.
Phys., 2015, 17, 8372-81; b) Z. C. Kramer, B.K. Carpenter, G.S. Ezra,
S. Wiggins, J. Phys. Chem. A, 2015, 119, 6611−6630; c) B.K.
Carpenter, J.N. Harvey, A. J. Orr-Ewing, J. Am. Chem. Soc., 2016, 138,
4695–4705.
[47] a) R. C. Tolman, Proc. Nat. Acad. Sci. 1925, 436-439; b) R. L. Burwell,
R.G. Pearson, J. Phys. Chem. 1966, 300-302; c) S. Chandresekhar,
Research on Chemical Intermediates, 1992, 173–209; d) D. Blackmond,
Angew. Chem. 2009, 121, 2686–2693; Angew. Chem. Int. Ed. 2009, 48,
2648–2654.
[48] For example: “In a reversible reaction, the mechanism in one direction
is exactly the reverse of the mechanism in the other direction”. IUPAC
Compendium of Chemical Terminology, 2nd ed. (the “Gold Book”)
Compiled by A. D. McNaught, A. Wilkinson. Blackwell Scientific Pub,
Oxford (1997). This and other formulations are discussed extensively in
refs 47b and c.
[49] a) E.R.; Kay, Leigh, D.A. Pure Appl. Chem. 2008, 80, 17–29; b) S.
Erbas-Cakmak, D.A. Leigh, C.T. McTernan, A.L. Nussbaumer, Chem.
Rev., 2015, 115, 10081−10206.
[50] a) R.A. van Delden, M.K.J. ter Wiel, V. Pollard, J. Vicario, N. Koumura,
B.L. Feringa, J. Am. Chem. Soc., 2000, 122, 12005-12006; b) J.C.M.
Kistemaker, P. Stacko, J. Visser, B. L. Feringa, Nature Chem. 2015,
890-896.
This article is protected by copyright. All rights reserved.