Dibridgehead diphosphines that turn themselves inside out

Article abstract of DOI:10.1002/anie.201100893

Molecular contortionists can lurk in unexpected places. The title compounds undergo equilibria that appear to involve straightforward pyramidal inversions at the phosphorus atoms, but in reality the stereoisomers interconvert by turning themselves inside

Full text of DOI:10.1002/anie.201100893

DOI: 10.1002/anie.201100893
Macrobicyclic Phosphanes
Dibridgehead Diphosphines that Turn Themselves Inside Out**
Michael Stollenz, Michał Barbasiewicz, Agnieszka J. Nawara-Hultzsch, Tobias Fiedler,
Ryan M. Laddusaw, Nattamai Bhuvanesh, and John A. Gladysz*
Molecules and macromolecules that undergo topologically
complex dynamic processes—such as knot-forming[1] and
multistep folding[2] sequences—have attracted considerable
attention from numerous standpoints. However, there is
much less awareness that certain types of molecules, including
but not limited to macrocyclic bicyclic (macrobicyclic) com-
pounds, are able to turn themselves inside out.[3] This has been
termed “homeomorphic isomerization”,[4] even though the
process can be degenerate. At the time of a 1996 review,[3]
only four (degenerate) cases had been rigorously established
by spectroscopic means.[5] Although we are unaware of
additional confirmed examples since, such equilibria have
been invoked to rationalize the NMR spectroscopic proper-
ties of other macrobicyclic compounds.[6]
Herein we demonstrate this type of dynamic behavior—in
both degenerate and nondegenerate manifestations—with
three stereoisomers of a macrobicyclic aliphatic dibridgehead
diphosphine (in,in, out,out, and in,out, referring to the
orientations of the lone pairs of electrons on the phosphorus
atoms[3]). In the nondegenerate case, the lone pairs of
electrons are alternately directed in a convergent manner
towards an interior domain (in,in) or directed externally
(out,out). Thus, such processes can potentially mediate the
sequesterization, transport, and delivery of Lewis acid guests.
Analogous dynamic behavior has recently been proposed for
a hexaaryl dibridgehead diphosphine[6b] and other types of
aromatic dibridgehead diphosphorus compounds.[6a,c]
Our story begins with the platinum dichloride complex
trans-1 (Scheme 1), in which three (CH2)14 chains connect the
trans-arranged phosphorus atoms.[7] This complex exemplifies
a class of compounds termed “gyroscope like”, because of the
rapid rotation of the caged MLn moieties in suitably sized
systems on the NMR time scale, and their structural
Scheme 1. Synthesis and complexation of the dibridgehead diphos-
phine 2.
similarities to common toy gyroscopes.[7–9] Treatment of
ꢀ
trans-1 with an excess of the nucleophiles NaC CH, Li-
ꢀ
ꢀ
C CPh, or KC N afforded the macrobicyclic dibridgehead
diphosphine 2 as an analytically pure, moderately air-sensitive
white powder in 66–91% yield. The dianionic platinum
tetrakis(acetylide) complex 3 could also be isolated (35%)
ꢀ
when LiC CPh was employed.
The three in/out stereoisomers of 2 are depicted in
Scheme 1 (middle). In the PtCl2 adduct 1, both lone pairs of
electrons of the dibridgehead diphosphine ligand are directed
inwards. Treatment of 2 with PtCl2 in C6D6 regenerated 1,
which constitutes an overall retention of configuration at the
phosphorus atoms. For this reason, it was originally thought
that only in,in-2 was produced. For small bicycles, out,out
isomers are energetically much more favorable, but computa-
tional data for analogous hydrocarbons indicate that in,out
isomers become most stable at medium ring sizes, and that
in,in isomers become most stable at larger ring sizes.[10] DFT
calculations (see the Supporting Information) indicated (as
did preliminary molecular mechanics calculations) that in,in-2
was considerably more stable than out,out-2 (6.98 kcalmolÀ1),
and somewhat more stable than in,out-2 (1.59 kcalmolÀ1).
The longest aliphatic dibridgehead diphosphine reported
previously features one (CH2)3 and two (CH2)4 chains,
which are much shorter than the (CH2)14 linkers in 2.[11]
The isomers of 2 can be regarded as configurational
diastereomers that are interrelated by pyramidal inversions at
[*] Dr. M. Stollenz, Dipl.-Chem. T. Fiedler, R. M. Laddusaw,
Dr. N. Bhuvanesh, Prof. Dr. J. A. Gladysz
Department of Chemistry, Texas A&M University
PO Box 30012, College Station, TX 77842-3012 (USA)
Fax: (+1)979-845-5629
E-mail: gladysz@mail.chem.tamu.edu
Homepage: http://www.chem.tamu.edu/rgroup/gladysz/
Dr. M. Barbasiewicz, Dr. A. J. Nawara-Hultzsch
Institut fꢀr Organische Chemie and
Interdisciplinary Center for Molecular Materials
Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Henkestrasse 42, 91054 Erlangen (Germany)
[**] We thank the US National Science Foundation (CHE-0719267), the
Humboldt Foundation (Fellowship to M.B.), and Johnson Matthey
(platinum loans) for support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100893.
Angew. Chem. Int. Ed. 2011, 50, 6647 –6651
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6647

Communications
the phosphorus atoms. In classic studies, Baechler and Mislow
species, assigned as in,out-2, slowly formed. After 40 h, a 51:49
equilibrium mixture (in,in/out,out versus in,out) was present.
The data gave a DG423K value of 0.034 kcalmolÀ1 and a
established that such inversions in simple trialkylphosphines
require 29–36 kcalmolÀ1,[12] which corresponds to a very slow
process at room temperature. This is consistent with the many
chiral PRR’R’’ species that can be generated in enantiopure
form.[13] Thus, we did not expect to encounter any facile
configurational equilibria involving 2.
DG°
value of 33.8 kcalmolÀ1. The latter is in good
423K
agreement with the results of Baechler and Mislow.
As shown in Scheme 2, this sample was treated with
excess Me2S·BH3. Chromatographic separation on silica gel
gave the bis(borane) adducts (in,in/out,out)-2·2BH3 as a
slowly solidifying oil and in,out-2·2BH3 as a colorless viscous
liquid in yields of 43 and 42%, respectively. In a second route,
31P NMR spectra of 2 were recorded in various solvents at
low-temperature. After some time, we became convinced that
a small signal reproducibly appeared in [D8]toluene (area
ratio 97:3; Figure 1).[14]
A
31P EXSY experiment established
the phosphine borane H3B·P((CH2)6CH CH2)3[18] was treated
=
that the species responsible for the two signals are in
equilibrium (DG193K = 1.33 kcalmolÀ1). Line-broadening
analyses indicated DG°193K values of 11.5 and 10.4 kcalmolÀ1
(major to minor and minor to major isomers, respectively),
which are much lower than those for pyramidal inversions of
trialkylphosphines.
with the Grubbs catalyst. Such an alkene metathesis would be
expected to yield much oligomer, polymer, and other by-
products. However, subsequent hydrogenation (Wilkinson
catalyst) and column chromatography afforded (in,in/
out,out)-2·2BH3 and in,out-2·2BH3 in yields of 2 and 4%,
respectively. Although these are poor yields, the route is not
stoichiometric in platinum.
The bis(borane) adduct in,out-2·2BH3 could be depro-
tected in neat pyrrolidine at reflux. Workup afforded in,out-2
in 56% yield as an analytically pure, moderately air-stable,
colorless oil. Importantly, in,out-2 exhibited a single signal in
the 31P NMR spectrum, although from symmetry consider-
ations two would have been expected. This implies that a
degenerate in,out/out,in homeomorphic isomerization is rapid
on the NMR time scale (Scheme 2, top right). Accordingly, a
solution of in,out-2 in CH2Cl2 was cooled, and 31P NMR
spectra were recorded. As shown in Figure 2, two signals of
equal intensities separated (Tc = À738C). The data yielded a
DG°200K value of 8.5 kcalmolÀ1.
The adducts 2·2BH3 illustrate additional nuances of these
topological equilibria. First, the presence of progressively
larger Lewis acids on the lone pairs of electrons of the
phosphorus atoms should eventually render out,out isomers
more stable than in,in isomers. Indeed, when samples of
(in,in/out,out)-2·2BH3 were crystallized from hexanes or
methylcyclohexane, out,out-2·2BH3 was obtained, but
always with a guest molecule in the cavity. A representative
structure, which features methylcyclopentane, a known com-
ponent of hexanes, is given in Figure 3. The phosphorus–
phosphorus distance expands to 13.22 ꢀ, from 4.61 ꢀ in trans-
1, thereby underscoring the conformational flexibility of the
diphosphine.
Similarly, progressively larger phosphorus-bound Lewis
acids should eventually render an in/out isomer untenable. In
this case, alternative conformations featuring out/out phos-
phorus substituents and crossed (CH2)14 chains may become
preferred. The homeomorphic isomerization of in,out-2
presumably involves crossed-chain species, illustrated by IV
and VI in Scheme 3, which summarizes the principal equi-
libria in our system.[19] There remain many subtle unresolved
issues, such as whether the equilibration of in,in-2 and out,out-
2 involves a concerted disrotatory bridgehead motion or an
intermediate with crossed chains generated by a “half turn”.
Experiments to address such points are in progress.
Figure 1. 31P{1H} NMR spectrum of 2 in [D8]toluene at À808C (the
arrows denote exchanging species; * and ** denote unknown and
known impurities, respectively).
With the help of models, we then realized that in,in-2 and
out,out-2 can be interconverted by a purely conformational
process not involving phosphorus inversion or a “homeomor-
phic isomerization”. This process corresponds to pulling one
of the (CH2)14 chains connecting the phosphorus atoms
through the macrocycle defined by the other two chains
(Scheme 2, top left). The minor signal would correspond to
out,out-2. This constitutes the first time such a process has
been established for an in,in/out,out pair of isomers.[3] In this
context, it is relevant that cis-1 (Scheme 1) can be independ-
ently prepared.[15] The idealized 908 angle between the lone
pairs of electrons on the phosphorus atoms demonstrates the
inherent flexibility of 2.[16]
This mechanistic model would be strengthened by a
sample of in,out-2. If in,out-2 were stable with respect to the
other two isomers at room temperature, a pyramidal inversion
sequence with anomalously low barriers would be definitively
excluded.[17]
Various extensions of the above concepts are shown in
Scheme 4. First, similar inside-out conformational equilibria
should be possible with structures of the types VII and VIII.
The former could operate in the case of 1,3,5-cyclophanes
In the most direct approach, 2 was heated in mesitylene at
1508C and monitored by 31P{1H} NMR spectroscopy. A new
6648 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6647 –6651

Scheme 2. Isomerization of 2 and syntheses of BH3 adducts.
Figure 3. Crystal structure of out,out-2·2BH3·(C5H9CH3). The carbon
atoms C1–C14 and C15–C18 exhibited disorder, which could be
modeled; the dominant conformation is depicted.
with sufficient bridge lengths.[20] Type VIII structures have
four bridges, with isomerization involving pulling two of them
through the macrocycle defined by the remaining two. Many
dibridgehead diamines exist, but in/out isomers preferentially
interconvert through pyramidal inversion at the nitrogen
atom, because of the much lower energy barriers. The
corresponding protonated diammonium salts isomerize
through deprotonation/inversion sequences.[3,4] However, the
complexation of metal ions by cryptands may involve initial
Figure 2. Low-temperature 31P{1H} NMR spectra of in,out-2.
Angew. Chem. Int. Ed. 2011, 50, 6647 –6651
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6649

Communications
[1] J. Guo, P. C. Mayers, G. A. Breault, C. A. Hunter, Nat. Chem.
2010, 2, 218 – 222, and references therein.
[2] D. J. Hill, M. J. Mio, R. B. Prince, T. S. Hughes, J. S. Moore,
Chem. Rev. 2001, 101, 3893 – 4011.
[3] R. W. Alder, S. P. East, Chem. Rev. 1996, 96, 2097 – 2111.
[4] C. H. Park, H. E. Simmons, J. Am. Chem. Soc. 1968, 90, 2429 –
2431.
[5] a) A. H. Haines, P. Karntiang, J. Chem. Soc. Perkin Trans. 1 1979,
2577 – 2587; b) R. S. Wareham, J. D. Kilburn, D. L. Turner, N. H.
Rees, D. S. Holmes, Angew. Chem. 1995, 107, 2902 – 2904;
Angew. Chem. Int. Ed. Engl. 1995, 34, 2660 – 2662; c) M.
Saunders, N. Krause, J. Am. Chem. Soc. 1990, 112, 1791 – 1795;
d) R. W. Alder, E. Heilbronner, E. Honegger, A. B. McEwan,
R. E. Moss, E. Olefirowicz, P. A. Petillo, R. B. Sessions, G. R.
Weisman, J. M. White, Z.-Z. Yang, J. Am. Chem. Soc. 1993, 115,
6580 – 6591.
[6] a) F. Dꢁbritz, A. Jꢁger, I. Bauer, Eur. J. Org. Chem. 2008, 5571 –
5576; b) F. Dꢁbritz, G. Theumer, M. Gruner, I. Bauer, Tetrahe-
dron 2009, 65, 2995 – 3002; c) earlier papers in this very
interesting series, and related works by others, have been
reviewed: I. Bauer, W. D. Habicher, Collect. Czech. Chem.
Commun. 2004, 69, 1195 – 1230.
Scheme 3. Summary of key equilibria.[19]
[7] A. J. Nawara, T. Shima, F. Hampel, J. A. Gladysz, J. Am. Chem.
Soc. 2006, 128, 4962 – 4963.
[8] a) T. Shima, F. Hampel, J. A. Gladysz, Angew. Chem.
2004, 116, 5653 – 5656; Angew. Chem. Int. Ed. 2004,
43, 5537 – 5540; b) L. Wang, F. Hampel, J. A. Gladysz,
Angew. Chem. 2006, 118, 4479 – 4482; Angew. Chem.
Int. Ed. 2006, 45, 4372 – 4375; c) L. Wang, T. Shima, F.
Hampel, J. A. Gladysz, Chem. Commun. 2006, 4075 –
4077; d) G. D. Hess, F. Hampel, J. A. Gladysz, Orga-
nometallics 2007, 26, 5129 – 5131; e) K. Skopek, J. A.
Gladysz, J. Organomet. Chem. 2008, 693, 857 – 866.
[9] See also J. E. Nuꢂez, A. Natarajan, S. I. Khan, M. A.
Garcia-Garibay, Org. Lett. 2007, 9, 3559 – 3561, and
references therein.
[10] M. Saunders, J. Comput. Chem. 1989, 10, 203 – 208.
[11] R. W. Alder, C. P. Butts, A. G. Orpen, D. Read, J. M.
Oliva, J. Chem. Soc. Perkin Trans. 2 2001, 282 – 287,
and references therein. The conjugate acid of out,out-
1,6-diphosphabicyclo[4.4.4]tetradecane—the
ana-
logue of out,out-2 with four (CH2)4 chains—has
been reported, but attempted deprotonations afford
deep-seated rearrangements.
Scheme 4. Other relevant reactions and structures.
[12] R. D. Baechler, K. Mislow, J. Am. Chem. Soc. 1970,
92, 3090 – 3093.
[13] A. Grabulosa, J. Granell, G. Muller, Coord. Chem. Rev. 2007,
251, 25 – 90.
binding to an out lone pair of electrons, followed by a
homeomorphic isomerization, as illustrated by IX–XI. To our
knowledge, this pathway has not been considered previously
in the literature. An analogous mechanism is likely for the
reconstitution of trans-1 from PtCl2 and 2 (Scheme 1). Finally,
other families of isomeric macrobicyclic dibridgehead diphos-
phorus compounds, as exemplified by XII, have been
reported,[6] and it should be possible to demonstrate analo-
gous dynamic behavior by appropriate NMR investigations
and preparative experiments.
[14] Analogous separation of the signals is not observed in CDFCl2 at
À1008C (precipitation begins at lower temperatures, thus
degrading the signal). Very slight signal broadening gradually
occurs between À10 and À708C (w1/2 = 7.6–22.2 Hz), which then
drops (15.5 and 13.9 Hz at À80 and À1008C). In contrast, the
signal in [D8]toluene broadens dramatically between 27 and
À408C (w1/2 = 5.8–93.0 Hz), and then sharpens (72.4, 26.2, 9.7,
and 6.3 Hz at À50, À60, À70, and À808C). Thus, it is possible
that solvent adducts analogous to that in Figure 3 might play a
role in these equilibria.
[15] K. Skopek, M. Barbasiewicz, F. Hampel, J. A. Gladysz, Inorg.
Chem. 2008, 47, 3474 – 3476.
Received: February 4, 2011
[16] Since trans-1 and cis-1 do not interconvert within 14 h at 1808C,
trans/cis isomerization cannot play a role in Scheme 1.
[17] Certain reactions of dibridgehead diphosphines based upon
smaller rings can afford species with phosphorus–phosphorus
bonds, for which inversions at phosphorus atoms have been
documented: a) R. W. Alder, D. Read, Angew. Chem. 2000, 112,
Published online: June 3, 2011
Keywords: olefinmetathesis· phosphines·platinumcomplexes·
.
pyramidal inversion · variable-temperature NMR spectroscopy
6650 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6647 –6651

3001 – 3004; Angew. Chem. Int. Ed. 2000, 39, 2879 – 2882; b) see
also R. W. Alder, C. P. Butts, A. G. Orpen, D. Read, J. Chem.
Soc. Perkin Trans. 2 2001, 288 – 295.
projections down the phosphorus–phosphorus vectors, are
eclipsed (I, III, and V) or staggered (II, IV, VI). There is a C2
axis in IV and VI that leads to exchange of the phosphorus
atoms, which are inequivalent in V.
[18] A. J. Nawara-Hultzsch, K. Skopek, T. Shima, M. Barbasiewicz,
G. D. Hess, D. Skaper, J. A. Gladysz, Z. Naturforsch. B 2010, 65,
414 – 424.
[20] For in/out isomers in 1,3,5-cyclophanes tethered to bridgehead
atoms, see R. A. Pascal, Jr., Eur. J. Org. Chem. 2004, 3763 – 3771.
[19] The point groups given in Scheme 3 are a function of whether
the phosphorus–carbon bonds, as viewed in Newman-type
Angew. Chem. Int. Ed. 2011, 50, 6647 –6651
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6651