Direct evidence of a multicentre halogen bond: Unexpected contraction of the P-XXX-P fragment in triphenylphosphine dihalides

Article abstract of DOI:10.1039/c2cc38363b

Triphenylhalophosphonium halides, Ph₃PX₂, form crystals comprising bridged linear cations [Ph₃P-X-X-X-PPh ₃]⁺ where the X₃ bridge is shortened from 6.56 A in Cl-Cl-Cl to 6.37 A in the Br-Br-Br system. It is proposed that this structure is stabilised by five-centre/six-electron (5c-6e) hypervalent interactions.

Full text of DOI:10.1039/c2cc38363b

ChemComm
View Article Online
View Journal
COMMUNICATION
Direct evidence of a multicentre halogen bond:
unexpected contraction of the P–XXX–P fragment
Cite this: DOI: 10.1039/c2cc38363b
in triphenylphosphine dihalides†
Received 20th November 2012,
Accepted 21st December 2012
Kirill Nikitin,* Helge M u¨ ller-Bunz and Declan Gilheany*
DOI: 10.1039/c2cc38363b
www.rsc.org/chemcomm
Triphenylhalophosphonium halides, Ph
3 2
PX , form crystals comprising
+
bridged linear cations [Ph P–X–X–X–PPh ] where the X bridge is
3
3 3
shortened from 6.56 Å in Cl–Cl–Cl to 6.37 Å in the Br–Br–Br system.
It is proposed that this structure is stabilised by five-centre/
six-electron (5c–6e) hypervalent interactions.
The idea of multicentre hypervalent interaction, such as the
three-centre, four-electron (3c–4e) bond, developed alongside
1
those of classical covalent bonding. Extended versions of such
2
systems e.g. 5c–6e cases were identified later and the area is
3
,4
now of substantial current interest. For example, Nakanishi
3
and co-workers have extensively developed extended hypervalent
Chart 1 The principal dihalophosphorane structures (orange – nominal positive
bonds in pre-designed naphthalene and anthracene structures
involving strings incorporating moderately polar functional
charge; blue – nominal negative charge).
groups as individual components. Related to this are the extensive
4
studies by Kilian et al. of sub van der Waals contacts in peri- bipyramidal species, 1D. However a third spoke-geometry B was
substituted naphthalenes. In the case of the halogens, hypervalent subsequently identified for species R P–X–X (Chart 1, X = Br, I)
3
interactions have helped to rationalise the nature of the halogen bearing strong electron-donating alkyl substituents and the
9
bond (XB) connecting a covalently bonded halogen atom X to an structure of B was studied theoretically.
electron-rich electronegative atom. This extremely active current
5
We now report that we have identified a self-organised extended
area has seen XB established in the formation of solid state 5c–6e PXXXP motif. Our observations are based on an earlier report
10
structures, molecular association, medicinal chemistry and by Godfrey, McAuliffe and co-workers of yet another unique
5
,6
supramolecular processes. However relatively little is known bridged structure for dichlorophosphoranes (Chart 1, C). They
7
about such interactions involving a P–X moiety.
found that 1 crystallised from dichloromethane (DCM) shows
+
Dihalophosphoranes R PX show remarkable structural an unexpected linear cationic species [Ph P–Cl–Cl–Cl–PPh ] , and
3
2
3
3
À
8
,9
À
+
diversity
and they have been the subject of investigation chloride anions Cl rather than individual [Ph PCl] and Cl . They
3
especially owing to the possible co-existence and interconversion also noted that, while this structure is essentially ionic, the ClÁÁÁCl
of both ionic and covalent species with, in many cases, a fine distance (3.28 Å) between the adjacent bridging chlorines in 1C is
8
energy balance between them. In the earliest studies, solid state markedly shorter than the sum of van der Waals radii (3.56 Å). Since
3 2
Ph PX (1, X = Cl; 2 X = Br) had commonly been thought to occur the bridged structure C, similarly to B, can be viewed as a case of
in just two structural types (Chart 1): the type A tetracoordinate polycentre XB, we were interested to establish whether the spatial
+
À
salt [Ph
3
PX] X 1A and the type D pentacoordinate trigonal organisation in C is driven by the propensity of 1 to form such
bridged structures. Here we present new data for the dibromide 2
Centre for Synthesis and Chemical Biology, School of Chemistry and Chemical Biology,
University College Dublin, Ireland. E-mail: kirill.nikitin@ucd.ie,
and mixed chloride–bromide Ph PClBr, 3 (Scheme 1). Our findings
3
suggest that the configuration of bridged species is indeed
driven by polycentre covalent interactions and that phosphane
halides show an example of robust hypervalent interactions
capable of maintaining their integrity within a range of geometrical
parameters of the crystal structure.
declan.gilheany@ucd.ie; Fax: +353-01-7162127; Tel: +353-01-7162316
†
Electronic supplementary information (ESI) available: Preparation of 1, 2 and 3,
crystallographic, computational and NMR data. CCDC 887828 and 887829. For
ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/
c2cc38363b
This journal is c The Royal Society of Chemistry 2013
Chem. Commun.

View Article Online
Communication
ChemComm
Scheme 1 Preparation of dihalides 1C, 2C and mixed halide 3C.
10
Fig. 1 X-ray crystal structure of the dichloride 1C showing linear spatial
arrangement of the bridge moiety P–Cl–Cl–Cl–P, anions Cl(3) and solvent mole-
cules (all hydrogens are omitted for clarity).
Fig. 2 X-ray crystal structure of: (a) dibromide 2C showing linear bridge
P–Br–Br–Br–P; (b) mixed halide 3C showing a linear P–Cl–Br–Cl–P unit (isolated
counteranions and solvent molecules are omitted).
Dihalides 1 and 2 were previously prepared by halogenation of
9,10
Table 1 Geometric parameters (Å) of 1C–3C
phosphines.
which arose from their central role in our asymmetric Appel-type
We had an interest in alternative preparations,
Cl–Cl–Cl
15.995
Cl–Br–Cl
15.974
Br–Br–Br
11
processes for P-stereogenic phosphorus compounds. Specifi-
a
d
16.162
2.17
cally, we wished to establish that the reaction of phosphine oxides
(P–X)
(XXX)
Exp
Calc
Exp
Calc
2.01a
2.00a
1
2
b
b
b
b
a
b
b
with oxalyl halides (COX)
2
(X = Cl, Br) leads to the same halide
1.97 /2.06
2.06 /2.06
2.17 /2.26
13
d
6.56a
6.52a
6.37
6.47 /5.85
8.64
structural type as shown in Chart 1. When a solution of Ph PO
a
3
6.11 /5.58
6.53 /5.93
c
c
c
in DCM (0.1 M) was treated with (COCl) vigorous effervescence
2
d(
8.56
8.57
PÁ Á ÁX)
was followed by precipitation of colourless symmetrical crystals.
The decoupled P NMR spectrum in DCM (sharp singlet at
a
b
c
À
HF calculations. DFT B3LYP. Isolated counteranions X .
3
1
8
,14
1
13
6
2.1 ppm),
with the tetracoordinate chlorophosphonium cation structure the isolated anionic occupancies were 0.62 and 0.38 for Cl and
see ESI† section). The X-ray diffraction analysis of this material Br, respectively, which are very close to their statistical distribu-
clearly indicated that its cubic crystal structure (a = 15.995 Å) tion in solution. The unit cell of 3C is smaller (a = 15.974 Å) than
and the H and C NMR spectra were consistent
(
1
0
was identical to the earlier reported
chloride 1C and that of 1C making 3C the most densely packed of all three
contained the bridged linear moieties [Ph P–Cl–C–Cl–PPh ] as dihalides. In comparing 1C–3C (Table 1, Fig. 3), it is apparent
3
3
shown in Fig. 1. This confirmed that crystallisation of the that the X–X–X bridge interconnecting the two phosphorus
dichloride 1 from DCM leads to 1C regardless of the method centres contracts progressively as the sum of VDW sizes of the
by which 1 is generated in solution.
The analogous bromide 2 was prepared by treatment of dibromide 2C (6.37 Å) is the shortest, despite the fact that its
PO in DCM with oxalyl bromide (see Scheme 1). X-ray diffrac- unit cell is the largest (16.162 Å) and the P–X covalent bond is the
three halogens involved is increased. The trihalide bridge in the
Ph
3
tion analysis of the bright yellow crystals of 2 revealed a bridged longest (2.169 Å). This intriguing contraction phenomenon is, to
tribromo(bis)triphenylphosphonium bromide 2C (Fig. 2a) fully the best of our knowledge, exceptional. However, it can be
isostructural with 1C. The parameters of 2C exhibit several inter- rationalised by postulating that a polycentre covalent bond is
esting features. First, its cubic unit cell is somewhat larger (a = responsible for the formation of the bridged trihalide structures.
16.162 Å) compared to the chloride 1C (Table 1). Second, the
To examine the electronic structure of the P–X–X–X–P multi-
position of the isolated counteranions is significantly, by 1.1 Å, centre systems, we compared the experimental and VDW distances
disordered and the average distance to the phosphorus, d(PÁÁÁBr) is to quantum chemistry calculation results (Fig. 3). The equilibrium
longer, 8.64 Å. Third, and most striking, the distance between the geometries in vacuum were optimised at the most commonly used
two terminal bridge bromines Br(2) in 2C is shorter (6.37 Å), than HF and DFT levels using the 6-31G* basis set. Both approaches
in the analogous chloride 1C (6.56 Å) despite the bigger size provide linear symmetrical geometries of the P–X–X–X–P fragment
(1.15 Å) and greater van der Waals radius (1.85 Å) of bromine. and realistic P–X covalent bond lengths (Table 1). Clearly, the
We were able to go a step further. When solutions of 1 and 2 experimentally observed distances d(XXX) are significantly shorter
1
5
were mixed in a 2 : 1 ratio and the solvent was gradually than VDW prediction, as mentioned above, and they are inversely
removed, new pale yellow cubic crystals were isolated. X-ray proportional to the halide size (Fig. 3). Interestingly, quantum
diffractometry indicated that this material, 3C, contained only chemistry calculations give even shorter interatomic distances
one type of bridged fragment Ph
3
PCl–Br–ClPPh
3
(Fig. 2b) and
d
(XXX): for instance, the calculated d(ClClCl) in vacuum is shorter
Chem. Commun.
This journal is c The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
This finding is significant for a better understanding of
chemical bonding in main group chemistry.
Notes and references
1
(a) G. C. Pimentel, J. Chem. Phys., 1951, 19, 446; (b) R. J. Hatch and
R. E. Rundle, J. Am. Chem. Soc., 1951, 73, 4321; (c) E. E. Havinga and
E. H. Wiebinga, Recl. Trav. Chim. Pays–Bas, 1959, 78, 724;
(
d) R. E. Rundle, Surv. Prog. Chem., 1963, 1, 81; (e) J. I. Musher,
Angew. Chem., Int. Ed. Engl., 1969, 8, 54; ( f ) J. C. Martin and
E. F. Perozzi, Science, 1976, 191, 154; (g) P. A. Cahill, C. E. Dykstra
and J. C. Martin, J. Am. Chem. Soc., 1985, 107, 6359; (h) Chemistry of
Hypervalent Compounds, ed. K.-y. Akiba, Wiley-VCH, 1998.
Fig. 3 Experimentally observed (red) lengths of the XXX bridge in 1C–3C as
compared to the van der Waals prediction (green), HF (blue) and DFT (grey)
calculated geometries.
2 (a) W. B. Farnham, D. A. Dixon and J. C. Calabrese, J. Am. Chem. Soc.,
988, 110, 8453; (b) W. Nakanishi, S. Hayashi and S. Toyota, Chem.
Commun., 1996, 371; (c) W. Nakanishi, S. Hayashi and N. Itoh, Chem.
Commun., 2003, 124.
1
3
(a) W. Nakanishi, T. Nakamoto, S. Hayashi, T. Sasamori and
N. Tokitoh, Chem.–Eur. J., 2007, 13, 255 and references cited therein;
(
b) T. Nakai, M. Nishino, S. Hayashi, M. Hashimoto and
than in crystalline 1C. Importantly, the calculations predict
contraction of the XXX fragment in the bromide 3C when
compared to mixed halide 2C which is in good agreement with
the trend experimentally observed in the solid state (Fig. 3).
This indicates, possibly, an expected strengthening of the 5c–6e
bond with the introduction of the less electronegative bromines
at the 2nd and 4th positions as pointed out by Farnham, Dixon
and Calabrese. Nevertheless, calculations under these artificial
conditions (i.e. in vacuum and in the absence of counterions)
cannot be expected to fully replicate experimental data. The
discrepancies can possibly be accounted for by electrostatic
interactions with counter-ions and compression between inter-
secting linear chains in the crystal lattice leading to elongation
of relatively weak contacts.
W. Nakanishi, Dalton Trans., 2012, 41, 7485; (c) M. Buehl,
P. Kilian and D. J. Woollins, ChemPhysChem, 2011, 12, 2405.
4 (a) P. Kilian, F. R. Knight and J. D. Woollins, Chem.–Eur. J., 2011,
7, 2302; (b) P. Kilian, F. R. Knight and D. J. Woollins, Coord. Chem.
1
Rev., 2011, 255, 1387.
(a) Halogen Bonding: Fundamentals and Applications, ed. P. Metrangolo
and G. Resnati, Structure and Bonding, vol. 126, Springer, 2008;
5
(b) P. Metrangolo and G. Resnati, Chem.–Eur. J., 2001, 7, 2511; (c) A. C. C.
Carlsson, J. Grafenstein, J. L. Laurila, J. Bergquist and M. Erdelyi, Chem.
Commun., 2012, 48, 1458; (d) M. Tuikka, P. Hirva, K. Rissanen, J. Korppi-
Tommola and M. Haukka, Chem. Commun., 2011, 47, 4499.
(a) supramolecular chemistry: P. Metrangolo, F. Meyer, T. Pilati, G. Resnati
and G. Terraneo, Angew. Chem., Int. Ed., 2008, 47, 6114; A. Caballero,
F. Zapata, N. G. White, P. J. Costa, V. Felix and P. D. Beer, Angew. Chem.,
Int. Ed., 2012, 51, 1876; H.-J. Schneider, Angew. Chem., Int. Ed., 2009,
2a
6
4
8, 3924(b) anion recognition: G. Cavallo, P. Metrangolo, T. Pilati,
G. Resnati, M. Sansotera and G. Terraneo, Chem. Soc. Rev., 2010, 39,
772; (c) halocarbon–protein complexes: E. Parisini, P. Metrangolo,
T. Pilati, G. Resnati and G. Terraneo, Chem. Soc. Rev., 2011,
0, 2267(d) drug design: Y. X. Lu, Y. T. Liu, Z. J. Xu, H. Y. Li, H. L. Liu
3
Analysis of the calculated electrostatic charge distributions
appears to suggest that a linear PXXXP geometry is unlikely to
arise purely from electrostatic interactions. In 2C for example
4
and W. L. Zhu, Expert Opin. Drug Discovery, 2012, 7, 375(e) medicinal
chemistry: M. Z. Hernandes, S. M. T. Cavalcanti, D. R. M. Moreira, W. F. de
Azevedo and A. C. L. Leite, Curr. Drug Targets, 2010, 11, 303( f ) fluorous
chemistry: J. A. Gladysz and M. Jurisch, Top. Curr. Chem., 2012, 308, 1.
(a) W. I. Cross, S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard,
J. M. Sheffield and G. M. Thompson, Dalton Trans., 1999, 2795;
(b) J. Konu and T. Chivers, Dalton Trans., 2006, 3941.
(a) M. A. Al-Juboori, P. N. Gates and A. S. Muir, Chem. Commun.,
1
Soc. A, 1966, 598.
9 (a) N. Bricklebank, S. M. Godfrey, A. G. Mackie, C. A. McAuliffe and
R. G. Pritchard, Chem. Commun., 1992, 355; (b) S. M. Godfrey,
D. G. Kelly, C. A. McAuliffe, A. G. Mackie, R. G. Pritchard and
S. M. Watson, Chem. Commun., 1991, 1163; (c) S. M. Godfrey,
A. Hinchliffe and A. Mkadmh, J. Mol. Struct., 2005, 85; (d) S. M.
Godfrey, C. A. McAuliffe, R. G. Pritchard, J. M. Sheffield and
G. M. Thompson, J. Chem. Soc., Dalton Trans., 1997, 4823.
0 (a) S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard and J. M. Sheffield,
Chem. Commun., 1996, 2521; (b) S. M. Godfrey, C. A. McAuliffe,
R. G. Pritchard and J. M. Sheffield, Chem. Commun., 1998, 921.
1 (a) K. V. Rajendran, L. Kennedy and D. G. Gilheany, Eur. J. Org. Chem.,
(Fig. S2, ESI†), the terminal bromines of the bridge are essen-
tially electrically neutral as their Mulliken charges are +0.001.
We therefore propose that the observed tendency of the triphenyl-
phosphine dihalides to form linear PXXXP trihalide structures
can be attributed to a five-centre hypervalent bond (5c–6e)
involving three electron pairs: one is provided by the central
halide anion and other two by P–X bonds of the halophos-
phonium fragments.
Another intriguing feature of dihalides 1C–3C is that the
isolated anions are separated from the cationic phosphorus
centres by a solvent molecule. This is shown explicitly for 1C
7
8
991, 1270; (b) A. D. Beveridge, G. S. Harris and F. Inglis, J. Chem.
(
Fig. 1) where it can be seen that the arrangement mimics that
1
1
6
of a solvent separated ion pair whereby the molecules CH
2
Cl
2
À
and isolated anions X are aligned with the axis P–X–X–X–P.
This arrangement could be seen as a snap-shot of a process in
which the isolated anion is liable, upon loss of the solvent
molecule, to replace the covalently attached halogen in a
nucleophilic exchange process. A study of processes of this
type is underway in our group.
1
2
010, 5642; (b) E. Bergin, C. T. O’Connor, S. B. Robinson, E. M. McGarrigle,
C. P. O’Mahony and D. G. Gilheany, J. Am. Chem. Soc., 2007, 129, 9566.
2 (a) M. Masaki and K. Fukui, Chem. Lett., 1977, 151; (b) R. M. Denton,
J. An and B. Adeniran, Chem. Commun., 2010, 46, 3025.
3 (a) K. V. Rajendran and D. G. Gilheany, Chem Commun., 2012, 817;
(b) P. A. Byrne, K. V. Rajendran and D. G. Gilheany, Org. Biomol.
Chem., 2012, 10, 3531.
4 N. C. Gonnella, C. Busacca, S. Campbell, M. Eriksson, N. Grinberg,
T. Bartholomeyzik, S. Ma and D. L. Norwood, Magn. Reson. Chem.,
2009, 47, 461.
5 When equal amounts of the chloride 1 and bromide 2 were mixed
only yellow cubic crystals of 2C were formed.
6 S. Winstein and G. C. Robinson, J. Am. Chem. Soc., 1958, 80, 169; R. J.
LeSuer, C. Buttolph and W. E. Geiger, Anal. Chem., 2004, 76, 6395.
1
1
In conclusion, phosphonium dihalides tend to form crystals
comprising of complex bridged trihalophosphonium dications,
molecules of solvent and individual halide anions positioned
potentially to effect nucleophilic attack on the phosphorus.
1
1
We propose that the linear structure of the bridged cations
1
+
[
Ph
3
P–X–X–X–PPh
3
] is stabilised by hypervalent 5c–6e interactions.
This journal is c The Royal Society of Chemistry 2013
Chem. Commun.